Re: [eclipse-incubator-e4-dev] What I dislike about using EMF for e4...

[Eric Moffatt <emoffatt@xxxxxxxxxx>]

Tom wrote:

Does the following make sense?

If our workbench uses a model an IConfigurationElement is not needed any
more because you do your contribution using XMI (if we use EMF).

Tom, this is one of the important points
that is being missed; thank you for pointing it out. 

We -cannot- simply say we dont' need
to use 'x' anymore where 'x' is something that comes from our use of 'core'
infrastructure. IConfigurationElement is the integration point for our
extension point handling. We're a -consumer- of these objects (which are
supplied by equinox now) and unless we're going to re-write equinox they
will, by necessity, exist in e4. The same is true where 'x' is part of
our current API, at least IMemento and ScopedPreferenceStore fit into this
category...

This is why I've been proposing YAMI
(Yet Another Model Interface); if it were simply another API supporting
the same functionality I'd be justifiably shot (I'd do it myself...;-).
What I'm hoping to be able to do is to at least allow our users to have
a 'common' api through which they can access the underlying information
in a consistent manner; we make one more API but they see at least 3 -less-.

Please step back from EMF for a second
and take this into consideration when thinking about the 'common' API;
it's job isn't necessarily to support a single particular model but to
allow common access to many different model 'flavors' in order to simplify
the story for our clients.

At least that's my dream,
Eric