[rdfweb-dev] Re: Friendster

Bill Kearney wkearney99 at h...
Thu Dec 19 20:35:10 UTC 2002

> > Gee, ya mean like how MSN messenger and Passport already does this sort of
> > thing?
> And Gator and.......

Now yer walking into a minefield. Gator, what a joke, little more than spyware.
Completely unlike the Passport stuff, but the rabble rally 'round bashing
Microsoft so there's little point in going into an real discussion of it.

> DOS level? I can't think anyone has ever asked for my ICQ list of
> contacts (perhaps I'm just unpopular). My main concern was that it
> wasn't worth the effort as so few people would use it.

Correct, the balancing act of how much 'interaction' people want is really
tricky. On one hand it seems like a fine idea to have questions being pinged
back and forth at you. Until you start getting them so often you can't get any
work done. Not to mention what'd happen if someone deliberately signed up for
an account for the purpose of bugging others.

> This could be done if everything was centralised but it won't be so
> the request could be coming from all sorts of different places (if at
> all though - see above).

Yet the trendiest approach to these things is decentralization. I'm with you, a
little bit of centralizing is a good thing. And if it's not masquerading as yet
another way to assault me with advertising and marketing then I'm all for it.

> On the wrapping thing up (for people on the same server) I'd probably
> favour something like an online admin which people can visit and
> they'd get a note once a week if there were any requests they hadn't
> viewed. People could also specify if the didn't want to be contacted
> under this kind of system.

Yes, the GNU mailman's admin interface does a good job of this sort of thing.
As does the yahoogroups moderator interface.

> > Transitive trusts; now there's a king-size can of worms.
> Fancy expanding on that point?

Transitive trusts, extending the fact that userA trusts userB does not mean
userB trusting userC has any impact on the userA to userC relationship. And
extending it into a group raises a rats nest of 'consensus' issues. Next thing
you know somebody bings up Robert's Rules of Order and all hell breaks loose.

> This would still have the same problems you outlined above without
> some of the solutions available if things were kept between people on
> the server.

Yes. Getting developers to participate in what's largely a closed system is
always a challenge however. But yes, you're correct. Once it's running it
seems users will naively join up to just about anything.

> Quite - I wouldn't claim to know everyone in a community (although,
> of course, there is knowing and knowing so......).


> Quite possibly - I suspect the way FOAF will really take off is if it
> is made easier for people to produce and edit the files.
> I don't know - it will probably have to be a bit trail and error to
> start off with.

Yep, being the pioneers means wandering down the wrong trails now and then. The
trick is avoiding too many arrows in the back from angry natives. (playing on
your 'trail and error' typo a bit)

It'll be interesting to see how the users in non-computer related circles view
the use of foaf files.

-Bill Kearney

More information about the foaf-dev mailing list