Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[higgins-dev] Attribute Value Types
Title: Attribute Value Types
I thought about this idea too, but it fails for the same reason I wrote about in my email (quoted below). If we add restrictions such as you propose we will not be able to losslessly express existing data sets that take advantage of this freedom to mix different types of simple values for the same attribute.


On 6/17/08 5:11 PM, "Drummond Reed" <drummond.reed@xxxxxxxxxxxx> wrote:

[renaming this thread to stay on topic]

Paul, Jim:

Is there a way to have our cake and eat it too, i.e., could we go back to
saying that all values of an attribute have a single type, BUT if that type
is Entity, then each value (instance of an Entity) can have its own type.

That way, any multi-valued attribute that is NOT of type Entity ensures that
all values are of the same type, while multi-valued attributes of type
Entity preserve the rich multi-typed multi-value model that Daniel and Mike
and Paul don't want to lose.

Or is there something that prevents this happy duality?

=Drummond

> -----Original Message-----
> From: higgins-dev-bounces@xxxxxxxxxxx [mailto:higgins-dev-
> bounces@xxxxxxxxxxx] On Behalf Of Paul Trevithick
> Sent: Tuesday, June 17, 2008 10:25 AM
> To: higgins-dev
> Subject: Re: [higgins-dev] Higgins 1.1 M2 release status - 13-June-08
>
> I don't like the added complexity either, but it is good for another
> reason.
> Today the Higgins "Context Data Model" is expressive enough to be able to
> represent with minimal loss a wide range of data models that exist in the
> wild. Without this flexibility, for example, it could not represent many
> kinds of RDF-based data (e.g. from the semantic web, etc.).
>
>
> On 6/17/08 11:36 AM, "Jim Sermersheim" <jimse@xxxxxxxxxx> wrote:
> >
> >
> >  This is branching into another topic.
> >
> >  IAttributeModel can no longer have a method getValueModel.  Why?
> Because the
> > group decided that each value of an attribute is allowed to be of a
> different
> > type.  Daniel and Mike were the main proponents of this.  I don't like
> it and
> > I pretty much caved to group consensus.  If you'd like to re-raise that
> > argument, feel free.  To me, it's much simpler to say that all values of
> an
> > attribute are of the same type.
> >
> >  I only deprecated it for now, so that we could figure out how to fix up
> all
> > the old code that used it.  If we can reverse that decision, I'd love
> it!
> >
> >  Jim
> >
> <snip>
>
>
> _______________________________________________
> higgins-dev mailing list
> higgins-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/higgins-dev

_______________________________________________
higgins-dev mailing list
higgins-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/higgins-dev

Back to the top