> In adding RDFa, we are giving machines more information, and hopefully they can pass this onto their users / use it to facilitate better search results. If this takes next to no time then why not?
Yes, I think that if you generate your HTML resource descriptions from a
database of some kind, it should be pretty simple to also generate RDF
from the same data. I've no strong position on whether it should be RDFa
embedded in the HTML or a separate file with RDF (of any flavour) and
suitable use of <link> in the HTML header or content negotiation to get
machines from the HTML to the to the RDF.
If RDFa has traction at Google then that's great. But, to be honest, I
don't think that describing OERs as a product using a set of properties
that are clearly designed for advertising commercial offerings is a
great solution. (*) What I think I would prefer is for Google to add an
"Education Resource" rich snippet type that used properties that already
exist. As far as the UKOER tech requirements go it could be done as
Title: DC Title
Author/Owner/contributor: DC Creator/CC attributionName/DC Contributor
Date: DC date
URL: DC Identifier
Licence Information: CC [for non-OER stuff Google offer might be OK]
Description: DC description
Comments: Google review seems OK
Subject: DC subject
Language: DC Language
Tags: not sure. Brand maybe wasn't bad for UKOER, maybe doesn't need to
be supported.
(DC = DC Element Set, DC Terms doesn't seem as good a match to the
metadata we are likely to have. I mean, I'm all for URIs to identify the
Author, but if they're not there you don't get much benefit from making
them up based on the literal metadata you do have. Winston--did I get it
right?)
Phil
* Aside/question:
One factor that I'm not sure about is how well would this approach play
with other approaches to using RDFa? I put a CC licence statement in my
example, and it doesn't seem to have broken it, but it was pretty much
independent of the other statements, what would happen if there were
some overlaps? For example, can I use
<a xmlns:cc="http://creativecommons.org/ns#"
href="http://www.example.com"
property="cc:attributionName"
rel="cc:attributionURL">Phil Barker</a>
along with
<span itemprop="seller"> ...<span property="v:name">Phil
Barker</span>...</span>
(FWIW, I still don't think I've got the use of Google's "person" type as
a Seller right)
On 08/02/2011 09:54, Patrick Lockley wrote:
> I'd agree that it doesn't give us much directly and I'm trying to argue from a workload perspective as to what systems can do it easily (it'd be next to no time to get RDFa into Xpert for example) - however, also note that we still don't know how people search for OER.
>
> It might seem crazy to have price zero for example, but "Free educational resource" strikes me as something people would search for a long time before they put in OER.
>
> In adding RDFa, we are giving machines more information, and hopefully they can pass this onto their users / use it to facilitate better search results. If this takes next to no time then why not? If it takes an age this end up like a linked data discussion and I'll be off.
>
> -----Original Message-----
> From: Open Educational Resources [mailto:[log in to unmask]] On Behalf Of Phil Barker
> Sent: 07 February 2011 17:20
> To: [log in to unmask]
> Subject: Rich snippets for OER was Re: Discovering DC elements in HTML docs
>
> On 07/02/2011 10:29, Andy Powell wrote:
>> It would be interesting to think about how much ofhttp://blogs.cetis.ac.uk/lmc/2010/12/03/oer-2-technical-requirements/ can be carried in the RDFa vocabularies currently understood by Google. Probably quite a lot.
> Yep, it would.
>
> My first shot at working out which properties match (posted here so that
> folk can tell me what I've got wrong). I think Pat is right in
> suggesting that an OER as a product is the best fit (though I'm not sure
> it's a great one).
>
> See https://www.google.com/support/webmasters/bin/answer.py?answer=146750
>
> =Mandated Metadata=
> Programme tag = Brand?
> Project tag = Brand?
> Title = name
> Author / owner / contributor = seller?
> Date =
> URL = offerURL (but not on OER page itself)
> Licence information [Use CC code] price=0
>
> =Suggested Metadata=
> Language =
> Subject = category
> Keywords = category?
> Additonal tags = category?
> Comments = a review
> Description = description
>
>
> According to the guidance from Google the Seller property can contain a
> person or organization, so you could follow
> https://www.google.com/support/webmasters/bin/answer.py?hl=en&answer=146646
> to describe the Author/owner/contributor with a sort of RDF-ified vCard
> (yeuch, vCards. I hate vCards)
>
>
> I've tried all of that at http://www.icbl.hw.ac.uk/~philb/temp/gglSnip.html
>
> You can see the result at on the Google rich snippet testing tool at
> http://bit.ly/fZFveJ
>
> (I'm not sure that I got the Person as Seller property bit right).
>
>
> Thoughts? Myself, I'm not sure that it gives us much, though I guess
> what is more important is what google does with what it gives them.
>
> Phil
>
--
Please note new email address: [log in to unmask]
--
Heriot-Watt University is a Scottish charity
registered under charity number SC000278.
|