General technical and community questions about Rei

9 views
Skip to first unread message

justincc

unread,
Oct 20, 2009, 8:10:28 AM10/20/09
to 3di-viewer-rei-dev
Hi guys. I just gave Rei a spin and it looks nice - congratulations
on the project so far! I think that viewers embedded in webpages are
going to be very important in the future.

I do have a few broad technical and philosophical questions about the
project, so thanks in advance for your patience in looking through
these :) Here goes

1. Is the plugin supported in Firefox through something like NPAPI
rather than ActiveX? Does this mean that different browsers have to
be directed to different plugin pages?

2. Is 3di planning on spending development resources of their own to
implement Second Life avatars or textures or will support for this
come from community plugins?

3. Along the same lines as above, will 3di be implementing other
features found in Linden Lab's SL viewer (e.g. groups, inventory,
etc.) or will this again come from community plugins?

4. Is there any timeframe at all on Linux and Mac support? I
appreciate that it may be impossible to say :)

5. How does 3di intend to sustain the Rei project? Is it just
through cross-subsidy from sales/support of the 3di enterprise server
or are there other business models planned (e.g. support of the viewer
with vanilla OpenSim). Again, I appreciate that it might not be
possible to answer this :)

6. If Rei gathers an open-source community, will 3di give trusted
non-3di developers access to the git master branch and a voice in
project decisions?

Best Regards,

--
justincc
Justin Clark-Casey
http://justincc.org

Johan Berntsson

unread,
Oct 20, 2009, 9:45:17 PM10/20/09
to 3di-viewer-rei-dev
Hi Justin,

We just had a discussion that touches on several of your questions, so
let me give a general reply on those points in order to get a
discussion started.

We're hoping for this viewer to be an open source project in its own
right, and not just a sales tool for 3Di OpenSim enterprise, so 3Di
will participate but not mandate the future development of the code.
Perhaps the main obstacle for mainstream adoption of the viewer is the
missing SL avatar and texture support, and we would welcome and
support any help in getting this implemented. Unfortunately we don't
know too much about the SL formats in-house, so we are currently
looking for more information. Any help here would be appreciated.

Groups and inventory control are planned for the future, but things
will of course move more quickly if more people get involved.

The viewer is based on multiplatform tools so it should be fairly easy
to port to Linux and Mac.

Best regards,
Johan

Norman Lin

unread,
Oct 20, 2009, 10:19:14 PM10/20/09
to 3di-viewe...@googlegroups.com
Hi Justin,

justincc schrieb:

> Hi guys. I just gave Rei a spin and it looks nice - congratulations
> on the project so far! I think that viewers embedded in webpages are
> going to be very important in the future.

Thanks for the encouragement and glad to hear your interest!

>
> I do have a few broad technical and philosophical questions about the
> project, so thanks in advance for your patience in looking through
> these :) Here goes
>
> 1. Is the plugin supported in Firefox through something like NPAPI
> rather than ActiveX? Does this mean that different browsers have to
> be directed to different plugin pages?
>

Yes, separate plugins are needed for IE and Firefox. The Firefox plugin
is implemented using the XULRunner SDK.

> 2. Is 3di planning on spending development resources of their own to
> implement Second Life avatars or textures or will support for this
> come from community plugins?
>

This will probably depend on community support to get realized. Rei was
at first developed for use cases needing 3D mesh-based avatars generated
with external 3D modeling tools. However, we realize that for general
community use, existing users will want to use their existing SL-style
avatar appearance. We don't know much about the technical implementation
of SL avatars. We've started to look for some information about what it
would take to get implemented and plan to document this on our wiki, but
again, community support would be needed to really get this rolling.

> 3. Along the same lines as above, will 3di be implementing other
> features found in Linden Lab's SL viewer (e.g. groups, inventory,
> etc.) or will this again come from community plugins?
>

This too will likely come from community plugins, though 3Di as a
community member may develop those plugins. :) By this I mean that our
business use cases don't require inventory or groups right at this
moment, but our open source resources can potentially be utilized to
realize these features, but with lower priority. In the end, as with
much open source development, what will get implemented depends on the
time and interests of the individual developers.

As a technical note, Rei can interact with elements on the web page via
JavaScript. This means for example that the GUI for inventory could be
implemented in the browser.

> 4. Is there any timeframe at all on Linux and Mac support? I
> appreciate that it may be impossible to say :)

As a proof of concept it has been possible to compile and start a
standalone version of the viewer in Ubuntu Linux. It runs rather
slowly. However, this at least demonstrates that the core components are
cross-platform. Getting it into the browser is another issue that needs
further investigation, though. Zaki, could you comment more on this?

> 5. How does 3di intend to sustain the Rei project? Is it just
> through cross-subsidy from sales/support of the 3di enterprise server
> or are there other business models planned (e.g. support of the viewer
> with vanilla OpenSim). Again, I appreciate that it might not be
> possible to answer this :)

To sustain the project, we intend to continue developing value-added,
QA-checked features as proprietary viewer plugins on top of the BSD
viewer core. These proprietary plugins can support our business use
cases for our software sales and for our SaaS/PaaS service offerings.
Our vision is that Rei can become a core (BSD) + plugins (any license)
model like OpenSim itself, with anyone being able to use the core as-is
or build on the core for any purpose.

