Simon et al
"sounds like it would be high time to take it on board" may be a bit
premature!!
I agree that Stephen's paper is very good but it makes a series of points
not a single one and these are not all of equal merit. It would be a mistake
to accept the flaws just because it is a great and stimulating read.
He does accurately point out many of the weaknesses in current metadata
practice and makes some excellent recommendations - in particular about the
need for greater normalisation and identification.
But there are weaknesses in his suggestions:
Stephen's first rule"
"metadata for a given entity should never be stored in more than one place.
If we have, for example, metadata about a given person (say, me), then it
should be stored in one and only one location. (That does not mean that it
cannot be aggregated or mirrored, but it does mean that there is one and
only one location that would constitute the source of information about this
person, and that aggregators and mirrors would update on a regular basis
from this source."
I strongly oppose this.
In addition to the generic danger of "eggs in baskets", learning records
about a person are complex and managed variously by the person, the
state(s), by learning organisations and by awarding bodies etc. These
authorities have various responsibilities to maintain information and
protect levels of privacy. To centralise this process is not a good idea.
However: to suggest that a particular piece of information has a single
authority makes more sense. As long as there can be many authorities for
information about an entity.
If the information is objective (definitive) then having a single source has
advantages as pointed out by Stephen. If the information is subjective
(non-definitive) then there should be the possibility for many sources for
information, for example regarding curriculum use, expressing views in other
languages etc.
In other words to argue that it is possible to have, or even that there
should ideally be, a single source (one giant metadata aggregation) for an
entity is both naive and un-democratic.
(I do not think that is Stephen's real intention but he did write " there is
one and only one location that would constitute the source of information
about this person")
Stephen's second principle (sic)
"metadata for a given entity should not contain metadata for a second
entity"
This seems reasonable as an ideal to *tend towards*. There are however some
reservations in practice if taken too literally.
Not everyone in the world has internet access all the time - for many
reasons. Reliance on universal perpetual connectivity is unrealistic even
though it may be a valid assumption when participating in an online
discussion.
There are situations when it is useful to have a bit of information about a
second entity. The combined indentifier/location
"dc:creator resource="http://www.downes.ca/my.foaf" does not give the name
of the person in the metadata instance. I would not therefore be able to
sensibly interrogate a set of metadata instances on a disconnected laptop or
in a closed off environment.
Taking the argument to the extreme: Assume A and B are related and both have
metadata. In a perfect minimalist world where would the relationship be
stored? Possibly in both instances.
If this relationship is expressed in a way that goes beyond simply
isRelatedTo then it necessarily carries information about the other entity
above just its identity.
Given that titles, terms, names, dates, descriptions, concepts, keywords etc
are all valid entities (in a context) then taking the axiom "metadata for a
given entity should not contain metadata" (metadata = information other than
identity??) to the extreme would produce a metadata record that is simply:
Identifier of entity plus a bag of identifiers for related identities.
Very neat but not very informative.
On identifiers:
I would strongly support the arguments in favour of better unique,
persistent and global identification of things (object, terms, competencies,
vocabulary values etc). I think though that Stephen wrongly promotes a
mixing of location and identification that is very common in the metadata
world.
He says
"For after all, an identifier is no use if it is not also a pointer of some
sort, and in particular, a pointer to information about the entity it
identifies."
An identifier IS useful even if it just identifies something (as long as it
does it well). That is what its primary role is and this is extremely
valuable in itself. e.g. ISBN is useful wherever the store or library
records are. Knowing that two things are the same or different is vital.
Many people mix location and identification into a single URL. This is often
fine but is not universal and if adopted as mandatory, or even best
practice, will preclude the ability to multiply resolve identifiers.
So some bits onboard OK, the rest hopefully overboard.
Cheers
Mike 7:-D
-----------
Mike Collett, Schemeta
+44 7798 728 747
------------
www.schemeta.com
email: [log in to unmask]
> From: Simon Grant <[log in to unmask]>
> Reply-To: Simon Grant <[log in to unmask]>
> Date: Mon, 11 Jul 2005 16:21:22 +0100
> To: <[log in to unmask]>
> Subject: Re: OLDaily on metadata
>
> At 16:01 2005-07-11, Robert John Robertson wrote:
>> At the risk of provoking a Friday afternoon discussion at the start of the
>> week, has anyone else read the OLDaily section on metadata?
>>
>> "Metadata
>> This was intended to be a note to myself, partially to comment on
>> discussion list metadata and partially to frame some throughts for my talk
>> in Colorado in August. But it makes some points that bear wider
>> consideration, and in particular, two principles of metadata: metadata for
>> a given entity should never be stored in more than one place; and metadata
>> for a given entity should not contain metadata for a second entity. By
>> Stephen Downes, Half an Hour, July 7, 2005 "
>>
>> http://halfanhour.blogspot.com/2005/07/metadata.html
>
> I hadn't read it. Thanks for pointing it out.
>
> On reading it, I am struck with extensive agreement. I'd be curious, at
> least, to hear of any arguments rationally refuting SD's position. If no
> one can refute it, sounds like it would be high time to take it on board.
>
> 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
|