on Wed, 6 Aug 2003
23:47:04 PDT7
> I'm not sure what Lynn's project goal is, and I'm worried that his
> Warpstock audience won't understand it either.
The goal is to improve programmer productivity. To do this Lynn has taken
features from some languages and wants to combine them into one. Then he has
this programmers assistant, the middle parts of the system that let you know
what parts of the code have incorrect syntax and which parts of the code are
incomplete.
> If he wants to incorporate various already-in-use language features into
> a new HLL, how will he accomplish this?
You create a new language.
> Does he have a game plan, a
> critical path, a copy of the language spec of every language he wants to
> draw from?
Not that I know of but it does not all have to be in place at the start.
> Will he start with a new language spec and then move on to a
> compiler or is he more intrigued with the open typing and code
> optimization and will first write the compiler, then write the language
> specs based on what the working compiler can do?
As I see it the compiler has the least benefit for improving a programmers
productivity.
You do not need a native compiler to start with. You can use any assembler or
compiler and have the "compile" part generate code to go into any current
compiler. It is an extra step and probably not the most efficient but that is
not where the programmer productivity problem is.
> But this is *good*. Better to be confused in an understandable way. I
> just *hate* it when I write something in Rexx and get unanticipated
> results and have to spend 30 minutes looking for an if-then-else group
> that's outside the logic block I thought I'd created.
Some thing will not change. ;-)
> Hmm. Should the programmer interface be designed *before* writing the
> HLL spec?
I don't think so. I think some of the interface will depend on how the
language is handled.
> Lynn: How will you cohesively bring PL/I and Prolog together in one
> HLL? I have only a cursory understanding of Prolog, but nevertheless I
> don't see a full overlap.
Why do you want full overlap? Take what is best from each and use that.
> -- Suppose you're creating a typical match-merge program (read two
> sorted files and create a single sequenced file). Do you use PL/I-ish
> statements written in a text editor, or do you instead open a
> boilerplate match-merge template, import the file formats for the two
> source files, click on the fields to be matched, and let the PI do the
> rest?
Either way, it is up to the programmer.
> In short, what exactly is does the new PI encompass and is it, or is it
> not, a front-end for generating a line-by-line source code file?
It generates the executable. It does not generate source code, that is what
the programmer does.
Time to sign off for the night.
Goodnight.
--
Robert Blair
=====================================================
To unsubscribe from this list, send an email message
to "steward@scoug.com". In the body of the message,
put the command "unsubscribe scoug-programming".
For problems, contact the list owner at
"rollin@scoug.com".
=====================================================
<< Previous Message <<
>> Next Message >>
Return to [ 07 |
August |
2003 ]
The Southern California OS/2 User Group
P.O. Box 26904
Santa Ana, CA 92799-6904, USA
Copyright 2001 the Southern California OS/2 User Group. ALL RIGHTS
RESERVED.
SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group.
OS/2, Workplace Shell, and IBM are registered trademarks of International
Business Machines Corporation.
All other trademarks remain the property of their respective owners.
