- About Scala
- Documentation
- Code Examples
- Software
- Scala Developers
How I failed to convince another programmer that Scala is good - too many problems getting started
Mon, 2011-11-14, 22:18
Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool.I don't want to complain, this is just how it happened. I love Scala but I think thiscould happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as testedwith sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4:He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7:He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google.And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
this is story of how I failed to convince a fellow programmer that Scala is cool.I don't want to complain, this is just how it happened. I love Scala but I think thiscould happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as testedwith sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4:He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7:He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google.And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
Mon, 2011-11-14, 22:47
#2
Re: How I failed to convince another programmer that Scala is g
I see Eclipse mentioned in about N places. Surely people have stopped using that old piece of junk by now? No?
It is not a Scala problem that half of the articles and blog entries out there about this or that programming related subject use 50% of their content for the project setup in Eclipse and/ or Maven. I mean wtf. To me that reeks of:
1. What are you doing reading a programming related articles if you are unable to deal with your build environment or IDE?2. What's going on with the development of build environments and IDEs out there? No-one appears to be able to figure them out.
On Nov 14, 2011, at 10:18 PM, Markus Jais wrote:
It is not a Scala problem that half of the articles and blog entries out there about this or that programming related subject use 50% of their content for the project setup in Eclipse and/ or Maven. I mean wtf. To me that reeks of:
1. What are you doing reading a programming related articles if you are unable to deal with your build environment or IDE?2. What's going on with the development of build environments and IDEs out there? No-one appears to be able to figure them out.
On Nov 14, 2011, at 10:18 PM, Markus Jais wrote:
Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool.I don't want to complain, this is just how it happened. I love Scala but I think thiscould happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as testedwith sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4:He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7:He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google.And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
Mon, 2011-11-14, 22:57
#3
Re: How I failed to convince another programmer that Scala is g
Hi Markus,
I'm sorry to hear that your experience was less than stellar.
We created the Typesafe Stack to make it easy to start making awesome software using Scala:
http://typesafe.com/stack
Cheers,
√
On Mon, Nov 14, 2011 at 10:18 PM, Markus Jais <markus.jais@yahoo.de> wrote:
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
I'm sorry to hear that your experience was less than stellar.
We created the Typesafe Stack to make it easy to start making awesome software using Scala:
http://typesafe.com/stack
Cheers,
√
On Mon, Nov 14, 2011 at 10:18 PM, Markus Jais <markus.jais@yahoo.de> wrote:
Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool. I don't want to complain, this is just how it happened. I love Scala but I think thiscould happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4:He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7:He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google.And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
Mon, 2011-11-14, 23:27
#4
Re: How I failed to convince another programmer that Scala is g
On 11/14/2011 10:18 PM, Markus Jais wrote:
I can't help to notice; There seems to be many problems related to getting Scala projects working using sbt+eclipse. I don't use neither myself, and the above certainly doesn't encourage me to.
How hard is it to build a working Maven project and import it in IDEA? I've done this since Scala-2.8.0-pre days and have had very little trouble. I hear people bashing Maven all the time in the Scala community but stories like this tell me the grass isn't much greener in the sbt/eclipse-camp.
Some sbt+eclipse-users care to shed some light on what sbt+eclipse provides that Maven doesn't (I'm not interested in the "it compiles faster" or "continuous testing" arguments).
I'm posting this because the trend in the Scala-world is to favor sbt/eclipse before Maven+(other Maven and Scala aware IDEs), and I'm still wondering what the benefits are?
1321305480 [dot] 76322 [dot] YahooMailNeo [at] web132309 [dot] mail [dot] ird [dot] yahoo [dot] com" type="cite"> Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool. I don't want to complain, this is just how it happened. I love Scala but I think this could happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4: He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6: I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7: He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google. And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
I can't help to notice; There seems to be many problems related to getting Scala projects working using sbt+eclipse. I don't use neither myself, and the above certainly doesn't encourage me to.
How hard is it to build a working Maven project and import it in IDEA? I've done this since Scala-2.8.0-pre days and have had very little trouble. I hear people bashing Maven all the time in the Scala community but stories like this tell me the grass isn't much greener in the sbt/eclipse-camp.
Some sbt+eclipse-users care to shed some light on what sbt+eclipse provides that Maven doesn't (I'm not interested in the "it compiles faster" or "continuous testing" arguments).
I'm posting this because the trend in the Scala-world is to favor sbt/eclipse before Maven+(other Maven and Scala aware IDEs), and I'm still wondering what the benefits are?
-- Andreas Joseph Krogh- mob: +47 909 56 963 Senior Software Developer / CTO - OfficeNet AS - http://www.officenet.no Public key: http://home.officenet.no/~andreak/public_key.asc
Mon, 2011-11-14, 23:47
#5
Re: How I failed to convince another programmer that Scala is g
On 2011-11-14 2:19 PM, Andreas Joseph Krogh wrote:
If you really want people to use your stuff, or think your creations are cool, you have to make the learning curve for newbies as short and shallow as possible.
Anyway - that is my pet peeve, but sadly it does not solve your problem Markus
4EC193F5 [dot] 2040402 [at] officenet [dot] no" type="cite"> On 11/14/2011 10:18 PM, Markus Jais wrote:Um, this kind of problem is not unique to Scala, SBT, Eclipse, Akka or anything else. All too often I find that when people develop something cool, that because they can demonstrate something works, then their job is done. They then take one of two approaches (1) they try to document how to repeat what they did, but fail to test their documentation or fail to update it later, or (2) they leave behind a terse recipe of technologies and figure if you don't know any or all those ingredients then you ought not to be cooking.1321305480 [dot] 76322 [dot] YahooMailNeo [at] web132309 [dot] mail [dot] ird [dot] yahoo [dot] com" type="cite"> Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool. I don't want to complain, this is just how it happened. I love Scala but I think this could happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4: He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6: I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7: He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google. And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
I can't help to notice; There seems to be many problems related to getting Scala projects working using sbt+eclipse. I don't use neither myself, and the above certainly doesn't encourage me to.
How hard is it to build a working Maven project and import it in IDEA? I've done this since Scala-2.8.0-pre days and have had very little trouble. I hear people bashing Maven all the time in the Scala community but stories like this tell me the grass isn't much greener in the sbt/eclipse-camp.
Some sbt+eclipse-users care to shed some light on what sbt+eclipse provides that Maven doesn't (I'm not interested in the "it compiles faster" or "continuous testing" arguments).
I'm posting this because the trend in the Scala-world is to favor sbt/eclipse before Maven+(other Maven and Scala aware IDEs), and I'm still wondering what the benefits are?
-- Andreas Joseph Krogh andreak [at] officenet [dot] no" rel="nofollow"><andreak@officenet.no> - mob: +47 909 56 963 Senior Software Developer / CTO - OfficeNet AS - http://www.officenet.no Public key: http://home.officenet.no/~andreak/public_key.asc
If you really want people to use your stuff, or think your creations are cool, you have to make the learning curve for newbies as short and shallow as possible.
Anyway - that is my pet peeve, but sadly it does not solve your problem Markus
4EC193F5 [dot] 2040402 [at] officenet [dot] no" type="cite">
Mon, 2011-11-14, 23:57
#6
Re: How I failed to convince another programmer that Scala is g
Whether or not you use sbt has nothing to do with whether or not you use eclipse.Many people like sbt because it has a very powerful and concise system for configuring builds and writing plugins. For small projects with a straightforward structure that you just need to compile and run, this may not matter.
On Mon, Nov 14, 2011 at 5:19 PM, Andreas Joseph Krogh <andreak@officenet.no> wrote:
On Mon, Nov 14, 2011 at 5:19 PM, Andreas Joseph Krogh <andreak@officenet.no> wrote:
On 11/14/2011 10:18 PM, Markus Jais wrote:Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool. I don't want to complain, this is just how it happened. I love Scala but I think this could happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4: He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6: I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7: He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google. And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
I can't help to notice; There seems to be many problems related to getting Scala projects working using sbt+eclipse. I don't use neither myself, and the above certainly doesn't encourage me to.
How hard is it to build a working Maven project and import it in IDEA? I've done this since Scala-2.8.0-pre days and have had very little trouble. I hear people bashing Maven all the time in the Scala community but stories like this tell me the grass isn't much greener in the sbt/eclipse-camp.
Some sbt+eclipse-users care to shed some light on what sbt+eclipse provides that Maven doesn't (I'm not interested in the "it compiles faster" or "continuous testing" arguments).
I'm posting this because the trend in the Scala-world is to favor sbt/eclipse before Maven+(other Maven and Scala aware IDEs), and I'm still wondering what the benefits are?
-- Andreas Joseph Krogh- mob: +47 909 56 963 Senior Software Developer / CTO - OfficeNet AS - http://www.officenet.no Public key: http://home.officenet.no/~andreak/public_key.asc
Mon, 2011-11-14, 23:57
#7
Re: How I failed to convince another programmer that Scala is g
> Some sbt+eclipse-users care to shed some light on what sbt+eclipse provides that Maven doesn't (I'm not interested in the "it compiles faster" or "continuous testing" arguments).
Drop the eclipse part and i am interested. After years of Maven quirks I refuse to subject myself to it on personal projects. I hear 3.x remedies a lot of my issues however (like suddenly upgrading my plugins because an automatically configured one has a newer version or not being able to dynamically set my parent pom version).
Negatives of sbt are the lack of mojos for many activities that scala devs seem not to care too much for (flex, site and websphere would be things on my work list) and the massive pace of change. For site I have made a plugin just needs porting to 0.11.
Positives are fine grained control when you want it, a truly great parent project model (mostly because it isn't one) and rarely having to leave it to run tests or package or run a console with the correct classpath.
I use Emacs and Ensime by preference because they are fast and get out of my way until I need them. If either sbt or Ensime get slow I kill em and start them up quickly again. In the rare cases I need to debug there is jswat to cover my needs. My spare time (in the train commute) is precious to me and both Maven and Eclipse rob me of it.
Tue, 2011-11-15, 00:57
#8
Re: How I failed to convince another programmer that Scala is g
>
> My spare time is precious to me and both Maven and Eclipse rob me of it.
Is this the part of the thread where we're supposed to start the 'sbt' vs 'Maven' flame war?
Tue, 2011-11-15, 01:27
#9
Re: How I failed to convince another programmer that Scala is g
I think that when the tools start getting in the way we need to change
the tools. I had my share of problems with eclipse and it's plugins
and I don't recommend and we can't blame Scala or Akka for eclipse
issues.
cheers
Tue, 2011-11-15, 08:47
#10
Sv: Re: How I failed to convince another programmer that Scala
På mandag 14. november 2011 kl 23:37:14 skrev Naftoli Gugenheim <naftoligug@gmail.com>:
Whether or not you use sbt has nothing to do with whether or not you use eclipse. Many people like sbt because it has a very powerful and concise system for configuring builds and writing plugins. For small projects with a straightforward structure that you just need to compile and run, this may not matter.
