What is happening and where are we heading? - was "RE: Hello?"

Hello Burcin,
Your points are certainly valid, so I feel I need to explain why a few
things have been done as they have been done.
It was always the idea that Zynot and also Xeta should be open and community
driven projects. Right now it does not seem very open, I agree. Once the
fork was announced we got "slashdotted" in a way and there was a huge wave
of initial interest. Many wiki pages were made, much talk was made in forums
and on mailing lists where people discussed what could be cool to do and
what they would like. 
Now, looking back at all that; there was no controlling organ in Zynot back
then to take this wave of creative force and drive it in one single
direction. This was imo a forgivable mistake since noone in advance could
have predicted what dimensions it would take. So creativity became
confusion, confusion became frustration, and frustration became lost
momentum. I know this well because I also felt it myself. Yet, as I was put
on the board of directors I felt obliged to stay. The board eventually
granted me right to call myself CEO here and thus leader of all staff for
the time being.
When I was granted this authority it was time to start picking up some
pieces. At that time the number of idlers in #zynot had already shrinked to
about 30 so there wasn't a lot left.
The first thing I did was to gather a few devs for the Xeta project. We
decided to dump a portage based distribution because it became apparent that
having Xeta support ebuilds would not be worthwhile for us. Besides,
speaking for myself, I had a desire to shed the Gentoo-is-our-big-brother
image and make Zynot less of a fork. This desire came because of some
less-than-nice things being said about us in #gentoo-dev while even the
present leadership did nothing to stop it, and because of letters sent to
our attorney that at least imo seemed more like harassment than an actual
wish to rectify some things. All this is hopefully behind us now, but also
today it seems to me that breaking lose from gentoo was the right thing to
do.

Seeing that the community interest had vaned rather fast I instead asked a
member of the Xeta team to go through the archives of all mailing lists and
look at all forum threads and basically find any idea and suggestion that
was made as to the functionality of a Portage replacement. This uncovered a
rather large amount of possible and impossible ideas and wishes and we wrote
them all down and considered them.

This leads me to a key point here: The development of Xeta might right now
seem closed, but it is almost exclusively being built on wishes and ideas
put forth in the community in the past.

This fact actually makes me worry quite a bit less about the community
momentum at this moment. When the community gets what they ask for they will
use it, right? Build it and they will come.
We gathered all the ideas in a requirements specification. You can see the
latest version here:
http://jupiter.zynot.org/~zllow/RequirementsSpecification/x104.html
Whole Xeta team uses any opportunity to spam this link around. Even so, the
feedback on it has been limited. Either people are satisfied or they just
don't want to read long requirement specifications. I don't know :)

Your question about why don't we open up and let people help is a good one.
Since we were in a situation where we started from scratch there just has
been a lot of things where we haven't been ready to accept community
contributions. Logically, we need a functional package manager before we can
accept packages for instance. Strangely enough volunteers tend to turn up at
the wrong time always, but I do keep an archive of email addresses here and
I do intend to send out "Now YOU can contribute too" kinds of emails out
directly to the people whom I talked to, but where it just wasn't the right
time yet. :)
Reason number 2 is that most of the work done on Xeta so far has been
analysis and modeling. If you are good at this we would love to hear from
you still. Lots of open source projects have suffered from a code first and
architect later syndrome and it always comes back and kicks them in the
butt. We have been doing some crucial foundation work to avoid this mistake
and perhaps even more importantly, this work will put you and anyone else in
a much better position to contribute.

So what kind of help can we actually use right now?
Well, if you are itching to contribute then we need the following profiles:
1) C++ code monkeys. Should at least have some familiarity with STL and
libxml2 or be fast at learning them. Look at the requirements spec and
especially at the items tagged as "core". Those are what we need implemented
right now. Also check the XBuild specification and read the DTD. This is
essentially what Xeta must parse.
2) UML experts. Would you like to put our requirements spec and architecture
into UML? Give us a shout then.
3) First line testers. Be very aware that you need to be a patient soul if
you volunteer to test the very first versions of Xeta. You must not mind if
your system gets hosed cause hosed is what it most likely will get. Also,
you will be extra ideal as first line tester if you have some obscure arch
that you can test the Xeta code on.
4) XBuilders. If you are just as brave as the first line testers then you
could be suitable for this job. In about a month there will be an IDE ready
for assisting in making XBuilds. An XBuild is our version of what is called
ebuilds in gentoo speak. You can help by testing this tool, point out all
its shortcomings, and assist in making a basic tree for what will eventually
be the Zynot Linux distribution. Basically what you will be doing is testing
and dry-running the XBuild format before we freeze it. Be aware that you
will be using only an IDE. You wont be able to actually use the XBuilds it
produces yet, but believe me, it would be a great help and a noteworthy
contribution to Xeta team if you actually sat down and did it still.

