[Uni-verse] PCC Meeting 20050905, log+notes

[Emil Brink]

Hey all.

Here are my notes from the (refreshingly high-speed) PCC meeting, and 
also the verbatim IRC log, for those of you who want it but didn't save 
it in your local client.

Enjoy, and any comments and/or corrections are of course welcome.

Regards,

/Emil
-------------- next part --------------
emil I'd suggest upgrading to patch versions as they become available: patch versions (such as r5p1, etc) don't change the API so should be safe
peter yes
LauriS ok
Gert OK
PLu y
Marcus yes, bugs should always be fixed....
emil is anyone still using r4?
Gert OK I find that the propsal is accepted
PLu not us
* emil has changed the topic to: Uni-Verse PCC Meeting -- Current Item: 3. Changes in the Project Plan
emil gert?
peter item 4
Gert We had a proposal there as well
Gert The Verse - Blender integration is added as an deliverable
Gert The FHG work is concentrated on making a Verse server with disk based storage and access control.
Gert Less work is directed towards the radiosity module and the the proxy reduction server is removed 
emil oh? it's item three to me, I joined the sota+market analyses to one :)
emil perhaps that's confusing, though
* emil has changed the topic to: Uni-Verse PCC Meeting -- Current Item: 4. Changes in the Project Plan
Gert The Verse - Blender intregartions we have already decided on a provious PCC
Marcus yes we will integrate a backup functionality into the server as well as a user management as far as verse supports it
eskil that could be far....
ton that would be cool
LauriS Marcus, Sascha: Is the proposal OK for you?
Marcus doing this we decided to drop the built-in reduction and the proxy functionality...
eskil are we cuting the radiosity module?
eskil we should.
Marcus the radiosity module will be developed anyway but with less effort
eskil so where does that efort go?
Marcus we re-use most of the radiosity calculations from another project
Marcus to the server implementations
eskil but server implementation was doen instead of proxy and reduction right?
Marcus user managment and disk storage are more effort than proxy and reduction-built-in
Gert Right as far as I understand
eskil i would not agree, removing so many things and only geting a server is not right.
eskil We need to get more done.
eskil I have a list of very needed things that doesnt cost much in time.
Gert A good server is really some work!
Marcus i agree ;)
eskil still i dont think it enough.
eskil I have written a server i know.
Marcus for the user management we have to update nearly all functions belonging to verse...
SaschaS in short: effort of dropped proxy functionality, build in reduction and shrinked radiosity goes to user management and disk based server now
eskil A server brings a level of stability to the project for people who already use verse, but it doesnt bring in new users. and we need to focus on that.
eskil a photo shop plugin would do that, for instance.
PLu I think it will be a gigant leap forward to be able to have persestent models, we need this to be able to build the demonstrators. It has been a great hasle to not been able to load geometry into verse
Gert I'm not so sure! Stability is very important to keep the users that try verse
Gert happy.
eskil so we already have persistant models. I have written a saver that can dump the entire state every 10 seconds, and emil has written a loader, so thats not realy any new functionality.
PLu where do I find it?
eskil the verse server works great for first time users. the only time you would need it is when having very large projects.
eskil Elmil?
eskil Emil?
emil the VML tools are, for silly technical reasons, a bit ... under-released, at the moment
PLu :-)
emil I guess I can commit what I have to CVS though, it's mainly some issues with code-sharing that have kept me
eskil Im not saying a disk suported verse server is a bad thing, im saying we need to get the most possible out of the funds we have.
emil (the loader uses parts of Purple, but is an unrelated project. you can't symlink across CVS repositories, and I don't want to make an officia "library" out of Purple's utility code)
emil make that "CVS modules"
eskil I have been testing the vml stuff this week end, and i have hundereds of objects and maybe 50 geometrys and it works.
emil the saver (which Eskil wrote) uses Enough, and right now lives in the Quel  Solaar module, it's been committed for quite a while
eskil Things we realy need, and that would push verse forward: Photoshop plug, XSI plug, Purple interface, RIB exporter... all thees would take less then one man year and would give good return.
asch_minusplus i think we need a stabile system that can be worked on
PLu The loader and saver uses some private format i guess?
emil PLu: yes, a simple XML-based format called "VML"
emil pretty much a 1:1 mapping of the Verse data model to xml, really nothing fancy
emil rather ... costly for e.g. bitmaps, heh
Gert The sort of open question in the project plan is the game/lego like demonstrator
Gert The last word is that Martin likes to continue this idea
emil eskil: do you mean they would take less than a man-year *together* ?
emil Gert: he does? heheeh
eskil emil: no
emil eskil: ah
