Hi Aaron, thanks for your reply.

Le dimanche 6 août 2006 21:47, Aaron J. Seigo a écrit :
> On Sunday 06 August 2006 10:30, Olivier Goffart wrote:

> by keeping the data coupled to the view, you will never be able to
> experiment too wildly with views and uses of that data because the effort
> of doing so will be too high.
>
> this approach also introduces complexity in the form of having to manage
> both data and view in one place. the complexity of the MV approach is party
> a perception thing due to being new (meaning we have to learn things) and
> only partly due to actually being more complex.

I think understand this. But I think this is already the case in Kopete, 
without the need of a QAbstactItemModel stuff and all theses QModelIndex.

Or what's the difference ?  (I'm asking to Matt, Michel and Michaël)


> > Note that our storage data is already separated from it's representation
> > in the Kopete::ContactList
>
> and the contactlist data structure ought to be able to feed directly into
> views.

What do you mean ?


> one of the use cases kopete will run into is being able to bridge the
> contact data to lists that appear on the desktop or in other applications.
> identity is moving more and more towards sharing and coordinating this data
> between multiple apps.

And that's one important topics.

But I fail to see how QAbstractItemModel helps.

What we need here is a defined dbus interface. 

Attachment: pgp57QIKbI84X.pgp
Description: PGP signature

_______________________________________________
kopete-devel mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/kopete-devel

Reply via email to