When showing the photo app demo in the two talks I did, one question that 
came up was (an obvious one), "Why do you need to keep restarting the RCP 
app after every model change?". 

I don't think we've discussed this aspect, but I assume that we eventually 
want to get to the point where model changes can be realized against the 
live app. This matches the web DOM approach where the way I change what is 
on screen is via DOM element creation/deletion.  My understanding is that 
EMF can provide the change notification, its a matter of us listening and 
doing the appropriate element deletion and running the factories on the 
new nodes. 

Btw, a super cool demo would be to modify the structure of the live app 
via javascript, which you hacked live.

Some questions:

1) Do we agree this is the behaviour we want?

2) When do we we think we'd provide it?

3) I think I can imagine with the current code base how we handle 
on-the-fly creation and deletion, but what about modification?  Right now 
the factories 'make', they don't 'modify existing'. 

4) Do we need batching?  Does every model change result immediately in 
on-screen changes?  I'm not sure how the web handles this; it doesn't seem 
to require explicit batching but the changes appear to come out as one 
change (might just be it all happens so fast it looks like one change).

Because of the implications for (3), I think we'd like to do this work 
sooner rather than later.  It would also help to tell the story of why the 
modelled UI is more interesting than just "I don't need this all in 
plugin.xml anymore".

Regards,
Kevin
_______________________________________________
eclipse-incubator-e4-dev mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/eclipse-incubator-e4-dev

Reply via email to