Gert maybe change some of the lego idea but still do a similar game
Gert That is right Peter?
eskil Photoshop: maybe 3-6, Rib: max 3, XSI maybe 9 - 12, purple interface 6 - 9?
peter Gert: yes, thats is part of the necro-verse integration plan
* Kaito (n=ton at nivm250.xs4all.nl) has joined #uni-verse
peter the step after intergration is doing some game demos
Kaito sorry, wifi was reset here
eskil I think the lego ide should be droped, too much work that cant be re used by other, and isnt a particular strength of verse.
Gert I think most of the work will be done by Martins students.
peter I don't really think there will be a lego demo, it will change, but can't at the moment specify exaclty how
eskil so what, make them do something usefull instead.
* ton has quit (Read error: 104 (Connection reset by peer))
emil networking seems stable in Ton's end, heh
eskil too bad.
LauriS Are these two demonstrators really able to show the potential of Verse and the work done in the project?
eskil we love him so....
Kaito today is first day in orange studio, we're setting things up :)
emil Kaito: ah, right
Gert I was also inclined to skip the game demo but two things make me change my mind
* Kaito is now known as ton
Gert 1) Martin likes to keep it
eskil we should have too demonstrators, one game content creation demo and one architect demo.
Gert 2) There is no other demonstrator to replace it with
eskil I can provide the data for the game content creation demo, then that is done and we can spend te money on something better.
LauriS Martin's demo != game content creation demo ?
Gert 3) also it tested some part of the system whoch not will be tested otherwise
Gert Lauri: martin's demo is using verse to run the game
asch_minusplus it would be nice to have something common in the game and arch demo
peter As far as I understand, it will be a game demo more of plug-in type, that is how an existing game is verse-enabled and this way demoing dynamic capabilities of verse
eskil but in wont be fully versifyed.
eskil A content creation demo is so much more afit to A what we are/should do, B what people are most likly to use verse for.
Gert What did Eskil mean by "I can provide the data ..."
eskil it means that, a content creation demo is free in terms of development time.
peter Are there any more suggestions - and resources - for demos
peter I would like to see hte movie project as a kind of demo of content production line changing through verse
eskil "hte"?
emil "the" :)
eskil you are talking about orange? thats up to ton.
eskil that would be an other possibility for a "free" demonstrator.
emil free for us, at least :)
eskil well obviusly a content creation demo int fee either, ist just free to the project.
Gert Orange would indeed be an interesting demonstrator
emil I spoke to the Orange technical lead ("antont") last week, he sounded positive but hasn't looked into Verse much yet. we'll talk again later this week. and of course, it's up to them to decide
LauriS Orange?
emil LauriS: orange.blender.org :)
Gert orange = movie
>Gert< vore det omöjligt att erbjuda rendering på något kluster här, som morot för att få med orange?
peter I suggest we stick to our architecture and game demos, and blender movie on hope-to-see llist
aplohman I think this is best 
Gert Sounds OK for me
eskil i dont like the game demo at all, its going to take up too many resources.
Gert Hopefulle mostly students
eskil it will just become something that no on can ever get running on thir machine because it will need 5 apps to set up correctly, and it will take alot of time.
emil I'd rather make a Photoshop plug-in than a lego physics engine :)
eskil I think its bad to say let some students do it. I will take alot of resources, and will not be very impressive.
eskil This demo will not show case anything that hasent been done in games years ago, and wont look nerly as good.
SaschaS perhaps this should be discussed with martin, when he is back online?
peter eskil: no new resources, actually it will be done mainly under the wp5.4 framework
peter and yes, saschas prop good
eskil Yes he should be here, but that doesnt stop us from recognizing that this is a bad idea.
eskil doessent matter what frime work, it will take up alot of time, time that could have been used on better things.
Gert But we have to replece it with some other demonstrator.
eskil Yes: game content creation!
peter the point is, there are no big resource saving possible here, since this is bonus woork from students
eskil its a demostrator thert the reviewers even asked for.
eskil that is not how it was budgeted form the start.
Gert Eskil: who is doing the game content creation?
eskil Emil is sceduled to work on this for over a year.
LauriS But if that doesn't consume any project resources, is it really the best demonstration we can show?
eskil Gert: I am, i can send you the content today.
ton we will check on verse as communuty based cooperation tool, but dont intend to have it as part of universe mentioned
ton we did all work & financing for it ourselves now, if uni-verse likes to be official involved they should sponsor it :)
eskil In the year to the end of this project, my guess is that emil could and would like to do maybe even two of the things ive liste as alternative apps.
emil sure I would, if it were up to me :)
LauriS Instead of something else, what?
Gert No one has expected Emil to spend a year on the game demo
eskil we need to make some changes, We need to focus our project, the reviewers where very clear about this.
eskil Thats what he is sceduled to do.
ton but; for our own validation & testing part we will mention our movie project yes
emil Gert: I can probably spend a year doing the web-plugin for Verse, mentioned in (more or less) the first sentence of the game demo writeup, heh
peter the kth resources in wp 8, are 5 mmonths, in total for evaluation incl architeture and "digital media"=game and others
eskil In order to fucus, we need to find our strengths and fous everything on that, and cut away as much other stuff as possible.
eskil That is why Froun should be doing something more then the server, and Kth should do a few more tools.
Gert It feels that it is hard to take any decsions on the game demo today
emil yes, next item perhaps
peter i suggest we put game demo on hold, for a new proposal after discussion where also martin is involved,
Gert There are to many unknow factors:
peter but there are no huge resources planned here
Gert Martin not here
Gert Discussion with orange on going
eskil NO, DONT PUSH THIS IN TO THE FUTURE; WE NEED TO GET ON WHIT THE PROJECT TODAY!
emil the demo text, last time I read it, *sounds* like it requires rather large resources
eskil It will.
Gert Eskil has some ideas about a game contect creation demo which is not specified
eskil We need to make changes and we need to make them today, we should have done then 6 months ago.
emil Gert: eskil is doing a game in his own time, which is (I guess) why it would be easy for him to put together such a demo
Gert Agree that it would be nice to decide today
eskil Listen, im working 20% at KTH form now on , the other 80% i will work on a content creation pipeline, that that if you needit, its much better then spending time on something we dont know hos much efort its going to be.
eskil I already have such a demo.
Gert it's not the demo as such that is valuable it is more the experince from real users
Gert ising the demo
eskil if we want real users, we should defenetly do tools.
peter Gert: That is why a movie project in blender community is sucha good demo
Gert Peter: I agree that would be an excellent demo but obviously it required some more
emil um, could we perhaps do the remaining two items so I can take the notes, and the go back to this discussion? that would sort of free me up a little, I might need to run out on an errand at 12 (CET) if at all possible
ton i doubt you can call it 'demo' since in best case verse is only used in the production pipeline, you cant demo that really\
Gert discussion with Blender which cannot be done in this session
Gert demo is maybe not the right word, evaluation is better
eskil ton: well that is what people are goin to use it for, Our main customer are the users of blender and other content creation tools.
emil eskil: aren't you going to talk at GDC about game pipes? or is that just submitted, not accepted?
eskil Submitten not accepted yes.
emil right
eskil a talk like that would be a big difference if we had say a photoshop plug.
emil yup
eskil The bigst thing that has happend to verse the last year in my mind is jiris work on blender-verse integration, it hasent even been properly anounced yet and still we are getting new users from it.
Gert Are you not talking of principals of the pipeline?
eskil The max and maya plug, will probably do the same.
eskil Im saying that thees people are the once who will find and use verse, and thats why we sould A mak tools for them, B demosntrate how they can make use of it.
ton the google projects have deadlines closed now, we are going to have nice publicity for it
ton i think its cooler to do them all together
ton the fluid simul is awesome :)
eskil Im not complaining that it hasent been released yet, that up to you an jiri, im just saying what an impact it has had.
peter Are we not talking of two levels of demos, the one demonstrating the capabilities ofuni-verse  tools, like maya plug-in, audio, dynamic mesh or other, adressed to developer, the other of integrating into application adressed to more of end users, like architects
Gert Peter:
Gert we are talking about two levels
eskil no they are all demos for content creator, maya plug is not for programmers its for users.
eskil (i call architects content creators in this statement)
Gert In the WPs themselves there are of course demos to demonstrate the individual pieces
Gert The WP8 stuff is really larger demos covering several WP:s
eskil I think fraunhofer should pick one tool more to do and emil should doo maybe 2 more, and the audio team should make sure that there are some audio management tools. Then we can make this project come together.
ton is paregos still going to coordinate 8?
ton part of that is also preparing everything for online demos & releases
Gert Yes paregos will coordinate WP8
LauriS So, what is their proposal?
Gert No one from Paragos is present today
LauriS You got the point!
PLu :-(
emil it costs $195 to get access to the photoshop sdk, it seems. *sigh*
