[rdfweb-dev] foaf:publications

Peter Mika pmika at cs.vu.nl
Wed Feb 22 14:25:16 UTC 2006


Dear All,

I was wondering lately about the foaf:publications property. My scutter runs
show that it's not much used (so this email may not be of interest to a
great many people), but when it's used it is used in quite ambiguous ways.

Use#1 Some people use it to point to individual publications, which is not
what the spec suggest, but imho it is quite useful, because it can be used
even for disambiguation. A common problem with CiteSeer et al. is the
conflation of people with the same name, while you could distinguish authors
by their publications. ("I'm the Peter Mika who wrote 'Ontologies are
us:...') BTW, in this usage foaf:publications is inverseOf swrc:author,
which connects publications to authors. 

Use#2 Other people use it in the spirit of the spec to point to an HTML
publication page:

"The foaf:publications property indicates a foaf:Document listing (primarily
in human-readable form) some publications associated with the foaf:Person.
Such documents are typically published alongside one's foaf:homepage."

Personally, I don't think this is very useful. Not only that I might not
have a separate publications page in HTML, but even if I do it will be
human-readable only and most likely a subpage of my homepage. (So why list
separately in my profile? Where is the semantics?)

Now, here comes the surprise: I'm actually here today to suggest a third
use. 

Use#3 In our little not-yet-so-public openacademia project (sneak peek at
vu.openacademia.org) we would like people to reference their bibtex files
from their foaf profiles. The reason is that BibTeX is structured and
quasi-standard format and almost everyone in our target group (CS
researchers) has one already. If you read the spec loosely, this would even
be allowed: a BibTex file is "primarily human readable" and lists some
publications associated with the person. Yet, it's not desirable to mix
references to bibtex files with references to html pages.

So my question what should the foaf community do in terms of disambiguating
these cases and supporting some/all of them?

Some alternatives:

a) Do nothing. 
b) Sanction all uses (Use#1-#3) in the specification. 
c) Stick to the current specification (Use#2), but advise people on how to
implement the other two use cases.

Let me know what you think!

Best, 
Peter

 




More information about the foaf-dev mailing list