[Opensim-dev] future rexviewer merger

Sun Dec 7 14:01:06 UTC 2008

I didn't say that any discussion should be reserved to core, I said things shouldn't be forced on developers just because others decide its a good idea.

If the tools/system was more developed than I would be all for us moving to GIT/HG/etc. But they aren't, and by switch to them we are forcing extra effort on developers. People have limited time to devote to opensim. So we shouldn't do things that put extra strain on developers or raises the entry level for new developers.

And again I don't buy the argument that we should use alpha tools because opensim is alpha. Again we advise people to not use opensim in production enviornments. But of course we are happy that people test it and try it out and use it. But for developing opensim we want to use stable, user friendly tools that makes development easiest. 

Of course its great that some opensim developers have decided to use GIT etc, and are helping to improve it. But we should not force it on everyone. So yes if any developer wants to use GIT, then thats great. But as you said they are already doing so. But we shouldn't be talking about it being the main system we use until its more developed. 

As for the learning curve well again what I mean is it rasies the effort and entry level for opensim development, which isn't a good thing. They most likely are way over exacturating, but comments like "if you switch to GIT, don't expect to get any work done with it for the first month" are common from people who switch to it.

James Stallings II <james.stallings at gmail.com> wrote: Well I never meant to suggest that we should use these tools if they arent cross platform. I certainly did not suggest that, if they weren't, no one should bring the subject up. I believe I also suggested that such adoption needent be full-scale in the offing.

What I *did* say remains true: we have an obligation to move these tools forward if we hope to benefit from them. And for those that can't afford the time to assist with moving their native clients forward, there are alternatives such as HomerH has just suggested.

As for complaints about the learning curve? They seem a bit tedious - there *are* those in core who use these tools, and can and most likely will assist in supporting them. The individuals in question certainly dont fail to satisfy wrt any other requests for assistance. OpenSim has a learning curve that is profound, and no other single application even remotely resembling it is in common usage, so that curve pretty well starts at ground level. Philosophically speaking, It feels a bit awkward to shy away from the kind of effort we routinely require of others.

Reserve such decisions to core as you will - it certainly hasnt seemed to impact the decision-making process of the other core members who have adopted these tools successfully, in at least one case nearly a year past.

Just my *non-core* .02$L

Cheers
James

On Sun, Dec 7, 2008 at 7:18 AM, Michael Wright <michaelwri22 at yahoo.co.uk> wrote:
 I think its a bit different because we are talking about toolchains that we want to use in production enviorments. We always say opensim is still Alpha and not ready for production level use. I think it would be very bad to use alpha tools to try to develope opensim. Its just increasing the problems.

Also people jump up and down and shout loudly when something might not work on Linux. So why can't people complain when things are unusable on windows? I know that if the linux tools were completely unusable then everyone would shout that we shouldn't even consider using them until it had better cross platform support. Sorry but thats just a personal little annoyance that I get from some of the things that are said/moaned about.

But anyway I think the bigger problem for everyone is just the steep learning curve that (from most feedback) those system take to even be able to start using them.

Its easy to say everyone should swap to Git/HG, but when you are  the one who has to use tools that just don't really work very well as a result, its a different matter. So I think this is something for the core team to decide. 

And at this time my vote would be certainly -1 to swapping to them. 

James Stallings II <james.stallings at gmail.com> wrote:

  Forgive me Adam, if I think this argument against Hg/GIT is a little leaky ;)

The reason I say this is that 1) we depend on 'very very very alpha code' every day - in the corpus of the opensim work itself, and with the employment of custom clients like hippo and meerkat. 2) if we hope to see improvements in these toolchains that work to our benefit, we have an obligation to adopt, test, and report, just as we ask of the broader opensim community wrt the opensim project. This developement model works brilliantly for us - why wouldnt we  participate in it's employment for a different important project?

That being said, such an adoption needen't be simultaneously end-to-end - incremental and progressive adoption by some of our more adventurous windows devs might serve to inventory the trouble to be anticipated as adoption progresses, perhaps filing bug reports to get showstopping issues fixed, and blazing the trail forward into the more progressive and far more productive workflows supported by these tools.

I think the benefits for the community are strong, and should be examined in this light, rather than dismissed because they are nacent. If we all took such a position, where would opensim be now?

Cheers!
James

On Sun, Dec 7, 2008 at 6:45 AM, Frisby, Adam <adam at deepthink.com.au> wrote:
                  I *wish* I could say that OpenSim's core dev team is getting the message about the toolchains - this is something that hobbles many innovators. Some have quietly adopted these newer toolchains for themselves, but much of the benefits of this are lost because core still sits in an SVN repo (yes, I am aware that Hg and GIT can work with SVN repos, but to do things this way rather dilutes their strengths).

GIT/Mercurial have completely retarded clients for platforms-other-than-linux. The windows version for instance is commandline-only  which does not work anywhere near as efficiently as TortoiseSVN/etc. (I am aware TortoiseHg/TortoiseGit do exist – but they are in very very very alpha stage).

_______________________________________________
 Opensim-dev mailing list
 Opensim-dev at lists.berlios.de
 https://lists.berlios.de/mailman/listinfo/opensim-dev

-- 
===================================
The wind
scours the earth for prayers
The night obscures them

http://osgrid.org
  http://del.icio.us/SPQR
http://twitter.com/jstallings2
http://www.linkedin.com/pub/5/770/a49
   _______________________________________________
Opensim-dev mailing list
Opensim-dev at lists.berlios.de
https://lists.berlios.de/mailman/listinfo/opensim-dev

_______________________________________________
 Opensim-dev mailing list
 Opensim-dev at lists.berlios.de
 https://lists.berlios.de/mailman/listinfo/opensim-dev

-- 
===================================
The wind
scours the earth for prayers
The night obscures them

http://osgrid.org
 http://del.icio.us/SPQR
http://twitter.com/jstallings2
http://www.linkedin.com/pub/5/770/a49
  _______________________________________________
Opensim-dev mailing list
Opensim-dev at lists.berlios.de
https://lists.berlios.de/mailman/listinfo/opensim-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20081207/0d9fe529/attachment-0001.html>