ton for the website we would really like to see 1 coordinator for all of the releases and demonstrators
ton gert can do that too of course :)
eskil Emil -ith a budget of 3.8 milion we should be able to do that....
PLu We need to start the process of finalizing the spec for the achitecture demo
Gert Ton; What do you mean?
emil eskil: in a perfect world, and so on
Gert Ton: what releases are you talking about?
eskil are peregos qualifyed to make a release? isnt that something that we need people like emil to do?
ton all deliverables actually
Gert That is not paregos work 
Gert They lead the evaluation
peter yes, and all partners have part in wp 8, kth 5 mmonths, FIDG 4, II 5 etc
Gert This has been an interesting discussion but I no consensus seems to be possible at the moment
eskil we cant stop, 
Gert I think some more discussion is needed with Martin and Ton abou the best way
eskil This must be resolved today. 
eskil I dont care, it he is not here thats bad, but wen cant stop the project because of that.
ton what is 'this'
eskil We must come to the conclution of:
eskil A: are the two demonstrators
eskil B: what are the applications we are going to spend our remaining time on developing.
eskil I refuse to have this pushed forward. The rewiewers said they would kill the project if we didnt do this, and if we dont get started that is what is going to happen.
Gert Obviously we need two demonstrators.
eskil Emil and others are almost siting on their hands, because we cant come to a dessision on what to do. that is unacceptable.
ton eskil; we can decide without proposal what to decide
Gert On the applications we are of course allowed to develop more than promised but it is really no reason 
ton cant!
eskil And i am saying we can have two demonstrators, the arcitectiual one that we have always planed and a content creation one, it can contain either Orange, or my stuff or both.
Gert to do that it seems hard to fullfill what we alreay has promised
eskil Gert: that is not the issiue, the issiue is what are we doing with in the project.
ton I also miss the clue what a 'demonstrator' is really for. we can also call the upcoming blender-verse a demonstrator
SaschaS It seems that this point canŽt be resolved now. Perhaps we should shift this. Might it be possible switch to the next agenda point please? I think emil (and maybe others as well) have to leave at 12:00?
emil SaschaS: thanks
eskil We should keep at this all day untill we get it resolved.
* emil has changed the topic to: Uni-Verse PCC Meeting -- Current Item: 5. Reminder of mid term internal review ?? ?? ?? ?? ??
PLu I like to see a written proposal
eskil We should have taken thees desissions in germany. so we are already too late.
Gert I will try to come with a written proposal within a weeks time
eskil There are about amilion written proposals and we have had monts amnd monts to do rewiews and stuff. time is up.
Gert we can decide about it over email.
eskil no lets do it now when every one is here!
Gert Mid term review:
* emil has changed the topic to: Uni-Verse PCC Meeting -- Current Item: 5. Reminder of mid term internal review
eskil why bother with a review if we are not going to do anything?
Marcus we uploaded the report of our work already to the web page
Gert For the mid term review I will nead a status report from each WP as we did in the first review
Marcus comments are welcome...
Gert Marcus: Great!
aplohman for WP9 I really need updates from everyone on the dissemination activities undertaken
aplohman so far I have only received info from Lauri
PLu when is the dead-line ?
Gert Dead line Sept 15
Gert Next point
SaschaS wasnŽt it 01.08.?
Gert No ose has followed that so we need some time to get it done
* emil has changed the topic to: Uni-Verse PCC Meeting -- Current Item: 6. Next Meeting
Gert For the next meeting 5-7 oktober I and Peter would suggest to have it in Stockholm
LauriS If you come to Finland, you will hear an audio demo :-)
Gert to save budget for the entire project, Most person are located in STockholm
LauriS But I'm happy to travel to Stockholm
Marcus no problem for us... ;)
Gert Great!
emil yeah, great ... should you continue with the demontrators talk, or what?
Gert Thanks for today!
Marcus we have to leave now....
eskil yes
* emil has changed the topic to: Uni-Verse PCC Meeting -- Current Item: Closing?
Gert Sorru for the long discussion on the game stuff
emil okay, bye to those who have to leave, I'd very much like to sneak out myself, heh
eskil hey we didnt get aconclution on anything!
Gert we really have to solve that really soon now
PLu LauriS: think we can set up a demo in Stockholm too
eskil how about now!!
eskil dod we even talk about deliverables?
Marcus i think the better way for discussion is mail ... at least for the missing time right now....
eskil did we even talk about deliverables?
LauriS PLu: join #uni-verse-audio and we can continue over there
emil cool! more irc power :)
eskil we dont get any conclution there either. jut passing docs back and fourth.
Gert Thanks everyone!
* Gert (i=Gert at c-2015e255.1325-1-64736c10.cust.bredbandsbolaget.se) has left #uni-verse
* peter has quit ("Chatzilla 0.9.68.5.1 [Firefox 1.0.6/20050717]")
-------------- next part --------------


