On Mon, Feb 19, 2007 at 04:10:28PM -0800, Ken Taylor wrote:
> Not just donors/investors, but I'd say you'll also want to attract new
> users, contributers, collaborators, and content-developers. It's still
> fundamentally an open-source project and can only benefit from more eyes :)
Yes! Absolutely. However the problem is that new users need to be
supported -- and my time right now is extremely tight, which is why I'm
working towards the goal of funding VOS development full-time, so that I
can devote the time required to get new users get up to speed. (Even
writing documentation to help people get up to speed is pretty time
consuming...) I am a strong believer in community, but that has to be
balanced against the pragmatic goals of actually making sure the
software gets developed :-) It's hard to build a user community when
the software isn't quite functional enough to be useful to anyone
beyond hard-core hackers.
On the topic of community, I should also mention there is a #vos irc
channel on irc.freenode.net that has been continously occupied for about
four years. Come visit!
> A lot of activity happened recently getting a requirements document together
> for "Interreality 3D" -- and I'm assuming this is driving the design of "s5"
> that you're working on. But I don't think this should mean that s4 dies or
> goes into stasis. I think it would be interesting to flesh out a
> roadmap/requirements kind of plan for the rest of s4's lifespan, until s5 is
> in a state where it can replace it. I guess what I'm saying is... is
The goal now for s4 will be to identify a set of demos that are
relatively easy to achive with the current codebase, and implement
those. This will serve as a prototype to try and spur interest that
will support development of the next version.
> http://interreality.org/cgi-bin/moinwiki/moin.cgi/VosRoadMap still accurate
> as to your plans, or does it need to be modified with the new s5 plan?
The roadmap isn't up to date. The s5 plan is still in progress,
although the intention for it to be a much more detailed development
schedule than a vague list of stuff we want to work on.
> As for what should go in it, here are my thoughts:
>
> S4 can act as the beacon of "hey, this is what we're doing, and this is what
> it's capable of" for the purpose of attracting the people listed above, and
Right. Some of the challenges are mainly in the design of A3DL which
are a separate issue from the underlying VOS design. We were planning
an overhaul of A3DL anyways, because the current design doesn't
accomodate personalized views of the world space.
> at the same time be a kind of test bed for ideas in order to refine the s5
> design. Obviously, s4 will have some limitations due to its design, and some
The three basic limitations that spurred the move towards the s5 design
are scalability, scripting and threading. Scalability is a big problem
because s4 stores all objects in memory, and not very efficiently at
that, so beyond 20,000-30,000 objects overhead becomes an issue. For
scripting, we decided that we didn't want scripts to be "second class
citizens" but rather would have the same (or better, due to being a
dynamic language) access to vobjects than you would get in C++. In
particular, being able to use scripts to attach behaviors and message
handlers brought up a slew of issues that seemed to be best addressed by
redesigning the architecture from scratch to support cross-language
method calls. Finally, the threading design is s4 is a terrible mess,
imposing a huge burden on the programmer to keep all his ducks (and
locks) in a row. The s5 design will significantly simplify that.
There's a some of other changes that are planned as well (mostly
redesigning things with the benefit of hindsight of what works and what
doesn't), but let me reassure you that the fundamental approach is still
by and large the same. One of the luxuries of open source compared to
commercial development is the ability to throw code that doesn't work so
well, so as to more quickly converge on the ideal design :-)
> things will have to be "thrown out" in the switch-over to s5, but there are
> still a lot of areas that can be worked on:
> - fleshing out a3dl
> - improving content-development tools/documentation
> - improving client gui and capabilities
> - writing server plugins that do "neat" things for demo purposes
Yes, I agree those are all valuable tasks. That's quite a lot of work,
though, so I hope you're volunteering to do some of it :-)
> And some things that can be worked on at the same time, but which aren't
> strictly s4 related:
> - better default demo worlds / included avatars / animations / sample
> content
Yes. One thing we need to think about is what the content creation
pipeline will look like. From my experience with the Gamebryo engine,
"real" modeling tools focus on making the trip between the modeler and
target 3D environment as sho