Kokua Viewer Released for OpenSim

The Kokua team made an announcement. They will be spitting development of their viewer, forking in geek speak. They will make a viewer for OpenSim and one for SL.

Pathfinding and various other things work only on one grid or the other. Now that Linden Lab is going to a CDN system the incompatibilities between is such that OpenSim servers and hosting system will have to change. This is going to be significant incompatibility. The OpenSim host will have to think about how they will handle it. Will OpenSim grids move to CDN? I have never looked into the cost of CDN service, but it is a cost. In the low income realm of OpenSim any additional cost is a problem.

So, as of Kokua Viewer the Kokua team will be putting out an SL and an OpenSim version of the viewer. This latest one is for OpenSim.

There was a duel use viewer released, I think, as 3.7.12. But, it was having problems on some grids. The numbering indicates a bit of a regression for the OpenSim versions.

It isn’t surprising that viewer developers are splitting their viewers into multiple single use viewers. It reduces the work and time it takes to get updates out. Some no doubt will continue to make duel use viewers. But, with SL2 coming a third type of viewer is going to come into existence. The open source third party developers are getting spread pretty thin.

For a time there will be no third party SL2 viewers. But, I suspect some developers will start reverse engineering a viewer. That time will come out of time for SL1 viewer work.

The Firestorm team has said they’ll try to release a viewer about every third month (other month? I forget and I’m  too tired to stop and look it up). The Lab was releasing a viewer every other week. But, the Lab doesn’t have to make changes and test against a second grid. I suspect the workload on volunteers is a major factor in the FS release schedule.

I’ve speculated that third party developers would have a hard timekeeping up. But, developers like amazing Henri Beauchamp make it look easy and do amazing work (Cool VL Viewer). But, others have dropped production of viewers and joined larger teams to lighten the work load.

The Firestorm team is an odd mix of behind and ahead of the Lab’s developments. In some areas the FS team lags behind the Lab. For instance the FS Viewer does not have the newer AISv3 code for handling inventory. That is a good and bad thing. Bad that the viewer doesn’t use the new processes and server calls. Good that the team is fixing stuff for more reliable performance before distributing it in their viewer.

But, the work load seems to keep increasing. Right now the Lindens are rebuilding their tools for building/compiling the viewer and its support software. Once that is done, the Lab will be recompiling all their viewer code.

This gives the third party developers time to catch up. But, then they will have to deal with all the code changes and compile processes that change. For them it is probably a smaller task than for the Lab. Many of the libraries the Lab will compile are used by third party developers. But, they use the Linden compiled versions and can avoid compiling their own.

Still the workload is changing the face of the third party development world in the SL community.

3 thoughts on “Kokua Viewer Released for OpenSim

  1. CDN is no reason for forking the viewers… CDN servers work just fine with current viewers and even if LL provides slightly better viewer-side code to deal more smartly with CDN servers, it would be a piece of cake to just add a test to decide which code path (the current one or the CDN-optimized one) to take based on which grid you are connected to…

    The only real problem with SL & OpenSim is for TPVs that have got LL’s license to include Havok stuff (namely, pathfinding visualization tool and advanced mesh decomposition code); this license does not extend to OpenSim grids and viewers including the Havok code are therefore not allowed to connect to OpenSim grids.

    For viewers not using the Havok code (such as the Cool VL Viewer or Singularity, which only lack the pathfinding visualization and the most advanced options for mesh physics hull decomposition at mesh upload time), there is no problem whatsoever and no need for a fork.

    • I think the significant problem in the area of CDN is with the OpenSim grids having to set it up. So, depending on what the grid operators do, it can become another special condition that has to be handled. Whatever they do, it starts out as a problem to be handled for viewer teams. Either way it can be handled in viewer code. But, it is another thing to do. Those special ‘things’ seem to be piling up.

      While you seem to be able to quickly program around these special ‘things’, I suspect for some they are reaching their last straws. I see it as not so much what can be done as what people are willing to do. The Kokua/Imprudence viewers were being neglected for a time. I suspect that is the visible indication of a lack of time to do ‘things’ to keep Kokua up to date. Fortunately Kokua seems to be getting much more attention these days.

      • Again (and please, do believe me, for I know the viewer code pretty intimately), the CDN stuff is not an issue at all; and no, OpenSim grids will not *need* to use CDN, just like viewers will not *need* to be modified beyond some very marginal and optional optimizations to support CDN, which they actually already ‘support’ since using CDN or not is just a matter of using the URL passed by the sim server via the existing capabilities.

Leave a Reply

Your email address will not be published. Required fields are marked *