Scala Incubator wiki page on Scala XML Conformance

This might be a good point at which to ask whether we have to do interpolation exactly the way it works at present. For example,

seems to me to have the advantage over

of being well-formed XML, and of looking more like XSL-type interpolation, and the bloat of two characters per attribute interpolation probably wouldn't create a RAM supply crisis. (The conciseness argument seems stronger to me in the case of

{myValue]

rather than

and the first example is at least well-formed XML, although I'm not sure whether it corresponds to xsl:value-of or xsl:copy-of or both.)

More generally, if we are considering implementing XPath, and assuming that not everyone will want to use XPath, do we want to consider ways of having different interpolation vocabularies? Could we, for example, have

as well as

or could we provide libraries to overload the {} ?

-----Original Message-----
From: "Anthony B. Coates (Londata)"
Sent: Saturday, January 2, 2010 3:47pm
To: "scala-xml List"
Subject: [scala-xml] Scala Incubator wiki page on Scala XML Conformance

I have added a new "scala-xml" page to the Scala Incubator wiki, this one
discussing how best to approach XML conformance in Scala:

http://wiki.github.com/scala-incubator/scala-xml/scala-xml-conformance

I would be very interested to have your comments on whether I've described
the current situation with Scala XML support correctly, and whether my
proposal seems workable or not. Thanks a lot in advance,

Cheers, Tony.

Lots of good thoughts here. Comments below. Thanks, Cheers, Tony.

On Sat, 02 Jan 2010 19:22:27 -0000, Mark Howe wrote:

> This might be a good point at which to ask whether we have to do
> interpolation exactly the way it works at present. For example,
>
>
>
> seems to me to have the advantage over
>
>
> of being well-formed XML, and of looking more like XSL-type
> interpolation, and the bloat of two characters per attribute
> interpolation probably wouldn't create a RAM supply crisis. (The
> conciseness argument seems stronger to me in the case of
>
> {myValue]
>
> rather than
>
>
>
> and the first example is at least well-formed XML, although I'm not sure
> whether it corresponds to xsl:value-of or xsl:copy-of or both.)

I agree that would be more natural to XSLT users
than is. What I hadn't had time to look into, as yet,
was whether this is a limitation of the Scala approach of defining the XML
markup characters to be Scala operators. Not that I personally consider
it a big limitation, and it would need to continue to be supported for a
while for the usual legacy reasons, but a more XPath-like and XSLT-like
syntax would definitely lower the learning curve for people who are
already processing XML, just not yet with Scala.

> More generally, if we are considering implementing XPath, and assuming
> that not everyone will want to use XPath, do we want to consider ways of
> having different interpolation vocabularies? Could we, for example, have
>
>
>
> as well as
>
>
>
> or could we provide libraries to overload the {} ?

I would certainly like to experiment whether we could use standard XPath
syntax, i.e. "/" rather than "\", etc. Once again, the older syntax would
need to be supported for a while, but convergence with XPath syntax would
be useful. I also have an interest in whether the same path sequence
could be applied to non-XML structures, e.g. JSON, as that would be a
definite bonus for some uses.

Thanks, Cheers, Tony.