Subject: Re: Caldera will publish DR DOS source code
From: "L. Peter Deutsch" <>
Date: Fri, 20 Sep 96 08:08 PDT

> They got all the DOS technologies from Novell, which got them from
> Digital Research.  Their model for DOS isn't GPL, but they're getting
> closer.  Full source, free for personal use, small fee for commercial
> redistribution.  They have noticed how quickly Linux is evolving, and
> are hoping to bring that innovation to DR DOS.  The overall strategy
> appears to build up a free software base that can keep up a faster
> momentum than Microsoft and eventually unseat them from OS domination.
> Not a bad idea.  Ray Noorda is behind it all.  See for
> details.

This is pathetic.  The problem is not the technical quality of the bottom
level of the software pyramid: Linux is vastly better than either DR DOS or
MS-DOS, and has been around for years.  The problem is that the free
software community has never succeeded at (1) establishing reasonably stable
API standards above those of Unix and X, (2) creating user-attractive
applications in any significant numbers, or (3) working with the unsightly
de facto standard file formats (such as Microsoft Word and Excel) essential
for use of existing investments in data.  Of these, (2) is the most
significant; in my opinion, (2) is the reason why Unix blew a 15-year
technology lead over the PC.

A free (I'm using the word in its common meaning of "cost-free" here) OS is
irrelevant (i.e., not attractive enough to draw users) if the applications
aren't cost-free, because Windows 95 is bundled with essentially every PC
sold today (i.e., 90% of all hardware units), and, from the purchaser's
point of view, costs nothing.  The fact that it isn't "free" in the sense of
source-code-available means nothing to users: I think the record shows that
users, other than a small community of hackers like the people who read this
list, would much rather have one-stop shopping for support (a small amount
of which is even often cost-free) than the freedom to shop around among a
small number of tiny (and often not particularly competent) companies for
separately paid support.

Beyond the issues of support and cost, DR DOS, or Linux, has enormously less
capability than W95.  Take a look on what comes with W95 one of these days.
At the bottom, neither DR DOS nor Linux even has the equivalent of DLLs, let
alone DDE, OLE, the W95 widget or icon set, a metafile architecture for
graphics, or most of the dozens of bundled applications.  The bundled Linux
applications that do make an attempt at a better-than-ASCII UI are wildly
disparate in their appearance, and almost all clunky-looking compared with
the glossy W95 apps.  DR DOS is starting way behind Linux in these respects,
as well as in the technical aspect (DOS is at least a 10-year-old
technology, and doesn't have any form of hardware-based memory management,
protection, or scheduling).

Maybe Caldera's intention is simply to build up DR DOS to the point where it
can run Windows applications.  In theory, this could draw some OS $$ away
from Microsoft.  The problem with this approach is that (1) Microsoft keeps
raising the bar on what constitutes Windows faster than any modest-size
outside team of programmers can clone it (e.g., today it would have to be
W95), (2) Microsoft doesn't document its APIs nearly well enough to make
good cloning feasible without time-consuming reverse engineering, and (3)
Microsoft now makes most of its money from things other than the OS, and is
obviously heading even further that way.  Because of these considerations,
DR DOS might be a good strategy for Caldera, whose goal is to make money
rather than change the industry: if they can get 10% of Microsoft's current
OS revenue by producing a W95 clone without having to pay royalties to
Microsoft, they'll be a wealthy company, but they won't have changed
anything significant in the industry culture.  My guess is that this *is*
their strategy, because they are keeping the right to charge for

I find all of this somewhat amusing because, of course, my own business
model for Ghostscript is exactly the one that I just discussed.  It's only
been possible for me to do this successfully because (1) Adobe has been
raising the bar just slowly enough for me to keep up (although I haven't
seen the PostScript Level 3 specs yet, and this may be the iteration that
knocks me over), (2) Adobe documents PostScript and PDF amazingly well, and
(3) I don't want to change the industry, just have fun, make lots of people
happy, and put away a pot of money for retirement.


L. Peter Deutsch         |       Aladdin Enterprises ::::
203 Santa Margarita Ave. | tel. +1-415-322-0103 (AM only); fax +1-415-322-1734
Menlo Park, CA 94025     |
          "Implementation is the sincerest form of flattery."