3 applications of "indexed composition" as a language design principle

All,

At this blog entry i describe an idea for the auto-generation of some pretty cool extensions to for/do comprehensions. i have built some sample applications in Scala+Lift to explore the ideas on 3 non-trivial examples: relations, graphs and processes. There are goodies -- proposals and ideas around which to have discussion and debate -- at every level.

• language design
  
• is indexed composition a good language feature
• is a better feature when restricted to DSL generation
  
• domain specific ideas that fall out of this analysis
• would SQL-like languages benefit from a richer 'notion of collection' if the corresponding connectives could be generated
• would SQL-like languages benefit from a lazier evaluation strategy
• would a compositional graph description language be better suited to various applications than, say GraphML
• would a query on graphs as a basic feature of the language be a useful feature
• would a process language that enables the use of codes of processes as channels and data be pragmatically useful
• would a process language that enables search on processes on the basis of structure and behavior be useful
  
• application generation possibilities
• essentially i am hand building Lift wrappers around parsers i'm auto-generating from BNFC -- i'm pretty sure the whole procedure can be automated directly from the BNFC grammar -- i can imagine a Lift application generator for this context and would love to work with someone to make a simple prototype
  

Best wishes,

--greg

--
L.G. Meredith
Managing Partner
Biosimilarity LLC
806 55th St NE
Seattle, WA 98105

+1 206.650.3740

http://biosimilarity.blogspot.com