Re: is this valid to make a named graph in RDFa?

On 10 Mar 2008, at 15:30, Golda Velez wrote:

>
> On Monday 10 March 2008 2:48, Tim Berners-Lee wrote:
>>>> But if you really need to play with that use N3. The best is to try
>>>> to
>>>> keep things as simple as possible.
>>>
>>> Hm - I might use N3 on the back end eventually, right now its all
>>> mysql and perl - but the presentation end is what I'm working on
> currently,  and that has to be something browser-parsable.
>>
>> If you have the Tabulator extension for Firefox, it will allow a  
>> user to
>> read the data an N3 graph, with subgraphs in bubbles.  We have used  
>> it
>> to look at justifications of the form 'that document with these
>> properties led us to believe the following ...'
>>
>> timbl
>>
>
> Got it - I'd meant to install Tabulator, now I have.  But, when I said
> 'browser-parsable', I should have specified 'for the average user'  
> since my
> sites are directed at non-technical audience, and I want them to help
> annotate the data.

You are in a bit of a dilemma here. You want a non technical audience  
to generate RDF directly? That would be working at the wrong level of  
abstraction: a bit like asking a conusmer desktop user to program in  
C. True, most programs that the end user writes are written in C, but  
no one asks the end user to know that fact, let alone be able to work  
with it. The same argument would of course be made with Java.

If you want to help them annotate data, then you have to do that in a  
way that end users understand, by creating simple tools such as say  
jPhoto, that would be a bit like iPhoto, but allow one to annotate  
photos anywhere on the web. In that case you can use N3 since you  
control the output format that the end user never sees.

Still I don't know of a simple user interface that a non technical  
user would understand as being a way to express propositional  
attitudes, other than that he believes the world to be a certain way.  
I imagine this would make sense in a user interface dealing with  
historical research.  Clearly such tools will be developed. But to  
start with those is to start with some of the most difficult problems,  
as you have to solve the user interface problem, and the complex  
reasoning problems that come from dealing with this level of  
abstraction.

As I said previously, it is much better to start out by seeing how far  
one can go with simple factual assertions.

Later we can look at propositions such as  "If kangaroos had no tails  
they would topple over. "

> Do you think that a tool like Tabulator is necessary to write into the
> semantic web?   I was envisioning a portal site that presents data  
> to the
> user in a very low-entry-bar way, and uses old fashioned forms to  
> get users
> to enter annotations in a precise way. This would include making  
> annotations
> about other URIs that are not inside the portal, hopefully including  
> URIs of
> individual statements or graphs.

You can write to the Semantic Web using any tool you like.
FTP, SCP, HTTP PUT, WEbDAV are all good well established protocols  
your tools can use.

> If you use Tabulator, can you 'write' data into/onto/about any page  
> that has
> RDF in it?

No, since you don't have write access to every page.

> Do the original server(s) where the data is stored have to accept
> the update, or can you write it into some other friendly data  
> repositories
> somewhere?

No of course not. Do people around you have to accept your point of  
view on the world?
What we could develop btw, which would be RDF friendly, would be  
simple propositional attitudes towards contents of resources
such as

:me dontBelieve :joesFoafFile .

That would be a little bit like the <a rel="nofollow"  
href="blahblah" ...> link .

A simple vocabulary like that would be useable with the new quad  
stores that are being developed, and may even not require too much  
work developing.


> I'll try to read more precisely but I don't want to draw the
> wrong conclusions and start spreading rumors ;-)
>
> Thanks!
>
> --Golda
>