Quantcast

Hash URIs and RDFa documents

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
7 messages Options
Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Hash URIs and RDFa documents

Stéphane Corlosquet
Hi,

Imagine an RDFa document describing a person. The foaf:Document URI is <> and the foaf:Person URI is <#person> because you want to be able to distinguish between the two; also foaf:Document and foaf:Person are disjoint.

<>        a foaf:Document .
<#person> a foaf:Person .
<> foaf:primaryTopic <#person> .

If the document is describing an online account, it might have
<>        a sioc:User .
<#person> foaf:account <>.

My concern is about the #person fragment with regard to the HTML document. If the page is only about one person, there might not be a tag with id="person" in the page. Is this a problem? Should I have a tag with such id, or, on the contrary, should I avoid this as to ensure the resource being described is not confused with the actual HTML tag contained in the page?

cc'ing Dan since this message is related to his point #5 at http://danbri.org/words/2010/01/14/549

regards,
Stéphane.
Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hash URIs and RDFa documents

Mark Birbeck-4
Hi Stéphane,

As you imply, there is actually no need for an @id value. For some
reason many examples that people have created in the past have tried
to align @about and @id, but it really isn't necessary.

Whilst it doesn't hurt to have an @id though, my preference would be not to.

As things stand today it's unlikely that there would be a confusion,
but I don't know if you are familiar with @role (another W3C standard
which myself, Shane and Steven have been involved in), but with that
you really are making statements about an HTML element. I'm hoping
that the RDFa/@role story is properly fleshed out at some point, so I
think it would be good to try to keep the boundaries clear, ready for
this.

So I'd vote for your second option. :)

Regards,

Mark

--
Mark Birbeck, webBackplane

[hidden email]

http://webBackplane.com/mark-birbeck

webBackplane is a trading name of Backplane Ltd. (company number
05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
London, EC2A 4RR)



On Mon, Feb 1, 2010 at 9:45 PM, Stephane Corlosquet
<[hidden email]> wrote:

> Hi,
>
> Imagine an RDFa document describing a person. The foaf:Document URI is <>
> and the foaf:Person URI is <#person> because you want to be able to
> distinguish between the two; also foaf:Document and foaf:Person are
> disjoint.
>
> <>        a foaf:Document .
> <#person> a foaf:Person .
> <> foaf:primaryTopic <#person> .
>
> If the document is describing an online account, it might have
> <>        a sioc:User .
> <#person> foaf:account <>.
>
> My concern is about the #person fragment with regard to the HTML document.
> If the page is only about one person, there might not be a tag with
> id="person" in the page. Is this a problem? Should I have a tag with such
> id, or, on the contrary, should I avoid this as to ensure the resource being
> described is not confused with the actual HTML tag contained in the page?
>
> cc'ing Dan since this message is related to his point #5 at
> http://danbri.org/words/2010/01/14/549
>
> regards,
> Stéphane.
>

Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hash URIs and RDFa documents

Michael Hausenblas
Stephane,
 
I by and large agree with Mark, however ...

> Whilst it doesn't hurt to have an @id though, my preference would be not to.

Hm. Not so sure about this one. Let's look at the following example,
namespace declarations assumed to be done already, and the base URI is
"http://sw-app.org/"

1: <div about="#me" typeof="foaf:Person">
2:  <a rel="foaf:homepage" href="http://sw-app.org/about.html">my
homepage</a>
3: </div>
4: <div id="me">
5:  yada yada
6: </div>

What I am saying here is that the *part of the document (4-6)*, identified
by "http://sw-app.org/#me, is of type foaf:Person. Richard explained it much
more elegant a while ago [1]. I don't think this is what we want to express.

In any case, I'm gonna take this discussion into account for updates on [2].

