Not The Right Opinion

Yesterday, i posted a message to kde-core-devel. I happen to disagree with moving away from dcop and adopting dbus instead it. However, i was rejected by moderation, a cryptic reason of “Peter, this was about the same as you said in your previous post…”. No idea what i said where, but my previous post to the list was about svk i think. Whatever, i am somewhat disappointed that possibly unpopular opinions are discarded at moderation time… I used to like KDE, but lately, i am more and more disappointed.

Whatever, below is the mail i sent and was deemed inappropriate for the list…

On Sunday 28 August 2005 19:09, Maks Orlovich wrote:
> On Sunday 28 August 2005 07:00, Stephan Kulow wrote:
[snip]
> >
> > The first one was IPC. We once again summarized the benefits of KDE
> > switching to DBUS (among the lines of 'well maintained', 'support from
> > toolkits and other desktops', 'distribution support already very high')
> > and what bothers us with it ('C API', 'unsolved performance problems',
> > 'unknown upgrade path').
>
> Here, I have to express disappointment (most of the points have to do with
> hype and not anything technical), though I am hardly surprised (see
> 'hype'). Perhaps I am mistaken, but has D-Bus reached any sort of API, ABI,
> and wire compatibility stability? Has the protocol been properly specified?
> (If the wire compatibility is stable, and the spec is finally reasonably
> complete, I can probably do a client implementation in a weekend).
>
[snip]
>
> Oh, and has anyone volunteered to port the zillions of calls, and to do it
> in a way that doesn't totally break everything for a month?
That is going to be an interesting exercise yes. And actually, what is the 
real benefit here? Noone stops us from talking to system dbus... and does 
session dbus buy us anything over dcop? Apart from a hypothetical "kde and 
gnome apps can talk to each other" -- i mean, how exactly is this going to be 
useful? I may be missing something... but last time i checked, all the 
interfaces with kde/gnome, looking away from minor details like .desktop 
files, were pretty different -- having different interfaces on top of a 
single protocol won't help all that much anyway... unless we want to share 
dbus interfaces with gnome (and this will probably amount to some nasty 
surprizes), there is actually little point in making half of the step but 
stopping there... When we actually have an idea on how to specify/share the 
interfaces, it starts to be worth thinking about. I suspect that the wire 
protocol is actually least of the problems.

If dbus is similar enough to dcop (as i hear everywhere), it shouldn't be hard 
to translate dcop calls to dbus and vice versa for the few places where it 
could make sense. Switching all of kde to avoid one extra process and some 
processing on the (quite rare, apparently) rpc calls sounds like hell of a 
premature optimisation to me... (And premature optimisation is root of all 
evil, right).

[snip]

Yours,
    Peter.