OOOS dynamics (was Re: ARGON)

>>>: Alaric
>>: Fare
>: Jecel

>>> If you wanna come eye the competition, I've recently updated my ARGON
>>> page with the most recent incarnation of my plans. It's at
>>> "http://www.abwillms.demon.co.uk/os/"...
>> I cheerly welcome competition (BTW, where are you, Jecel?).

> About OS competition, the way I see things there is too
> much redundant effort going on. Please see this neat OS
> list -
> http://www.lfbs.rwth-aachen.de/~sven/OS-Projects/index.html
>
Yeah, that's the most comprehensive list available on the 'net,
which is why it's the first meta-link on the Tunes OS Review page,
that contains a lot of pointers, too...
http://www.eleves.ens.fr:8080/home/rideau/Tunes/Review/OSes.html

> This is to be expected as all the advanced systems are either
> in the design phase or are some tiny assembler thing that only
> boots and prints "hello world" from 386 protected mode. Minor
> preference differences are enough to make another little group
> or individual start their own OS project.
>
Yes, I've joined several project lists before,
and that's also my conclusion from personal experience.
Most of the time, the new OS is begun without real ideas,
as a toy for the author(s).
They are indeed spending lots of time booting and writing basic drivers,
when their ideas would be better first promoted and refined
as a layer above an existing OS (Linux, VSTa, or the HURD).
They're not very advanced and don't go anywhere precise,
hence they can't attract developers and support;
the authors ought to join another project instead
(e.g. GGI for graphics, VSTa or HURD for free message-passing OS, etc).
I think the same partly applies to Tunes,
because of management problems of which I'm responsible )-:
we're nowhere in implementation,
and the goal isn't clear to potential developers.
Maybe I should advertise the LLL page here:
http://www.eleves.ens.fr:8080/home/rideau/Tunes/LLL/LLL.html
It defines enough things to
Feel free to refine it.

> The situation from Unix clones was a little like this a few
> years back. Now we have practically two choices: Linux and
> FreeBSD. They are not necessarily better than others which
> have disappeared, but the got to a point where they were
> working enough that it was better for others to abandon
> their own versions and contribute to one of these instead.
>
Exactly. Once you have a stable supported free system,
why build a concurrent version of the very same system?
But the advantage Linux had is that its specs were clear: POSIX.
Linus never had any design problem, only implementation problems.
*BSD started before/faster, but was left behind Linux,
because it had lots off licensing problems to begin with.
Linux originally attracted lots of support,
and these are self-catalytic phenomena.

> Something like this will happen in the OOOS area as well.
> Once one of our systems is actually doing neat things then
> others will prefer to try to implement their ideas on top
> of it rather than doing everything themselves.
>
> Of course, one Unix clone is pretty much like another while
> there are significant differences in the systems we are
> designing. So we might end up with four or five "real"
> OS projects, instead of two. But I am pretty sure that most
> of the list I meantioned will quickly fade away once the
> first projects start working.
>
There really ain't any clear way for OOOSes:
people don't really know what they "OO" is or should be
(see the very unclear OO FAQ -- lots of jargon and propaganda);
rather there are a lot of school wars
and people fighting for (buzz)words,
because "OO" is a fad and attracts lots of money.
Most of these "OO" approaches can't do any better
than what CORBA already does:
unreliable (time,space)-coarse-grained objects implemented in C/C++.
   I don't think these "OO" wars can stop before
some OOOS becomes popular.
I don't think the current kind of semi-OO OSes
(like NeXTStep or BeOS) can be successful against traditional OSes,
because they're not free, not standard, and don't bring enough
advantages to counter-balance that.
   As for all the above pet projects, their motivation seems
programming fun more than anything else.
Even when they "work", they have no goal to share;
they both can't die and can't live: they are one-user OSes.
On the other hand, every research projects tackles
a particular implementation problem, never a whole-system approach
so again these systems can't die and can't live: they are one-idea OSes.
No one seems to have the guts to develop a free OS based on a novel design,
whereas a novel design can't succeed over tradition unless it's free.
   Anyway, they are far from a clean reflective system as Tunes proposes.
I'm convinced that Tunes holds lots of strong points
such that the project will survive until either it succeeds,
or these points are jointly tackled by another project.

> So good luck to us. We may be one day working together.
Surely we will!

== Fare' -- rideau@ens.fr -- Franc,ois-Rene' Rideau -- DDa(.ng-Vu~ Ba^n ==