Re: Re: Generating valid unique IDs for hAtom

On 22 Jun 2007, at 07:07, Toby A Inkster wrote:
> What's wrong with using Permalinks as an id?
>
> If you need to make several entries onto the hAtom feed
referencing  
> the
> same URL, then you can just add
"#ref-20070722", "#ref-20070723"  
> and so on
> to the end of the URL to make it unique.

The best starting point about identifiers for Atom was
written by  
Mark Pilgrim:

• http://diveintomark.org/archives/2004/05/28/howto-atom-
id

It includes instructions on generating Tag URIs and also
explains why  
permalinks are sub-optimal as IDs.

Ben

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On 6/28/07, Ben Ward <listsben-ward.co.uk> wrote:
> On 22 Jun 2007, at 07:07, Toby A Inkster wrote:
> > What's wrong with using Permalinks as an id?
> >
> > If you need to make several entries onto the hAtom
feed referencing
> > the
> > same URL, then you can just add
"#ref-20070722", "#ref-20070723"
> > and so on
> > to the end of the URL to make it unique.
>
> The best starting point about identifiers for Atom was
written by
> Mark Pilgrim:
>
> • http://diveintomark.org/archives/2004/05/28/howto-atom-
id
>
> It includes instructions on generating Tag URIs and
also explains why
> permalinks are sub-optimal as IDs.

Can you push this into the Wiki Ben? This may make a very
good
recommendation for hAtom 0.2.

Regards, etc...
David


-- 
David Janes
Founder, BlogMatrix
http://www.blogmatrix.com
http://blogmatrix.bl
ogmatrix.com

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Ben Ward wrote:

> The best starting point about identifiers for Atom was
written by Mark 
> Pilgrim:
> 
> • http://diveintomark.org/archives/2004/05/28/howto-atom-
id
> 
> It includes instructions on generating Tag URIs and
also explains why 
> permalinks are sub-optimal as IDs.

Thanks Ben, but I have already read Mark's comments there.

Mark talks about unique IDs for Atom, which will not do for
hAtom 
because the suggested tag: syntax cannot generate IDs that
validate as 
HTML - the id attribute won't allow it:
htt
p://www.w3.org/TR/html401/types.html#type-name

Jon

-- 
gr0w.com
dotjay.co.uk
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Ben Ward wrote:

> http://diveintomark.org/archives/2004/05/28/howto-atom-
id
> It includes instructions on generating Tag URIs and
also explains why  
> permalinks are sub-optimal as IDs.

And the reasons he gives are:

	1. In case developers of Atom clients are stupid; and
	2. In case your permalinks change.

#1 is an exercise in making things foolproof, which is not
even worth
bothering trying, as those fools can be so ingenious. As for
#2, it's
fixing the wrong problem: if changing permalinks causes the
Atom ID to
change, then you should fix your permalinks which are not
supposed to
change!

-- 
Toby A Inkster BSc (Hons) ARCS
[Geek of HTML/SQL/Perl/PHP/Python/Apache/Linux]
[OS: Linux 2.6.12-12mdksmp, up 25 days, 13:44.]

                          demiblog 0.2.0 Released
          http://tobyinkster.co.uk/blog/2007/06/28/demiblog-0.2.0/


_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On 7/16/07, Toby A Inkster <mailtobyinkster.co.uk>
wrote:

> #1 is an exercise in making things foolproof, which is
not even worth
> bothering trying, as those fools can be so ingenious.
As for #2, it's
> fixing the wrong problem: if changing permalinks causes
the Atom ID to
> change, then you should fix your permalinks which are
not supposed to
> change!

There's lots of reasons a permalink may change -- for
example, another
site may be reposting the message, in which case it has a
different
permalink but a different Atom ID. The beauty here is that
an
intelligent feed reader won't show it to you twice.

-- 
David Janes
Founder, BlogMatrix
http://www.blogmatrix.com
http://blogmatrix.bl
ogmatrix.com
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

David Janes wrote:

> There's lots of reasons a permalink may change -- for
example, another
> site may be reposting the message, in which case it has
a different
> permalink but a different Atom ID. The beauty here is
that an
> intelligent feed reader won't show it to you twice.

Why repost? Why not just link to the original article? In
which case it
has the same Permalink and the same Atom ID. The beauty here
is that an
intelligent browser won't hit the server twice (it will
retrieve the
article from the cache) and it will mark the link purple to
indicate that
you've already read the page.

-- 
Toby A Inkster BSc (Hons) ARCS
[Geek of HTML/SQL/Perl/PHP/Python/Apache/Linux]
[OS: Linux 2.6.12-12mdksmp, up 25 days, 17:29.]

                          demiblog 0.2.0 Released
          http://tobyinkster.co.uk/blog/2007/06/28/demiblog-0.2.0/


_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On 7/16/07, Toby A Inkster <mailtobyinkster.co.uk>
wrote:
> Why repost? Why not just link to the original article?
In which case it
> has the same Permalink and the same Atom ID. The beauty
here is that an
> intelligent browser won't hit the server twice (it will
retrieve the
> article from the cache) and it will mark the link
purple to indicate that
> you've already read the page.

Perhaps the appropriate forum for this question is the Atom
mailing
list. Note that they've spent about a 1000 person years
figuring this
stuff out and have been working with syndication etc. for a
significant portion of their professional lives.

Regards, etc...

-- 
David Janes
Founder, BlogMatrix
http://www.blogmatrix.com
http://blogmatrix.bl
ogmatrix.com
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On 16 Jul 2007, at 11:08, Toby A Inkster wrote:
> As for #2, it's
> fixing the wrong problem: if changing permalinks causes
the Atom ID to
> change, then you should fix your permalinks which are
not supposed to
> change!

Of course _perma_links are not supposed to change. But I
think the  
point that Mark Pilgrim makes, and one that I agree with, is
that  
principal does and will collapse in real-world tools.

Migrating between CMS systems, rebranding to a new domain
name,  
switching advocacy to support the ‘no-WWW’ movement? All
perfectly  
likely and reasonable things to do, which CMS's generating
permalinks  
have not been designed to support and a problem which,
regardless of  
the ideal, many (most?) authors simply won't account for.

There may not be a solution to this. Fitting a tag URI into
page  
content seems clumsy for sure, and those same CMS systems
that make  
permalinks less-than permanent are publishing dynamically
generataed  
permalink URLs as IDs all the while. I'm not sure that we're
in a  
position to ‘fix’ the problem. But I think we should try to
exhaust  
the non URL options thoroughly.

Ben
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

> Note that they've spent about a 1000 person years
figuring this
> stuff out and have been working with syndication etc.
for a
> significant portion of their professional lives.

And I don't disagree with them -- just with Mark Pilgrim's
claim than a
(true) Permalink is a poor Atom ID.

-Toby
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss