OpenPublicCommons uploads: Support automatic conversion of Microsoft Office to LibreOffice/OpenOffice.org formats
- Assigned To
- T44725: Multimedia file format support (tracking)
- Blocked By
- T37607: MediaWiki incorrectly detects OOXML types of files saved by OpenOffice
T19497: Add oasis opendocument and oo.o legacy document to mime.types
T4089: Whitelist OASIS OpenDocument file format
- Sj, Krenair, Petrb and 4 others
As with bugs 43149-51: Allow DOC, XLS, and PPT uploads to Commons: transparently
converting to the preferred free media format and returning that as the final file (and filename extension) on Commons.
This will add support for the vast majority of all office files in existence,
and will allow people who don't understand what "transcode" means or how to run OOo to sharing their existing (freely licensed) documents, datasheets, and presentation decks with the world.
I see the other bugs in this series were reclassified as a MediaWiki bug: the solution does not necessarily have to work for all MW installations.
For one thing, this is a very Commons-specific problem. Most non-WMF wikis can simply turn on acceptance of all file types: problem solved :)
For another thing, the easiest solution is to pipe such requests through the Internet Archive converter, and then read back the results. IA is willing to do this for all uploads to Commons; but would not do so for all uploads to any MW install.
The set of bugs is a request to provide the end-result functionality via the Commons "upload file" interface - the obstacle blocking the majority of all people from sharing the majority of all non-image files - not any particular implementation.
We don't allow open/libre office formats on commons at the moment.
Problems with this:
*Commons doesn't allow office formats, free or otherwise [Primary problem. Part of this is historical with security issues in zip based formats, but nonetheless the politics of allowing this type of format needs to be sorted out]
*Conversion programs aren't perfect [unless things have been improved significantly]. Its one thing to tell a user to convert and they do it imperfectly, when we do conversions wrong, people yell at us.
As for conversion programs, there are command line converters to (wvConvert and stuff). I'd personally be opposed to depending on a third party,
This should possibly be moved to an extension requests. I am almost certain we won't be doing office format conversions in core.