🚀 go-pugleaf

Greg Lindahl (gl8f@fermi.clas.Virginia.EDU) wrote:
: I doubt it's that serious. If Wolfram won't provide Mathematica on
: Linux, Maple probably will, since their policy used to pretty much be
: that if someone did the port, they'd support it -- and they are much
: nicer about copy protection. Support sane software companies and
: ignore the twits.

Maple V is now available for Linux. Obviously there is money to be made.
There appear to be many vendors supporting Linux, and the list
continues to grow. One interesting example involves X Inside, which
offers a commercial X11 package. Originally X Inside estimated that
Linux would account for 10% of their business--but amazingly it turns
out that Linux users now account for 75% of their business. Linux/GNU
may now be the most popular desktop Unix system ever. Here is a
recent post by Jeremy Chatfield of X Inside that covers some of the
issues of concern to commercial vendors:

: From: jdc@crab.xinside.com (Jeremy Chatfield)
: Newsgroups: comp.os.linux.development
: Subject: Re: Linux on Alpha's ?
: Date: 11 Feb 1995 11:09:39 GMT
: Organization: X Inside Incorporated
: Message-ID: <3hi5tj$3e2@potogold.rmii.com>
: 
In article <3h4hd3$es@senator-bedfellow.mit.edu>,
Jason Mondanaro  <jtmondan@mit.edu> wrote:
> 
>> The other main drawback of Linux is that GNU pretty much kills any 
>> hope of real commercial vendors getting involved.
>
>I disagree. EMACS and GNUPLOT are neat and all (as are ez and the ATK)
>but I dump them all for a version of MS word or Wordperfect or Excel
>MATLAB and a variety of other commercial products which plain work 
>better and are easier to use.
>
>I think a lot of people feel the same way out there. The only reason 
>I still have a DOS partition running windows is for WORD 6.0 and
>Cakewalk Pro 3.0

As a commercial software vendor, I disagree, too.  We distribute a
binary X Server to replace that of XFree86.  We get (rare) gripes
about not making the sources available and the occasional sniping
shot about having to pay for an X Server when the XFree86 Server is
free.

We sell into a lot of different markets (e.g. BSDI bundles our Server
in the OS price, a few other vendors do the smae to a greater or 
lesser extent).  This time last year we estimated that Linux would
account for about 10% of our direct-sales to end-users (the
onesy-twosy purchases, rather than the hundreds or thousands of
units).  As of December last year, Linux end-user sales accounted for
nearly 75% of our direct end-user sales, and the volume keeps growing.

Linux users have clearly proven that there is a worthwhile commercial
market.  After all, we're selling a replacement for something that
you can get for free, admittedly with a lot of usability and
performance tweaks, but functionally the same kind of software.  If
with this niche-of-a-niche product, we get a sizable chunk of
business, someone with a more horizontal software component should be 
able to make a damn fine income stream...

We're also working on other products, and Linux is now ahead of most
other Intel UNIX OS's for certain product classes.  The response has
been *that* good.

What the GPL kills, is interest in using GNU C libs and OS sources.
Back when I was learning UNIX, it was really helpful to be able to
cobble together a program out of bits of other programs.  This is
inhibited for commercial software developers, but since most of them
will have generated their code on another OS already, there should be
no need for them to snatch code from Linux to do things... and if
their coding proficiency was still at the level where they had to
snarf components of OS sources to make anything work, you probably
wouldn't want the product anyway :-)  We actually compile more than
90% of our product for Linux, on UnixWare (because our OS neutral
dynamic linker is derived from tricks we developed under SVR4.0 and
migrated to UnixWare).

The GPL certainly has an effect when considering the kernel.  We have
a potential product that would require a device driver.  Our
customers seem to keep bang up to date with Linux versions.  If we
release a binary-only device driver, it will be unusable within a
week of release, because there'll have been yet another kernel
version issued... We can't make the sources available, because of the
non-disclosure agreements that we have willingly signed.  Without the
sources, users can't make the patches required to keep a binary driver
usable.  We certainly can't devote someone to just following the
kernel revs and making the device driver avaialble in binary form for
each and every new kernel issued.  And every line of code we use,
must demonstrably be our own, not snitched from some other line of
GPL'ed code...  You want to prove that " i++; " is your own work, eh?

Commercial vendors will be a lot happier to get into Linux, if Linux
users can get out of the mindset of asking whether the product has
been tested on the latest kernel.  Our test suite, for all the
graphics boards that we support, takes about three weeks to complete.
We can't finish before there's another new target.  Fortunately, our
interface to the OS is robust enough that we rarely see our X Server
affected by any versions from 1.08 to 1.1.88 (with the exception of
the obvious reasons that drive this rate of change).

Summary:

Linux offers a very tempting commercial opportunity.  Linux users are
willing to buy commercial software.  The GPL holds a few fears, but
the rate of change of Linux is more of a problem for a commercial
software development and support effort.

Cheers, JeremyC.
-- 
Jeremy Chatfield, +1(303)470-5302, FAX:+1(303)470-5513, email:jdc@xinside.com
        X Inside Inc, P O Box 10774, Golden, CO 80401-0610, USA.
   Commercial X Server - for more information please try these services
http://www.xinside.com            info@xinside.com            ftp.xinside.com