I know sbt+eclipse are separate things. It's just that the scala-community seems to push/market sbt+eclipse as a bundle for newcommers and my concern is that it results in more problems than it solves. It seems that more than half of the advices setting up a scala-project using sbt+eclipse fail for newcommers and that certainly doesn't motivate them. I just wonder why Maven doesn't get pushed as I'm very pleased with it and I've yet to have less than 20 modules in my projects. Setting up a Maven-project in all major IDEs is just a matter of importing the poms, seems pretty bullet-proof to me, instead of relying on poor sbt-integration and what not.
I'm not interessted in starting Maven vs. sbt or IDEA vs. Eclipse, just offering newcommers advice which is robust and works.
Just my 2¢
--
Andreas Joseph Krogh <andreak@officenet.no> - mob: +47 909 56 963
Senior Software Developer / CTO - OfficeNet AS - http://www.officenet.no
Public key: http://home.officenet.no/~andreak/public_key.asc
Tue, 2011-11-15, 08:57
#11
Re: How I failed to convince another programmer that Scala is g
Sounds like it's about whether Scala can be sold in department stores.
Maybe not yet.
But those who appreciate it, they just know.
If you saw how our interns pick up everything, not having even heard of Scala beforehand.
They do produce production code, scala/lift/scalacheck, all going pretty fast.
It's a new generation, they don't feel they are entitled to use Java till they retire.
Thanks,
-Vlad
On Mon, Nov 14, 2011 at 1:18 PM, Markus Jais <markus.jais@yahoo.de> wrote:
Maybe not yet.
But those who appreciate it, they just know.
If you saw how our interns pick up everything, not having even heard of Scala beforehand.
They do produce production code, scala/lift/scalacheck, all going pretty fast.
It's a new generation, they don't feel they are entitled to use Java till they retire.
Thanks,
-Vlad
On Mon, Nov 14, 2011 at 1:18 PM, Markus Jais <markus.jais@yahoo.de> wrote:
Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool. I don't want to complain, this is just how it happened. I love Scala but I think thiscould happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Problem 1:
$ sbt eclipse
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Problem 4:He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Problem 7:He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google.And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
Cheers,
Markus
Tue, 2011-11-15, 10:27
#12
Re: How I failed to convince another programmer that Scala is g
Just to be clear, I don't wanted to blame Scala or Akka for the problems we had. I think they are someof the coolest things that ever happened to the JVM.
Markus
Markus
Von: Gilberto Garcia <giba.dmb@gmail.com>
An: Brian Schlining <bschlining@gmail.com>
Cc: scala-user@googlegroups.com
Gesendet: 1:19 Dienstag, 15.November 2011
Betreff: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
I think that when the tools start getting in the way we need to change
the tools. I had my share of problems with eclipse and it's plugins
and I don't recommend and we can't blame Scala or Akka for eclipse
issues.
cheers
Tue, 2011-11-15, 16:07
#13
RE: How I failed to convince another programmer that Scala is g
THUMBS UP
From: scala-user@googlegroups.com [mailto:scala-user@googlegroups.com] On Behalf Of Vlad Patryshev
Sent: November-15-11 2:48 AM
To: Markus Jais
Cc: scala-user@googlegroups.com
Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
Sounds like it's about whether Scala can be sold in department stores.
Tue, 2011-11-15, 17:37
#14
Re: How I failed to convince another programmer that Scala is g
On 15 November 2011 00:19, Gilberto Garcia <giba.dmb@gmail.com> wrote:
I think that when the tools start getting in the way we need to change
the tools.
That's fine once you're already using scala, as you've already made the judgement that scala is worth enough that you will change the tools you use. For people who wish to just give it a spin, having to change IDE, build system and whatever else in their tool-chain just to try out another JVM-hosted language is too much. If the intent is to allow scala the best opportunity of being adopted (where it makes sense) by as wide a community as possible, expecting people to change their entire tool stack is a non-starter.
Matthew
cheers
--
Dr Matthew PocockIntegrative Bioinformatics Group, School of Computing Science, Newcastle Universitymailto: turingatemyhamster@gmail.com gchat: turingatemyhamster@gmail.commsn: matthew_pocock@yahoo.co.uk irc.freenode.net: drdozerskype: matthew.pococktel: (0191) 2566550mob: +447535664143
Tue, 2011-11-15, 18:07
#15
Re: How I failed to convince another programmer that Scala is g
I agree with you Matthew. But if people start getting trouble with the
tools that we, as a community recommend, when they just want give it a
spin, well, that's not good either.
People will not adopt Scala or whatever language they choose to try if
the environment advertised is to complex and brings more troubles and
headaches than solutions.
Things should be simple for the newcomers.
I don't know, maybe we, as a community, should work to make sbt
integration with eclipse smoother or start thinking in how can we make
newcomers life's easier.
Tue, 2011-11-15, 18:47
#16
Re: How I failed to convince another programmer that Scala is g
On 11/15/2011 05:49 PM, Gilberto Garcia wrote:
> I agree with you Matthew. But if people start getting trouble with the
> tools that we, as a community recommend, when they just want give it a
> spin, well, that's not good either.
>
> People will not adopt Scala or whatever language they choose to try if
> the environment advertised is to complex and brings more troubles and
> headaches than solutions.
>
> Things should be simple for the newcomers.
>
> I don't know, maybe we, as a community, should work to make sbt
> integration with eclipse smoother or start thinking in how can we make
> newcomers life's easier.
That's why I recommend Maven+IDEA to people asking me how to start with
Scala development. It's familiar for many.
Tue, 2011-11-15, 19:07
#17
Re: How I failed to convince another programmer that Scala is g
On Mon, Nov 14, 2011 at 19:18, Markus Jais <markus.jais@yahoo.de> wrote:
Building things you want to show is like seeing how the food you are going to eat is prepared. It should be avoided for a more pleasant experience. :-)
It is a bit weird, but sbteclipse won't add to the classpath things that have not been downloaded it. You have to make the project at least once, so that it downloads the jar files, before that will work. Also, maybe you want:
sbt "eclipse with-sources same-targets"
ant doesn't create a project. maven doesn't create a project. make and gnu make don't create a project. Why should sbt act different than other tools? You can use the sbteclipse plugin, though:
sbt "eclipse create-src with-sources same-targets"
Too little information at this point. Where you running this on the Akka project you downloaded or what?
Try running test on Eclipse. :-)
Too little information. If you keep the project build on sbt, running "sbt update eclipse" after any change to the project helps, though.
There was one big breaking change between 0.7.7 and 0.10.0 (as the missing versions attest). Marc decided that the fundamental structure of SBT was wrong, and changed it radically. Like if Sun decided to launch C# instead of Java 1.5. Though this is a major break in compatibility, SBT projects are usually small, which makes the process of completely rewriting the configuration less painful.
IntelliJ works, Eclipe doesn't. Sad but true.
Maven wasn't fundamentally broken. Or, depending on were you stand in relation to maven, Maven 3 didn't fix the fundamental brokenness. :-)
Well, that's not quite the case. SBT 0.7 wasn't really fundamentally broken, but to proceed in the direction he wanted, Marc *had* to make very fundamental changes.
Get him off Eclipse and on IntelliJ. I fail to see how can one work in Scala with Eclipse and think Scala is production-ready: the environment is conspiring against him. I'll be very happy when ScalaIDE for Eclipse becomes a rock-solid platform for Scala development, but, up to now, that is not the case at all.
It *can* be much simpler, but, really, you went through the forest instead of taking the nice road that cuts through it.
First of all, you built, instead of used, projects -- and some pretty heavy at that. Second, you picked an unstable development environment. Third, you completely avoided the "shortest-distance-path" by not picking Typesafe.
--
Daniel C. Sobral
I travel to the future all the time.
Hello,
this is story of how I failed to convince a fellow programmer that Scala is cool. I don't want to complain, this is just how it happened. I love Scala but I think thiscould happen to more people and turn them away before they get started.
I wanted to show him Akka:
First whe checked out akka and built it using sbt. That worked fine.
Building things you want to show is like seeing how the food you are going to eat is prepared. It should be avoided for a more pleasant experience. :-)
Problem 1:
$ sbt eclipse
It is a bit weird, but sbteclipse won't add to the classpath things that have not been downloaded it. You have to make the project at least once, so that it downloads the jar files, before that will work. Also, maybe you want:
sbt "eclipse with-sources same-targets"
(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
Then he looked at the Akka tutorial and tried a sample project:
He tried to create the sample project as described here:
http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
Problem 2:
sbt didn't ask him the question about creating a project. A little google research showed that we
had to create the project folders ourservels (why?)
ant doesn't create a project. maven doesn't create a project. make and gnu make don't create a project. Why should sbt act different than other tools? You can use the sbteclipse plugin, though:
sbt "eclipse create-src with-sources same-targets"
Problem 3:
He did that, but then sbt claimed that it couldn't find "PluginDefinition".
Too little information at this point. Where you running this on the Akka project you downloaded or what?
Problem 4: He ran sbt test on the Akka source: java.lang.OutOfMemoryError
Try running test on Eclipse. :-)
Problem 5:
He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
Too little information. If you keep the project build on sbt, running "sbt update eclipse" after any change to the project helps, though.
Problem 6: I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
There was one big breaking change between 0.7.7 and 0.10.0 (as the missing versions attest). Marc decided that the fundamental structure of SBT was wrong, and changed it radically. Like if Sun decided to launch C# instead of Java 1.5. Though this is a major break in compatibility, SBT projects are usually small, which makes the process of completely rewriting the configuration less painful.
Problem 7: He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
IntelliJ works, Eclipe doesn't. Sad but true.
Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google. And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
Can I blame him?
Maven wasn't fundamentally broken. Or, depending on were you stand in relation to maven, Maven 3 didn't fix the fundamental brokenness. :-)
Well, that's not quite the case. SBT 0.7 wasn't really fundamentally broken, but to proceed in the direction he wanted, Marc *had* to make very fundamental changes.
I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
Get him off Eclipse and on IntelliJ. I fail to see how can one work in Scala with Eclipse and think Scala is production-ready: the environment is conspiring against him. I'll be very happy when ScalaIDE for Eclipse becomes a rock-solid platform for Scala development, but, up to now, that is not the case at all.
I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
It *can* be much simpler, but, really, you went through the forest instead of taking the nice road that cuts through it.
First of all, you built, instead of used, projects -- and some pretty heavy at that. Second, you picked an unstable development environment. Third, you completely avoided the "shortest-distance-path" by not picking Typesafe.