Cheers,
      Michael

[1] http://lists.w3.org/Archives/Public/semantic-web/2007Dec/0157.html
[2] http://ld2sd.deri.org/lod-ng-tutorial/#checklist-fragid

--
Dr. Michael Hausenblas
LiDRC - Linked Data Research Centre
DERI - Digital Enterprise Research Institute
NUIG - National University of Ireland, Galway
Ireland, Europe
Tel. +353 91 495730
http://linkeddata.deri.ie/
http://sw-app.org/about.html



> From: Mark Birbeck <[hidden email]>
> Date: Mon, 1 Feb 2010 22:55:12 +0000
> To: Stephane Corlosquet <[hidden email]>
> Cc: RDFa TF list <[hidden email]>, Dan Brickley
> <[hidden email]>
> Subject: Re: Hash URIs and RDFa documents
> Resent-From: RDFa TF list <[hidden email]>
> Resent-Date: Mon, 01 Feb 2010 22:56:50 +0000
>
> Hi Stéphane,
>
> As you imply, there is actually no need for an @id value. For some
> reason many examples that people have created in the past have tried
> to align @about and @id, but it really isn't necessary.
>
> Whilst it doesn't hurt to have an @id though, my preference would be not to.
>
> As things stand today it's unlikely that there would be a confusion,
> but I don't know if you are familiar with @role (another W3C standard
> which myself, Shane and Steven have been involved in), but with that
> you really are making statements about an HTML element. I'm hoping
> that the RDFa/@role story is properly fleshed out at some point, so I
> think it would be good to try to keep the boundaries clear, ready for
> this.
>
> So I'd vote for your second option. :)
>
> Regards,
>
> Mark
>
> --
> Mark Birbeck, webBackplane
>
> [hidden email]
>
> http://webBackplane.com/mark-birbeck
>
> webBackplane is a trading name of Backplane Ltd. (company number
> 05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
> London, EC2A 4RR)
>
>
>
> On Mon, Feb 1, 2010 at 9:45 PM, Stephane Corlosquet
> <[hidden email]> wrote:
>> Hi,
>>
>> Imagine an RDFa document describing a person. The foaf:Document URI is <>
>> and the foaf:Person URI is <#person> because you want to be able to
>> distinguish between the two; also foaf:Document and foaf:Person are
>> disjoint.
>>
>> <>        a foaf:Document .
>> <#person> a foaf:Person .
>> <> foaf:primaryTopic <#person> .
>>
>> If the document is describing an online account, it might have
>> <>        a sioc:User .
>> <#person> foaf:account <>.
>>
>> My concern is about the #person fragment with regard to the HTML document.
>> If the page is only about one person, there might not be a tag with
>> id="person" in the page. Is this a problem? Should I have a tag with such
>> id, or, on the contrary, should I avoid this as to ensure the resource being
>> described is not confused with the actual HTML tag contained in the page?
>>
>> cc'ing Dan since this message is related to his point #5 at
>> http://danbri.org/words/2010/01/14/549
>>
>> regards,
>> Stéphane.
>>
>


Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hash URIs and RDFa documents

Mark Birbeck-4
Hi Michael,

I disagree with the point made in the email you refer to, from Richard.

Only one triple is generated in your sample, so nowhere do we have an
indication that #me is both a person and an HTML element in a
document.

My point is that it would be good practice to keep these things apart,
but I think it's going too far to say that we create some kind of
contradiction if we don't.

Regards,

Mark



On Tue, Feb 2, 2010 at 8:38 AM, Michael Hausenblas
<[hidden email]> wrote:

