At 14:58 2005-07-12, Robert John Robertson wrote:
>Aside from the commitment, e-literacy, and technical competency involved
>in every author creating and maintaining a .foaf file, and a consistent
>website address, the problem with this approach seems to be that the
>service referring to the my.foaf relies entirely on the accuracy of the
>metadata created by the (external) author.
>For example, if the author misspells the name of his institution (or uses
>a different variant of it) there is little the service can do about it.
I suspect SD says this, anyway I'll give my understanding - the actual name
of an institution should belong to the institution and be listed in its own
metadata maintained by a suitably authorised official of that institution.
The individual would just have a link to the institution's metadata - OK,
many individuals can get even the URL of their home institution wrong, but
is this a majority fault? I could imagine tools to help people to check
URLs and get them right (I guess they exist already), particularly if the
URL points to machine-readable metadata.
>Consequently the service's reliability and consistency suffers. Service's
>may also not be able to depend on the completeness of any given foaf -
>some authors may include email addresses, some not.
I haven't had the time/energy to get into foaf, so I can't comment. But a
good principle might be: don't assume foaf is perfect, or obeys SD's laws,
already.
>I think that managing consistency within a distributed service ideally
>requires some level of a centralised name authority file (within a
>given service or a given community). Even if this is only on the level of
>being able to standardise the name format (John Brown vs. Brown, John),
>the service needs a degree of control to normalise its metadata.
I agree with the need to standardise representations of, e.g. names and
addresses. Neither of the forms of name given here would seem a good standard.
Simon
--
Simon Grant http://www.simongrant.org/home.html
Information Systems Strategist http://www.inst.co.uk/
Please continue to use my established e-mail address
a (just by itself) (at) simongrant.org
|