To: stik@ON-Luebeck.DE
Subject: The true STiK 2 origins AFAIK (Ronald)


On Tue, 21 Jan 97 18:55:32 G atari@bright.net (Lonny P wrote: 
>
>  Peter Rottengatter <perot@pallas.amp.uni-hannover.de> wrote:-

----- snip ----- re: lots of stuff not concerning me directly


>>With the single issue causing a lot of friction, I assume you mean the 
>>module loading ?? Well, that's another example. If everybody is concerned 
>>about it, why did nobody discuss on it ?? I saw only Martin, Ronald, and 
>>myself discussing it. Within this group of three, Ronald and I held the 
>>majority on a single opinion, and along with Ronald I'm still convinced 
>>that the Pexec scheme is better than a custom design. So why should we 
>>abandon it, if apparently nobody else cares, or supports Martin's opinion ?
>
>NO!  Not the dam module loading scheme.  How many times do I have to say it?
>You removed everyone from the decision making process, except perhaps
>Ronald who you have been coding with in secret so it seems.

'Seems' ?   Well that 'seeming' happens to be quite incorrect.

What happened was this:

Peter and I have long had discussions on various topics both in and out of
the list, since some people have voiced a dislike about overly technical
speculations on the list.  In one such discussion I suggested various things
that would be very hard to implement in STiK 1.  I fully expected Peter to
say that these things could not be done now (if ever), but instead he choose
to tell me about some ideas he had been working on for a while.  This was of
course what has now become the STiK 2 project, which will indeed allow the
implementation of such things  as I suggested.

Apparently Peter (with Dan's approval) had for some time been working on ways
to improve STiK 1 by adding stuff to his planned INETD ACC so as to achieve
some of the goals he later set up for STiK 2.  After working on this for a
while he realized how much better those routines could work if used with some
TCP/IP stack designed from the start for more flexible uses.  As programmers
tend to do, he sketched some ideas for how the kernel of such a system should
work, and how modularization could be implemented and so on.

That was the real birth of the STiK 2 project, and as I hope you have noticed
there was no 'secrecy' involved.  He was just doodling at that stage, though
he soon realized that, by actually making a project of it, it should be quite
possible to achieve an improved modular STiK version with the improvements
planned (via INETD.ACC) for STiK 1.  And to do so in shorter time due to the
modularization allowing authors to work efficiently in parallel.

When he had spent enough time on this to be sure that his ideas and conclusion
were valid he decided to contact Dan and talk things over, to gain his support
(both morally and as a module programmer) for a new design and to make sure
that Dan had no direct objections of course.

But this was the period when Dan lost his e-mail account, so contacting him
was not possible for a while.  To make sure that Dan should not misconstrue
what Peter had done and wanted to do, Peter decided to take no chance that
Dan should hear of it by rumour, which meant that he (Peter) would have to
keep silent on it towards the STiK list as well as to others.  That is the
only form of secrecy that was ever involved in the STiK 2 project.

AFAIK he only broke this decision once, telling me about it, which was done
mainly because he wanted to ask me for alternative ways of reaching Dan, and
to ask for my advice on what to do about the situation.

Unfortunately I could not help him in contacting Dan, but I did confirm his
opinion that this design concept should not be presented to the list in the
abscence of Dan, and that secrecy should be maintained until reaching Dan.
This was achieved not long after (when Dan got e-mail access again), and
since Dan approved (or seemed to) the result was that Peter went ahead and
told the list about this, inviting opinions and suggestions from all.
To this there was, however, surprisingly little response, until Dan's quite
shocking mail of some days ago.

Note that the only secrecy involved at all was absolutely necessary due to
Dan's lost e-mail account, and the necessity for Peter to get his approval.

As for my supposed secret conspiring and programming, it's pure fiction.
I did get some days start on you others in the list for absorbing information
about the STiK 2 design, and did in that time discuss future plans for code
optimization and such things, but no real programming was done by me for the
STiK 2 project in that short time.  Since then I have helped improve and
optimize the kernel's scheduler code, but that's about it so far.


>Like I said before the motivation you have is a good thing, but I am really
>unhappy with your methods and assumptions.  

I think that is mainly because of some false assumptions of your own, as I
hope you will agree now that you know more of the true story.


>I was unware of the fact that CSLIP was closely releated to PPP and so
>holding up CSLIP was a bad suggestion on my part.  I retract that statement.
>However I must agree with Flinny, we would probably have a PPP version STiK
>already if your energy had been channeled in that direction rather than
>dictating specs and coding in secret.

That final comment of yours is unwarranted, as you should realize from the
STiK 2 story given above.  You also oversimplify the ease with which anything
can be added to a non-modular program by someone not having the source of it.
It is the modular design that would enable STiK 2 to have functional PPP
before STiK 1, but that is now moot, since the other module authors seem to
have lost all interest in participating.


I am cutting all text below where you (Lonny) refer (yet again) to the
supposed secrecy of Peter.  Now you have the facts, and I certainly hope
that you too realize that the accusations are not true.  Under the given
circumstances no other sensible course of action was open for Peter.

-------------------------------------------------------------------------
Regards:  Ronald Andersson                     mailto:dlanor@oden.se
                                               http://www.oden.se/~dlanor
-------------------------------------------------------------------------
