[opensource-dev] The Plan for Snowglobe

Henri Beauchamp sldev at free.fr
Fri Sep 10 15:30:19 PDT 2010

On Fri, 10 Sep 2010 12:22:25 -0400, Oz Linden (Scott Lawrence) wrote:

>   On 2010-09-09 7:15, Aidan Thornton wrote:
> > On 9/8/10, Oz Linden (Scott Lawrence)<oz at lindenlab.com>  wrote:
> >>          * Take down the Snowglobe subversion repository
> > That's going to be kinda obnoxious, because it means non-Linden Labs
> > developers won't be able to look back at what was changed in
> > Snowglobe, when and why. There will still be mainstream 1.23-based
> > viewers by then unless Linden Labs does something incredibly brilliant
> > or something incredibly stupid, and they'll still have a use for the
> > Snowglobe version control history.
> Viewer developers should be thinking now about how to migrate to the new 
> code base so that you can support your features and interfaces with the 
> new underlying capabilities (and there are more coming)

Viewer developers will not consider migrating to the viewer 2 code base
for many months, because it's simply easier for them to port the few
interesting new features of viewer 2 to viewer 1, than to redo the whole
UI of viewer 2 to match their user base expectations and needs.

I already ported the Tattoo and Alpha wearables months ago to the Cool
VL Viewer (and most TPVs now reuse my patch), and the inventory item
links support a few weeks ago. I'll work next on multiple attachments
per point.

With Snowglobe v1.5 as the code base and the above cited backports,
you already have a better viewer than viewer 2, with a better stability
and higher frame rates...

> and superior stability that the new code base has.


You are kidding, aren't you ?... If not, then please try using a good
TPV and see how many times it crashes in a week... 0 for the Cool VL
Viewer (and I'm using it every day). Fact is that TPVs got fixes that
v1.23.5 doesn't have and that makes them MUCH stabler than 1.23.5, and
v2 (which is even worst, stability-wise, than v1.23.5 !). The reason
is simple: should I crash, I trace the crash down and fix the code.
Crash gone !

> Eventually (and there is _no_ plan for when this will be - certainly 
> longer than 3 months), it will no longer be possible for us to continue 
> to support viewers based on the 1.x code base (including our own), and 
> we'll stop.

By then, all the required changes will be ported to the v1 codebase
and migrating to v2 will still be unnecessary... This could go like
that for at least one or even two years before the backports become
too cumbersome to be worth staying with v1. I know it first hand,
since I did just that with the Cool VL Viewer v1.19.

> Well before this happens, we'll have a public discussion 
> about it, and about what must be supported to remain compatible.  If you 
> have moved to and stayed reasonably current with the 2.x code base, then 
> it will be a non-event for you.

Again, TPV developers are not going to bother with v2 unless YOU, Linden
Lab, change your stance on the UI and do accept reversals to the v1 way,
where needed.


More information about the opensource-dev mailing list