Cheers,
Markus
--
Daniel C. Sobral
I travel to the future all the time.
Tue, 2011-11-15, 21:47
#18
Re: How I failed to convince another programmer that Scala is g
I think this would be important. I also had trouble when I started to do more sbt with Scala and Eclipse.sbt is awesome, no doubt but most people I know use Eclipse (or Netbeans) and they want to use the samieIDE for other languages like Scala or Clojure.If the Eclipse plugin has full integration with sbt, that is run sbt from eclipse, etc, it would be a great helpfor beginners.
Markus
Markus
Von: Gilberto Garcia <giba.dmb@gmail.com>
An: scala-user <scala-user@googlegroups.com>
Gesendet: 17:49 Dienstag, 15.November 2011
Betreff: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
I agree with you Matthew. But if people start getting trouble with the
tools that we, as a community recommend, when they just want give it a
spin, well, that's not good either.
People will not adopt Scala or whatever language they choose to try if
the environment advertised is to complex and brings more troubles and
headaches than solutions.
Things should be simple for the newcomers.
I don't know, maybe we, as a community, should work to make sbt
integration with eclipse smoother or start thinking in how can we make
newcomers life's easier.
Tue, 2011-11-15, 22:37
#19
Re: How I failed to convince another programmer that Scala is g
>
>
>On Mon, Nov 14, 2011 at 19:18, Markus Jais wrote:
>
>Hello,
>>
>>
>>this is story of how I failed to convince a fellow programmer that Scala is cool.
>>I don't want to complain, this is just how it happened. I love Scala but I think this
>>could happen to more people and turn them away before they get started.
>>
>>
>>I wanted to show him Akka:
>>
>>
>>First whe checked out akka and built it using sbt. That worked fine.
>
>
>Building things you want to show is like seeing how the food you are going to eat is prepared. It should be avoided for a more pleasant experience. :-)
>
>
>>
>>Problem 1:
>>
>>
>>$ sbt eclipse
>
>
>It is a bit weird, but sbteclipse won't add to the classpath things that have not been downloaded it. You have to make the project at least once, so that it downloads the jar files, before that will work. Also, maybe you want:
>
>
>sbt "eclipse with-sources same-targets"
Thanks, I didn't know that. Just tried that myself. sbt did indeed download lots of stuff. But eclipse still doesn't work with
the generated .classpath.
>
>>
>>(latest sbt, latest eclipse plugin) it created the .classpath and .project file but in the .classpath no src folders were inserted and Eclipse couldn't load the akka source. The plugin seems to work generally, as tested
>>with sceletor (a Cassandra library for Scala) but id doesn't create correct .classpath files for projects like Akka or spray - at least his (and my) Eclipse couldn't use them.
>>
>>
>>
>>Then he looked at the Akka tutorial and tried a sample project:
>>
>>
>>He tried to create the sample project as described here:
>>
>>
>>http://akka.io/docs/akka/1.2/scala/tutorial-chat-server.html
>>
>>
>>Problem 2:
>>
>>sbt didn't ask him the question about creating a project. A little google research showed that we
>>
>>had to create the project folders ourservels (why?)
>>
>
>
>ant doesn't create a project. maven doesn't create a project. make and gnu make don't create a project. Why should sbt act different than other tools? You can use the sbteclipse plugin, though:
>
>
>sbt "eclipse create-src with-sources same-targets"
>
>
>
>
>>
>>Problem 3:
>>
>>He did that, but then sbt claimed that it couldn't find "PluginDefinition".
>
>
>Too little information at this point. Where you running this on the Akka project you downloaded or what?
Yes, we tried it on the Akka source code (master branch). sbt couldn't find that class.
>
>
>>
>>Problem 4:
>>He ran sbt test on the Akka source: java.lang.OutOfMemoryError
>
>
>Try running test on Eclipse. :-)
I have rarely problems with Eclipse, it is probably years since I had an java.lang.OutOfMemoryError.
Netbeans ist better in some cases (e.g. for JEE 6).
>
>
>>
>>Problem 5:
>>
>>
>>He created a scala project in Eclipse (2.0 RC1 plugin) from the Akka source code. Eclipse found many
>>
>>source folders but not all. I helped and we added manually and tried to run ChatServer (it has a main method) but Eclipse didn't recognise that and we couldn't run the file.
>
>
>Too little information. If you keep the project build on sbt, running "sbt update eclipse" after any change to the project helps, though.
sbt builds fine. but still no useful .classpath.
>
>
>>
>>
>>
>>Problem 6:
>>I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
>
>
>There was one big breaking change between 0.7.7 and 0.10.0 (as the missing versions attest). Marc decided that the fundamental structure of SBT was wrong, and changed it radically. Like if Sun decided to launch C# instead of Java 1.5. Though this is a major break in compatibility, SBT projects are usually small, which makes the process of completely rewriting the configuration less painful.
>
>
>>
>>Problem 7:
>>He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
>
>
>IntelliJ works, Eclipe doesn't. Sad but true.
I know. But many people use Eclipse - some even like it. It is not as bad as some people say. I also use it for
Python, Clojure and most of my Java work. The same for the guy I tried to show Scala. IintelliJ may be better
but not everyone wants to change and IDE, particularly not after having become really productive with Eclipse.
I can understand when someone says he wants to continue his using his tools if they work really well for him.
>
>
>>
>>
>>
>>
>>
>>Why am I writing this? I don't want to bitch about Akka, the Eclipse plugin, etc. I love Akka and I think the typesafe guys did a fantastic job with the Eclipse plugin. And all the problems above can be solved with a little trying and Google.
>>And I know sbt 0.11 is different than older versions but he said "Maven 3 works well with Maven 2 files in Java".
>>
>>Can I blame him?
>
>
>Maven wasn't fundamentally broken. Or, depending on were you stand in relation to maven, Maven 3 didn't fix the fundamental brokenness. :-)
>
>
>Well, that's not quite the case. SBT 0.7 wasn't really fundamentally broken, but to proceed in the direction he wanted, Marc *had* to make very fundamental changes.
I like sbt. I think it is the best build tool out there. Just not all easy for a beginner to get started. But on the other side, Maven was a lot more pain :-)
>
>
>>
>>I think it is too hard for the beginner to get this stuff running. And after about 2 hours he was frustrated. I helped him and he is still trying to learn Scala but at the moment he thinks that "all the Scala stuff is not production ready".
>
>
>Get him off Eclipse and on IntelliJ. I fail to see how can one work in Scala with Eclipse and think Scala is production-ready: the environment is conspiring against him. I'll be very happy when ScalaIDE for Eclipse becomes a rock-solid platform for Scala development, but, up to now, that is not the case at all.
I will try IntelliJ myself but I can't force others and as I wrote, I can understand when people want to stay with Eclipse (or Netbeans). I personally find the Eclipse Scala Plugin pretty cool. When I don't want to interact with sbt, it works pretty well.
So this is why I recommend it to new Scala developers - but I will check out IntelliJ and maybe I will change my mind.
>
>
>>
>>I told him how hard it was to get Maven and Ant projects running when we started Java and the trouble we had when we compiled our first C programs or when we started with Ruby. There is always a ladder to climb once you get started with a new technology.
>>
>>But when you are used to Java stuff working (at least most of the time) and Scala stuff doesn't work as you expected, this can keep programmers away from such a wonderful technology as Scala.
>>
>>
>>This is just one of my observations I had today. Again, I am not bitching about Scala, I love it but I think it could still be simpler to get started for someone coming from Java land.
>
>
>It *can* be much simpler, but, really, you went through the forest instead of taking the nice road that cuts through it.
>
>
>First of all, you built, instead of used, projects -- and some pretty heavy at that. Second, you picked an unstable development environment. Third, you completely avoided the "shortest-distance-path" by not picking Typesafe.
This all worked for me when I first tried it so I thought it might also work for him. I will check out the Typesafe stack.
The good news is that he now has an Eclipse + Scala setup running for him and he still likes plays with Scala. He told me that the collections convinced him that Scala is really cool.
Markus
>
>
>
>
>>
>>Cheers,
>>
>>
>>Markus
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>
Wed, 2011-11-16, 03:47
#20
Re: How I failed to convince another programmer that Scala is g
On Tue, Nov 15, 2011 at 19:24, Markus Jais wrote:
>
> >sbt "eclipse with-sources same-targets"
>
> Thanks, I didn't know that. Just tried that myself. sbt did indeed download lots of stuff. But eclipse still doesn't work with
> the generated .classpath.
Did you try "sbt compile" before running that? In my experience, what works is:
1. Create the project, with all dependencies
2. "Compile" the project (no source code yet, but it will download all
dependencies).
3. Then use sbteclipse to create the source directories and eclipse
configuration.
> >Too little information at this point. Where you running this on the Akka project you downloaded or what?
>
> Yes, we tried it on the Akka source code (master branch). sbt couldn't find that class.
Well, Akka build problems is with the Akka people.
> >>Problem 4:
> >>He ran sbt test on the Akka source: java.lang.OutOfMemoryError
> >
> >
> >Try running test on Eclipse. :-)
>
> I have rarely problems with Eclipse, it is probably years since I had an java.lang.OutOfMemoryError.
> Netbeans ist better in some cases (e.g. for JEE 6).
I mean running the unit tests of the Eclipse project. After all, you
picked a fairly heavy project in Akka, considering all the modules and
dependencies. Akka is meant to be used, not compiled -- just like
Eclipse. :-)
> >Too little information. If you keep the project build on sbt, running "sbt update eclipse" after any change to the project helps, though.
>
> sbt builds fine. but still no useful .classpath.
Strange and unfortunate.
> >>Problem 7:
> >>He played with the Scala collection in Eclipse. He liked it a lot. I showed him the parallel collections and other cool stuff. After a while Eclipse showed errors where there weren't any. Only after several Eclipse restarts did the code work.
> >
> >
> >IntelliJ works, Eclipe doesn't. Sad but true.
>
> I know. But many people use Eclipse - some even like it. It is not as bad as some people say. I also use it for
> Python, Clojure and most of my Java work. The same for the guy I tried to show Scala. IintelliJ may be better
> but not everyone wants to change and IDE, particularly not after having become really productive with Eclipse.
> I can understand when someone says he wants to continue his using his tools if they work really well for him.
I don't *want* him to change IDEs, but we must be honest and admit
up-front that the plugin for Eclipse is not production-ready. It seems
the first release candidate is already out, but I'll reserve my
judgement for when 2.0 is finally released.
