Yeah, I stole the title from the Halloween series. It seemed fitting.
So here we are about a year later. Guess what? Revision3 has killed off a few more shows, started a bunch more, and redesigned their site.
A few days ago, Moonlit and I got into an interesting Twitter conversation with Revision3, the results of which can be found here and here. Basically, with their site redesign they lost a bunch of shows which they then regained after we brought it up. They also claimed to fix some dead links, so some possible good news there. On the negative, they did say that some shows have been removed purposefully. Thankfully I believe I have a copy of everything they’ve taken down to date. On the whole, it felt as though they danced around issues I brought up, but at this point I find it unsurprising. It’s probably a good thing they didn’t Google me while we were talking.
I checked my download script from last summer, and sure enough it didn’t work. Upon doing a few trials, it’s a simple fix to change “small” to “medium.” So, I went ahead and updated it. This now works perfectly again (woohoo!).
I went through the wiki page and added in the shows that they had cancelled while I was out of the loop. I’m just going off of their “Archive Shows” page, which may be incomplete, but I don’t know enough about this age of Revision3 to tell you if something is missing or not. It doesn’t look too much as if they throw out shows altogether these days.
I count 10 more dead shows.
On a similar topic, I’m also finally getting work done with Hack College (One of the baby Rev3 Beta shows) and am in the process of uploading it. Let’s walk through how I did this finally because you might be curious.
I know that they have the series up on http://blip.tv/hackcollege. This is good for starters. I also know that youtube-dl supports Blip.tv. Excellent. After a little trial and error, I settled on this command:
youtube-dl -c -i -t http://blip.tv/hackcollege
This downloads all the videos from the hackcollege account, puts titles in the file names, continues incomplete downloads, and skips errors. I only added that last part because one video gave me an error (I eventually just downloaded it manually). Then, I forgot I wanted descriptions for each video, so I ran this:
youtube-dl -i -c -t –write-info-json http://blip.tv/hackcollege
See how easy that was? JSON descriptions in just a few seconds. Okay. A few people still know a little trick for getting RSS from Blip.tv: just add “/rss” after the account url. You can go ahead and load it yourself at http://blip.tv/hackcollege/rss. Now, if you’ve been downloading along, you might notice that youtube-dl only snags the .m4v files and there are these beautiful giant .mov files in the RSS feed. Wouldn’t it be nice to grab these? Unfortunately, you can’t fine-tune youtube-dl for these just yet (maybe if I hunt around in the source code I can set this up, but I honestly didn’t think of that until right now). How do we get these files? I came up with this one-liner:
curl http://blip.tv/hackcollege/rss | grep -o ‘http.*mov’ | sort | uniq > out.txt
Pretty self explanatory. Grab the RSS, filter for .mov links, and sort it to get rid of duplicates. Then, save it to a text file. You might notice that the RSS feed doesn’t contain links for every video. There really isn’t anything we can do about that, but it does appear that the .mov files were a semi-recent addition to these videos and earlier videos most likely don’t have the option. I’d say it’s a pretty safe bet that we’ve got them. Next, we will want to download those links. You can probably pipe that one-liner into wget, but I wasn’t ready to download when I wrote it, so I saved the text for later. Here’s how I eventually used it with wget:
wget –user-agent=”iTunes/10.6.1″ -c -i out.txt
It appears that Blip.tv white-lists your user-agent, so it knows to not allow wget, or probably a slew of other bots and/or browsers. You can find this out if you can download a file in your browser but running something automated on the site results in redirects and file fragments. There’s always a workaround. If you trick it into thinking you are iTunes, you not only get access to the files, but also get them faster than if you just manually downloaded from your browser (throttle-free!). Very nice. So after feeding the text file into wget, the .mov files download rather quickly. You don’t get the nice fancy file names like with youtube-dl, but you do get the files.
So there. Only took an hour or two to work everything out.
Well, as you can see, I’m finishing up a few parts of SaveRev3. Unfortunately, there is still work to be done and the list keeps on growing.
Here’s to another summer of hard drives and bandwidth. It’s going to be a hot one.