On 4 September 2011 21:44, Michael Dale mdale@wikimedia.org wrote:
It will be a lot easier to import from YouTube once Timed media handler adds support for webm to commons. If you check out the wikivideo-l and commons lists for some recent example YouTube to commons scripts. I know this is not super useful info right this second, but there is hope on the horizon.
How's Timed Media Handler (which will also allow WebM/VP8) going? ETA?
- d.
On Sun, Sep 4, 2011 at 1:49 PM, David Gerard dgerard@gmail.com wrote:
On 4 September 2011 21:44, Michael Dale mdale@wikimedia.org wrote:
It will be a lot easier to import from YouTube once Timed media handler
adds support for webm to commons. If you check out the wikivideo-l and commons lists for some recent example YouTube to commons scripts. I know this is not super useful info right this second, but there is hope on the horizon.
How's Timed Media Handler (which will also allow WebM/VP8) going? ETA?
The code exists and has been revamped a few times in response to reviews, but I'm not sure whether there are actually any assigned resources for pushing it to production at this time.
Robla, Erik, can you clarify state on this? Is TMH deployment on the radar for the next few months, or does it still need to get put on the map? There's some notes on http://www.mediawiki.org/wiki/TimedMediaHandler/statusabout earlier discussion of a September-ish production test, but not sure if that's current.
-- brion
On Tue, Sep 6, 2011 at 11:02 AM, Brion Vibber brion@pobox.com wrote:
The code exists and has been revamped a few times in response to reviews, but I'm not sure whether there are actually any assigned resources for pushing it to production at this time.
Yes, there are. Ian and Neil are scheduled to do a code review of TMH, once remaining high priority issues with UploadWizard have been resolved, later this month. Before we've done an initial assessment of the code, it's hard to give a realistic deployment estimate -- there may be parts that need to be rewritten or taken out. So I won't commit us to a public date just yet, just to say that it's definitely something I'd like to see user-visible progress on this calendar year.
Whatever remaining bugbears are lurking in the code, TMH definitely represents the key set of features that are needed to make video in Wikimedia projects suck significantly less (multi-codec and multi-bitrate derivatives generation; a non-ugly player skin; subtitle support). This is really a baseline feature set that we have to get done (this and better large file upload support) to not give users an embarrassingly bad video experience.
On Tue, Sep 6, 2011 at 3:15 PM, Erik Moeller erik@wikimedia.org wrote:
On Tue, Sep 6, 2011 at 11:02 AM, Brion Vibber brion@pobox.com wrote:
The code exists and has been revamped a few times in response to reviews, but I'm not sure whether there are actually any assigned resources for pushing it to production at this time.
Yes, there are. Ian and Neil are scheduled to do a code review of TMH, once remaining high priority issues with UploadWizard have been resolved, later this month. Before we've done an initial assessment of the code, it's hard to give a realistic deployment estimate -- there may be parts that need to be rewritten or taken out. So I won't commit us to a public date just yet, just to say that it's definitely something I'd like to see user-visible progress on this calendar year.
All good news!
The player support is definitely a lot nicer -- and I think we've been running extra JS stuff from that on Commons for a while.
The generation & handling of derivative files looks like the biggest candidate for potential breakage / rewriting from what I recall (it's still pretty basic in terms of managing processes, so if anything goes wrong it may be tricky to recover).
Simply being able to upload larger files will be a big help as well; audio and video clips from eg conference talks are routinely hard to upload to Commons because of the 100mb limit we still have in place for plain HTTP uploads and limited access to upload-by-URL.
The upload size limit could be bumped a bit more, but can only be bumped so far before it starts using too much memory during upload -- but there's been work on incremental uploads too, which for many modern browsers will be able to kick in automatically and will bypass that side of the limits, provide better upload progress feedback, and be more reliable in the case of flaky connections or having to put your laptop to sleep!
IIRC the incremental uploads aren't part of TMH directly, but is also in Michael's sphere of awesome projectness. :)
-- brion
wikitech-l@lists.wikimedia.org