Meanwhile, just explain that IntelliJ works reliably with Scala, and
Eclipse doesn't. Eclipse is not being ignored -- much on the contrary
Wed, 2011-11-16, 04:47
#21
Re: How I failed to convince another programmer that Scala is g
> I mean running the unit tests of the Eclipse project. After all, you
> picked a fairly heavy project in Akka, considering all the modules and
> dependencies. Akka is meant to be used, not compiled -- just like
> Eclipse. :-)
Actually it is *very* easy to set up Akka using Eclipse (2.0-RC01) and sbt (0.11.0) and sbteclispe (1.4.0 or 1.5.0-SNAPSHOT):
1. In sbt execute eclipse skip-parents
2. As Akka uses continuations, go to Preferences > Scala > Compiler and on the Standard tab add continuations:enable for "P"
3. In Eclipse use Import > Existing Projects into Workspace, select the Akka root directory, then all sub projects will be selected by default, just click OK
Heiko
Wed, 2011-11-16, 11:27
#22
Re: How I failed to convince another programmer that Scala is go
Eclipse) I get tired of people keep complaining about Eclipse. Eclipse
is literally the easiest IDE I've ever worked with. Eclipse is an
ecosystem as well as being an IDE. The stuff that came from IBM, the
Java and C++ functionality - is rock solid. The crappy PHP plug-in you
downloaded last week is not their responsibility.
1) That is the greatest crock of **** I've ever read. Programming is
about getting stuff done, doing computations, producing information
from data. It is not about figuring out build environments and
multiple incompatible library versions and the current angle of that
magical wheel of reinvention. You don't know the magic incantations to
join our club... Get real, mate!
2) Complexity. Complexity. Complexity.
Wed, 2011-11-16, 11:27
#23
Re: How I failed to convince another programmer that Scala is go
Maven == Not invented here. Maven and Eclipse can be made to work.
It's hard to get it so that Maven doesn't trample the ScalaIDE output
classpath. I say it's hard in so far as I've had to change my settings
about 20 times to get it working right.
Wed, 2011-11-16, 11:37
#24
Re: How I failed to convince another programmer that Scala is g
On Wed, Nov 16, 2011 at 3:39 AM, Daniel Sobral <dcsobral@gmail.com> wrote:
I don't *want* him to change IDEs, but we must be honest and admit up-front that the plugin for Eclipse is not production-ready. It seems
the first release candidate is already out, but I'll reserve my
judgement for when 2.0 is finally released.
It's fine to reserve judgement, just that you didn't. I'm fine with criticism, but I'm much better with concrete, actionable issues. I couldn't find any ticket filed by you on the IDE space, and I'd really like to know what version you were using and what problems you experienced. I'm not saying the Scala IDE for Eclipse is perfect, but I know several people here (myself including) that are using it daily for 'production work'. We're not the only ones.
Bottom line, if you (in general, not you personally) want to help, please file tickets for the issues you see.
cheers,iulian
Meanwhile, just explain that IntelliJ works reliably with Scala, and
Eclipse doesn't. Eclipse is not being ignored -- much on the contrary
Wed, 2011-11-16, 11:47
#25
Re: How I failed to convince another programmer that Scala is go
> Drop the eclipse part and i am interested. After years of Maven quirks I
> refuse to subject myself to it on personal projects. I hear 3.x remedies a
> lot of my issues however (like suddenly upgrading my plugins because an
> automatically configured one has a newer version or not being able to
> dynamically set my parent pom version).
Specify the version, as well as the plug-in you wish to use. Then, it
doesn't auto-update.
Wed, 2011-11-16, 12:37
#26
Re: How I failed to convince another programmer that Scala is g
On Wed, Nov 16, 2011 at 10:26 AM, iulian dragos <jaguarul@gmail.com> wrote:
It's definitely worth filing tickets. The Scala IDE team is responsive and has been doing a great job dealing with various issues that are, as far as I can see, a real pain to debug. It's obviously best if you can file issues with test cases, but if you can't create a test case and have a good description of the problem, they may also be able to fix it. Iulian fixed one such issue that I filed recently (stale errors after changing git branch). Thanks by the way. :)
Best,Ismael
Bottom line, if you (in general, not you personally) want to help, please file tickets for the issues you see.
It's definitely worth filing tickets. The Scala IDE team is responsive and has been doing a great job dealing with various issues that are, as far as I can see, a real pain to debug. It's obviously best if you can file issues with test cases, but if you can't create a test case and have a good description of the problem, they may also be able to fix it. Iulian fixed one such issue that I filed recently (stale errors after changing git branch). Thanks by the way. :)
Best,Ismael
Wed, 2011-11-16, 15:47
#27
Re: How I failed to convince another programmer that Scala is g
On Wed, Nov 16, 2011 at 08:26, iulian dragos wrote:
>
>
> On Wed, Nov 16, 2011 at 3:39 AM, Daniel Sobral wrote:
>>
>> I don't *want* him to change IDEs, but we must be honest and admit
>> up-front that the plugin for Eclipse is not production-ready. It seems
>> the first release candidate is already out, but I'll reserve my
>> judgement for when 2.0 is finally released.
>
>
> It's fine to reserve judgement, just that you didn't. I'm fine with
> criticism, but I'm much better with concrete, actionable issues. I couldn't
> find any ticket filed by you on the IDE space, and I'd really like to know
> what version you were using and what problems you experienced. I'm not
> saying the Scala IDE for Eclipse is perfect, but I know several people here
> (myself including) that are using it daily for 'production work'. We're not
> the only ones.
> Bottom line, if you (in general, not you personally) want to help, please
> file tickets for the issues you see.
It is interesting that if you go to the Scala IDE site you'd be hard
pressed to find a release. Well, there is something that says
"RELEASE", and then it points to "2.0.0-RC01", that is, release
*candidate*.
Why isn't the true last release visible at all? That's what I had to
work with for two years, with stability being "just around the
corner". And just like Markus, I got burned by people who tried to use
that. Opening/closing the project/perspective/eclipse itself was a
standard troubleshooting technique, just like rebooting Windows. So,
for the past year, I just don't use Eclipse with Scala anymore, except
to keep my environment up-to-date.
WHEN 2.0 *release* comes out, I'll give it a shot again, with a clean
slate. That's what I mean by reserving my judgment. Until then, I have
plenty other things to do with my time, and I'm *not* recommending
Eclipse to anyone who's trying out the language, given that a
perfectly workable alternative exists.
If you feel I'm being unfair, see problems 1 and 7 of the original
report. Whether they are easily fixed or not, whether the version of
the Eclipse plugin (not reported beyond "latest") isn't good or not,
whether there were some missing steps or pre-configuration to be done,
all that is irrelevant. The precise amount of problems someone being
introduced to Scala should see is ZERO.
Wed, 2011-11-16, 16:57
#28
Re: How I failed to convince another programmer that Scala is g
> It is interesting that if you go to the Scala IDE site you'd be hard
> pressed to find a release. Well, there is something that says
> "RELEASE", and then it points to "2.0.0-RC01", that is, release
> *candidate*.
True. We haven't yet pushed a final release out, but the Scala IDE plugin for
Eclipse steadily improved in the past year, beta after beta, to the point where
it is now, i.e., a release candidate for a final 2.0 release.
I honestly think the old Eclipse plugin and the new one have diverged a
lot, so it is unfair to judge the current state of the plugin based on the 1.0
version, for which, by the way, no final has ever been released (Miles, please,
correct me if I am wrong).
To have a honest opinion, I think you should check what the current status
is, and then criticize. Going the other way around is simply unfair and a real
turn down for all contributors that are investing time and energy in improving
the current state of the Scala IDE.
For my own curiosity, why a final version would make you any more confident
than an unreleased one? I can understand the argument when it comes from
an enterprise environment, but I was genuinely surprise to hear that from you.
(And I'm not talking about using nighties, but a release candidate should give
you enough confidence, no?).
> Why isn't the true last release visible at all?
As I said above, no true last release exist. The two versions have diverged to
the point that you should consider the 2.0 release as a replacement, and not
as an improvement of the 1.0 plugin. As a matter of fact, many aspects of the
plugin have been re-designed underneath (for instance, the way the presentation
compiler is used by the plugin, or the fact that we use SBT for tracking
dependencies instead of the -- buggy -- refined manager builder. That makes a
*huge* difference).
Sure, without the 1.0 plugin our work would have been tremendously harder
(and it would have taken way more time to get to where we are now), all the
insights that Miles and all contributors had in the Eclipse internals have been
leveraged.
But, I still believe (and I'd like to make clear that this is my own opinion), that
2.0 is simply too different from 1.0 to be compared. That is why we push people
into using 2.0 plugin, instead of using the old and discontinued 1.0 plugin.
> If you feel I'm being unfair, see problems 1 and 7 of the original
> report.
I'm sorry, but they were simply not Eclipse issues. For sure we can write a plugin
on top of the Scala IDE to make integration of SBT (and sbteclipse) smoother, but
that has simply nothing to do with Eclipse and the Scala IDE itself.
> Whether they are easily fixed or not, whether the version of
> the Eclipse plugin (not reported beyond "latest") isn't good or not,
> whether there were some missing steps or pre-configuration to be done,
> all that is irrelevant. The precise amount of problems someone being
> introduced to Scala should see is ZERO.
Sure, but all the blame shouldn't be on the Scala IDE ;)
I think we all agree that we need to improve and make the whole Scala experience
for newcomers as simple and straightforward as possible. This is definitely one of
the main Typesafe's goals with the Typesafe stack, and I'm confident this will go a
long way into lowering the barrier to Scala adoption.
Cheers,
Mirco
Wed, 2011-11-16, 17:37
#29
Re: Re: How I failed to convince another programmer that Scala
On Wed, Nov 16, 2011 at 4:24 AM, bryan hunt <sentimental.bryan@gmail.com> wrote:
This is the behavior I want, which is why I don't use Maven/sbt at all.
> Drop the eclipse part and i am interested. After years of Maven quirks I
> refuse to subject myself to it on personal projects. I hear 3.x remedies a
> lot of my issues however (like suddenly upgrading my plugins because an
> automatically configured one has a newer version or not being able to
> dynamically set my parent pom version).
Specify the version, as well as the plug-in you wish to use. Then, it
doesn't auto-update.
This is the behavior I want, which is why I don't use Maven/sbt at all.
Wed, 2011-11-16, 17:47
#30
Re: How I failed to convince another programmer that Scala is g
On Wed, Nov 16, 2011 at 13:49, Mirco Dotta wrote:
>
> I honestly think the old Eclipse plugin and the new one have diverged a
> lot, so it is unfair to judge the current state of the plugin based on the 1.0
> version, for which, by the way, no final has ever been released (Miles, please,
> correct me if I am wrong).
>
> To have a honest opinion, I think you should check what the current status
> is, and then criticize. Going the other way around is simply unfair and a real
> turn down for all contributors that are investing time and energy in improving
> the current state of the Scala IDE.
