FunctionalRelationalProgramming
From SPA Wiki
This session discussed a potential approach to building large systems in a way to minimize their complexity.
Please get in touch with us - Ben Moseley (benATmoseley.name) and Peter Marks (publicATindigomail.net) if you have any private comments / questions (good, bad or ugly - we don't mind!), or add stuff below for public viewing.
Contents |
Please join the mailing list if you're interested
Future Plans
If people are interested we'll probably organise some meet-ups (likely in the London area) - let us know if you'd be interested in that kind of thing.
Web Resources
http://ben.moseley.name/frp/frp.html - includes the paper which was handed out at the session as well as the slides.
Mailing List
Homepage: http://groups.google.com/group/frp-discuss
Group email: frp-discussATgooglegroups.com
Description: Discussion of ideas related to Functional Relational Programming
Postscript
This topic came up last night in the discussion after SPA-196e. John Everhard had explained how he views most IT systems as being expressible in terms of rules, for the most part - well, isn't this just what functional programs are? It turns out that a rule system like Pega 1 is very similar to the proposed functional relational paradigm. It also eliminates unnecessary state to a very large extent, removes the perceived need for developers to optimise the code, ceremonious incantations etc. It took a number of years to develop to the current stage of sophistication. Suggest close investigation of this approach.
For what it's worth, here are my notes of the session File:FunctionalRelationalProgramming1.doc. -- ImmoHuneke
Thanks very much for the feedback Immo. Will try to go through your notes when I get a chance.
I personally have a number of concerns about rule-based systems - eg how are choices made between multiple applicable rules, and what can the rules _do_ - ie in many systems they're able to update some kind of "working memory" - and I'm concerned that this could start getting into the classic problems of state. I haven't looked at Pega though so it is possible that that does something different.
I guess at the high level though rule-based systems have a very similar objective to ours - of separating out business logic, and that's definitely something that we're in favour of. Basically I think my personal concerns are just about exactly how that is done. -- BenMoseley
I've put the Slides on the FRP website too now. -- BenMoseley