Re: let's drop model support

Samuele Pedroni <[email protected]>:
>
> From: Greg Ewing <[email protected]>
> > Magnus Lie Hetland <[email protected]>:
> >
> > > In-place modification was one of the main reasons for push/pull,
> > > and if we drop that (which I think well we might) we might end
> > > up with an architecture that is much easier to understand,
> > > explain, and maintain.
> >
> > This would be good. This would be very, very good. You need
> > simplification. You need it very badly. For the last few months, I
> > haven't been able to follow *any* of these discussions...
> 
> What can I say, the only way I was able, after my hiatus, to make
> sense of the discussions was to read the code.

Sounds like a reasonable approach ;)

[snip problems with models]
> Now given that it seems we don't have a strong use case for models, and that
> layering models on top of an architecture, whose point is decoupling from
> toolkit widgets [and models would problably need the the contrary, that means
> strong coupling to work effectively],

This is a good point, I think.

> is not trivial and does not produce a mechanism simple to
> understand, I think we all agree that making models the only
> operative mode would be overkill.

I definitely think so.

> So maybe we should simply drop model support ...

It is a possibility. (We should, in that case, take care not to
through out the baby with the bathwater -- do we, for instance, still
want in-place modification? If we allow modifiable attribute values,
then we probably should. If we introduce copy semantics then we should
probably not.)

> we already have canvas, tagging for text, events, menus, backends to
> complete and to develop... ang bugs to fix

Yes.

On the other hand, I'm reluctant to closing the door on MVC
completely. Even if (big if!) we decide to drop MVC at the moment, it
would be good to have the option of reintroducing it later without a
complete redefinition of attribute semantics etc.

> I have not researched whether I was a proponent for supporting
> models in the past,

Doesn't matter. We're all allowed to change our minds on this list
anyway :)

> I know I have not written related code and ask pardon to the people
> whose code I'm proposing to drop...
> 
> I don't know what Magnus or the other think ...

I don't really know myself, yet. But I think I have observed (and
mentally formulated) four alternative approaches at this point. I'll
try to summarise them (with pros and cons) in a separate post.

> For the rest, if we come, instead of dropping it, to a next
> incarnation of an approach for supporting models I offer to
> implement it ...

Hopefully, if we drop models now and reintroduce them later we should
be able to use quite a bit of the current model code -- or?

> regards.

-- 
Magnus Lie Hetland        Practical Python          The Anygui Project
http://hetland.org        http://ppython.com        http://anygui.org


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf