Inclusion of documentation in introspection data?

[John (J5) Palmieri]

And someone can write a program to send huge data across the bus now or
run in a continuous loop and take up your CPU.  We really can't solve
for these problems so I'm not worrying about it too much.  What I don't
want to do is mandate or approve something that in the end turns out to
be a bad idea.  Since we are so close to 1.0 it would be especially
painful if it turns out to become a bottleneck. Remember it is easier to
add things in the future than to take them away.    

On Tue, 2005-10-25 at 01:53 +0200, Tako Schotanus wrote:
> Havoc Pennington wrote:
> 
> >What John says about experimenting with it as a custom annotation then
> >we could add it post-1.0 seems sensible. If it uses too much bandwidth
> >we could add some kind of introspect method that lets you specify
> >annotations to exclude or something like that later on.
> >
> >  
> >
> Ok, but that's less backward compatible, not that it will actually break 
> anything but if future introspection data suddenly start passing 
> megabytes of annotations programs compiled against older versions would 
> suddenly become a burden on system resources, wouldn't they?
> 
> So maybe introspection should by default leave out all annotations or 
> something like that? Probably not, I imagine.
> 
> Another option would be to distinguish between annotations that add 
> actual machine-usable information and annotations that are only there  
> for the benefit of us humans (like documentation).
> 
> Of course in the end it all depends on how much the introspection 
> feature is going to be used, but it seems logical that  if we start 
> passing around MBs of data (even while filtering out the annotations you 
> know are  bandwidth eaters there could still be new ones you didn't know 
> about) some possible uses of introspection will made more difficult. 
> (Don't know, you want to expand your spam mail filtering script with 
> some call to a DBus service which has to generate a proxy on the fly 
> using the introspection data for example ;-)
> 
> Cheers,
> -Tako
> 
> _______________________________________________
> dbus mailing list
> dbus at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dbus
-- 
John (J5) Palmieri <johnp at redhat.com>