[Opensim-dev] OpenSim's direction after Linden cutting support, and the possibility of an official OpenSim viewer

[Justin Clark-Casey]

It's true that the viewer situation is becoming more difficult and I this is going to be an issue for quite a while. 
I'm grateful to the viewer developers who continue to support OpenSimulator and I understand why some cannot due to the 
lack of resources to work around Havok licensing restrictions.

I'm not a fan of having an 'offical' OpenSim viewer that is tightly coupled to the OpenSim server software or team. 
Whilst this might allow for faster progress in the short term, I think that philosophically it moves us away from the 
idea of a broad metaverse with multiple server and viewer projects.

I would rather have a situation where we document and make protocols and data extensible so that many different systems 
can interoperate and innovate (admittedly we are a long, long way off that right now).

Realistically, I think that we will continue to be pulled along by LL for a long time yet - I think non-programmers can 
vastly underestimate the amount of work required to do significant design and fix difficult core bugs.  Following LLs 
work is the path of least resistance.  Only when and if the balance of effort changes between LL and third party 
projects (viewers and servers) will this really start to change.

That said, I'm personally very happy to see server and viewer folks work together (or for server folks to work directly 
on viewer projects and vice versa).  New features that do not originate from LL can certainly be incorporated into 
opensim core as appropriate where they are well designed and useful to many people.  And even where this is not the 
case, hopefully it will become easier to put more specific functionality into modules as the module facilities and 
underlying opensim capabilities improve and become more flexible.

On 10/12/12 23:08, Mircea Kitsune wrote:
> Hey everyone. I've been away from OpenSim for a while but got back. Since last time I've been around, I read about
> Linden's decision to cut OpenSim support from their viewer due to some crazy licenses with their physics engine. The
> rights and wrongs have been discussed and are not the point of this email. But this decision means that in some senses,
> OpenSim will become completely separate from Linden's SL, and some fundamental things might change. I wanted both to ask
> what exactly is going to happen, but also post my own suggestion. Obviously I'm not someone who can say what's good and
> bad, but personally this is what I strongly believe OpenSim should do and go for.
>
> The way I see things, OpenSim has primarily been a server-side for Second Life during its existence. Even if its purpose
> is a general virtual worlds platform, SL was the only usable viewer in practice. Now that LL cut its support, it's in
> the situation of having no exact client to be used with. There are many third-party viewers that will continue
> supporting OpenSim, but IMHO they can't be considered a reliable source in the long term. I assume most of them have
> their own developers which take decisions independently from the OpenSim team. If one of those viewers dies for
> instance, it's up to the user to go looking for another one that's still under development. Apart from the fact that
> people have to hunt for a viewer, this situation also kept OpenSim from being able to make changes that would require
> viewer modifications as well.
>
> My opinion is that OpenSim won't get far if it relies on random viewer forks to be used with at this point. We are not a
> clear standard technology, unlike web browsers for instance where you can use Apache to host and FireFox to browse, both
> unrelated and a variety of choices available for each. This is not to say OpenSim should be unusable with viewers
> unrelated to SL, since that would go against its purpose. But the SL viewer is a very large and complex thing, and there
> will surely never be anyone making a client from scratch which will implement all of its features, look as good, be as
> fast, as bug free, etc. The building and prim editing tools, the terrain editor, the avatar and mesh system, the
> rendering features, the GUI... it's unlikely anyone will properly re-write all that from zero when the SL viewer exists
> and works fine.
>
> This is why I believe we need our own official viewer, developed by and with the OpenSim server, and based on one of the
> Second Life viewers. Apart from the fact that people will know they don't depend on someone else to make them a viewer,
> it would allow client + server changes to be done for the first time, rather than having to stay within the SL viewer's
> limits. If that doesn't happen, I don't believe we'll ever become a better virtual worlds platform outside of SL's
> shadow. I remember the days of RealXtend (I heard it's dead now) which took the SL viewer in one hand, OpenSim in
> another, and created its very own project which was completely amazing for those days (SL now has mesh support and
> better graphics so it makes RealXtend less special at this day). My personal opinion is that it's time OpenSim does
> something similar.
>
> If it was me, I'd say grab the latest SL viewer without the restrictive Havoc library and make OpenSim viewer from that.
> If Linden adds a nice feature to theirs which we can copy over, sure thing, but otherwise it can go its own separate
> way. On the other hand, it would be a good aim to allow previous SL viewers to still connect to OpenSim, though they
> wouldn't recognize some of its specific features then. I know OS and SL and under two different licenses, but that
> doesn't make it wrong to distribute both on the same website and as part of the same project.
>
> What does the core team of OpenSim think about this? Are there any plans or will to go in this direction? Or does anyone
> believe that instead we should continue providing support for SL's features and have OpenSim users find their own
> clients like until now? I believe this is an important question, and would like to know what to expect in the future.
> Personally I really hope something in this direction will be decided, but I'm not one to know best.
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev
>


-- 
Justin Clark-Casey (justincc)
OSVW Consulting
http://justincc.org
http://twitter.com/justincc