[opensource-dev] Snowglobe 2.0^H^H^H1.3 way forward?

Ardy Lay ardylay at gmail.com
Fri Mar 12 08:45:34 PST 2010

Argent Stonecutter wrote:
> So you think the way forward to Snowglobe should be to ignore 2.x and  
> see about backporting the 2.x functional improvements to 1.3.x?
> What would Linden Lab do if people actually started doing that?
I remember several instances when people asked what Viewer 2.0 would 
bring to Second Life and them getting back the answer "You will be 
amazed!" Well, I am. Not particularly happy though. I see some bug fixes 
we have been hoping for, for a long time, made it into Viewer 2.0, but I 
cannot use the UI. All the winking, blinking and bouncing gives me 
seizures. Hell, I can play stuff like Battlezone II and Portal without 
problems but something about Viewer 2.0's UI drives me over the edge. 
Most games do make me uncomfortable but Second Life has never been an 
issue until now. I can kill particles. I can kill face lights. I can 
return spinning flashy garbage and I can teleport home avatars that 
cause me grief but I cannot do anything about the UI but avoid it. Back 
to Snowglobe 1.4.0 for me.

I have heard the excuse that "It tests well with new users" several 
times but I really don't care what a person that doesn't care about 
Second Life and doesn't know diddle squat about what's missing or broken 
in the UI thinks about it. "Play testing" only works when you make sure 
the "player" is involved in ALL ASPECTS of the game. In Second Life 
that's a VERY TALL ORDER. No way are you going to get the uninitiated 
man off the street involved deeply enough to properly gauge your UI. All 
you can do with him is fiddle with the few things he does whine about 
until he stops whining about it. Something I learned long ago, give a 
man a task to perform and when he fails to perform satisfactorily his 
first excuse amounts to him blaming the tools. This has resulted in many 
broken golf clubs and damaged power tools. Imagine what would happen if 
sporting equipment and tool manufacturers failed to properly train their 
product testers and blindly followed their suggestions?

I know I am going to catch heat for this post. Somebody is going to tell 
me "That's not the way worked" or some such rubbish. Well, when you 
sequester your User Interface design and development away from the users 
for a year while making massive changes you have to expect a large 
number of those users to dislike the results and suspect you are not 
acting in their best interests. After all, we have to use this thing, 
not the random people that are not interested in Second Life that 
answered some ad to be a product tester for some unnamed product.

Argent, to answer your question, I think the proper thing to do is 
distance the UI from the rest of the code. Maybe LL has started doing 
this. A decent API for user interface functions that pushes the UI code 
out of the render loop is gonna have to happen eventually. I think now 
is the time. I only wish I had the mental capacity to dive into this 
task and achieve results. God knows I need the job.

What would Linden Lab do if people started enhancing Snowglobe 1.x with 
features added to Viewer 2.0 by Linden Lab? Rub their hands together and 
cackle with glee as their user support costs go down because more and 
more users are using an unsupported viewer would be my first guess. 
Until there is a public announcement, in the Linden Blog FEATURED 
section from "up stairs" telling us insignificant little existing users 
that this is not the case I am sticking with my opinion that "up stairs" 
only cares about the real world population that IS NOT in Second Life 
already. All argument against it here will be ignored as usual as I know 
for a fact that many Linden Lab employees do in fact care a great deal. 
They will have to let their upper management fend for themselves in this 

We are being fed conflicting information by Linden Lab employees. It may 
not be deliberate. It may just be people repeating stuff they hear 
without really thinking about how it will formulate in our minds. We 
have some people telling us Linden Lab cannot foresee the future and we 
have people telling us that Viewer 1.x will be supported for the 
foreseeable future. This fires off both of my doublespeak detectors. 
(That's bullshit detectors for the laymen out there.)

This is Ardy Lay and turn the lights off on your way out.

More information about the opensource-dev mailing list