[Uni-verse] Review comments: discussion and actions

Emil Brink emil at obsession.se
Fri Jun 2 11:08:34 CEST 2006


Gert Svensson wrote:
> Dear Friends,
> 
> here is some excerpts from the review report and some discussion how 
> we should react on them:

Oh, good idea.

>> At this point of project progress, all remaining efforts should be
>>  targeted at demonstrating a significantly increased level of 
>> integration, a sound second evaluation case with 
>> quantative/qualitative measurements and the clear description and 
>> demonstration of benefits for users.
> 
> To speed up the integration a special meeting has been scheduled June
>  20-21.

Check.

>> An additional technical review meeting should be scheduled, as soon
>>  as a full integration can be demonstrated.
> 
> We should plan to do this in early fall.

It sounds like we have lots of freedom in picking a date, here. I guess
that's nice. Fall is after Siggraph, which means there is hope for lots
of (non-technical, but still) positive reports on how it went. :)

>> In addition, it is recommended to set-up an active and excellent 
>> quality open source oriented website that demonstrates the benefits
>>  of the project and motivates newcomers to join the development or 
>> adopt the developed technologies.
>> 
>> In addition, the website does not indicate a huge activity with 
>> potential users of Verse, thus discouraging most of them to join
>> the initiative. In addition, the website does not provide the basic
>>  services one can expect from a project targeting the open source 
>> domain. Both issues MUST be addressed urgently.
> 
> I agree that the web site has to be improved.  I, Peter Becker and 
> Niklas Forslund and Ton Roosendaal have scheduled an IRC meeting next
>  week to discuss this. All ideas are welcome!

Okay ... I think the complaints are still a bit vague, but of course
it could be better. It's a bit hard to require it to be "active", since
it cannot be that on its own.

I am not aware of many external users, so it's hard to provice, for
instance, a nice glossy list of links to real-world use-cases of a
decent size.

[...]
>> Provide guidelines and  thoughts, how other functional components
>> of the content creation pipeline such as AI could be integrated
 >> into a “versed” system.
> 
> A report should be developed by Eskil and Emil.

Oh? Fun! :) Actually, I don't know a lot about this off the top of my
head, and "AI" is a very vague component in a pipeline, in my opinion.

In my (limited, but still) experience with real game-development, AI
often requires extraneous information in the world, that tells it what
things "are" and what can happen where. I guess the layered way we
store geometry makes it easy to express that, but other than that ...

Text nodes can be used to store rules or even code that controls AI,
but I don't know if there's a ready-made "solution" for AI in Verse.

AI today is often implemented by 3:rd parties, in commercial middle-
ware that the game developer licenses. It would be nice to show how
the data (whatever it is) such a middleware package needs can be stored
inside Verse, but doing so is hard without actual access to the middle-
ware. And they're not generally freely available (or cheap). :/

[...]
>> The consortium is strongly encouraged to collaborate with the 
>> Khronos group and support the development of a Verse/Collada 
>> passerelle.
> 
> Who has contact with Khronos?

And who knows what "passerelle" means? I had to look it up, heh. Not
sure I found a working definition yet, either. I found this:
<http://www.doubletongued.org/index.php/citations/passerelle_1/>,
stating "It’s the jargon name given to the clause that allows the EU to
increase its power without a new treaty", which feels ... odd. I guess
my French is simply not good enough for this, heh.

There was an external project started to do ... something in the area
of Verse+Collada, but it failed for various (non-technical, afaik)
reasons, and the guy who initialized it has now abandoned the idea.

It's pretty vague how the consortium could "support" such a development;
should we hire new developers to do something, or just help an existing
project, for instance? Perhaps something for a thesis or so ... But time
is running out.

[...]

Regards,

/Emil


More information about the Uni-verse mailing list