[Oberon] V4 versus S3 survey

[Andreas Pirklbauer]

> hmm, I am not sure. I guess a growth story would have
> suffered even more  from the forking than Oberon System
> suffered already (V4/S3/ETHOS/AOS/A2).

Only if no proper software planning and release management
process is in place (sadly, Oberon never had any to begin
with, and yet it is indispensable in professional software
development). There is more to software than just software.

But I doubt that a professional release management would have
done the trick for Oberon. But developing a “real” product
with a clear vision and product roadmap would at least have
been a good start - similar to what the Borland line of
products did for Modula-2 about a decade earlier, or the
Pascal compilers available on Apple another decade before
that (80s), or the P-code compilers available (mostly to
academia) a few years prior (70s).

But in the end it is never the tool (here: the language and
the developer tools that come with it) that drives adoption,
but always the underlying platform it is running on.

Unix is the reason for the success of C (as it is written in
it). When Apple switched from Pascal to C in the early 80s,
it was (more or less) the end of Pascal on the then-current
version of the Mac (the platform). If Borland had stuck with
Turbo Modula-2 on CP/M (the platform), Modula-2 might have had
a somewhat better future (although not much, since CP/M itself
had no great future, but that’s another story..).

Oberon never had any such sponsor to begin with.

For Oberon to have succeeded, it would have needed to be
embraced by *at least* one large platform provider. Microsoft
in the late 1990s would have been such a case, but ETH
completely missed out on that - even though they were
quite involved with Oberon on .Net around that time. In
the end, Microsoft ended up releasing C#. That closed
that window of opportunity.

Today, I don’t think there is much chance that Oberon will
ever come back to life again, as all major players (Microsoft,
Google, Apple, Linux/Unix..) now already have their own
“modern” languages (C#, Go, Swift, Java,..). Despite the
apparent weaknesses of some (all?) of these languages, they
won’t go away anytime soon. The investment is too high,
the code base too large. The bandwagon has left.

Perhaps someone has the idea to build a LLVM frontend for
LLVM one day. But even with that it would only stand a chance
if (a) it is promoted to a first-class language by at least
one platform provider (more is better), and if (b) it is fully
integrated with the “rest” - and all of that would require
building a professional software development organization.

That closes the circle. It won’t happen.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.inf.ethz.ch/pipermail/oberon/attachments/20171121/a2061b41/attachment.html>