Re: ODF and semantic web

Mark,

Mark Birbeck wrote:
> Hi Ivan,
> 
> GRDDL is a necessary hack to allow legacy mark-up to be made
> 'semantic'. But I don't think anyone would seriously suggest that you
> can build a 'semantic web' on such a flaky framework. Which means that
> it's not a good idea to design languages on the basis that 'it doesn't
> matter what I do, because I can always GRDDL it'.
> 

I friendly, respectfully, but firmly disagree:-)

Indeed, this characterization of GRDDL is a bit dismissive and unfair.
Of course GRDDL can be used for what you describe. However. Applications
may define complex XML formats for their own application areas, which
may have all kinds of constraints on XML that we do not know about. If
they want to bind that application area, and hence the XML format, to
the Semantic Web, they may have two different routes to follow:

1. either they extend their XML application with additional RDF
attributes, for example by adopting the RDFa style attributes; or

2. they define the mapping from the XML format to RDF separately, and
put adequate references to that mapping in the namespace document of the
XML application; this is, essentially, the GRDDL approach

I agree that if approach #1 is chosen, then it is probably better to
re-use attributes that might have been used elsewhere and RDFa may
become a proper focal point for that (although at present we concentrate
on the XHTML1 version only...). But I would not dismiss the second
possibility. (A typical example might be XBRL[1] that people are working
on).

I do not think we should make a judgement on which of these approaches
are better in general; this should be left to the specialized
application. I also cannot judge at this point which of the two
alternatives are appropriate for ODF; it is not for me to say.

My fundamental goal is that the format, ie, the data it describes _is_
integrated to the Semantic Web; the application area/community has to
decide which route to take. Whichever their choice is, the integration
does merit a yey!:-)

