[TUHS] Source code abundance?

[Nick Downing]

Yes. And I just want to point out the systems vendor's worst nightmare:
Competition from an earlier version of their own product. History is
littered with examples where something was deliberately left to wither and
die for this reason.

Apple II and IIgs: We all know that the IIgs was deliberately crippled, and
then discontinued in favour of the IIc+, as it presented a viable
alternative to the 68000-based Macs.

680x0 Macs: Apparently some licensees had 68060 Macs and accelerators in
the works, but Apple refused access to the ROMs to add the 68060 support
code, because it would have been a viable alternative to the PowerPC 603.

IBM OS/2: Was heavily DOS based (I believe it used the INT 21h API with
modifications for protected mode), but in fact was eclipsed by later
versions of DOS/Windows that were retrofitted with things like DPMI
support, hacky but effective in providing a viable alternative to OS/2.

BSD and SysIII: For a while it looked like the 32V-derived BSDs were going
provide a viable alternative to AT&T's official developments of the same,
and it took some heavy handed legal and political manouevring and backroom
deals to make sure that did not happen in the end.

AMD64 and Itanium: Enough said, a very expensive egg on face episode for
Intel. 8086/8088 and iAPX432: Same thing except it was actually Intel's own
product that provided a viable alternative to the "official" new version
rather than a competitor's development of it. Of course a similar story can
be told about 8080/Z80/8085/8086, Intel faced stiff competition from an
enhanced version of their own product before wresting back control with the
much improved 8086. A nightmare for them.

That's the real reason vendors won't open source.

Nick

On Mar 4, 2017 12:02 PM, "Henry Bent" <henry.r.bent at gmail.com> wrote:

On 3 March 2017 at 18:56, Wesley Parish <wes.parish at paradise.net.nz> wrote:

>
> And since the central Unix source trees have been static - I don't think
> Novell was much more than a
> caretaker, correct me if I'm wrong - and the last SysVR4 release of any
> consequence was Solaris - has
> Oracle done anything with it? - I think the best thing for all would be
> the release of the Unix SysV
> source trees under a suitable open source license.


There was an SVR5, even if it was not nearly the popular product that its
predecessors were.  While development certainly slowed, it contained some
amount of technological progression.  Obviously at this point development
has stopped completely and it probably does make sense to open source that
code base.


> (I've made a similar argument for the IBM/MS OS/2,
> DEC VAX VMS, and MS Windows and WinNT 3.x and 4.x source trees on various
> other Internet forums:
> the horse has bolted, it's a bit pointless welding shut the barn door now.
> Better to get the credit for
> being friendly and open, and clear up some residual bugs while you're at
> it ... )


Equating VMS, old versions of Windows, etc. isn't quite the same.  Even old
versions of those products may well include source that contains, or is
believed by its owners to contain, novel ideas or novel implementations of
existing ideas that may have survived relatively unchanged in newer
versions.  And because there is at least a reasonably sized user base for
all of the products you mentioned, corporate customers have an interest in
protecting their investment, and the software creators have an interest in
responding to the desires (or perceived desires) of their customers.

Don't get me wrong - I'd love to see a legal release of the VMS 5 source,
or Windows 3 source, or classic Macintosh source.  I'm just not holding my
breath.  I think the community's time would be better spend advocating for
source releases of products that are truly dead or all but dead.

-Henry
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20170304/8bf8aec7/attachment.html>