Stefan Andersson wrote: > > Resizing the terseupdate packet size is going to be placed in to > ModRex at least, that's why in the patch SendAvatarTerseUpdate method is > marked as virtual. > > > Just a very quick short one; (yeah right) > > Not talking about this patch in particular, just realized there's > something that app devs probably should be made aware of: > > In OpenSim, there's been an implicit understanding (that we should > probably discuss and make explicit) that; > > 1) Marking stuff as 'virtual' because of the need to subclass in a > particular proprietary application should never be a problem. This > strenghtens the API.
+1 > > 2) Going from 'private' to 'protected' because of the need to > subclass in a particular proprietary application should never be a > problem. This strenghtens the API. This is not always the case. See http://opensimulator.org/mantis/view.php?id=3072. Sometimes there are good reasons for keeping methods private. > > 3) Introducing extra layers of indirection because of the need to > re-direct control flow in a particular proprietary application should > never be a problem. This strenghtens the API. +1 > > 4) Introducing stuff directly into core that is not used by core because > a particular proprietary feature demands it, will always be a > problem. Instead, do 1) or 2) or 3). +1 > > 5) Going from 'protected' to 'public' because a particular proprietary > feature demands it, will always be a problem. Instead, do 1) or 2) or 3). +1 In other words, I broadly agree. I would also add that none our APIs can be regarded as 'stable' until we near 1.0 - everything is subject to change. Indeed, I would say that at the moment virtually all of our APIs are 'accidental' APIs in that they were originally formulated for internal use. Therefore, they are often not well documented or thought through in terms of their functionality and usability for non-core code. > > Best regards, > Stefan Andersson > Tribal Media AB > > > From: mikko.pall...@adminotech.com > > To: opensim-dev@lists.berlios.de > > Date: Thu, 12 Feb 2009 14:26:26 +0100 > > Subject: Re: [Opensim-dev] Avatar movement packet optimization > > > > Hi, > > > > Yes, 'dramatic' as in positive. > > > > The changes introduced in the patch work for LL clients also and > propably for OpenMetaverse clients too, but I haven't tested this patch > with them. > > > > Resizing the terseupdate packet size is going to be placed in to > ModRex at least, that's why in the patch SendAvatarTerseUpdate method is > marked as virtual. > > > > Sean Dague wrote: > > > Very cool to see this patch. There is some feedback in the mantis > tracker about adjusting the way types are used. It would be great if you > could respin the patch based on that so we could get this in. > > > > Already done. > > > > Cheers, > > Mikko > > > > -----Original Message----- > > From: opensim-dev-boun...@lists.berlios.de > [mailto:opensim-dev-boun...@lists.berlios.de] On Behalf Of Dirk Krause > > Sent: 12. helmikuuta 2009 13:36 > > To: opensim-dev@lists.berlios.de > > Subject: Re: [Opensim-dev] Avatar movement packet optimization > > > > Hi, > > > > - dramatic in terms of 'positive'? > > - to take advantage of the client side changes one would need to > change the viewer source (or take the Rex Viewer) then? > > I also take it that OpenMetaverse based viewers (like IdealistViewer) > would need a patch for OpenMetaverse? > > > > > > even with the 'server only' optimization, it still is a great > contribution. Should be tested though, and I wonder what would be a > smart way to do this. > > > > Best, > > Dirk/Bart > > > > -----Ursprüngliche Nachricht----- > > Von: opensim-dev-boun...@lists.berlios.de > [mailto:opensim-dev-boun...@lists.berlios.de] Im Auftrag von Mikko Pallari > > Gesendet: Donnerstag, 12. Februar 2009 12:16 > > An: opensim-dev@lists.berlios.de > > Betreff: Re: [Opensim-dev] Avatar movement packet optimization > > > > That optimization did this and also reduced the terseupdate packet > size. That needed changes to protocol, and because we can't change the > protocol with Linded client, the effect won't be so dramatic here. > > > > Cheers, > > Mikko > > > > -----Original Message----- > > From: opensim-dev-boun...@lists.berlios.de > [mailto:opensim-dev-boun...@lists.berlios.de] On Behalf Of Jeroen van Veen > > Sent: 12. helmikuuta 2009 11:15 > > To: opensim-dev@lists.berlios.de > > Subject: Re: [Opensim-dev] Avatar movement packet optimization > > > > i saw that robotdemo a while ago at > http://community.rexdeveloper.org/. It's > > the same optimization? That was really some impressive cool stuff. > keep up the > > good work! > > > > Jeroen > > > > On Wednesday 11 February 2009 18:59:19 Tommi Laukkanen wrote: > > > Nice! > > > > > > We need more work like this. > > > > > > Tommi > > > > > > On Wed, Feb 11, 2009 at 1:21 PM, Mikko Pallari > > > <mikko.pall...@adminotech.com > > > > > > > wrote: > > > > > > > > Hi, > > > > > > > > > > > > > > > > I have made a patch to optimize number of packets to enable even more > > > > users in simulator. I have submitted this patch to mantis and you can > > > > review it here: http://opensimulator.org/mantis/view.php?id=3136 > > > > > > > > > > > > > > > > And as described in mantis this packet should increase the max > number of > > > > avatars the server can handle. It combines many terse update > packets in > > > > to one packet. > > > > > > > > > > > > > > > > Our initial tests show that: > > > > > > > > - two avatars moving in a circle with optimization: 30-60 packets per > > > > second on client > > > > > > > > - two avatars moving in a circle without optimization: 40-70 > packets per > > > > second on client > > > > > > > > > > > > > > > > Cheers, > > > > > > > > Mikko Pallari > > > > > > > > _______________________________________________ > > > > Opensim-dev mailing list > > > > Opensim-dev@lists.berlios.de > > > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > > > _______________________________________________ > > Opensim-dev mailing list > > Opensim-dev@lists.berlios.de > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > _______________________________________________ > > Opensim-dev mailing list > > Opensim-dev@lists.berlios.de > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > > > > > _______________________________________________ > > Opensim-dev mailing list > > Opensim-dev@lists.berlios.de > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > _______________________________________________ > > Opensim-dev mailing list > > Opensim-dev@lists.berlios.de > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > > ------------------------------------------------------------------------ > > _______________________________________________ > Opensim-dev mailing list > Opensim-dev@lists.berlios.de > https://lists.berlios.de/mailman/listinfo/opensim-dev -- justincc Justin Clark-Casey http://justincc.wordpress.com _______________________________________________ Opensim-dev mailing list Opensim-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/opensim-dev
