[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Software Development Progress
- To: info-dylan@cambridge.apple.com
- Subject: Software Development Progress
- From: farrow@stephanie.hc.ti.com (Rob Farrow)
- Date: Thu, 3 Dec 92 10:58:10 CST
The conversation about Dylan/Newton was really interesting. Thanks to
the participants for some good debate. I'd like to use that as a
springboard for some issues about Dylan I've been pondering.
One of the nice things about Smalltalk is that it is has a well
conceived development environment BUILT-IN. Development environments
for C/C++ are big business. That's because the systems of any
complexity written in these languages are huge and require tools just
to keep track of them. But Lisp is not that much better. I have been
a Lisp programmer on very large systems and it is quite apparent that
we could really use a better strategy for maintaining and modifying
these systems.
The Smalltalk approach is a good one. But I'm not convinced it's the
best one. There are plenty of good ideas for how to make software
easier and more modular. Look in some of the Software Technology
papers or in the Electronic CAD industry where they have even worse
headaches than C/C++ programmers.
I don't think Dylan can be more than "yet another language" if these
issues aren't addressed. In my estimation, the MAC revolutionized the
operating system world. It took years, but finally the rest of the
industry is getting in line -- as evidenced by MicroSoft Windows, Open
Windows, etc. (Unix is glitch in that progress. But I hope "Pink"
can fix that.) In retrospect, it seems to me (though I'm an invalid
historian) like Apple just took its cues from the existing research
(ala Smalltalk) and built on the best ideas.
Not just their operating system's easy functionality but it's COMPLETE
interconnectedness with the machine. If Apple can do the same thing
for software development that it did for operating systems then, it
may still take years, but hopefully the rest of the industry will
again get in line. The latch-key technology of the new DUO-DOCK
system is spawned from this mold. How about a latch-key style
technology for software development?
-- Rob.
PS. You can somewhat gather my position on patents from my comments.
Which needless to say does not reflect the opinions of the company I
work for.
PSS. What about latch-key OODBs and Interface Managers, too?