[rdfweb-dev] Re: <foaf:community>? (Groups are subtle and tricky, keep the format simple!)

[Bill Kearney]

> Us geeks are kind of poor at guessing good models for how to represent
> groups/communities etc - we keep making appeals to set theory or OO
> inheritance models when actual groups are much more tricky.

I liken it to having a deaf person make you a violin. No disparagement of
anyone with disabilities intended. And indeed with the right tools it might
well be possible for anyone to make anything. But lacking the time and budgets
that analogy seems apt. A crucial point to consider is how geeks form groups
may or may not model that of any number of larger subsets of the population.
Design decisions would benefit from factoring in larger issues.

I can still hear Meg Hourihan saying that when newsblogger was created people
actually complained they were making it "too easy" for people to become more
informed. The ability to form groups that exclude and it's impact on the
desires of non-members to join is really something.

> I don't want to stand on anyones toes, but based on the ways we've
> studied the problem and what we've read from anthropology and social
> psychology about what's understood (and not understood) about groups -
> almost any "intuitive", set-based, hierarchical model just doesn't
> work for a whole bunch of real-world cases.

And lacking visualization tools for overlapping sets makes it even harder.

> Sure some groups can be related using specialisation/generalisation
> links, but lots of other groups have relationships that just don't fit
> that model - "split over this issue", "on a similar topic", "just like
> that group, but about this topic, not that one".
>
> Similarly, to use Bill's example of cliques, many groups never have a
> definitive "roster" or a coherent topic or even a name (what do you
> call that crowd of people you meet in the bar on a Friday every so
> often? Who's in it? How do you think the rest of the group would
> answer those questions?).

That's the big question, how does the 'real world' see this?

> I'm definitely not saying ignore community or groups, but I am saying
> that, based on what we've observed, it's good to keep it simple and
> under-specify. I believe if we keep it to simple things that can be
> used in lots of places and then only extend it by using it and
> observing the use (rather than by exercising our creativity on
> thinking up what might work), we'll get more useful results.
>
> Just my $0.02. I'm not sure I've expressed myself clearly enough, so
> if this seems silly or just wrong to anyone, please just bellow at me

Clear enough to me, +1 on all points. If you're in the DC area on a Friday
night I'll buy you a beer.

-Bill Kearney