Re: [Vrs-development] Re: [DotGNU]pnetlib and the VRS, SEE and other server concepts

[Bill Lance]

--- Norbert Bollow <address@hidden> wrote:
> 
> Rhys Weatherley <address@hidden> replied:
> 
> > The hardest thing is GUI's, since every widget set
> has its
> > own peculiarities.  The next hardest is audio and
> other forms
> > of multimedia.  Usually it isn't worth trying to
> solve this
> > problem unless you have something like Java AWT
> around
> > to take care of the icky details.
> 
> But unless another project does this work for us,
> we'll
> eventually need to address the hard issues of
> creating
> portable free libs for these things.
> 


hummm  ...

Maybe I'm asking three questions.

One is good practices for crossplatform development. 
As Rhys said abstraction .. keep the machine away from
the design till the last layer. These are good answers
to that question.  

The question I was getting at is more architectural, I
guess.  In Unix/Linux we build systems from
plug'n'play like parts.  Esamples are common, like cd
recording on Linux involves a number of frontends, all
using cdrecord.  The VRS Services Manager, and several
versions of the SEE idea may need to rely heavily on
this feature in order to stitch together all the
protocol servers that we are talking about.  But that
whole environment of parts is missing in the MS world.
 Do we have to rebuild their whole ecology for them
before they can be functional?

And the third question is one of general long term
strategy.  How do we see dotGNU interacting with the
MS world in the future?


__________________________________________________
Do You Yahoo!?
Yahoo! Sports - live college hoops coverage
http://sports.yahoo.com/