Ivan

[1] http://www.xbrl.org

> So, I'm going to save my 'yey' for later. I'm hoping that there will
> be some serious coordination on this issue, and anything less is a
> missed opportunity.
> 
> It will be interesting to see if the two standards organisations can
> rise to the challenge.
> 
> All the best,
> 
> Mark
> 
> On 15/10/2007, Ivan Herman <ivan@w3.org> wrote:
>> Elias, Mark, Bruce, & al
>>
>> First of all: I think Elias is right. 'Our' (if one can define this
>> 'our', which is not always easy) first reaction should be (and believe
>> me, is!): yey! Having _some_ syntax to store RDF metadata in ODF _is_
>> major and good news. In some ways, _what_ the exact serialization syntax
>> is remains of a secondary importance as long as it is clearly defined
>> and transformable (via GRDDL or anything else, although GRDDL comes to
>> one's mind first) into other formats. So yes, yey!:-)
>>
>> As for Mark's concerns: yes, if RDFa could be used, that would be even
>> better, because it would reduce the number of overlapping serializations
>> and would therefore help in a quicker integration of ODF metadata into
>> the SW world. It would be good _if_ it is possible and meets the
>> constraints that ODF has. At this point, the obvious question and
>> comment is: what can be done to help improve this? There are some
>> (probably solvable) technical issues; and there are also 'social', ie,
>> the 'how to do it?', 'where and how to comment?' part. I think Elias'
>> and Bruce's advise on that would be really welcome. We can then try to
>> take it from there...
>>
>> Sincerely
>>
>> Ivan
>>
>> Elias Torres wrote:
>>>
>>>
>>> Mark Birbeck wrote:
>>>> Hi Bruce,
>>>>
>>>>> I've mentioned this here before, but more on RDF-in-OpenDocument.
>>>>>
>>>>> <http://www.robweir.com/blog/2007/10/odf-enters-semantic-web.html>
>>>>>
>>>>> The OpenOffice project is now starting to looking into implementing it,
>>>>> so people here might be interested.
>>>> I note that the attributes used in ODF are 'inspired' by RDFa [1]--but
>>>> why not just incorporate RDFa as is?
>>> First reason is because RDFa is still not finished, published,
>>> recommended etc. We are currently working on a XHTML 1.1 module and that
>>> I know of there's no work in progress for a recommendation on how to
>>> host RDFa in other XML languages. I understand that you have
>>> ideas/vision/plans, but just like every other standards group or task
>>> force, you can't depend on too many working drafts/vision/plan unless
>>> deadlines are of no concern. This is a very common practice at the W3C
>>> as well, so I hope that's enough for you to understand why we couldn't
>>> embed RDFa as is.
>>>
>>> Secondly, we are just getting to a point of good coverage of the issues
>>> that surround adding metadata to XHTML, but unfortunately, although you
>>> might see very little differences between XHTML and any other XML
>>> vocabulary, there's a lot of things to both work out and build from
>>> scratch in some XML vocabularies where something as basic as the
>>> document location/hyperlink not being defined as it's the case for ODF.
>>>
>>> I think we are too critical when we make these statements especially
>>> when I indicated many times in our calls that I was involved in this
>>> work and I didn't hear anyone volunteering to help. For example, several
>>> times I went as far as asking you personally for suggestions in some
>>> aspects of the RDFa spec before it was even brought up to the RDFa task
>>> force for sake of progress in the ODF metadata specification.
>>> Unfortunately, the task forces were working in parallel and it just
>>> wasn't feasible to combine both groups, learn each others requirements
>>> and deliver a single document. I wish things were as simple as me
>>> saying: hey guys let's use the RDFa spec from the W3C and put an
>>> OASIS/ODF rubber stamp on it and everyone just said: great, let's do that.
>>>
>>>> It's especailly confusing for authors when this 'inpiration' seems to
>>>> involve copying some RDFa attributes, but changing the names of
>>>> others. For example, @about is used, but @datatype has been renamed to
>>>> @data-type!
>>> I would first hope that there's no a problem with us getting
>>> inspiration/copying the RDFa attributes. It was a long and arduous
>>> process to get where we are today. Bruce and I put in a LOT of time and
>>> patience until the group passed from storming to performing.
>>> Svante/Patrick put in an amazing effort with the documents (very similar
>>> to what you have done with the RDFa documents). We are now a
>>> happy/loving bunch and Bruce and I are grateful to our colleagues for
>>> putting up enough with us to the point that they now share our vision
>>> for metadata in office applications. It really took a lot of listening
>>> on their part for us to share everything we thought was great about the
>>> RDFa work. The ODF Metadata group was so much more welcoming to our
>>> perspective as opposed to other non-SW bred groups and us arguing about
>>> 'dash' felt to me disrespectful, if not rude.
>>>
>>>> This lack of alignment is a shame, especially when the proponents of
>>>> ODF are generally critical of the confusion that can be caused by
>>>> companies and organisations pursuing alternate document formats. There
>>>> is a fantastic opportunity here for creating tools and search engines
>>>> that could leverage a 'standard' way of incorporating metadata into
>>>> HTML, XHTML, ODF, and other mark-up languages. That opportunity now
>>>> looks like it is going to be missed.
>>> I'm not as intimate with the ODF organization, but I would not confuse
>>> this sub-committee/task force with the rest of the organization.
>>> Besides, I think the issues surrounding OOXML and ODF are orthogonal to
>>> what you claim is happening in this 'divergence' of formats. Of course,
>>> I believe that there's a fantastic opportunity here for creating tools
>>> and blah blah into HTML, XHTML, ODF, etc. But please don't blame us for
>>> the fact that not everyone in the world wants to adhere to our
>>> views/technology of the Semantic Web. I think that this 'standard' way
>>> of thinking has hurt us more than helped to reach the goal. I totally
>>> disagree that one parser will  be capable to address the issues of
>>> metadata in ODF vs HTML. I caught myself making those arguments to later
>>> change my mind and understand that in the end it's just code that gets
>>> written and overwritten every other day, but a consensus to work
>>> together as individuals and put our differences aside is much harder to
>>> develop, no pun intended.
>>>
>>> We were hoping to receive a warm welcome for the work we put into the
>>> ODF Metadata for the purpose of advancing the Semantic Web, but as
>>> always, you can't please everyone. Fortunately, I still believe ODF
>>> Metadata + RDF/XML is making the case for extensibility, flexibility,
>>> linked data, openness and so on, independently of whether we used the
>>> same parser or not. We need to keep examining ourselves in the likes of
>>> Bijan [1] so we assess what are the real problems hindering progress on
>>> the Web by our standards and do more showing/telling and
>>> implementation/adoption before rushing to standardizing. At least I
>>> partially felt that way with ODF Metadata and towards the end of the
>>> first draft, I agreed that less was better given that this was the first
>>> introduction of RDF to the ODF world. Look at Mozilla for example and I
>>> hope that we start small and prove the value before forcing things
>>> without immediate benefits.
>>>
>>> DISCLAIMER: At the risk of sounding schizo, here it goes. Mark, you know
>>> we are cool and I'm not at all targeting everything towards you only but
>>> to the larger community. We are colleagues, have been working together
>>> for a while now and share a lot in common when it comes to RDFa, but I
>>> had been meaning to reply to Bijan's email and vent a little on some of
>>> the issues surrounding many groups/technologies on the W3C and you had
>>> to push me over the edge :D
>>>
>>> [1] http://lists.w3.org/Archives/Public/semantic-web/2007Oct/0039.html
>>>
>>>> Regards,
>>>>
>>>> Mark
>>>>
>>>> [1]
>>>> <http://www.oasis-open.org/committees/download.php/25055/Metadata_22August2007.txt>
>>>>
>>>>
>> --
>>
>> Ivan Herman, W3C Semantic Web Activity Lead
>> Home: http://www.w3.org/People/Ivan/
>> PGP Key: http://www.ivan-herman.net/pgpkey.html
>> FOAF: http://www.ivan-herman.net/foaf.rdf
>>
>>
> 
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf