CLIM mail archive
[Prev][Next][Index][Thread]

contents of clim-library as of 92/09/09

   Date: Wed, 16 Sep 92 19:27 MDT
   From: Simon Leinen <simon%lia.di.epfl.ch@lucid.com>
   References: <19920916150835.6.SWM@SUMMER.SCRC.Symbolics.COM>
	   <19920914151218.3.SWM@SUMMER.SCRC.Symbolics.COM>

   Scott McKay writes:

   > I would also like to know the answers to these questions...  In my
   > opinion, a major reason for the success of X Windows is the sheer
   > extent of the stuff publicly available for X Windows.  I don't think
   > CLIM can be successful without a similar effort.  Of course, CLIM
   > itself may be too immature or limited to support such an effort.  I
   > don't know, but I would like to.
   [...]
   > If anybody wishes to write good CLIM demos and donate them to the
   > cause, that would be terrific.

   I don't see the point in contributing anything to CLIM.

   Programs like the X Window System or GNU Emacs have been contributed
   to me, so I contribute back by spending time porting and debugging
   them and making available improvements or X clients that I write.

X Windows had major sponsorship from portions of the computer industry
to underwrite the development project.  CLIM had no such support.
When we started the CLIM effort in 1988, ILA was a 4-person Lisp
consulting and products company.  The project was an attempt to
overcome what we thought was a substantial barrier to the development
of commercial Lisp software, namely the lack of a portable UI toolkit.
In an attempt to get some financial support and to send a message to
the Lisp community that the issue was being worked on, we signed up
some of the Lisp vendors as early licensees.  Our "market research"
showd that Lisp customers weren't interested in a random software
package from a random small company (leaving them vulnerable to the
same kind of single-point failure that they were already worried about
when using any one vendor's proprietary toolkit).  They wanted some
kind of assurance that the Lisp vendors were going to endorse a common
standard.

CLIM was never intended to make ILA money (although it was intended to
not lose nearly as much as it did :-).  The hope was that a
multi-vendor UI toolkit standard would help open up the market for
other Lisp products and services.  We also thought that a modest
ongoing CLIM royalty stream would help to fund the development of
future versions.

   For CLIM, we have to pay $1000 per license even though the software is
   not very stable yet (cf. number of problem reports in this list, for
   example).  

The various Lisp vendors set the CLIM price for their platforms.
There is a deep issue with pricing a product like CLIM: should it be
priced to generate additional revenue (to be used in part to continue
its development), or should it be freely distributed in the hopes of
generating good second-order effects?  I don't claim to have the
answer to this one.  It is tough for an industry as small as this one
to support much uncompensated resource investment.
	      
	      We also have to invest a lot of time to learn how to use
   it.  Sure, we could send someone to a CLIM course in Germany but we
   don't have the money for that.

This argument would apply to any new toolkit, so unless there was
already a generally-in-use, portable, endorsed-by-all-the-vendors
toolkit that you already knew how to use (and there wasn't) you would
have had to invest the learning effort.

   When I write an X application in C using the X Toolkit, I can look at
   the source for many existing clients as well as for the libraries I
   use.  If you send me the CLIM sources, I promise to write a couple of
   utilities for your FTP archive!

Again, each CLIM distributor chose its own source distribution policy.
Most put CLIM under their existing source policy.  There are
legitimate reasons for withholding sources, among them the fact that
it is hard to offer good, professional support for software that is
evolving.  Maybe these policies should be re-evaluated.  At minimum,
much more CLIM source code (in the form of example applications, etc.)
should be made available.

[which brings us back to the initial topic: send in your spare CLIM
code to the user library.  Even if you don't have a generally-useful
piece of substrate code, just providing some simple CLIM applications
that others can look at for examples would help.]

   If I write a program in CLIM, only people who can afford CLIM will be
   able to use it.  So for maximum usefulness I prefer to use a
   not-so-nice UI library that is free (such as Garnet or CLUE).  Almost
   everybody with a Unix machine (including 386 PCs) can use either AKCL
   or CMU CL, CLX and one of these free toolkits (CLM is also free, but
   currently depends on Motif which is not, LispView is now also in the
   public domain but only works under Lucid).

We had a customer discuss his LispView dilemma with us at AAAI.  On
the one hand he was really pissed that Lispview was now an unsupported
product and he couldn't rely on getting bugfixes, etc.  On the other
hand, he was really happy 'cause he had the source and was making all
sorts of substantial modifications to make Lispview meet his needs.  I
couldn't see how we could resolve this conflict.

   For the same reasons there are much more contributed X clients using
   the Athena Widgets or XView than using OSF/Motif or OLIT.

   Transform ILA into the CLIM Consortium and give the CLIM sources to
   everybody who is interested.  Then you will soon see AKCL and CMU CL
   ports and a lot of contributions from us Lisp hackers.

   Or keep it restricted to the lucky few who can afford CLIM licences or
   Symbolics software maintenance contracts - but don't be surprised if
   it remains in its market niche.

Again, the issue is one of resources.  The Lisp community was lucky
with PCL/CLOS: Xerox was willing to fund a team of N superb hackers
for M years to work on PCL.  This effort clearly led to the advanced
state of OOP tools in Lisp today.  In the CLIM case we all weren't so
lucky.  ILA's Lisp "division" went bust trying to see CLIM through to
completion.  A multi-company coalition without a clear, binding set of
operating rules is not a very stable platform from which to launch a
software product.

   Sorry if this message sounds aggressive - I'm just tired of seeing
   CLIM presented as a noble effort to help us poor CL users out of the
   lack of user interface toolkits, and to which we should all contribute
   (money, maintenance time, demos, utilies) in our own interest.  CLIM
   is a piece of commercial software that happens to be proprietary to a
   couple of vendors rather than one.  So it is open (the spec is
   available on FTP, and maybe you wouldn't even sue me for
   reimplementing it).  For me, CLIM is exactly like OSF/Motif except
   that it is technically more mature and lacks critical mass.

I don't take any offense at your general comments above.  In fact I
basically agree with the idea that CLIM will be successful to the
degree that it is widely available.  I must, however, emphasize that
in fact the original CLIM effort WAS an attempt to make things better
for the Lisp community as a whole, and that there was an altruistic
component despite fact that we tried to find ways to fund the effort.
I personally worked for over a year on CLIM on my own dime, and other
ILA principals did likewise.

0,,
References:
Main Index | Thread Index