OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

oic message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: RE: [oic] Interpretation of terminal "/" in URIs and Zip


There is already a URI scheme in the OOXML Open Package Conventions for
internal references from one Package Part to another, from outside of a
package to inside of a package, and from outside of one package to inside of
another. 

I don't see us specifying a mapping between ODF packages and file systems
even though we talk about "as if" that is done with regard to certain
notions.  The problem is, which file systems?  Or for that matter, which
pseudo-web server (case sensitive, not case sensitive, Unicode explicit or
IRI encoded, which character set and reserved characters, etc., etc.).

Since we don't need it for the ODF format specification, I would hope such
matters are not priority concerns while there are so many other features
that have a direct bearing on ODF usage and interoperability.

 - Dennis

PS: I have no idea where schema name odf came from.   the problem is
doc/prefix versus doc/prefix/ where doc/prefix/ is for a subdocument (and
doc/ might or might not be).

PPS: It looks to me that we have already defined a "new" protocol, it is
just based on mapping from relative IRIs.  But the Zip end is a fixed thing
and the coupling between it and manifest:full-path requires protocol,
protocol that IRI/URI will not resolve for us.  Hey, I didn't make this ODF
relative URI stuff up.  It was done before I got here.  

PPPS: I must remember to bug Larry Masinter about this now that he is done
with his HTML scope-creep objections for now.  He is working on a grand
unification of URI/URN/IRI/etc. into a single RFC that binds them all, so it
would be useful to take a look at any drafts they've got so far.

PPPPS: And I should point out that as much fun as this discussion is in
getting a look at all the considerations, this is probably not going to be
resolved here on OIC.

-----Original Message-----
From: Svante.Schubert@Sun.COM [mailto:Svante.Schubert@Sun.COM] 
Sent: Monday, March 01, 2010 11:54
To: Hanssens Bart
Cc: dennis.hamilton@acm.org; OIC TC List
Subject: Re: [oic] Interpretation of terminal "/" in URIs and Zip

On 03/01/10 19:30, Hanssens Bart wrote:
> Well, I agree with Dennis that manifest:full-path="prefix/" (with trailing
slash)
> feels more natural.
> 
> Whether or not a server will be able to map odf://www.acme.org/doc/prefix
> to prefix/ is something I'm not that worried about (mod_rewrite is your
friend)
> 
Unfortunately natural is not specified ;-)

Honestly, when this is only about accessing resources within the ZIP 
document than I still believe there is any problem at all with slashes..

[ ... ]

Parallel to this on the SC34 WG4 Murata-San owns a task to solve the 
inner ZIP access for OOXML. I discussed with him last year on Skype the 
problem, requirements and possible solutions and we agreed to try a 
common approach.

It seems certain (to most) that the solution is not to invent a new 
protocol that no client is yet supporting.

[ ... ]



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]