Re: [Dltk-dev] DLTK for Scheme

[Andrey Platov <andrey@xxxxxxxxx>]

Hi William,

Thank you for appreciating DLTK, as for first question:

At the time of presentation you mentioned, we were planning to have a powerful framework for type inference based on Lex Spoon Demand-Driven Analysis http://www.lexspoon.org/ti/ combined with abstract code interpretation, which will rely on other DLTK Core services. As we think type inference is a must do for powerful IDE and our idea was to minimize efforts for engineers implementing forward type inference (based on type info propagation through execution flow). 

According to that plans we initially stated that all ML-alike languages (Hindley-Milner type system) will not fit into DLTK (I know that LISP is not HM); other languages that we thought would not benefit a lot from DLTK Core Frameworks were marked not well-suited as well.

Right now I must admit that there is no good and generic TI framework in the DLTK yet, so when considering some language to be good or not to be good for DLTK we must understand if *existing* DLTK services could be helpful for language IDE implementers.

Scheme IDE probably may benefit from DLTK using Structural Code Model: DLTK will help you to maintain your codebase decomposed down to minial language elements like function definitions and variable declarations, track structural changes, index program elements and search for elements definition/references. Of course there is generic UI covering most aspects of the navigation through language elements like code outlines, open XYZ dialogs, hierarchies (call, type) views, etc. All of this you're getting almost for free :) The rest of implementation (and still a huge part) is up to IDE implementer.

Unfortunately I'm not familiar with Scheme and would be interested to know if it makes sense to use DLTK as a foundation for Scheme IDE. Please share your thoughts.

Kind Regards,
Andrey Platov
http://www.xored.com

----- Original Message -----
From: "William Cook" <william@xxxxxxxxx>
To: dltk-dev@xxxxxxxxxxx
Sent: Wednesday, June 4, 2008 8:36:20 PM GMT +06:00 Almaty, Novosibirsk
Subject: [Dltk-dev] DLTK for Scheme

Hi everybody! DLTK looks great... i have some questions:

In one of the presentations on DLTK is asserts that DLTK is not 
well-suited for Lisp, Scheme or prolog. I can understand why prolog 
might be a problem, since it has a different notion of stack frames. But 
what is wrong with Lisp or Scheme? these are traditional stack-based 
languages (even if they do have closures.. that doesn't matter).

Second question:
It seems that the code for DBGp is in the Ruby branch (and maybe 
duplicated in other branches?) Why isn't it in the core?

-- 
William Cook
Assistant Professor
UT Austin Computer Sciences
_______________________________________________
dltk-dev mailing list
dltk-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/dltk-dev