Daniel-Mietchen
commented
11 years ago Hi, thanks for chiming in.
We would certainly be interested in seeing the importer's videos more widely shared than to Wikimedia Commons (which is what we have) and YouTube (in the making, see https://github.com/erlehmann/open-access-media-importer/issues/82 ), and Internet Archive would seem like an excellent choice. However, we do not currently have the capacity to do this ourselves, so we would appreciate if you could take a lead on this.
codersquid
Create a pipeline to export videos to Internet Archive (IA).
If you like this feature idea, I can work on it or help by reviewing if you want to do the work. Some things of note.
IA has an s3 type api, which will have improved documentation, but for now you can see it at abouts3.txt. I have used boto to submit videos to IA and can provide help. You could also forgo a separate library and use the RESTful interface.
Based on helping to push videos to IA from conferences: make a bucket per item that you want to have with its own metadata. I made a bucket for a conference, and so the bucket has metadata for that conference, but it doesn't have metadata per video, which I would have also like to have.
Re: business constraints. Based on experience, video services under business constraints can end up changing how the service works to the detriment of someone trying to archive video to be a part of the commons. For example, back in 2009, we chose to archive creative-commones friendly conference videos to blip.tv due to some nice features they had. Mainly, they had an api to allow clients to upload multiple file formats per each video listed, and made it easy for users to download the files or watch online. This meant we could provide formats like ogv, as well as formats friendly for iOS devices. Their UI has changed so much so that this is no longer true. We have stopped using it for archiving videos. We started using youtube, but in the past year we've started also archiving on IA.
One thing you might want to do is make some sort of overlay that indexes the videos and points to all versions -- this is how we operate right now. Our friend has richard and steve that go in to making a site like pyvideo.org work. Here is an example of a video, which has metadata, a link to youtube, and a link to IA to download a version. I was also able to have a description with a link to related URLs, such as an ipynb. Server Log Analysis with Pandas