Hi Cees!

Cees de Groot <cg at cdegroot.com> wrote:
> On Mon, 2003-03-10 at 10:45, goran.hultgren at bluefish.se wrote:
> > Btw, Cees - I haven't had the time to check out trove in detail yet -
> > but what are the features/differences in trove that you would like to
> > have?
>
> Trove has a hierarchy of classifications (mostly orthogonal - platform,

Hierarchy of classifications = SM hierarchy of categories. What is the
difference?

> audience, level, ...); you can add a package in a whole bunch of
> classifications. For example, look at

Yes, you can add an SM package in a whole bunch of categories too. What
is the difference?

> http://freshmeat.net/projects/squeak under 'Trove Categories'.
> 
> What I like about it:
> 1. Classificying is a no-brainer - all categories are presented as a
> flat list (with '::' separating the hierarchical levels), and the
> classifier just runs down the list and checks relevant items;

Isn't that exactly what is done in Package Loader (and in the SM web UI
too)? What is the difference?

> 2. Navigation can be done tree-wise, or by various drill-down techniques
> (for example move to 'Audience :: End-User' and then narrow down to
> 'Status :: Stable').=20

AFAICT - again, what is the difference? This exact "drilling" was
intended for SqueakMap Browser, but I never got around adding it. But
there is nothing preventing it in the model.

> It's not perfect, but I think it is practical. It seems to work for the
> tens of thousands of packages on FM and SF...

As you are probably aware of I did get inspiration from for example
GNOME software map when I did SM. I didn't look at Trove specifically
though.

But I still need answers to two questions:

1. What are the differences with "Categories"? I can't see any.
2. What exactly IS Troves? Just an idea? Some XML schema? A piece of
software? Some form of standard regarding what kind of fields we have?

regards, Göran

PS. The tone in this reply is not meant to sound annoyed - I just want
to know the *differences*.