Been there, done that for a long while. It might be unfair to the
current IDE, but everyone has a limit, and mine was crossed. I'm
sorry, but I can't do that game anymore. For now, that's a cross
others will have to carry.
> For my own curiosity, why a final version would make you any more confident
> than an unreleased one? I can understand the argument when it comes from
> an enterprise environment, but I was genuinely surprise to hear that from you.
> (And I'm not talking about using nighties, but a release candidate should give
> you enough confidence, no?).
Basically, because I'm trusting you. I *know* Odersky wants a rock
solid Eclipse plugin, and I know he wants this as soon as it can be
accomplished. All the efforts being spent on it is proof enough of
that.
So, if you think the plugin is good enough to be released, then I'll
give it a shot. Conversely, until you *do* demonstrate confidence
enough in the plugin to get it released, I'll bind my time.
And I'll *not* recommend it to anyone until then. When people ask, I
explain that there is an Eclipse plugin, but my present recommendation
for a smooth experience is IntelliJ. When you say ScalaIDE 2.0 is
ready, then I'll try it again.
>> If you feel I'm being unfair, see problems 1 and 7 of the original
>> report.
>
> I'm sorry, but they were simply not Eclipse issues. For sure we can write a plugin
> on top of the Scala IDE to make integration of SBT (and sbteclipse) smoother, but
> that has simply nothing to do with Eclipse and the Scala IDE itself.
"After a while Eclipse showed errors where there weren't any. Only
after several Eclipse restarts did the code work."
That sounds, to me, like a plugin problem. Worse yet, the kind of
problem we had with the previous plugin.
>> Whether they are easily fixed or not, whether the version of
>> the Eclipse plugin (not reported beyond "latest") isn't good or not,
>> whether there were some missing steps or pre-configuration to be done,
>> all that is irrelevant. The precise amount of problems someone being
>> introduced to Scala should see is ZERO.
>
> Sure, but all the blame shouldn't be on the Scala IDE ;)
This is not so much a question of what is to blame. The question is:
which IDE gives the most smooth experience for a user new to Scala
today? This isn't even about what IDE will be most productive, much
less about what will be the future.
Wed, 2011-11-16, 18:47
#31
Re: How I failed to convince another programmer that Scala is g
On Wed, Nov 16, 2011 at 5:39 PM, Daniel Sobral <dcsobral@gmail.com> wrote:
On Wed, Nov 16, 2011 at 13:49, Mirco Dotta <mirco.dotta@gmail.com> wrote:
>
> I honestly think the old Eclipse plugin and the new one have diverged a
> lot, so it is unfair to judge the current state of the plugin based on the 1.0
> version, for which, by the way, no final has ever been released (Miles, please,
> correct me if I am wrong).
>
> To have a honest opinion, I think you should check what the current status
> is, and then criticize. Going the other way around is simply unfair and a real
> turn down for all contributors that are investing time and energy in improving
> the current state of the Scala IDE.
Been there, done that for a long while. It might be unfair to the
current IDE, but everyone has a limit, and mine was crossed. I'm
sorry, but I can't do that game anymore. For now, that's a cross
others will have to carry.
> For my own curiosity, why a final version would make you any more confident
> than an unreleased one? I can understand the argument when it comes from
> an enterprise environment, but I was genuinely surprise to hear that from you.
> (And I'm not talking about using nighties, but a release candidate should give
> you enough confidence, no?).
Basically, because I'm trusting you. I *know* Odersky wants a rock
solid Eclipse plugin, and I know he wants this as soon as it can be
accomplished. All the efforts being spent on it is proof enough of
that.
So, if you think the plugin is good enough to be released, then I'll
give it a shot. Conversely, until you *do* demonstrate confidence
enough in the plugin to get it released, I'll bind my time.
And I'll *not* recommend it to anyone until then. When people ask, I
explain that there is an Eclipse plugin, but my present recommendation
for a smooth experience is IntelliJ. When you say ScalaIDE 2.0 is
ready, then I'll try it again.
All your points are fair. But then you should probably start by saying 'Last time I checked, one year ago...'. Like Mirco, I believe the IDE has improved a lot since then. One year may be a long time in this game.
>> If you feel I'm being unfair, see problems 1 and 7 of the original
>> report.
>
> I'm sorry, but they were simply not Eclipse issues. For sure we can write a plugin
> on top of the Scala IDE to make integration of SBT (and sbteclipse) smoother, but
> that has simply nothing to do with Eclipse and the Scala IDE itself.
"After a while Eclipse showed errors where there weren't any. Only
after several Eclipse restarts did the code work."
That sounds, to me, like a plugin problem. Worse yet, the kind of
problem we had with the previous plugin.
A quick check at issue trackers show that other IDEs suffer from exactly the same problems. It's not an excuse, of course.
cheers,iulian
>> Whether they are easily fixed or not, whether the version of
>> the Eclipse plugin (not reported beyond "latest") isn't good or not,
>> whether there were some missing steps or pre-configuration to be done,
>> all that is irrelevant. The precise amount of problems someone being
>> introduced to Scala should see is ZERO.
>
> Sure, but all the blame shouldn't be on the Scala IDE ;)
This is not so much a question of what is to blame. The question is:
which IDE gives the most smooth experience for a user new to Scala
today? This isn't even about what IDE will be most productive, much
less about what will be the future.
--
Daniel C. Sobral
I travel to the future all the time.
--
« Je déteste la montagne, ça cache le paysage »
Alphonse Allais
Wed, 2011-11-16, 18:57
#32
RE: How I failed to convince another programmer that Scala is g
I did switch to Emacs/ensime in protest (or following a nervous break-down,
I don't remember which came first) at the beginning of the year - but I came
back and I am impressed with the evolution of the plugin in recent months...
I uncrossed my threshold, as it were ?
Give it 2G and it works fine... a bit slower and with less functionality
than the Java plugin, but the basics work fine!
Cheers,
Razie
-----Original Message-----
From: scala-user@googlegroups.com [mailto:scala-user@googlegroups.com] On
Behalf Of Daniel Sobral
Sent: November-16-11 11:40 AM
To: Mirco Dotta
Cc: iulian dragos; Markus Jais; scala-user@googlegroups.com
Subject: Re: [scala-user] How I failed to convince another programmer that
Scala is good - too many problems getting started
On Wed, Nov 16, 2011 at 13:49, Mirco Dotta wrote:
>
> I honestly think the old Eclipse plugin and the new one have diverged
> a lot, so it is unfair to judge the current state of the plugin based
> on the 1.0 version, for which, by the way, no final has ever been
> released (Miles, please, correct me if I am wrong).
>
> To have a honest opinion, I think you should check what the current
> status is, and then criticize. Going the other way around is simply
> unfair and a real turn down for all contributors that are investing
> time and energy in improving the current state of the Scala IDE.
Been there, done that for a long while. It might be unfair to the current
IDE, but everyone has a limit, and mine was crossed. I'm sorry, but I can't
do that game anymore. For now, that's a cross others will have to carry.
> For my own curiosity, why a final version would make you any more
> confident than an unreleased one? I can understand the argument when
> it comes from an enterprise environment, but I was genuinely surprise to
hear that from you.
> (And I'm not talking about using nighties, but a release candidate
> should give you enough confidence, no?).
Basically, because I'm trusting you. I *know* Odersky wants a rock solid
Eclipse plugin, and I know he wants this as soon as it can be accomplished.
All the efforts being spent on it is proof enough of that.
So, if you think the plugin is good enough to be released, then I'll give it
a shot. Conversely, until you *do* demonstrate confidence enough in the
plugin to get it released, I'll bind my time.
And I'll *not* recommend it to anyone until then. When people ask, I explain
that there is an Eclipse plugin, but my present recommendation for a smooth
experience is IntelliJ. When you say ScalaIDE 2.0 is ready, then I'll try it
again.
>> If you feel I'm being unfair, see problems 1 and 7 of the original
>> report.
>
> I'm sorry, but they were simply not Eclipse issues. For sure we can
> write a plugin on top of the Scala IDE to make integration of SBT (and
> sbteclipse) smoother, but that has simply nothing to do with Eclipse and
the Scala IDE itself.
"After a while Eclipse showed errors where there weren't any. Only after
several Eclipse restarts did the code work."
That sounds, to me, like a plugin problem. Worse yet, the kind of problem we
had with the previous plugin.
>> Whether they are easily fixed or not, whether the version of the
>> Eclipse plugin (not reported beyond "latest") isn't good or not,
>> whether there were some missing steps or pre-configuration to be
>> done, all that is irrelevant. The precise amount of problems someone
>> being introduced to Scala should see is ZERO.
>
> Sure, but all the blame shouldn't be on the Scala IDE ;)
This is not so much a question of what is to blame. The question is:
which IDE gives the most smooth experience for a user new to Scala today?
This isn't even about what IDE will be most productive, much less about what
will be the future.
--
Daniel C. Sobral
I travel to the future all the time.
Wed, 2011-11-16, 18:57
#33
Re: How I failed to convince another programmer that Scala is g
I don't know if I'm unusual in this regard, but for me the number one
reason to use idea, more than anything else, is the syntax
highlighting - I find the amount of additional information I get from
the formatting to be so important that I am essentially unable to
function properly in eclipse!
On Wed, Nov 16, 2011 at 12:43 PM, Razvan Cojocaru wrote:
> I did switch to Emacs/ensime in protest (or following a nervous break-down,
> I don't remember which came first) at the beginning of the year - but I came
> back and I am impressed with the evolution of the plugin in recent months...
> I uncrossed my threshold, as it were ?
>
> Give it 2G and it works fine... a bit slower and with less functionality
> than the Java plugin, but the basics work fine!
>
> Cheers,
> Razie
>
> -----Original Message-----
> From: scala-user@googlegroups.com [mailto:scala-user@googlegroups.com] On
> Behalf Of Daniel Sobral
> Sent: November-16-11 11:40 AM
> To: Mirco Dotta
> Cc: iulian dragos; Markus Jais; scala-user@googlegroups.com
> Subject: Re: [scala-user] How I failed to convince another programmer that
> Scala is good - too many problems getting started
>
> On Wed, Nov 16, 2011 at 13:49, Mirco Dotta wrote:
>>
>> I honestly think the old Eclipse plugin and the new one have diverged
>> a lot, so it is unfair to judge the current state of the plugin based
>> on the 1.0 version, for which, by the way, no final has ever been
>> released (Miles, please, correct me if I am wrong).
>>
>> To have a honest opinion, I think you should check what the current
>> status is, and then criticize. Going the other way around is simply
>> unfair and a real turn down for all contributors that are investing
>> time and energy in improving the current state of the Scala IDE.
>
> Been there, done that for a long while. It might be unfair to the current
> IDE, but everyone has a limit, and mine was crossed. I'm sorry, but I can't
> do that game anymore. For now, that's a cross others will have to carry.