If you are interested in any of this, or if you want to end up in my email
archive as future contributor then do not hesitate to either mail me, or
jump on Freenode IRC and talk to theDevil- 

A Xeta that can download, compile and install software for you should be
ready in Q1 next year. It will be alpha-ish and limited to the features
tagged with "core" on the requirements specification, but as soon as we have
this first executable then the rest of the work in setting up the
distribution finally starts for good. I hope that we can present an alpha
distribution also in Q1 next year. The purpose of this alpha distribution
definitely wont be enterprise stability, but rather it will give everyone a
chance to install and play and ask themselves the important questions "What
am I not happy with here? How can I make this better?"
It is from this point that I expect openness and contributions to happen and
accelerate. I am quite confident that the momentum will come back in full
force by that time.

The infrastructure we set up initially is still standing. Only one server
got lost along the way. Our infrastructure is ready to handle a rather large
amount of users rather fast. Also the more "soft" infrastructure is about
ready. We realize that with growth then management patterns and organization
structure also changes. But really this has all been planned already and
should not pose any problems.

As for Zynot in the future and the stability you are looking after. This
actually hasn't been totally decided yet. Right now I imagine we will do
community releases at something like every 6 months, and on top of that
there will be some subscription based enterprise releases every 18 months.
If we do it this way, the enterprise releases will only be open to active
contributors to the enterprise branch, or to people who want to pay for
access. This is much the same way as RedHat/Fedora works right now. However,
let me make it very clear: It has not been decided upon yet.
The scope and purpose of Zynot Linux and the technologies that powers it is
right now first and foremost: To be a good platform for cross development
that is suitable for easily and rapidly creating distribution images for
embedded devices, and secondly, to be a distribution that you can put in a
datacenter without having to pray and without spending tons of time on
nursing it. Furthermore, in community interest, Zynot Linux will also run as
a desktop. While desktop is not the main focus and is not where we will
spend the main bit of our energy, it is still important from a popularity
and educational point of view. We acknowledge that many admins want to use
the same software on their servers as on their desktop at home. If nothing
else, to reduce the daily complexity they are faced with. It has been said
that "if you own the desktop, then you own the corporation" and this has
certainly held true in many aspects. This partly explains how very
desktop-centric distributions such as Mandrake and Gentoo find their way
into the corporate datacenter.

I hope I have addressed all your questions and causes for concern. If you or
anyone have comments you are very welcome. 

Kind Regards,
Frantz Dhin
CEO
Zynot Foundation


-----Original Message-----

Is the codebase available anywhere? I'd like to do whatever I can to help.

I've been waiting too long to have a stable gentoo that I can install on
servers. Hearing the announcement of the fork made me think that dream was
really close, but unfortunately the excitement didn't last long.

I feel a great need for a "tame" portage. While the contributions done by
the gentoo community are great, I don't see much progress happening in terms
of having a portage tree that doesn't change so quickly. That behaviour
cannot be tolerated on servers, and I have a hard time keeping up with it
even on my laptop. 

I've spent a lot of time recently, to get portage to behave well so I can
rely on it for certain tasks. I could easily have spent the time on the xeta
package manager. My work is nowhere near satisfactory, and instead of
wasting more time on that I could work on xeta, if only the development
process was open. The main reason that I got frustrated with gentoo was
exactly that, development of main system components wasn't open enough, and
now zynot is going that way too?


It's a great shame that this project lost the initial momentum that came
with the first announcement. It would have been great if a release was made
in those first days. Even if it was all gentoo material, it would have given
people a base to work with and improve. This is no longer a fork now, we
seem to be starting from scratch and that's a lot harder.

Why don't you open up what's being done, update the wiki, and let people
help?

What happened to that subversion server and all the infrastructure that had
been setup?


Burcin

PS. I've been having serious email trouble in the last 10 days or so, and I
can't follow the mailing list, let alone keep track of personal emails.
Please excuse me if I missed anything that happened during that time.