Re: [vwrap] Some thoughts on Josh's Linden Lab Legacy Protocol discussion

[David W Levine <[e-mail]>]

"Hurliman, John" <[e-mail]> wrote on 04/28/2010 11:39:06 
AM:

> [image removed] 
> 
> Re: [vwrap] Some thoughts on Josh's Linden Lab Legacy Protocol 
discussion
> 
> Hurliman, John 
> 
> to:
> 
> David W Levine
> 
> 04/28/2010 11:40 AM
> 
> Cc:
> 
> "[e-mail]"
> 
> Does the method for exposing an API need to be different from the 
> method that exposes a user interface?
> 
> John
> 

I've been giving that exact question some serious thought. One simple
answer is to say, as we often do, "Content Negotiation Saves the day!"
If you want the programatic api, ask for application/xml, if you want 
something to present to the user, ask for text/html. If one wants to be
stronger, make the mime types more explicit. (I'm not thrilled with this
because if we can reuse existing, obvious mime types, I tend to think
its much better) 

But.. there are at least two things which make me go "maybe not." Part
one is, well, contextual. Do we expose things primarily as part of the
visual client centric scenegraph? If we tie things to closely to be
"logged in" to a region and having a client, then we end up making it
much harder for applications (and for that matter clients) which are
substantially different than the current client. We get a lot of power
by saying "This service/ui/api" is associated with "this portion of the
scenegraph." in terms of the immersive user experience. But.. the same
power gets in the way of web service like integration where one might well
want to simply expose a URI that lets you directly manipulate things in
a portion of the world. 

My current rough thinking is that if we do things correctly, we can expose
the same API via both "web external" facing discovery and 
"scenegraph internal" facing discovery. If we let the UI/API be chosen
by content negotiation, we may be done. (Tho, I'm not at all user exposing
a UI will help much, as you'll likely not have enough context to do 
anything
useful with it, but.. I can imagine clients which will) 

As I said. Very much partly baked thinking at the moment, ripe for 
discussion.

- David
~ Zha


> On Apr 27, 2010, at 2:42 PM, "David W Levine" <[e-mail]<
> mailto:[e-mail]>> wrote:
> 
> 
> As always, food for thought, probably partly baked.
> 
> Josh Linden's presentation on the Linden Legacy Protocol
[snip]