On https://meta.wikimedia.org/wiki/Wikidata/Notes/URI_scheme#Proposal_for_Wikid... I am missing the arguments why Wikidata needs the multitude of URI forms. The list needs commenting and arguments why
"URIs should be canonical within Wikimedia projects"
is given up.
----
(Of lesser importance, I wonder why the internal opaque ID has to be prefixed by a letter (Q) - why not a simple number?)
----
At them moment I believe a choice should be made between: http://%7Bsite%7D.wikidata.org/wiki/%7Btitle%7D and http://wikidata.org/title/%7Bsite%7D:%7Btitle%7D but perhaps the argument why both are needed could be added.
----
On the issue of interlanguage/interwiki linking, I believe the semantics should not be opposite in Wikidata and Wikipedias, see my comment https://meta.wikimedia.org/wiki/Talk:Wikidata/Notes/Wiki_links
Gregor
Hi Gregor,
thanks for your great comments! I am trying to answer them here.
2012/7/5 Gregor Hagedorn g.m.hagedorn@gmail.com:
On https://meta.wikimedia.org/wiki/Wikidata/Notes/URI_scheme#Proposal_for_Wikid... I am missing the arguments why Wikidata needs the multitude of URI forms. The list needs commenting and arguments why
"URIs should be canonical within Wikimedia projects"
is given up.
It is not. There is a canonical form. This does not mean it is the only one, but every other form can be canonicalized into the canonical one, and tools should work with that. (This is the same meaning as with titles in Wikipedia: whereas both https://en.wikipedia.org/wiki/Obama and https://en.wikipedia.org/wiki/Barack_Obama are URLs for the article on Barack Obama, only the latter one is canonical -- the first one is just a convenience URL).
(Of lesser importance, I wonder why the internal opaque ID has to be prefixed by a letter (Q) - why not a simple number?)
Originally this comes from a limitation in XML: the local part of a QName must not start with a number, so we prefixed it with a letter. As most export formats should also work in XML, we took that as an important enough restriction to make it move into our Identifiers. Furthermore, using a letter like Q which is rather seldom otherwise increases the likelihood of a Wikidata ID being recognized with little context (i.e. in a text, Q7237 will be in the end more readily recognized as a Wikidata ID than 7237 alone).
At them moment I believe a choice should be made between: http://%7Bsite%7D.wikidata.org/wiki/%7Btitle%7D and http://wikidata.org/title/%7Bsite%7D:%7Btitle%7D but perhaps the argument why both are needed could be added.
This is for merely technical reasons. Both URLs are convenience URLs anyway. The canonical one is the one with the ID.
On the issue of interlanguage/interwiki linking, I believe the semantics should not be opposite in Wikidata and Wikipedias, see my comment https://meta.wikimedia.org/wiki/Talk:Wikidata/Notes/Wiki_links
Added comment there. In short, the current proposal is actually more consistent over all the Wikimedia-projects, which means it is inconsistent how it is handled "internally" within the Wikipedia-projects. This means that the current suggestion is more or less what the software already does, and does not require much additional implementation.
Gregor
Cheers, Denny
Thanks Denny, I largely see your points. The distinction between convenience = webservice to redirect to canonical URL and canonical URL could perhaps be made clearer in the note. I read it as parallel URIs rather than as a redirecting service. To me the word "convenience" has a different implication, but this may be entirely my fault, I am not a native speaker either. I also agree on the choice of language prefixes, confusing as it may be, I should have know. The data plus wikidata is still confusing, but I guess you cannot avoid that one?
About the Q in front of identifiers: At the moment I see the item numbers being used in rdf:resource/about, but I understand that you may need them as element names? My understanding was that properties will be prefixed by Property: anyways.
In any event: I find the argument that a rare letter like Q is good branding not very convincing. I would suggest then a more memnonic choice, like WD2348972 or W2348972 instead. I believe the Q as prefix used in all canonical inbound links will be puzzling many people and end the explanation having to end up in the FAQ.
thanks again!
Gregor
Very delayed reply but think I'm still confused on this. Made a picture to clear my mind but not sure it works: http://smethur.st/wikidata
The bit I think I get: If I request http://en.wikidata.org/wiki/Berlin Or http://en.wikidata.org/title/Berlin I get a 301? to: http://wikidata.org/title/en:Berlin The html wiki page
But not sure I understand the machine readable part [1]
Bullet point 1 says http://wikidata.org/id/Q%7Bid%7D Resolves to the appropriate url depending on the request header
Does resolve mean a redirect? Is that a 303?
Or is there no redirect and the "thing" uri returns content?
What's the "appropriate url"? http://wikidata.org/data/Q%7Bid%7D Or http://wikidata.org/data/Q%7Bid%7D?format=%7Bformat%7D&language=%7Blangu... ?
Bullet point 2 says http://en.wikidata.org/item/Berlin Also resolves to the appropriate url. Is that a redirect? What's the appropriate url?
Is there content negotiation happening from http://wikidata.org/wiki/Q%7Bid%7D
Or just from http://wikidata.org/id/Q%7Bid%7D
What happens if I request http://wikidata.org/id/Q%7Bid%7D And accept only html?