>
>> For my own curiosity, why a final version would make you any more
>> confident than an unreleased one? I can understand the argument when
>> it comes from an enterprise environment, but I was genuinely surprise to
> hear that from you.
>> (And I'm not talking about using nighties, but a release candidate
>> should give you enough confidence, no?).
>
> Basically, because I'm trusting you. I *know* Odersky wants a rock solid
> Eclipse plugin, and I know he wants this as soon as it can be accomplished.
> All the efforts being spent on it is proof enough of that.
>
> So, if you think the plugin is good enough to be released, then I'll give it
> a shot. Conversely, until you *do* demonstrate confidence enough in the
> plugin to get it released, I'll bind my time.
>
> And I'll *not* recommend it to anyone until then. When people ask, I explain
> that there is an Eclipse plugin, but my present recommendation for a smooth
> experience is IntelliJ. When you say ScalaIDE 2.0 is ready, then I'll try it
> again.
>
>>> If you feel I'm being unfair, see problems 1 and 7 of the original
>>> report.
>>
>> I'm sorry, but they were simply not Eclipse issues. For sure we can
>> write a plugin on top of the Scala IDE to make integration of SBT (and
>> sbteclipse) smoother, but that has simply nothing to do with Eclipse and
> the Scala IDE itself.
>
> "After a while Eclipse showed errors where there weren't any. Only after
> several Eclipse restarts did the code work."
>
> That sounds, to me, like a plugin problem. Worse yet, the kind of problem we
> had with the previous plugin.
>
>>> Whether they are easily fixed or not, whether the version of the
>>> Eclipse plugin (not reported beyond "latest") isn't good or not,
>>> whether there were some missing steps or pre-configuration to be
>>> done, all that is irrelevant. The precise amount of problems someone
>>> being introduced to Scala should see is ZERO.
>>
>> Sure, but all the blame shouldn't be on the Scala IDE ;)
>
> This is not so much a question of what is to blame. The question is:
> which IDE gives the most smooth experience for a user new to Scala today?
> This isn't even about what IDE will be most productive, much less about what
> will be the future.
>
> --
> Daniel C. Sobral
>
> I travel to the future all the time.
>
>
Wed, 2011-11-16, 19:07
#34
Re: How I failed to convince another programmer that Scala is g
Hi,
Daniel Sobral wrote:
> Been there, done that for a long while. It might be unfair to the
> current IDE, but everyone has a limit, and mine was crossed. I'm
> sorry, but I can't do that game anymore. For now, that's a cross
> others will have to carry.
(snip)
> And I'll *not* recommend it to anyone until then. When people ask, I
> explain that there is an Eclipse plugin, but my present recommendation
> for a smooth experience is IntelliJ. When you say ScalaIDE 2.0 is
> ready, then I'll try it again.
without any offense meant, I find that kind of funny, because I have the exact same story to tell, just vice versa. The IntelliJ plug-in has been recommended and advertised to me over and over and again and again, and each time I tried it (over 1 and a half year, in 2-3 month periods distance), I could not get the most basic workflows to work (and I spent hours configuring, reading docs and asking people). In contrast, the Eclipse plug-in only gave me comparatively minor problems and all the tickets I reported got fixed in a timely manner, too. Thus, even though I acknowledge that things may have changed, I'm not going to "carry the cross" of IntelliJ anymore, while I'm pretty happy with 2.0-rc1 for Eclipse.
Cheers
Andreas
Wed, 2011-11-16, 19:47
#35
Re: How I failed to convince another programmer that Scala is g
Am 16.11.2011 18:54, schrieb Alec Zorab:
> I don't know if I'm unusual in this regard, but for me the number one
> reason to use idea, more than anything else, is the syntax
> highlighting - I find the amount of additional information I get from
> the formatting to be so important that I am essentially unable to
> function properly in eclipse!
+1
>
> On Wed, Nov 16, 2011 at 12:43 PM, Razvan Cojocaru wrote:
>> I did switch to Emacs/ensime in protest (or following a nervous break-down,
>> I don't remember which came first) at the beginning of the year - but I came
>> back and I am impressed with the evolution of the plugin in recent months...
>> I uncrossed my threshold, as it were ?
>>
>> Give it 2G and it works fine... a bit slower and with less functionality
>> than the Java plugin, but the basics work fine!
>>
>> Cheers,
>> Razie
>>
>> -----Original Message-----
>> From: scala-user@googlegroups.com [mailto:scala-user@googlegroups.com] On
>> Behalf Of Daniel Sobral
>> Sent: November-16-11 11:40 AM
>> To: Mirco Dotta
>> Cc: iulian dragos; Markus Jais; scala-user@googlegroups.com
>> Subject: Re: [scala-user] How I failed to convince another programmer that
>> Scala is good - too many problems getting started
>>
>> On Wed, Nov 16, 2011 at 13:49, Mirco Dotta wrote:
>>> I honestly think the old Eclipse plugin and the new one have diverged
>>> a lot, so it is unfair to judge the current state of the plugin based
>>> on the 1.0 version, for which, by the way, no final has ever been
>>> released (Miles, please, correct me if I am wrong).
>>>
>>> To have a honest opinion, I think you should check what the current
>>> status is, and then criticize. Going the other way around is simply
>>> unfair and a real turn down for all contributors that are investing
>>> time and energy in improving the current state of the Scala IDE.
>> Been there, done that for a long while. It might be unfair to the current
>> IDE, but everyone has a limit, and mine was crossed. I'm sorry, but I can't
>> do that game anymore. For now, that's a cross others will have to carry.
>>
>>> For my own curiosity, why a final version would make you any more
>>> confident than an unreleased one? I can understand the argument when
>>> it comes from an enterprise environment, but I was genuinely surprise to
>> hear that from you.
>>> (And I'm not talking about using nighties, but a release candidate
>>> should give you enough confidence, no?).
>> Basically, because I'm trusting you. I *know* Odersky wants a rock solid
>> Eclipse plugin, and I know he wants this as soon as it can be accomplished.
>> All the efforts being spent on it is proof enough of that.
>>
>> So, if you think the plugin is good enough to be released, then I'll give it
>> a shot. Conversely, until you *do* demonstrate confidence enough in the
>> plugin to get it released, I'll bind my time.
>>
>> And I'll *not* recommend it to anyone until then. When people ask, I explain
>> that there is an Eclipse plugin, but my present recommendation for a smooth
>> experience is IntelliJ. When you say ScalaIDE 2.0 is ready, then I'll try it
>> again.
>>
>>>> If you feel I'm being unfair, see problems 1 and 7 of the original
>>>> report.
>>> I'm sorry, but they were simply not Eclipse issues. For sure we can
>>> write a plugin on top of the Scala IDE to make integration of SBT (and
>>> sbteclipse) smoother, but that has simply nothing to do with Eclipse and
>> the Scala IDE itself.
>>
>> "After a while Eclipse showed errors where there weren't any. Only after
>> several Eclipse restarts did the code work."
>>
>> That sounds, to me, like a plugin problem. Worse yet, the kind of problem we
>> had with the previous plugin.
>>
>>>> Whether they are easily fixed or not, whether the version of the
>>>> Eclipse plugin (not reported beyond "latest") isn't good or not,
>>>> whether there were some missing steps or pre-configuration to be
>>>> done, all that is irrelevant. The precise amount of problems someone
>>>> being introduced to Scala should see is ZERO.
>>> Sure, but all the blame shouldn't be on the Scala IDE ;)
>> This is not so much a question of what is to blame. The question is:
>> which IDE gives the most smooth experience for a user new to Scala today?
>> This isn't even about what IDE will be most productive, much less about what
>> will be the future.
>>
>> --
>> Daniel C. Sobral
>>
>> I travel to the future all the time.
>>
>>
Wed, 2011-11-16, 20:17
#36
Re: How I failed to convince another programmer that Scala is go
1) That is the greatest crock of **** I've ever read. Programming isI really hope that's intended tongue-in-cheek. If not, I'd sure like to know what perfect-no-hassles-works-every-time IDE/build environment you've been using--because they don't work every time, mate, and when they don't you've got a huge productivity and time sink.
about getting stuff done, doing computations, producing information
from data. It is not about figuring out build environments and
multiple incompatible library versions and the current angle of that
magical wheel of reinvention. You don't know the magic incantations to
join our club... Get real, mate!
Ken
Wed, 2011-11-16, 20:47
#37
Re: How I failed to convince another programmer that Scala is g
Just curious about your problem #6. If you want to show **how to use** Finagle, why would you show *how to build* it?
All you need is the following in your build.sbt (and you can use sbt 0.10 or 0.11 for your projects. You don't need old version of SBT in order to *use* Finagle in your project).
scalaVersion := "2.8.1"
resolvers += "twitter-repo" at "http://maven.twttr.com"
libraryDependencies ++= Seq( "com.twitter" % "finagle-core" % "1.9.6", "com.twitter" % "finagle-http" % "1.9.6")
Thanks,-Durgesh
On Nov 14, 2011, at 4:18 PM, Markus Jais wrote:
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
All you need is the following in your build.sbt (and you can use sbt 0.10 or 0.11 for your projects. You don't need old version of SBT in order to *use* Finagle in your project).
scalaVersion := "2.8.1"
resolvers += "twitter-repo" at "http://maven.twttr.com"
libraryDependencies ++= Seq( "com.twitter" % "finagle-core" % "1.9.6", "com.twitter" % "finagle-http" % "1.9.6")
Thanks,-Durgesh
On Nov 14, 2011, at 4:18 PM, Markus Jais wrote:
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Wed, 2011-11-16, 22:47
#38
RE: How I failed to convince another programmer that Scala is g
Typesafe is a company. Typesafe is allocating resources to improve a product. If someone tells you that the product is not up to scratch wrt the competition, that is useful information to you, much as it may not be a message you would like to hear. Try using the competition for a bit to find out why he thinks this, rather than expecting to be told everything that is wrong with your product in detail.
Of course, if you were contributing to Eclipse for free, out of the goodness of your heart, then you can ignore the last paragraph.
Chris
From: jaguarul@gmail.com
It's fine to reserve judgement, just that you didn't. I'm fine with criticism, but I'm much better with concrete, actionable issues. I couldn't find any ticket filed by you on the IDE space, and I'd really like to know what version you were using and what problems you experienced. I'm not saying the Scala IDE for Eclipse is perfect, but I know several people here (myself including) that are using it daily for 'production work'. We're not the only ones.
Bottom line, if you (in general, not you personally) want to help, please file tickets for the issues you see.
cheers,iulian
Of course, if you were contributing to Eclipse for free, out of the goodness of your heart, then you can ignore the last paragraph.
