[Portland] Project Portland, RuDI and the Generic desktop adapter

[Bastian, Waldo]

>Hi,
>
>May i join in your discussion? :-)
>
>I have already written a poposal and a prototype for a cross desktop
>interface in between applications and desktops, called "The Generic
>Desktop Adapter". It seems very close to what "RuDI" wants to achieve,
>but using a traditional in-process library approach with dlopen().
>
>http://www.scheinwelt.at/~norbertf/dadapt/
>
>Perhaps you can give me some feedback on my proposal and tell me if it
>makes sense to continue working in that direction.
>
>What do you think?

I think that ISVs tend to have two requirements that may make your
approach less desirable.

The first requirement is that the API/ABI offered should be well defined
and stable in time. In-process approaches tend to be poor in this regard
because they bring along a lot of implicit ABIs. 

The second requirement is related to time to market. Ideally a solution
should be available now, and ideally it should work with all desktop
software that has been released already and that is in use today. That
may not be entirely realistic but the larger the changes required in the
upstream projects, the longer it will take for such solution to hit the
market, and then you need to add another 3 to 5 year in order for
pre-solution desktops to drop out of the installed base before it
becomes a solution that an ISV can rely on.

The above two points also apply to a certain extent to Rudi of course.

Cheers,
Waldo