[Dev] Collections document

[Philippe Bossut]

Hi,

I learned a lot reading this doc but it seems to me that the Collections 
design is overly complex. May be this complexity is required but it's 
not apparent why. The different kinds of collections are introduced but 
without justification of their existence and no apparent structure.

Some questions unanswered:
- Why is the result of set operations (intersection, union and 
difference) of a different kind than the base kind AbstractCollection?
- Why do we need this strange InclusionExclusionCollection at all? If it 
is the only kind that can be added to the sidebar (and then viewed by 
the user) it seems that there's some fundamental flaw in the design, or, 
put another way, why can't the base class AbstractCollection be usable 
as is in interaction with the user?
- IndexedSelectionCollection seems like a smart hack rather than 
something that fits neatly within the concept of Collections. Or may be 
it does fit well but  it's really not explained what are the advantages 
of this design.

I think one can make sense of the whole doc but it needs several rounds 
of reading to sink in. A more hierarchical presentation (making clear 
what derives from what) and a clear identification of the methods and 
attributes specific to each class would make understanding much easier.

Cheers,
- Philippe

John Anderson wrote:
> Hi:
>
> I recently wrote a developer document describing collections:
>
> http://wiki.osafoundation.org/bin/view/Projects/Collections
>
> I'd appreciate any feedback you might have.
>
> John
> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>
> Open Source Applications Foundation "Dev" mailing list
> http://lists.osafoundation.org/mailman/listinfo/dev