Newsgroups: gnu.misc.discuss,comp.lang.tcl,comp.lang.scheme,comp.lang.misc
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!cs.utexas.edu!uunet!s850.mwc.edu!hearst.acc.Virginia.EDU!murdoch!elvis.med.Virginia.EDU!sdm7g
From: sdm7g@elvis.med.Virginia.EDU (Steven D. Majewski)
Subject: Re: extensibility (was: Why you should not use Tcl)
Message-ID: <Cwuzp5.48n@murdoch.acc.Virginia.EDU>
Sender: usenet@murdoch.acc.Virginia.EDU
Organization: University of Virginia
References: <9409232314.AA29957@mole.gnu.ai.mit.edu> <364bq2$mjd@topaz.sensor.com> <36a5ri$kko@agate.berkeley.edu>
Date: Wed, 28 Sep 1994 21:29:29 GMT
Lines: 80
Xref: glinda.oz.cs.cmu.edu gnu.misc.discuss:18511 comp.lang.tcl:19582 comp.lang.scheme:10082 comp.lang.misc:17888

In article <36a5ri$kko@agate.berkeley.edu,
Brian Harvey <bh@anarres.CS.Berkeley.EDU> wrote:
>ron@myhost.subdomain.domain (Ron Natalie) writes:
>>However, how many of your applications are going to end up being kitchen
>>sink text editors?  99% of your applications are better designed to be
>>targetted at what the customer wanted to begin with
>
>Geez, I spend two days away from reading news and comp.lang.scheme
>takes all day to catch up on!
>
>Anyway, although I don't have anything to say about the main stream of
>this debate, the passage quoted above pushed one of my buttons.
>
>In my opinion it's a myth that the users want smoothly packaged
>software that does "what they want" and can't be customized.

We are constantly running into the problem of buying a package that
does 80% of what we want, but is not easily extensible. ( I enjoy 
programming, but a lot of other folks around here would rather be
doing science, and not be distracted. ) 

Sometimes the vendors answer is that we need to buy a $20K source 
code license if we want to modify something. 

Sometimes they say: write up your enhancement request (in triplicate,
so they have more copies they can loose! :-) and we'll consider whether
there is enough customer demand to add this feature. Oh - you're
willing to do it yourself rather than wait 8 months for an answer - 
sorry - that information is proprietary. ( Oh - you're willing to
sign a non-disclosure agreement ... well, we only share that
information with our "partners" - we don't want to be committed to
supporting the current interface/implementation. ) 

Sometimes that don't answer, or everyone who you are allowed to speak
with seems to know less about the internals of their product that you
do! 

>Users are so desperate to have extensible software that they pay
>good money for "macro" languages on the Mac that work by simulating
>mouse clicks!  This is the world's worst programming language, worse
>than Tcl and awk and C++ combined, because when the shape of some
>window changes in the next release of something, your macro stops
>working.  And of course it doesn't allow for iteration or naming
>or any of those ivory tower computer science things.

And a current vendor of our actually gave us a working interim
solution: Send mouse click events to their program from our program,
to simulate manual control of something we needed to be under program
control. 

But they were clearly getting customer demands for just the type of
extensibility that we were asking for, and they, and many other 
commercial vendors have planned a solution for the next generation 
of software:   ****VISUAL BASIC** and VBX modules** !  
( they are "looking into" OLE ... ) 


[ There you are, if you want a worthy target to flame at! ] 


One of the tenets of the GNU project and FSF has been that access to
the source code gives you the power to extend a program that almost
does what you want. Giving people this POWER, not giving out free
software, has been their primary aim. ( Making access to the source
code free is just the method to insure that no-one takes that power
away from you. ) 

I think a lot of people looking at using tcl/scheme/python/whatever
as extension languages share the same view, except they want to be
able to reach the same goal without making commercial profit
impossible or difficult.  I wish the some of the folks flaming RMS
so severely would consider that THAT is his real contribution, and if 
there is any one person who deserves credit for preaching the concept, 
long before Brad Cox's "software parts" or John Ousterhout's tcl - 
it is Stallman. 

-- Steve Majewski       (804-982-0831)      <sdm7g@Virginia.EDU> --
-- UVA Department of Molecular Physiology and Biological Physics --
-- Box 449 Health Science Center        Charlottesville,VA 22908 --
		 [ "Cheese is more macho?" ] 
