[Uni-verse] Software release (and more)

[Eskil Steenberg]

Hi

We Simply feel very un comfortable the direction this part of the project
has taken.

-Implementing any verse host from scratch is hard and takes a lot of
resources.
-The project you are describing sounds very bloated.
-It sounds very hard to port.
-We know that multitreading wont give anything substantial in terms of
performance.
-It is re-doing a lot of work already done.

But frankly this is about trust. We havent seen any verse code properly
written form fraunhofer. There is almost no comunication, and the things
we pick up point to that you are not 100% in to the way of thinking (like
writing bloated applications instead of modulizing). We have no idea what
you are doing, you are not comunicating, sharing code or asking may
querstions. You have eaten enourmus resources and we have seen very little
besides a rendering engine that doesnt work and that was written by a
student.

We just dont think that you implementing a server from scratch is a very
good idea given how deep understanding of verse you need to do so. Therfor
we suggest that we add security to our server, and you write the folowing
tools as stand alone apps:

-A tool for taking statistics from tags and posting them as web pages.
-Modefy loade and save tools so that thay can be cotroled by Methodcalls.
-Graphics interface using methodcalls to admin server.

On top of this you should implement 2 plugins for tools, Photoshop, and
one 3D application (Z-bruch, lightwave or XSI) to make up for the lost
work, and other things tht seems to have been droped.

Your explanations for you plans for implementation of several apps have be
disssatesfactory, and hasent proved that you have a good understanding of
verse. The geometry reduction code, was the first one we had doubts about
and now it seems no one will us it. It dint work for verse and it dint
work for audio. But then again it was writen before the project even
started, still you fealt it was ok to charge us for it. We have similar
feelings towards the rendering engine that claims that your existing
rendering engine can just snap together with verse, having written a
rendering engine i konw this is not realy possible. And when you talk
about pre-prosessing data for the engine that tells me also that you are
on the wrong track. Next in line is the failiure to bring the light weight
rendering engine to R5. This is the only tool written that we have seen,
but even back in r4 days it worked so-so, and since no one is maintaining
it, it is for now dead in the water. Next is the global illumination that
also has several problems, again it doesnt seem to work for dynamic data,
at least we have seen no evidence of a strategy for solving this, only a
"we can trow this together quickly" statement. It allso assumes that all
objects have texture cordinates in verse, and that the uvs are never
overlaping, something that isnt nearly true in verse.

Therfor we want you only to do small well defined things that we know is
usefull. wich is why we dont want you to write a server, a rendering
engine or anything else like that. You should write tool plugins because
thats what we need, and you should write some smaller admin tools. Or you
can just give us the money to do it.

It bothers me that Emil and I wrote the loader/saver in a week (plus two
more weeks of emil testing and fixing). a tool that many in the comunity
and the project needs where as if we ask you you take 300.000 Euros and we
see nothing.

I cant tell if this is a problem with fraunhofer or just the management of
fraunhofer, but it sure is a problem.

Cheers

E