Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!bellcore!decvax!spray!anson
From: anson@spray.CalComp.COM (Ed Anson)
Newsgroups: comp.sys.mac
Subject: Re: Criticisms of LSC (was Apple developers & programming on the Mac)
Message-ID: <2599@spray.CalComp.COM>
Date: 20 Feb 89 15:17:02 GMT
References: <1309@ndmath.UUCP> <3916@ece-csc.UUCP> <486@adobe.UUCP> <1195@husc6.harvard.edu> <70205@ti-csl.csc.ti.com>
Reply-To: anson@spray.UUCP (Ed Anson)
Organization: CalComp, A Lockheed Company, Hudson, NH, USA
Lines: 50

This thread has suggested several criticisms (opportunities for improvement)
of Lightspeed C. There have also been several defenses, based on cost of the
proposed features. I'd like to make an observation and a suggestion.

If every useful suggestion for improvement were converted to an enhancement
of LSC, the following would most likely happen:
  -- The package would become prohibitively expensive for many current users.
  -- Many people would not like all the extensions.
  -- Many people would still want more.

If none of the suggestions are taken, many users will be forced to move to
other development environments. I may be among them. Not because I don't love
LSC, but because of a few features which are becoming critical to me.

It thus seems that the best solution would be to make improvements in a
modular way, and unbundle them. Give people a choice of editors. Provide
extra support for specialized types of development. Many of us would be glad
to pay for these extras, even though not every user should be burdened with
the cost.

... But THINK probably doesn't have the resources to do all of the things
everyone might want. Why should they have to?

Why not add one basic feature: Open the architecture in a manner similar to
HyperCard. That is, provide hooks for third-party enhancements. MPW does this,
sort of. But I think the folks at THINK could do a better job. What we need
is a way to integrate new tools into the basic LSC environment.

With this approach, everyone can eventually have the features they want/need,
and THINK can concentrate on what they do best: creating a friendly user
environment and fast compilation. The rest of us can add the goodies that will
make LSC even more attractive than it is now.

For the record, I could care less about macro editors. What I'd like is:
  -- extension of the project concept to sets of related CODE resources
  -- an option to automatically check files for a Make, to be used when
     multiple projects share files (This should also apply to automatic
     makes, e.g., when I say Run.)
  -- version control for sets of source files
  -- C++ with MacApp compatibility
  -- optional optimized code generation

I realize that not everyone would want or need all of these features. But I do.
I would personally be willing to pay extra for each of these, if necessary.
	
[Usual disclaimers apply.]
-- 
=====================================================================
   Ed Anson,    Calcomp Display Products Division,    Hudson NH 03051
   (603) 885-8712,      anson@elrond.CalComp.COM