Re: Containers

[Tim Teulings <tim@edge.ping.de>]

Hallo!

> with others.  On the other hand, maybe too many people are writing
> code that is unsuitable for reuse.

I think I already lower your exspectation enough ;-) That OOCDTA ist just
a starting point not more.

> We had already had a discussion about linked lists.  The prose on the
> mailing list was a very good starting point, but needs to be
> formalized into a module interface for further discussion.  

Yes, the design already seemed supirior to mine. As told I suggest that
someone summerizes the results (I think Peter and the both Michaels had
written most of that thread) and then puts it in to a minimal
implementation.

> This is possible, if an elaborate framework exists and has been
> agreed
> upon.  For example, the OOC IO libs have been built this way. 
> Without
> a good framework (and good documentation), things won't work out.

Yes, we need a perfect prototype module. Most datatypes then could be
programmed just similar. However the baseclass problem is not yet fully
solved (important for e.g. persistence). Also datatypes and algorithms
interact (sorting a abitrary datastructure). STL solves this by using
iterators and that like, we should find a similar mechanism.

> Also remember my "godfather" posting.  It was not intended as a joke.
> In my experience, without a "godfather" nothing will happen beyond
> calls to "let's do this or that".

I can act as a hardware goodfather but not more. VisualOberon takes all
my time reserved for computer stuff.

> PF: PS: OOC and VisualOberon are still not in the Oberon Webring?
> Why's
> PF: that? :-)

I'll do it. But updating the VisualOberon webside is a little bit
tricky (making a tar archive, going online, making a ftp update of the
archive and then using a html mask for further installation). Adapting
to the webring takes time and the good weather and my job keep me too
busy for that. But I promise to do it, since it *is* a good idea!

-- 
Gru...
       Tim.