/ "Danny Ayers" <danny666@xxxxxxxxxxx> was heard to say:
|> / "Danny Ayers" <danny666@xxxxxxxxxxx> was heard to say:
|> | Initially, the identifier/locator will go in <id>. If there's a
|> change of
|> | domain, then the new 'preferred' identifier/locator will go in
|> <id>, an the
|> | old one shifted to <link>.
|>
|> Precisely. That is exactly, totally the wrong semantic for "id".
|>
|> identity: n. sameness of essential or generic character in
|> different instances.
|
| The resource identified in <id> has the same essential or generic character
| as the entry.
Either we're talking at cross purposes, or I'm confused.
Suppose I have an entry in my feed:
<entry>
<id>http://olddomain.com/blog/1234</id>
...
</entry>
Speaking of the case where this entry moves from olddomain.com to
newdomain.com, you said:
If there's a change of domain, then the new 'preferred'
identifier/locator will go in <id>, an the old one shifted to
<link>.
So I assume you mean I should now serve up this entry in my feed:
<entry>
<id>http://newdomain.com/blog/1234</id>
...
</entry>
The actual essay has an essential or generic character independent of
the entry(ies) that it has in one or more feed(s). That's what I want
to identify.
The only way that an <id> can be used to assert that
http://olddomain.com/blog/1234
and
http://newdomain.com/blog/1234
have the same essential or generic character is if THE SAME ID is used
to identify them in the respective entries.
I don't understand how you can speak of changing the ID while somehow
retaining a sense of identity. Those seem logically incompatible.
|> What you're saying is the "id" identifies the instance. That's just
|> not interesting.
|
| I could use the same argument about the URN. In both cases the benefit only
| comes knowing that the URI in <id> is equivalent to others.
It has *nothing* (I repeat NOTHING) to do with URNs. You can use HTTP URIs
or FTP URIs or any URI at all in there. What's important is that it not
change when the resource moves or is copied.
| If you don't want to have the concept of identity
|> independent of location, can we please at least change the tag to be
|> "uri" or "location" or anything that doesn't suggest what anyone
|> familiar with the English language would think "identity" meant.
|
| The identity is independent of location, though when you describe it as
| above it does sound confusing.
That's because it *IS* confusing. You're mangling the semantics of <id>.
| But what is happening in the spec at present is that the resource identified
| by the URI in <id> has no retrievable representations except that of the
| (different) resource in the <link rel="alternate"> URI. I think that's
| pretty confusing.
I don't. But I don't object to letting the ID be retrievable either.
Perhaps if there is an ID but no link, then the ID should be used as
the link. However, if one or more link elements exist, I don't want
anyone trying to do a retrieve operation on the ID.
| But ok, you think that <id> should act a single, unique identifier for the
| entry. At present I believe the spec's problematic largely because it talks
| about "relocated" representations (i.e. different URIs) while retaining the
| same id resource.
That's the point of having a persistent ID.
| Maybe that could be expressed in a less confusing fashion in the spec by
| taking a leaf out of the RDF book. We are saying that the resource
| identified in <id> is the entry, right? So why not drop the talk of other
| other representations being relocated, and make the entry behave more like
| the RDF/XML:
|
| <entry rdf:about="urn:blah">
| <content rdf:parseType="Literal">
| the content
| </content>
| </entry>
Ah, maybe this is the nub of our confusion. I never think about putting the
actual content in the Atom entry.
| So rather than "other representations", the <link rel="alternate"> could
| refer to other resources and have similar semantics as:
|
| <entry rdf:about="urn:blah">
| <content rdf:parseType="Literal">
| the content
| </content>
| <owl:sameAs rdf:resource="http://permalink.org"; />
| </entry>
|
| The 'definitive' representation of the resource is the inline literal.
That doesn't really help me any.
Be seeing you,
norm
--
Norman Walsh <ndw@xxxxxxxxxx> | It is as easy for the strong man to be
http://nwalsh.com/ | strong, as it is for the weak to be
| weak.-- Emerson
Attachment:
pgp00015.pgp
Description: PGP signature