Chris
From: jaguarul@gmail.com
It's fine to reserve judgement, just that you didn't. I'm fine with criticism, but I'm much better with concrete, actionable issues. I couldn't find any ticket filed by you on the IDE space, and I'd really like to know what version you were using and what problems you experienced. I'm not saying the Scala IDE for Eclipse is perfect, but I know several people here (myself including) that are using it daily for 'production work'. We're not the only ones.
Bottom line, if you (in general, not you personally) want to help, please file tickets for the issues you see.
cheers,iulian
Wed, 2011-11-16, 23:17
#39
Re: How I failed to convince another programmer that Scala is g
On Wed, Nov 16, 2011 at 10:31 PM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
Typesafe is a company. Typesafe is allocating resources to improve a product. If someone tells you that the product is not up to scratch wrt the competition, that is useful information to you, much as it may not be a message you would like to hear. Try using the competition for a bit to find out why he thinks this, rather than expecting to be told everything that is wrong with your product in detail.
You misunderstood my message. As it turns out, Daniel was referring to a rather old version of Eclipse. I'm glad it's clarified now.
Of course, if you were contributing to Eclipse for free, out of the goodness of your heart, then you can ignore the last paragraph.
:-)
cheers,iulian
Chris
From: jaguarul@gmail.com
It's fine to reserve judgement, just that you didn't. I'm fine with criticism, but I'm much better with concrete, actionable issues. I couldn't find any ticket filed by you on the IDE space, and I'd really like to know what version you were using and what problems you experienced. I'm not saying the Scala IDE for Eclipse is perfect, but I know several people here (myself including) that are using it daily for 'production work'. We're not the only ones.
Bottom line, if you (in general, not you personally) want to help, please file tickets for the issues you see.
cheers,iulian
--
« Je déteste la montagne, ça cache le paysage »
Alphonse Allais
Thu, 2011-11-17, 02:47
#40
Re: How I failed to convince another programmer that Scala is g
On Wednesday, November 16, 2011 3:31:10 PM UTC-6, Chris Marshall wrote:
Typesafe is a company. Typesafe is allocating resources to improve a product. If someone tells you that the product is not up to scratch wrt the competition, that is useful information to you, much as it may not be a message you would like to hear. Try using the competition for a bit to find out why he thinks this, rather than expecting to be told everything that is wrong with your product in detail.I'd like to also point out the the _perception_ of Scala on Eclipse may be a problem. I've tried using Scala/Eclipse at least, lessee...5 times in the past. Maybe more. In each case I was assured that major progress had been made. In each case, I spent a _lot_ of time, only to conclude that Scala/Eclipse not only did not meet promises, but was in fact totally unusable for even _simple_ work. IDEA for me has always been the "right" alternative.
Of course, if you were contributing to Eclipse for free, out of the goodness of your heart, then you can ignore the last paragraph.
Chris
The point of this is, I'm no longer willing to try using Scala/Eclipse, because I no longer believe the claims for it. I would love to see worthwhile evaluations of Scala/Eclipse from the people who are doing it, including summaries of what made past versions unusable. This would mean to me that Scala/Eclipse was being more forthcoming about its shortcomings.
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
I know there will be those who say "it worked fine for me". Just read the archives for those for who it didn't.
Ken
Thu, 2011-11-17, 07:27
#41
Re: How I failed to convince another programmer that Scala is g
Hi,
Ken McDonald wrote:
> I'd like to also point out the the _perception_ of Scala on Eclipse may be a
> problem. I've tried using Scala/Eclipse at least, lessee...5 times in the
> past. Maybe more. In each case I was assured that major progress had been
> made. In each case, I spent a _lot_ of time, only to conclude that
> Scala/Eclipse not only did not meet promises, but was in fact totally unusable
> for even _simple_ work. IDEA for me has always been the "right" alternative.
> The point of this is, I'm no longer willing to try using Scala/Eclipse,
> because I no longer believe the claims for it. I would love to see worthwhile
> evaluations of Scala/Eclipse from the people who are doing it, including
> summaries of what made past versions unusable. This would mean to me that
> Scala/Eclipse was being more forthcoming about its shortcomings.
>
I find it interesting that the situation for me was just the other way round.
Eclipse ran (and runs) fairly stable (in contrast to what others reported I
never had to do excessive restarts e.g.) whereas I found IDEA (almost) unusable.
My (personal and subjective) key points are: incremental build in the
background, correct error highlighting in the editor, a (sortable and
filterable) list of compile errors and warnings (and a clear distinction between
those) from which I can navigate to the source file, correct "rename" and "move"
refactorings (at least) project-wide, standard OS shortcuts for window
handling.
Whereas sophisticated semantic highlighting and an outline only help me very
little.
The Eclipse plug-in had its fair share of problems, especially in the "correct
error highlighting" and "rename/move refactoring" parts, but was (each time I
compared) considerably better than what I saw in IDEA, and it keeps improving
constantly.
>
> I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in
> that past has been truly unusable...
>
I respect your opinion (even though I can't accept it as "simple truth"), as it
truly makes me wonder why my experience has been so different, but I guess
different people expect different things from their IDE, and that makes the
whole issue anything but simple.
Kind regards
Andreas
P.S. It's sad that "discussions" about tools get so heated/emotional all the
time.
Thu, 2011-11-17, 07:57
#42
Re: How I failed to convince another programmer that Scala is go
To add to this. I had both experiences.
I started with eclipse about two years ago. I already used Eclipse at
work for Java, so it made sense to try it. Gave up after a few weeks
because of the sudden crashes, slow performance and nearly no
completion.
Switched to IDEA and stuck with it in spite of the erroneous error
reporting, slow performance on older laptop and although I had to
remember two sets of shortcuts it was worth it. The syntax
highlighting, navigation and the fact I could generate the project
definition from a working sbt project was enhancing my productivity.
Then a few months ago I decided to check the new eclipse plug in and
was impressed by the performance specially on laptops where I had 2G
or 3G of RAM and much improved stability and completion. So I switched
back again, because at this point it's good enough that I don't have
to run two IDEs. No more F3 and CTRL-B confusion!
Overall, I think the Eclipse plug in has progressed a lot in the past
year. Is the support for Scala as good as the one for Java. No, there
is still a long way to go. I will still recommend either IDEA Scala
plugin or Eclipse Scala plugin (with a few disclaimers).
My advice for the Eclipse plugin would be to focus on accurate error
highlighting, stable navigation, stable completion as well as tracking
down those odd errors that only show up for certain configurations
over new feature development. I think people are more forgiving of
lack of features than having lots of features working 95% of the
time.
--Jean-Laurent
On Nov 16, 9:53 am, Andreas Flierl wrote:
> without any offense meant, I find that kind of funny, because I have the exact same story to tell, just vice versa. The IntelliJ plug-in has been recommended and advertised to me over and over and again and again, and each time I tried it (over 1 and a half year, in 2-3 month periods distance), I could not get the most basic workflows to work (and I spent hours configuring, reading docs and asking people). In contrast, the Eclipse plug-in only gave me comparatively minor problems and all the tickets I reported got fixed in a timely manner, too. Thus, even though I acknowledge that things may have changed, I'm not going to "carry the cross" of IntelliJ anymore, while I'm pretty happy with 2.0-rc1 for Eclipse.
>
> Cheers
> Andreas
Thu, 2011-11-17, 08:17
#43
Re: Re: How I failed to convince another programmer that Scala
> Overall, I think the Eclipse plug in has progressed a lot in the past
> year. Is the support for Scala as good as the one for Java. No, there
> is still a long way to go. I will still recommend either IDEA Scala
> plugin or Eclipse Scala plugin (with a few disclaimers).
I agree. None is trying to sell the Eclipse Scala IDE as the best IDE for
Scala out there. It has its own issues (just like all others IDEs for Scala)
and in the end it is a personal judgment call to decide whether IDEA,
Eclipse or Netbeans is the best platform for doing Scala development.
If you compare any of the current IDEs with what we have in Java, then
there is simply no point discussing this further, we still need to work to
get there.
What we were trying to point out is simply that using the Scala IDE 1.0
plugin is a bad idea, as the current (in progress) 2.0 has improved in
so many ways.
> My advice for the Eclipse plugin would be to focus on accurate error
> highlighting, stable navigation, stable completion as well as tracking
> down those odd errors that only show up for certain configurations
> over new feature development. I think people are more forgiving of
> lack of features than having lots of features working 95% of the
> time.
That is exactly the focus of 2.0.0 (plus good Java/Scala interoperability).
We have been working on this for many months and since ScalaDays
(June) we have only been fixing defects (no features at all!). It is still not
100% bullet-proof, but we think that at the moment it works relatively well
to make people productive with Eclipse.
Thu, 2011-11-17, 09:17
#44
Re: How I failed to convince another programmer that Scala is g
He was interested in Finagle because he knows netty and he said lets check it out and build it. So we tried.
We gut Finagle running without problems. As far as I can see, it's an awesome framework.
Markus
We gut Finagle running without problems. As far as I can see, it's an awesome framework.
Markus
Von: Durgesh Mankekar <durgesh@gmail.com>
An: Markus Jais <markus.jais@yahoo.de>
Cc: scala-user <scala-user@googlegroups.com>
Gesendet: 20:40 Mittwoch, 16.November 2011
Betreff: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
Just curious about your problem #6. If you want to show **how to use** Finagle, why would you show *how to build* it?
All you need is the following in your build.sbt (and you can use sbt 0.10 or 0.11 for your projects. You don't need old version of SBT in order to *use* Finagle in your project).
scalaVersion := "2.8.1"
resolvers += "twitter-repo" at "http://maven.twttr.com"
libraryDependencies ++= Seq( "com.twitter" % "finagle-core" % "1.9.6", "com.twitter" % "finagle-http" % "1.9.6")
Thanks,-Durgesh
On Nov 14, 2011, at 4:18 PM, Markus Jais wrote:
Problem 6:I wanted to show him Finagle. We checked it out and tried to build it. It failed because it still needs sbt 0.7.5 or somthing similar. He asked "Why doesn't the latest version of sbt work as well?"
Thu, 2011-11-17, 11:27
#45
RE: How I failed to convince another programmer that Scala is g
+1
The only people whose "job" it is to carefully compile a list of everything that is wrong with the scala Eclipse plugin are those who are being paid to make it a better product.
I never really understood this obsession with eating your own dogfood. It's important, of course, but sometimes you need to eat other people's dogfood, so that you can find out what yours is up against.
Chris
Date: Wed, 16 Nov 2011 17:26:35 -0800
From: ykkenmcd@gmail.com
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
Ken
The only people whose "job" it is to carefully compile a list of everything that is wrong with the scala Eclipse plugin are those who are being paid to make it a better product.
I never really understood this obsession with eating your own dogfood. It's important, of course, but sometimes you need to eat other people's dogfood, so that you can find out what yours is up against.
