Quoting Jonathan Rochkind <rochkind_at_JHU.EDU>:
I _believe_
> that anything that can be expressed in microdata can be expressed in
> RDFa and vice versa.
http://schema.rdfs.org/ has translations between rdfa, microdata, and rdfa.
I suspect that it's not lossless in all instances, but that's not
unusual when translating from anything to anything else.
>
> The reason the article doesn't make a good argument for it's
> conclusion is not because microdata is only for SEO. It's because,
> like Matthew Beacom wrote in this thread, if microdata does win out
> against RDFa/RDF, it just means that the semantic web will be
> implemented using microdata.
Perhaps its best to say that SOME of the semantic web will be
implemented using microdata, and SOME will need more complex data. The
statement about thesauri is worth paying attention to: what the
semantic web can offer is richer relationships than purely descriptive
metadata. It can describe broader and narrower, based on, A cites B,
and other things that I think will "save the time of the reader."
Descriptive metadata forms the basic inventory of things that can
relate to other things, and any of these metadata formats that we are
talking about can do that.
I wonder if we (and others) aren't using "metadata" to mean
"descriptive metadata".
The one thing I wish they had done differently with schema.org would
be to define the data elements in relation to current known metadata
schemas, and to include guidance rules. I don't know if folks have
seen the RePec metadata (for papers in economics), but I really like
their documentation, e.g.
http://ideas.repec.org/t/booktemplate.html
kc
> microdata _is_ RDF-compatible triples, same as RDFa is. We will see.
> Things are certainly still in flux. Library community efforts to
> rationalize our metadata control for machine-actionability and
> contemporary standards -- which is what the RDA Vocabulary
> initiative is -- will bear fruit in any event.
>
> Jonathan
>
> On 10/12/2011 4:01 PM, James Weinheimer wrote:
>> On 12/10/2011 21:09, Diane Hillmann wrote:
>> <snip>
>>> Also, from the point of view of the search engine community, this
>>> optimization by embedding metadata is their one-and-only use case for
>>> metadata. Those of us who remember early attempts to embed DC properties in
>>> web pages also know some limitations of this approach: it assumes metadata
>>> is static and unchanging, and makes it virtually impossible to maintain or
>>> re-use. We have a lot more than that we hope to do, and that niche, should
>>> we chose to accept it, really uses our skills and experience to the utmost.
>> </snip>
>>
>> My own thoughts are that perhaps "Search Engine Optimization" (SEO)
>> really is the future for all metadata, for better or worse. Its
>> simplicity is going to make sense to a lot of people, and it can be
>> implemented now. But as I mentioned earlier, I don't see why
>> microdata would be any less susceptible to spam than "author
>> created metadata" was before, when people would put in practically
>> entire encyclopedias into the meta fields, and when search engines
>> disallowed information in the meta fields and took only on the text
>> in the body of the webpage, people came up with all kinds of clever
>> things. One I discovered that I thought was really neat, where
>> people would put in all kinds of spam using white text on white
>> backgrounds, so that you couldn't see it until you looked at the
>> page source.
>>
>> The only solution that was found was the Google-type solution that
>> rated pages by links *to* the page, and not by what was *on* the
>> page. This is liable to spam also, with Google-bombing and of
>> course, SEO, which is actually a type of Google-bomb but more
>> respectable.
>>
>> Nevertheless, library use of microdata could provide a level of a
>> predictability in the search result, which I think people want, at
>> least sometimes, as opposed to the endlessly "personalized" search
>> results that all of the information companies seem to be aiming
>> for: personalized by your own previous searches, by what your
>> friends searched, by what their friends searched, by what people
>> with similar profiles searched and so on and on....
>>
>> Libraries allow true *conceptual* searching, using Use Fors,
>> Broader/Narrower Terms, Related Terms, and scope notes, so that
>> people really can search by concept. Google-type searches will
>> remain as more and more complex variations on searching text, using
>> fuzzy searches for terms, thesauri and so on. But no matter what,
>> it will remain textual. Libraries and conceptual searching, no
>> matter the problems with it, is wanted by many once people begin to
>> understand it. But our conceptual searching is really weird for
>> most of the public today.
>>
>> I still say this is one place where libraries could provide
>> something nobody else does. But I don't know--maybe it's too late.
>> It would cost some rapidly-disappearing bucks that will be spent on
>> implementing RDA! That will certainly solve all of our problems! :-)
>>
>
--
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
Received on Wed Oct 12 2011 - 17:46:27 EDT