Performance: (was Re: [Accessibility] Re: [Accessibility-atspi] D-Bus AT-SPI - The way forward)

[Michael Meeks]

On Thu, 2007-12-13 at 15:03 +0000, Rob Taylor wrote:
> The interesting thing here is even when doing no marshalling and with
> validation disabled, its still at about 2x against ORBit. What other
> design decisions do you think might be influencing this?

	Wow - there must just be something silly going on (surely) :-) what
does kcachegrind say ?

> > in your mind is "there's a single-digit constant-factor overhead vs. raw
> > sockets" and "avoid round trips!!!!"

	Well said Havoc :-)

> There's only so much that can be done here, as in there's a lot of
> existing code out there already based on the current API. Some simple
> things like enabling cachable properties could make a huge difference,
> but even there there's a lot of assumptions in things like mozilla's ATK
> usage that break when doing that.

	Well; if we can make the protocol easily extensible [hint hint] we have
some more leg-room here I think. Particularly adding a more
transactional API to Atk is quite do-able, and we have a nice template
in the UIA thing: [ as an aside, they have a pretty pumped up thing
there to allow arbitrary queries of sets of properties & so on as a
single request to reduce IPC ].

> > If you need something raw-socket-like, then stick to CORBA, or
>> use D-Bus for discovery only then set up a custom socket channel,
>> or whatever.

	Lets not give up too early; CORBA is conceptually at least as
complicated as D-BUS, so there must be a lot we can do here.

> > There is no reason to drop CORBA when it is suitable. The reason for
> > dbus is that (in my opinion) CORBA was not suitable for many desktop
> > use-cases. That does not mean CORBA is not suitable for AT-SPI.

	Nah ;-) they're the same thing - and we should drop CORBA and move to
D-BUS if only to avoid wasting a chunk of memory & developer confusion
IMHO etc. Also, CORBA/ORBit2 is saddled with both the standad CORBA API
mangling, which people hate, and also some ugly back-compat
characteristics from the days of Gnome 1.x which are unfortunate (around
event-loops, 'main' threads etc.).

	HTH,

		Michael.

-- 
 michael.meeks at novell.com  <><, Pseudo Engineer, itinerant idiot