> 6. If Rei gathers an open-source community, will 3di give trusted
> non-3di developers access to the git master branch and a voice in
> project decisions?
>

As a basic rule, yes, 3Di will give trusted non-3Di developers access to
the git master branch and a voice in project decisions. This commit
access will be given out based on useful patches and ability to work
with the existing developers to improve the project.

Regarding code copyright issues, following the guidelines at
http://producingoss.com/en/copyright-assignment.html, 3Di asks for
contributors to Rei to submit a contributor license agreement before
their contributions can be accepted into the master branch. The
contributor agreement is identical to the Sun Contributor Agreement used
by Sun for their open source projects and is based on a joint-ownership
structure.

Best regards,
-Norman

justincc

unread,
Oct 22, 2009, 2:55:32 PM10/22/09
to 3di-viewer-rei-dev
Johan, Norman. Thanks very much for those answers, they're very
informative. For me, this is mostly for future reference at the
moment but it's very good to see that 3di is open to the community and
committed to sustaining Rei.

All the best,

Justin
> Regarding code copyright issues, following the guidelines athttp://producingoss.com/en/copyright-assignment.html, 3Di asks for

Dezso Zoltan

unread,
Oct 23, 2009, 3:35:25 AM10/23/09
to 3di-viewe...@googlegroups.com
Hi, Justin

1.  Is the plugin supported in Firefox through something like NPAPI
rather than ActiveX?  Does this mean that different browsers have to
be directed to different plugin pages?

Yes, Firefox is supported through NPAPI/NSAPI, however there is a javascript library in bin/ovwjs, that makes sure that you don't need to have different pages for different browsers. As we add supported browsers in the future, this library will also be updated if necessary to keep a one-page entry.

2.  Is 3di planning on spending development resources of their own to
implement Second Life avatars or textures or will support for this
come from community plugins?

Personally I would prefer to see LL avatar support in the core package (we have texture support in there) rather than in plugins, and if other developers could help with that, it would be awesome. We added mesh avatar support for two reasons: to showcase what is possible when you can freely control the server and the viewer and because our customers also wanted that feature. As the community wants LL avatar support, at some point, someone will create it :)

3.  Along the same lines as above, will 3di be implementing other
features found in Linden Lab's SL viewer (e.g. groups, inventory,
etc.) or will this again come from community plugins?

Yes*. Groups is also possible with JS, inventory is a little more complicated than that, but we love new features :)
* if by 3di you mean the community around Rei - "including but not limited to 3di" (hehe)

4.  Is there any timeframe at all on Linux and Mac support?  I
appreciate that it may be impossible to say :)

I started some work towards linux support, but realized, that rebasing to Irrlicht 1.6 will solve a couple of problems (mac support is now included for example), so I was doing that first. It took a little time to make it even compile, let alone run, but finally I pushed the first results to github. The rebase is not complete yet, but you can get an idea of what it would do. I talked with paupaw and he said he was very interested in adding some new features to .netcp, starting with the new lightmanager, so I guess he'll submit some patches later. With that, I can start to concentrate on Linux and Mac support next and as soon as I have something that at least crawls on my Ubuntu machine, I'll send the patches - it might take some time.

I have no idea if it is even feasible to run Rei (really mono) in Firefox on Linux or Safari on Mac, but I hope so. So with that tirade I guess the bottom line is: no, there is no clear timeframe :) hopefully quite soon the standalone will be able to start on linux with the new Irrlicht 1.6 libs.

5.  How does 3di intend to sustain the Rei project?  Is it just
through cross-subsidy from sales/support of the 3di enterprise server
or are there other business models planned (e.g. support of the viewer
with vanilla OpenSim).  Again, I appreciate that it might not be
possible to answer this :)

Rei is essential to 3Di's business, but the money is in our OpenSim distribution and our services related to that. We have an internal build that adds a few extra features that we deliver to customers who buy our OpenSim distro. I think we are also considering a way to help developers monetize on their Rei plugin development though. It's way too early to tell how that will look like.

6.  If Rei gathers an open-source community, will 3di give trusted
non-3di developers access to the git master branch and a voice in
project decisions?

Absolutely! In fact, anyone who wants to submit patches can already do so, and anyone can participate in the discussion on this list. We also allow (actually encourage) developers to join Rei development.

Personally, I believe that all of us would give OpenSim core members commit access without hesitation, as they are proving their skills day-to-day on OpenSim. I also believe that a BSD-license viewer, that is written in the same language as OpenSim itself should be very good grounds for innovation without having to switch context too much (C#->C++ for example)

Zaki

justincc

unread,
Oct 28, 2009, 3:45:46 PM10/28/09
to 3di-viewer-rei-dev
Thanks very much for the further answers Zaki - very interesting
information!
Reply all
Reply to author
Forward
0 new messages