[rdfweb-dev] Suggestion about FOAF vacabulary

James Moore jam at wirerimmed.com
Fri May 2 18:11:15 UTC 2003


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Friday, May 2, 2003, at 05:29  AM, Ian Davis wrote:

> On Friday, 02 May 2003 at 12:56, Libby Miller wrote:
>
>> You're right about people having several email addresses (and several
>> people having the same address, although that possibility is ruled out
>> by the definition of mbox). Another problem is that different people
>> over time may have the same mail address (this happens at Bristol
>> university, where they are recycled). This is ruled out too by the
>> definition of mbox.
>
> mbox has quite a narrow field of applicability - geeks like us that
> care about such things. Members of my family regularly change email
> addresses between the various free providers, even signing up several
> times to the same ones under new names over the course of several
> years. (I'm pretty sure I had iand at hotmail.com once upon a time but I
> can't remember the password...)
>

But if in each one of their FOAF files they only list their unambiguous 
addresses at the moment doesn't it reduce the danger of creating a bad 
connection? I suppose FOAF suffers from the same problem as every other 
kind of documentation namely, if it's not current it's worse than not 
having it at all.

> Uniquely identifying people is always going to be hard - how do I
> indicate that I know Nelson Mandela the former president of South
> Africa (I don't really)?
>

If you knew Nelson Mandela you'd likely have his email address as would 
all the other folks who knew him. I don't think it's important to know 
via your FOAF file that you know *the* Nelson Mandela just that you 
know a Nelson Mandela with an mbox of X and a FOAF descriptor located 
at Y. Now if he decides to describe himself fully at location Y then 
that makes FOAFSpace that much more lush but it's not necessary.

> One way is to identify some other characteristics that are unique,
> such as when he was president of South Africa. OWL and DAML-TIME could
> be used to assert that only one person can be president of a county as
> any one time.

True but you couldn't use this same scheme to uniquely identify 
everyone. Mboxes are universal amongst people who are online.

> What other alternatives are there I wonder?

Fully qualified phone number perhaps? Perhaps using the SHA1 to keep it 
private. If Nelson didn't use email but I had his direct line perhaps I 
could use that instead.

- -James

- ----
WireRimmed - Custom software for your small business
http://wirerimmed.com

-----BEGIN PGP SIGNATURE-----
Version: PGP 8.0
Comment: Let me know if you'd like to learn more about email encryption

iQA/AwUBPrK0wzNOZbajVDyKEQKLcgCfS6mnuY6l79Jvfk7BsQZOFG5S+vEAnRCx
V/VCKO1I4F//5LwUtNfhmWSq
=FTHn
-----END PGP SIGNATURE-----




More information about the foaf-dev mailing list