Is there content negotiation from http://wikidata.org/data/Q%7Bid%7D Or do I have to use parameters to get different representations?
Is there a better picture
Sorry to be thick Michael
[1] https://meta.wikimedia.org/wiki/Wikidata/Notes/URI_scheme#Machine-readable_a ccess
On 06/07/2012 18:20, "Gregor Hagedorn" g.m.hagedorn@gmail.com wrote:
Thanks Denny, I largely see your points. The distinction between convenience = webservice to redirect to canonical URL and canonical URL could perhaps be made clearer in the note. I read it as parallel URIs rather than as a redirecting service. To me the word "convenience" has a different implication, but this may be entirely my fault, I am not a native speaker either. I also agree on the choice of language prefixes, confusing as it may be, I should have know. The data plus wikidata is still confusing, but I guess you cannot avoid that one?
About the Q in front of identifiers: At the moment I see the item numbers being used in rdf:resource/about, but I understand that you may need them as element names? My understanding was that properties will be prefixed by Property: anyways.
In any event: I find the argument that a rare letter like Q is good branding not very convincing. I would suggest then a more memnonic choice, like WD2348972 or W2348972 instead. I believe the Q as prefix used in all canonical inbound links will be puzzling many people and end the explanation having to end up in the FAQ.
thanks again!
Gregor
Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
http://www.bbc.co.uk/ This e-mail (and any attachments) is confidential and may contain personal views which are not the views of the BBC unless specifically stated. If you have received it in error, please delete it from your system. Do not use, copy or disclose the information in any way nor act in reliance on it and notify the sender immediately. Please note that the BBC monitors e-mails sent or received. Further communication will signify your consent to this.
Hi Michael,
answers inline.
2012/7/26 Michael Smethurst michael.smethurst@bbc.co.uk:
Very delayed reply but think I'm still confused on this. Made a picture to clear my mind but not sure it works: http://smethur.st/wikidata
The bit I think I get: If I request http://en.wikidata.org/wiki/Berlin Or http://en.wikidata.org/title/Berlin I get a 301? to:
303, but otherwise correct.
http://wikidata.org/title/en:Berlin The html wiki page
But not sure I understand the machine readable part [1]
Bullet point 1 says http://wikidata.org/id/Q%7Bid%7D Resolves to the appropriate url depending on the request header
Does resolve mean a redirect? Is that a 303?
Yes, it is a 303 redirect.
Or is there no redirect and the "thing" uri returns content?
What's the "appropriate url"?
It says "the appropriate URL depending on the request header", i.e. it depends on the ACCEPT header if you get the HTML page or the data.
http://wikidata.org/data/Q%7Bid%7D Or http://wikidata.org/data/Q%7Bid%7D?format=%7Bformat%7D&language=%7Blangu... ?
The former I'd say, but this is not completely settled yet.
Bullet point 2 says http://en.wikidata.org/item/Berlin Also resolves to the appropriate url. Is that a redirect? What's the appropriate url?
As above. It depends on the ACCEPT header. Also, this is merely a convenience URI.
Is there content negotiation happening from http://wikidata.org/wiki/Q%7Bid%7D
Or just from http://wikidata.org/id/Q%7Bid%7D
Just from the latter. Not on the former.
What happens if I request http://wikidata.org/id/Q%7Bid%7D And accept only html?
303 redirect to http://wikidata.org/wiki/Q%7Bid%7D
Is there content negotiation from http://wikidata.org/data/Q%7Bid%7D Or do I have to use parameters to get different representations?
No content negotiation intended here. Maybe the latter. We are not sure yet about the supported formats and how this will work.
Is there a better picture
No, you made the best one so far.
Sorry to be thick Michael
Sorry for having explained it badly. I hope this helps! Denny
[1] https://meta.wikimedia.org/wiki/Wikidata/Notes/URI_scheme#Machine-readable_a ccess
On 06/07/2012 18:20, "Gregor Hagedorn" g.m.hagedorn@gmail.com wrote:
Thanks Denny, I largely see your points. The distinction between convenience = webservice to redirect to canonical URL and canonical URL could perhaps be made clearer in the note. I read it as parallel URIs rather than as a redirecting service. To me the word "convenience" has a different implication, but this may be entirely my fault, I am not a native speaker either. I also agree on the choice of language prefixes, confusing as it may be, I should have know. The data plus wikidata is still confusing, but I guess you cannot avoid that one?
About the Q in front of identifiers: At the moment I see the item numbers being used in rdf:resource/about, but I understand that you may need them as element names? My understanding was that properties will be prefixed by Property: anyways.
In any event: I find the argument that a rare letter like Q is good branding not very convincing. I would suggest then a more memnonic choice, like WD2348972 or W2348972 instead. I believe the Q as prefix used in all canonical inbound links will be puzzling many people and end the explanation having to end up in the FAQ.
thanks again!
Gregor
Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
http://www.bbc.co.uk/ This e-mail (and any attachments) is confidential and may contain personal views which are not the views of the BBC unless specifically stated. If you have received it in error, please delete it from your system. Do not use, copy or disclose the information in any way nor act in reliance on it and notify the sender immediately. Please note that the BBC monitors e-mails sent or received. Further communication will signify your consent to this.
Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l