Hi,
We need to identify ZIM files with the same content over the time to be able to provide update mechanisms.
Use case example: * User owns a ZIM file with Wikipedia in French from 2010. * New version is released in 2011 * User is automatically warned about that and the software propose to update (download) the content.
This is the reason why I propose to introduce an optional serieid in the ZIM format.
Comments?
Emmanuel
Can't we do that already with /M/Date?
On 15.07.2011 10:27, emmanuel@engelhart.org wrote:
Hi,
We need to identify ZIM files with the same content over the time to be able to provide update mechanisms.
Use case example:
- User owns a ZIM file with Wikipedia in French from 2010.
- New version is released in 2011
- User is automatically warned about that and the software propose to
update (download) the content.
This is the reason why I propose to introduce an optional serieid in the ZIM format.
Comments?
Emmanuel _______________________________________________ dev-l mailing list dev-l@openzim.org https://intern.openzim.org/mailman/listinfo/dev-l
On Fri, 15 Jul 2011 10:42:43 +0200, Manuel Schneider wrote:
Can't we do that already with /M/Date?
No, /M/Date is mandatory to sort multiple ZIM files of a serie on the timeline, but it can't obviously tell you (or to a system) that two ZIM file are of the same serie.
Emmanuel
Title or Source could help here maybe.
If we really need a new entity I would prefer if it was also defined by Dublin Core.
http://dublincore.org/documents/dces/
Identifiere looks useful.
On the other hand I think the information could be retrieved from combined metadata to keep it simpler and easier to understand (human readable):
Like Subject = Wikipedia and Language = en, plus Date should give everything we need.
/Manuel
On 15.07.2011 10:46, emmanuel@engelhart.org wrote:
On Fri, 15 Jul 2011 10:42:43 +0200, Manuel Schneider wrote:
Can't we do that already with /M/Date?
No, /M/Date is mandatory to sort multiple ZIM files of a serie on the timeline, but it can't obviously tell you (or to a system) that two ZIM file are of the same serie.
Emmanuel _______________________________________________ dev-l mailing list dev-l@openzim.org https://intern.openzim.org/mailman/listinfo/dev-l
On Fri, 15 Jul 2011 11:10:01 +0200, Manuel Schneider wrote:
Title or Source could help here maybe.
My opinion: * "Title" is a label, we should not base such a automatic mechanism on that. * "Source" could be better but I have a bad feeling. Currently nobody use it, this is not clear for me what should exactly be there (if you have for example a ZIM with WP and WS). We also though about it to be able to provide dynamically generated "go to online version" buttons on articles. Worth, I see now way to guaranty the uniqueness of this value over the different publishers.
If we really need a new entity I would prefer if it was also defined by Dublin Core.
http://dublincore.org/documents/dces/
Identifiere looks useful.
Why not? But "identifier" looks like our ZIM id. It does not seems to have something about "serie" or "collection".
On the other hand I think the information could be retrieved from combined metadata to keep it simpler and easier to understand (human readable):
Like Subject = Wikipedia and Language = en, plus Date should give everything we need.
Emmanuel