> Stephane,
>
> I by and large agree with Mark, however ...
>
>> Whilst it doesn't hurt to have an @id though, my preference would be not to.
>
> Hm. Not so sure about this one. Let's look at the following example,
> namespace declarations assumed to be done already, and the base URI is
> "http://sw-app.org/"
>
> 1: <div about="#me" typeof="foaf:Person">
> 2:  <a rel="foaf:homepage" href="http://sw-app.org/about.html">my
> homepage</a>
> 3: </div>
> 4: <div id="me">
> 5:  yada yada
> 6: </div>
>
> What I am saying here is that the *part of the document (4-6)*, identified
> by "http://sw-app.org/#me, is of type foaf:Person. Richard explained it much
> more elegant a while ago [1]. I don't think this is what we want to express.
>
> In any case, I'm gonna take this discussion into account for updates on [2].
>
> Cheers,
>      Michael
>
> [1] http://lists.w3.org/Archives/Public/semantic-web/2007Dec/0157.html
> [2] http://ld2sd.deri.org/lod-ng-tutorial/#checklist-fragid
>
> --
> Dr. Michael Hausenblas
> LiDRC - Linked Data Research Centre
> DERI - Digital Enterprise Research Institute
> NUIG - National University of Ireland, Galway
> Ireland, Europe
> Tel. +353 91 495730
> http://linkeddata.deri.ie/
> http://sw-app.org/about.html
>
>
>
>> From: Mark Birbeck <[hidden email]>
>> Date: Mon, 1 Feb 2010 22:55:12 +0000
>> To: Stephane Corlosquet <[hidden email]>
>> Cc: RDFa TF list <[hidden email]>, Dan Brickley
>> <[hidden email]>
>> Subject: Re: Hash URIs and RDFa documents
>> Resent-From: RDFa TF list <[hidden email]>
>> Resent-Date: Mon, 01 Feb 2010 22:56:50 +0000
>>
>> Hi Stéphane,
>>
>> As you imply, there is actually no need for an @id value. For some
>> reason many examples that people have created in the past have tried
>> to align @about and @id, but it really isn't necessary.
>>
>> Whilst it doesn't hurt to have an @id though, my preference would be not to.
>>
>> As things stand today it's unlikely that there would be a confusion,
>> but I don't know if you are familiar with @role (another W3C standard
>> which myself, Shane and Steven have been involved in), but with that
>> you really are making statements about an HTML element. I'm hoping
>> that the RDFa/@role story is properly fleshed out at some point, so I
>> think it would be good to try to keep the boundaries clear, ready for
>> this.
>>
>> So I'd vote for your second option. :)
>>
>> Regards,
>>
>> Mark
>>
>> --
>> Mark Birbeck, webBackplane
>>
>> [hidden email]
>>
>> http://webBackplane.com/mark-birbeck
>>
>> webBackplane is a trading name of Backplane Ltd. (company number
>> 05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
>> London, EC2A 4RR)
>>
>>
>>
>> On Mon, Feb 1, 2010 at 9:45 PM, Stephane Corlosquet
>> <[hidden email]> wrote:
>>> Hi,
>>>
>>> Imagine an RDFa document describing a person. The foaf:Document URI is <>
>>> and the foaf:Person URI is <#person> because you want to be able to
>>> distinguish between the two; also foaf:Document and foaf:Person are
>>> disjoint.
>>>
>>> <>        a foaf:Document .
>>> <#person> a foaf:Person .
>>> <> foaf:primaryTopic <#person> .
>>>
>>> If the document is describing an online account, it might have
>>> <>        a sioc:User .
>>> <#person> foaf:account <>.
>>>
>>> My concern is about the #person fragment with regard to the HTML document.
>>> If the page is only about one person, there might not be a tag with
>>> id="person" in the page. Is this a problem? Should I have a tag with such
>>> id, or, on the contrary, should I avoid this as to ensure the resource being
>>> described is not confused with the actual HTML tag contained in the page?
>>>
>>> cc'ing Dan since this message is related to his point #5 at
>>> http://danbri.org/words/2010/01/14/549
>>>
>>> regards,
>>> Stéphane.
>>>
>>
>
>

Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hash URIs and RDFa documents

Niklas Lindström
Hi!

I have been thinking the following regarding fragment identifiers.
Given an *information* resource (IR):

    http://example.org/about

Say it has the following representations:

    http://example.org/about.html
    http://example.org/about.n3

These are distinct resources.

When requesting <http://example.org/about> (an IR) with HTTP using
content negotiation (here asking for "text/html") allows the server to
send a 200 OK upon request, with the body being from
<http://example.org/about.html> (given in Content-Location).

(This is different from conneg on non-IR:s, which require 303:ing.)

I then stipulate that these two are owl:differentFrom each other (not
necessarily, but quite possibly):

    http://example.org/about#me
    http://example.org/about.html#me

The first identifies a non-IR, i.e. *me* (because I define it as that,
as (fictive) owner of example.org). The other, perhaps an element in
the html (not the bytes, but the DOM? this quickly gets slippery).

The HTTP + HTML (web browser) mechanics makes it work for me to
request <http://example.org/about#me>, receiving the html according to
above without any "change in the url bar" (i.e. "browser mechanics"),
and thus the browser view jumps to the element referenced by the
fragment identifier (the id alone, from the URI having been
*string-parsed* by the browser). But that is not the element's URI. It
*could* be, possibly, <http://example.org/about.html#me>. Those
semantics (or lack thereof) are reasonably under the provenance of the
html spec authors.

(I'm not really claiming anything about what
<http://example.org/about.html#me> identifies. I think that, even with
@id being present in the document, it's not actually defined. But for
this reason I think it's advisable to avoid using them, at least as
URI:s for non-IR:s.)

I think <http://www.w3.org/TR/cooluris/> support these claims, but I
could have mixed things up here.

What do you say?

(If these claims aren't just nonsense, getting the triples right in
RDFa for a document carrying triples about both </about> and
</about.html> (and fragments based on them) is an interesting next
exercise..)

Best regards,
Niklas



On Tue, Feb 2, 2010 at 11:13 AM, Mark Birbeck
<[hidden email]> wrote:

> Hi Michael,
>
> I disagree with the point made in the email you refer to, from Richard.
>
> Only one triple is generated in your sample, so nowhere do we have an
> indication that #me is both a person and an HTML element in a
> document.
>
> My point is that it would be good practice to keep these things apart,
> but I think it's going too far to say that we create some kind of
> contradiction if we don't.
>
> Regards,
>
> Mark
>
>
>
> On Tue, Feb 2, 2010 at 8:38 AM, Michael Hausenblas
> <[hidden email]> wrote:
>> Stephane,
>>
>> I by and large agree with Mark, however ...
>>
>>> Whilst it doesn't hurt to have an @id though, my preference would be not to.
>>
>> Hm. Not so sure about this one. Let's look at the following example,
>> namespace declarations assumed to be done already, and the base URI is
>> "http://sw-app.org/"
>>
>> 1: <div about="#me" typeof="foaf:Person">
>> 2:  <a rel="foaf:homepage" href="http://sw-app.org/about.html">my
>> homepage</a>
>> 3: </div>
>> 4: <div id="me">
>> 5:  yada yada
>> 6: </div>
>>
>> What I am saying here is that the *part of the document (4-6)*, identified
>> by "http://sw-app.org/#me, is of type foaf:Person. Richard explained it much
>> more elegant a while ago [1]. I don't think this is what we want to express.
>>
>> In any case, I'm gonna take this discussion into account for updates on [2].
>>
>> Cheers,
>>      Michael
>>
>> [1] http://lists.w3.org/Archives/Public/semantic-web/2007Dec/0157.html
>> [2] http://ld2sd.deri.org/lod-ng-tutorial/#checklist-fragid
>>
>> --
>> Dr. Michael Hausenblas
>> LiDRC - Linked Data Research Centre
>> DERI - Digital Enterprise Research Institute
>> NUIG - National University of Ireland, Galway
>> Ireland, Europe
>> Tel. +353 91 495730
>> http://linkeddata.deri.ie/
>> http://sw-app.org/about.html
>>
>>
>>
>>> From: Mark Birbeck <[hidden email]>
>>> Date: Mon, 1 Feb 2010 22:55:12 +0000
>>> To: Stephane Corlosquet <[hidden email]>
>>> Cc: RDFa TF list <[hidden email]>, Dan Brickley
>>> <[hidden email]>
>>> Subject: Re: Hash URIs and RDFa documents
>>> Resent-From: RDFa TF list <[hidden email]>
>>> Resent-Date: Mon, 01 Feb 2010 22:56:50 +0000
>>>
>>> Hi Stéphane,
>>>
>>> As you imply, there is actually no need for an @id value. For some
>>> reason many examples that people have created in the past have tried
>>> to align @about and @id, but it really isn't necessary.
>>>
>>> Whilst it doesn't hurt to have an @id though, my preference would be not to.
>>>
>>> As things stand today it's unlikely that there would be a confusion,
>>> but I don't know if you are familiar with @role (another W3C standard
>>> which myself, Shane and Steven have been involved in), but with that
>>> you really are making statements about an HTML element. I'm hoping
>>> that the RDFa/@role story is properly fleshed out at some point, so I
>>> think it would be good to try to keep the boundaries clear, ready for
>>> this.
>>>
>>> So I'd vote for your second option. :)
>>>
>>> Regards,
>>>
>>> Mark
>>>
>>> --
>>> Mark Birbeck, webBackplane
>>>
>>> [hidden email]
>>>
>>> http://webBackplane.com/mark-birbeck
>>>
>>> webBackplane is a trading name of Backplane Ltd. (company number
>>> 05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
>>> London, EC2A 4RR)
>>>
>>>
>>>
>>> On Mon, Feb 1, 2010 at 9:45 PM, Stephane Corlosquet
>>> <[hidden email]> wrote:
>>>> Hi,
>>>>
>>>> Imagine an RDFa document describing a person. The foaf:Document URI is <>
>>>> and the foaf:Person URI is <#person> because you want to be able to
>>>> distinguish between the two; also foaf:Document and foaf:Person are
>>>> disjoint.
>>>>
>>>> <>        a foaf:Document .
>>>> <#person> a foaf:Person .
>>>> <> foaf:primaryTopic <#person> .
>>>>
>>>> If the document is describing an online account, it might have
>>>> <>        a sioc:User .
>>>> <#person> foaf:account <>.
>>>>
>>>> My concern is about the #person fragment with regard to the HTML document.
>>>> If the page is only about one person, there might not be a tag with
>>>> id="person" in the page. Is this a problem? Should I have a tag with such
>>>> id, or, on the contrary, should I avoid this as to ensure the resource being
>>>> described is not confused with the actual HTML tag contained in the page?
>>>>
>>>> cc'ing Dan since this message is related to his point #5 at
>>>> http://danbri.org/words/2010/01/14/549
>>>>
>>>> regards,
>>>> Stéphane.
>>>>
>>>
>>
>>
>
>

Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hash URIs and RDFa documents

Toby Inkster-4
In reply to this post by Mark Birbeck-4
On Tue, 2010-02-02 at 10:13 +0000, Mark Birbeck wrote:
> I disagree with the point made in the email you refer to, from
> Richard.
>
> Only one triple is generated in your sample, so nowhere do we have an
> indication that #me is both a person and an HTML element in a
> document.

The graph generated from the example by an RDFa processor would not
contain an indication that #me is an HTML element, no. However another
graph, say, from a document outline generator, may contain something
like:

        <#me> a xhtml:Div ;
              xhtml:textContent "yada yada" .

It would be a perfectly sensible couple of triples to generate from the
given example. Merging the RDFa and document outline graphs would result
in a contradiction. (At least, it's a contradiction if we assume that
people and HTML elements are disjoint classes.)

--
Toby A Inkster
<mailto:[hidden email]>
<http://tobyinkster.co.uk>


Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hash URIs and RDFa documents

Mark Birbeck-4
Hi Toby,

You cut out my main point, so here it is again:

> My point is that it would be good practice to keep these things apart,
> but I think it's going too far to say that we create some kind of
> contradiction if we don't.

You can come up with all sorts of scenarios whereby contradictions are
created when triples from one document meet triples from another.

And it's also possible to get contradictions in the same document,
such as when using @role.

But the thread to that point was implying that just by using @id, a
contradiction was created, and that is simply not true.

So, I think my original points stand:

1. There is no *need* to have @about aligned with @id.

2. But there is no *inherent* contradiction created if you do.

3. However, because other scenarios will arise where we want to refer
to that element *as an element*, then we would do well to avoid
liberally adding @id values, because we could create contradictions
later.

Regards,

Mark


On Tue, Feb 2, 2010 at 7:07 PM, Toby Inkster <[hidden email]> wrote:

> On Tue, 2010-02-02 at 10:13 +0000, Mark Birbeck wrote:
>> I disagree with the point made in the email you refer to, from
>> Richard.
>>
>> Only one triple is generated in your sample, so nowhere do we have an
>> indication that #me is both a person and an HTML element in a
>> document.
>
> The graph generated from the example by an RDFa processor would not
> contain an indication that #me is an HTML element, no. However another
> graph, say, from a document outline generator, may contain something
> like:
>
>        <#me> a xhtml:Div ;
>              xhtml:textContent "yada yada" .
>
> It would be a perfectly sensible couple of triples to generate from the
> given example. Merging the RDFa and document outline graphs would result
> in a contradiction. (At least, it's a contradiction if we assume that
> people and HTML elements are disjoint classes.)
>
> --
> Toby A Inkster
> <mailto:[hidden email]>
> <http://tobyinkster.co.uk>
>
>

Loading...