Re: Determination of subjects/objects (was: ISSUE-42)

Hi,

Am 30.07.2007 um 17:15 schrieb Manu Sporny:

> Mark Birbeck wrote:
>> I haven't had a chance to re-read this thread, so I'm not going to  
>> say
>> anything on the substance. But if you don't mind, I'd like to comment
>> on a recurring theme, which seems exemplified by the following:
>>
>>> I have a visceral problem with about="_:", and that is that it makes
>>> bnodes explicit, which I really don't want to do to HTML authors.  
>>> That's
>>> just too much RDF.
>>>
>> I don't see the need to 'protect' authors who are not familiar with
>> RDF from RDF constructs that they will never use. If someone from the
>> RDF community thinks this is useful, and _if_ we can support it
>> without it getting in the way, then why not?
>
> Constructs such as "_:" are scary to non-RDF folks. :) From a
> historically RDF-unaware perspective (mine), I stared at the "_:"
> construct and had no idea what it does. It is not very intuitive.
>
> Even having seen it, I haven't taken the time to look up what it  
> means.
> It will probably make sense when I do, but to somebody that is not
> trained in CS/EE/ECE/etc., this is a scary construct. To the lay web
> page author, it is syntactic gibberish.
>
> There is already a very strong feeling in the Microformats community
> that RDFa is far too complicated for most web page authors. The last
> thing most of them want to learn is yet another language syntax for
> describing what they see as "corner-cases of the language".
>
> I see your argument: If they aren't going to use it, and if it doesn't
> cause any harm, then why not put it in there?
>
> I would argue that you shouldn't put things in there that aren't
> absolutely necessary. It complicates the RDFa specification. If  
> there is
> a need in the future, you can always add it in a later revision.

I would really argue to have bnodes in RDFa: I know they are very  
unpopular, and vocabularies like FOAF now recommend against using  
them. However, I hope that RDFa will, once it's finished, be a syntax  
for RDF that covers the whole specification. In fact, when I first  
heard about RDFa, I was put off, because I somehow thought it didn't  
allow you to express everything you can model in RDF.

In my particular use case (generating RDFa from desktop objects like  
address book or calendar entries) I often don't have a useful URI for  
the things I want to publish in RDFa: what is the URI of a person in  
my address book? What is the URI of an event in my calendar? I have  
the choice tof making up a bogus URI (that's bad, because it can't be  
resolved) or using a blank node. I find the latter option much more  
elegant.

I think that, once the RDFa specification is done, problems like  
scaring people off with concepts like bnodes can be handled on the  
documentation level. I mean, think of all the weird stuff you can do  
with Java (or any programming language). This hasn't stopped it from  
becoming very popular indeed. As long as you give people some good  
"Hello World" examples, and convince them that your technology is  
actually useful, they will still use it, despite its complexity.

Cheers,
Knud
-------------------------------------------------
Knud Möller, MA
+353 - 91 - 495086
Smile Group: http://smile.deri.ie

Digital Enterprise Research Institute
   National University of Ireland, Galway
Institiúid Taighde na Fiontraíochta Digití
   Ollscoil na hÉireann, Gaillimh