Chris
Date: Wed, 16 Nov 2011 17:26:35 -0800
From: ykkenmcd@gmail.com
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
Ken
Thu, 2011-11-17, 11:37
#46
RE: How I failed to convince another programmer that Scala is g
This statement is amazing.
I use the latest EAP builds of IDEA *because I have been using IDEA for 6 years and have built up a lot of trust in their capability*. If I were to use anything, *ANYTHING* for the first time, be it an IDE, a library, a piece of software or whatever, and all they had were release candidates and beta versions, I would simply walk away. Here's a few things:
1. It's unstable almost by definition (if it isn't, why don't they release it?)2. The design may substantially change between now and the release, potentially resulting in more work for me
Chris
> Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
> From: mirco.dotta@gmail.com
> For my own curiosity, why a final version would make you any more confident
> than an unreleased one?
I use the latest EAP builds of IDEA *because I have been using IDEA for 6 years and have built up a lot of trust in their capability*. If I were to use anything, *ANYTHING* for the first time, be it an IDE, a library, a piece of software or whatever, and all they had were release candidates and beta versions, I would simply walk away. Here's a few things:
1. It's unstable almost by definition (if it isn't, why don't they release it?)2. The design may substantially change between now and the release, potentially resulting in more work for me
Chris
> Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
> From: mirco.dotta@gmail.com
> For my own curiosity, why a final version would make you any more confident
> than an unreleased one?
Thu, 2011-11-17, 12:07
#47
Re: How I failed to convince another programmer that Scala is g
On Thu, Nov 17, 2011 at 11:24 AM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
+1
The only people whose "job" it is to carefully compile a list of everything that is wrong with the scala Eclipse plugin are those who are being paid to make it a better product.
Umm, are you serious? If people do not report problems, the people who are paid to fix it cannot fix it (simply because they are not psychic). If you don't open a ticket, you're a part of the problem, not the solution.
I never really understood this obsession with eating your own dogfood. It's important, of course, but sometimes you need to eat other people's dogfood, so that you can find out what yours is up against.
Chris
Date: Wed, 16 Nov 2011 17:26:35 -0800
From: ykkenmcd@gmail.com
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
Ken
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
Thu, 2011-11-17, 12:37
#48
RE: How I failed to convince another programmer that Scala is g
Yes, I am serious. If you have a product which doesn't work, your customer uses it, doesn't like it, walks away and uses a competitor instead, not telling you about the fault then who has the problem? You or your customer?
It's difficult enough to get internal customers of software *which is being built specifically for them* to report bugs. I cannot imagine how many bugs go unreported in software in general.
It's quite clear that you are in a minority - this thread is full of people who said "I used X, found it rubbish and walked away" whether X was Eclipse or IDEA. None of them have mentioned filing bug reports.
Chris
Date: Thu, 17 Nov 2011 12:05:45 +0100
Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
From: viktor.klang@gmail.com
To: oxbow_lakes@hotmail.com
CC: scala-user@googlegroups.com
On Thu, Nov 17, 2011 at 11:24 AM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
Umm, are you serious? If people do not report problems, the people who are paid to fix it cannot fix it (simply because they are not psychic). If you don't open a ticket, you're a part of the problem, not the solution.
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
It's difficult enough to get internal customers of software *which is being built specifically for them* to report bugs. I cannot imagine how many bugs go unreported in software in general.
It's quite clear that you are in a minority - this thread is full of people who said "I used X, found it rubbish and walked away" whether X was Eclipse or IDEA. None of them have mentioned filing bug reports.
Chris
Date: Thu, 17 Nov 2011 12:05:45 +0100
Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
From: viktor.klang@gmail.com
To: oxbow_lakes@hotmail.com
CC: scala-user@googlegroups.com
On Thu, Nov 17, 2011 at 11:24 AM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
+1
The only people whose "job" it is to carefully compile a list of everything that is wrong with the scala Eclipse plugin are those who are being paid to make it a better product.
Umm, are you serious? If people do not report problems, the people who are paid to fix it cannot fix it (simply because they are not psychic). If you don't open a ticket, you're a part of the problem, not the solution.
I never really understood this obsession with eating your own dogfood. It's important, of course, but sometimes you need to eat other people's dogfood, so that you can find out what yours is up against.
Chris
Date: Wed, 16 Nov 2011 17:26:35 -0800
From: ykkenmcd@gmail.com
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
Ken
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
Thu, 2011-11-17, 13:07
#49
Re: How I failed to convince another programmer that Scala is g
Now, this is only my highly personal take on things but here we go,
On Thu, Nov 17, 2011 at 12:35 PM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
The word "customer" implies that (s)he paid for the "product". I think we can all agree on that there is a widely accepted practice in OSS report problems, as both a good gesture and as a means to get things improved.
Absolutely, but I don't know if it's just me, but if I find a bug in a piece of OSS software, and don't report it, I don't expect it to get fixed either.
It's OK, I'm pretty used to be in a minority.
As a personal anecdote, when I started using Scala back in 2007 there was essentially _no_ IDE support, and during the years I've always switched between what was best at the moment, be it TextMate, Eclipse, Netbeans or IDEA, but I have, as I said, never expected any of them to improve without any feedback from the users.
Cheers,
√
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
On Thu, Nov 17, 2011 at 12:35 PM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
Yes, I am serious. If you have a product which doesn't work, your customer uses it, doesn't like it, walks away and uses a competitor instead, not telling you about the fault then who has the problem? You or your customer?
The word "customer" implies that (s)he paid for the "product". I think we can all agree on that there is a widely accepted practice in OSS report problems, as both a good gesture and as a means to get things improved.
It's difficult enough to get internal customers of software *which is being built specifically for them* to report bugs. I cannot imagine how many bugs go unreported in software in general.
Absolutely, but I don't know if it's just me, but if I find a bug in a piece of OSS software, and don't report it, I don't expect it to get fixed either.
It's quite clear that you are in a minority - this thread is full of people who said "I used X, found it rubbish and walked away" whether X was Eclipse or IDEA.
It's OK, I'm pretty used to be in a minority.
None of them have mentioned filing bug reports.
As a personal anecdote, when I started using Scala back in 2007 there was essentially _no_ IDE support, and during the years I've always switched between what was best at the moment, be it TextMate, Eclipse, Netbeans or IDEA, but I have, as I said, never expected any of them to improve without any feedback from the users.
Cheers,
√
Chris
Date: Thu, 17 Nov 2011 12:05:45 +0100
Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
From: viktor.klang@gmail.com
To: oxbow_lakes@hotmail.com
CC: scala-user@googlegroups.com
On Thu, Nov 17, 2011 at 11:24 AM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
+1
The only people whose "job" it is to carefully compile a list of everything that is wrong with the scala Eclipse plugin are those who are being paid to make it a better product.
Umm, are you serious? If people do not report problems, the people who are paid to fix it cannot fix it (simply because they are not psychic). If you don't open a ticket, you're a part of the problem, not the solution.
I never really understood this obsession with eating your own dogfood. It's important, of course, but sometimes you need to eat other people's dogfood, so that you can find out what yours is up against.
Chris
Date: Wed, 16 Nov 2011 17:26:35 -0800
From: ykkenmcd@gmail.com
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
Ken
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
Thu, 2011-11-17, 13:27
#50
Re: How I failed to convince another programmer that Scala is g
On 17 November 2011 11:35, Chris Marshall <oxbow_lakes@hotmail.com> wrote:
It's quite clear that you are in a minority - this thread is full of people who said "I used X, found it rubbish and walked away" whether X was Eclipse or IDEA. None of them have mentioned filing bug reports.
Personally, I only bother filing bug reports for software that I have an investment in. It takes time and effort to file the report, and then more time and effort when I'm asked for clarification or follow-up. It's easier to try multiple products until one works than it is to make the first one work if it fails, particularly if you're not sure if you're going to use what it provides in any case.
I understand your frustration about not getting feedback. I've been there. In the past we've resorted to running hour-long introductory sessions with 2+ staff members - one runs through some exercises to get users using the sw, and the other compiles a snag list, regardless of the things that come up are known issues, new, or mismatches between the user's expectation and what the program delivers. It's important that these issues are ranked by how they impact the user experience, and not by developer-oriented metrics like how easy to fix they are or if they contradict some internal project lore. Devs will say, "that's a known issue - it's in the bugtracker as #2134" and consider the feedback processed. But, the issue isn't dealt with until you've run a session and nobody has reported that issue. Things that come up in multiple user sessions get red-flagged as embarrassments.
On two academic projects I've worked on, we've had full-time posts for 'outreach', people who sit down with potential users and collect info about their use experience, as well as developing intro documentation and making it as easy as possible to download/install known working combinations and ensure that only 'power users' can download unsupported or deprecated builds.
Not much else I can suggest. First impressions matter disproportionately. So, you need to put disproportionate effort into making first impressions good. How long did apple spend optimizing boot times for the Apple ][?
Matthew
Chris
Date: Thu, 17 Nov 2011 12:05:45 +0100
Subject: Re: [scala-user] How I failed to convince another programmer that Scala is good - too many problems getting started
From: viktor.klang@gmail.com
To: oxbow_lakes@hotmail.com
CC: scala-user@googlegroups.com
On Thu, Nov 17, 2011 at 11:24 AM, Chris Marshall <oxbow_lakes@hotmail.com> wrote:+1
The only people whose "job" it is to carefully compile a list of everything that is wrong with the scala Eclipse plugin are those who are being paid to make it a better product.
Umm, are you serious? If people do not report problems, the people who are paid to fix it cannot fix it (simply because they are not psychic). If you don't open a ticket, you're a part of the problem, not the solution.
I never really understood this obsession with eating your own dogfood. It's important, of course, but sometimes you need to eat other people's dogfood, so that you can find out what yours is up against.
Chris
Date: Wed, 16 Nov 2011 17:26:35 -0800
From: ykkenmcd@gmail.com
I'm sorry to sound so harsh, but the simple truth is that Scala/Eclipse in that past has been truly unusable...from freezes almost immediately on entry to completely improper error highlighting to infinite loops. As far as I'm concerned, you need to earn my trust (and time) back, not the other way 'round.
Ken
--
Viktor Klang
Akka Tech LeadTypesafe - Enterprise-Grade Scala from the Experts
Twitter: @viktorklang
--
Dr Matthew PocockIntegrative Bioinformatics Group, School of Computing Science, Newcastle Universitymailto: turingatemyhamster@gmail.com gchat: turingatemyhamster@gmail.commsn: matthew_pocock@yahoo.co.uk irc.freenode.net: drdozerskype: matthew.pococktel: (0191) 2566550mob: +447535664143
Am 14.11.2011 22:18, schrieb Markus Jais:
never had such problems. (never used sbt or eclipse)