[mmox] xxSD working group?

[Catherine Pfeffer <[e-mail]>]

Infinity wrote:
> Catherine... Jon...
>
> can you create a proposal for a next generation XML serialization?

Yes I can.

> the existing serialization is in current use by SL, OpenSim and PyOGP,

Really? I am surprised to hear that, because I had once the curiosity to run
Wireshark to sniff packets exchanged by my Second Life client with the
simulators, and I saw that the binary serialization was used all the time,
even when using the new capacities system. Could you please point me to a
situation where I could test the existing XML serialization?

> so there's probably going to be a lot of resistance to changing
> something that currently works and is deployed.

Infinity, we need to make once for all clear whether this is just a Linden
Lab existing technologies rubber stamping work group, or if we are designing
vendor independent protocols for the 3 dimensional web of the future.

While I agree there is no reason to throw away existing stuff (like LLSD's
simplicity) when it is cool, I think that every bit of the technology should
be evaluated, and replaced when it has conceptual flaws in it.

If you disagree with this position, please state it clearly, and I would be
happy to withdraw from this mailing list and to go to other projects where I
might be more useful. Similarly, if Linden Lab's position is "let them speak
and then enforce our own stuff at the end", please also spare my time.

> keep in mind, however, that OGP does not use XML to represent it's
> PDUs, it uses LLSD. XML is one of three defined serializations of
> LLSD. it's a subtle difference, but important.

I have it in mind, even though the rationale for using three different
serializations is far from being clear to me.

> for instance... is there a benefit to explicitly adding support for
> namespaces to the XML serialization?

That's a very good question. What would be the purposes of such namespaces?

> there was a bit of discussion
> about this amongst some AWG members, and the consensus was... "why
> bother? the XML (presentation layer) is not the place where you want
> to extend the PDU... you want to extend it in the LLSD / LLIDL
> definition of the interaction."

Makes sense.

But this reasoning apparently assumes that the namespaces would be used to
extend the interoperability standard.

Normally, namespaces are rather used to mix data from different standards.

That could be *very useful* if we want to mix LLSD for example with a mesh
description expressed in another standard. Or to be able to include SVG
graphics. Etc. While this would be very powerful, that would not scale back
to the two other serializations, because it's pure XML mechanisms.

(That's not the only point where needing support for three serializations
has drawbacks. For example, LLIDL would become unnecessary if using only
XML, because the names and types of the fields could be stored in schema or
DTD fragments, or even in a WSDL dictionary. But never mind.)

> but, i still think it's an interesting idea as it might allow systems
> that use XML serialization exclusively to extend display and
> distribution options of various PDUs.


-- 
Cathy