1+2. State of State of the Art Review & Market Analyses [Angela, Ton]
	Still some documents missing. All of the market analyses done (need to
	restructure a bit), but	SotA docs are still in progress.
	Is something coming from KTH?
	Eskil: having problems with mail. Working on something.
	Plan is to include SotAs in D9.1, deadline is mid-September. SotA
	from II still missing?
	Peter: part of audio SotA done, but the 5.4 is still missing.
	Eskil: won't write "proper" SotA, but motivate why that's hard.
	Getting these done would be great.
	Peter: Martin still in Cube, having tech. problems. But it should be
	done soon. Market analysis will be revised, post-Lauri.

	Requested final changes to market analyses this week.
	Akos: comments on arch. analysis welcome.

	Plan to have branding things finished this week, by Friday, for web
	use.

3. Verse version to use [Gert]
	Gert & Peter suggests: PCC decides, and currently decides version 5.
	Some discussion here, but mostly agreement.
	Emil suggested people to upgrade to patch versions as they become
	available, since doing so should be safe; patch versions don't change
	the API.
	The proposal was accepted.

4. Changes in the Project Plan [Gert]
	* Verse-Blender integrations were decided on a previous PCC meeting.
	FHG will integrate backup functionality into the server, and user
	management as far as Verse supports it.
	Eskil wondered if the radiosity module is to be dropped.
	Marcus: it will be developed anyway, but with less effort. The freed
	up resources will go to the server implementations. User management and
	disk storage is more work than the previously planned proxy and geometry
	reduction.
	Eskil argued that this doesn't represent a good tradeoff, and that it
	would be better to spend resources on efforts that get new users into
	Verse, rather than stabilizing for existing users with a new server.
	Gert disagreed, arguing that stability is important to make people who
	do try Verse happy.
	Peter Lundén argued that persistance and being able to load things in
	would be very convenient to have. Eskil replied that we already have
	those, through the VML tools.
	Eskil suggested other deliverables that he felt would have more of an
	impact on the project: photoshop plug-in, xsi plug-in, purple interface,
	RIB exporter. All of which he claimed represented less than one man-year
	of effort, each (Photoshop: 3-6 mo, RIB: <= 3, XSI: 9-12, Purple: 6-9).
	Gert: Martin still likes the Lego-game idea, as part of the Necro3D
	integration plan. Next step after integration is doing some game demos,
	which is where it would fit in.
	Eskil felt the Lego idea should be dropped: too much work that can't
	be re-used, and not illustrating a strength of Verse.
	Gert was inclined to skip it, but changed his mind because of two
	things: 1) Martin likes and wants to keep it, and 2) there is no other
	demo to replace it with. Also, 3), it tests parts of the system that
	will not be tested otherwise.
	Eskil wanted game content production demo, and Lauri asked if the Lego
	demo isn't it. It's not, as Gert explained, since in the Lego demo, Verse
	is used to run the game itself. Eskil felt content creation demo is a
	better fit to what we "are/should" do, and what people are most likely
	to use Verse for. Eskil offered to provide data suitable for a game
	content creation demo.
	Peter raised the issue of using Orange as a demo of how a content production
	line can be changed trough Verse. After some talk, he then suggested we
	stick to our architecture and game demos, and have Orange on the list of
	hopeful possibilities.
	Again, Eskil raised objections that the game demo won't look competitive
	(or even comparable) to games from years ago.
	There was talk about whether or not the game demonstrator wastes resources
	that could be put to better use.
	The discussion continued about whether or not we should do something more
	"focused", or stay with the game demo as planned. No consensus reached.
	Gert said he will come with a written proposal within a few weeks time,
	Eskil opposed and wanted to continue the discussion until resolved. We
	didn't, meeting dissolved around 12:00 CET.

5. Mid-term review [Gert]
	Gert needs a status report from each WP, as we did in the first review.
	Marcus: report already uploaded.
	Angela: updates from everyone on dissemination activities taken, for WP9.
	These are needed by September 15.
	Marcus wondered if the deadline wasn't 1/8, but Gert said that since too
	few had delivered, more time was needed.

6. Next Meeting [Gert]
	Gert and Peter suggest October 5-7, in Stockholm.
	Lauri promised audio demo if we come to Finland. Gert replied that Sthlm
	is the cheapest option, since most of the people are there already. There
	were no protests.