W3C home > Mailing lists > Public > www-tag@w3.org > November 2008

Re: ZIP-based packages and URI references into them ODF proposal

From: Leo Sauermann <leo.sauermann@dfki.de>
Date: Mon, 10 Nov 2008 17:33:29 +0100
Message-ID: <49186258.40807@dfki.de>
To: Svante Schubert <Svante.Schubert@Sun.COM>
CC: www-tag@w3.org, David Carlisle <davidc@nag.co.uk>, Marcos Caceres <marcosscaceres@gmail.com>
Hi TAG,

I chime in only once (not disturbing you more than needed),

URI schemes for packages are also made up for the various "Virtual File 
System (VFS)" implementations out there.
For example there is one by Gnome, another by Apache VFS. Both lack 
documentation, and I think neither conforms to a VFS Uri standard.

some random links:
http://commons.apache.org/vfs/filesystems.html
http://www.usenix.org/events/usenix2000/freenix/full_papers/perazzoli/perazzoli_html/node9.html
http://library.gnome.org/devel/gnome-vfs-2.0/unstable/gnome-vfs-20-gnome-vfs-uri.html


Apache have more documentation and have done this uri scheme:|
zip:// /arch-file-uri/ [! /absolute-path/ ] |

These are the apache developers involved, I don't know anyone of them:
http://commons.apache.org/vfs/team-list.html

= My 2c =
New protocols won't help here, the mime-type of the downloaded file 
determines the container format (= mimetype zip means its a zip file, 
not zip:// at the beginning), and we have the fragment-identifier for 
just that: referencing elements inside the identified resource.

So,
* new schemes are bad
* anything after the fragment identifier "#" is interpreted by the 
client after downloading
* we have no "scheme" or type to say how to interpret the fragment 
identifier, only for HTML (or?)

so we would need to generate a standard to get more out of fragment 
identifiers:
http://someserver/package-file.abc#<missingpart to say that this is now 
a file within a zip>firstDirectory/someFile.xyz

best
Leo

It was Svante Schubert who said at the right time 03.11.2008 17:17 the 
following words:
>
> Hi,
>
> that was what we had in mind, when talking about a package URL 
> solution. Creating a schema for as many as possible packaged 
> application formats, which naturally would include OOXML as a file 
> format and the 'merge' with the pack: schema.
>
>
> Aside of a package scheme and a fragment identifier solution given by 
> Larry, there is a third approach. There were several requests (e.g. 
> W3C Widget Group, Marcos Caceres) for making it possible to point with 
> an URL into a package
>
> For example:
>    http://someserver/package-file.abc/firstDirectory/someFile.xyz
>
> Would this be a valid request for you?
>
> PS: As the RDF metadata feature of ODF 1.2 is in need to reference to 
> fragments of the ODF documents from outside the package via an URL, I 
> believe the mimetype given fragment identifer is the easiest solution 
> for now, no matter what additional (and compatible) ways later can be 
> agreed on in a larger circle.
>
> Regards,
> Svante
>
>
> David Carlisle wrote:
>>> The problem is is all about identifying resources within a package from
>>> outside a package by an URL.
>>>     
>>
>>
>> ODF/OOXML interoperability has a certain amount of history,
>> but shouldn't proposals in this area be merged with
>>
>>
>>
>> http://tools.ietf.org/id/draft-shur-pack-uri-scheme-04.txt
>>
>> which has abstract
>>
>> Abstract
>>    A package is a logical entity that holds a collection of parts.    
>> Given the URI for a complete package, the "pack" URI scheme provides
>>    for the construction and use of URIs referring to individual parts 
>>    within the package.  It also provides for the use of part's URIs 
>> as    base URIs for resolving relative references between the parts 
>> in a    single package.
>>
>> David
>>
>>
>> ________________________________________________________________________
>> The Numerical Algorithms Group Ltd is a company registered in England
>> and Wales with company number 1249803. The registered office is:
>> Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.
>>
>> This e-mail has been scanned for all viruses by Star. The service is
>> powered by MessageLabs. 
>> ________________________________________________________________________
>>
>>   
>
>


-- 
____________________________________________________
DI Leo Sauermann       http://www.dfki.de/~sauermann 

Deutsches Forschungszentrum fuer 
Kuenstliche Intelligenz DFKI GmbH
Trippstadter Strasse 122
P.O. Box 2080           Fon:   +49 631 20575-116
D-67663 Kaiserslautern  Fax:   +49 631 20575-102
Germany                 Mail:  leo.sauermann@dfki.de

Geschaeftsfuehrung:
Prof.Dr.Dr.h.c.mult. Wolfgang Wahlster (Vorsitzender)
Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats:
Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313
____________________________________________________
Received on Monday, 10 November 2008 16:35:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:07 GMT