Searching for doubles

March 25, 2013 4:52 PM

I've noticed that the search function doesn't catch YouTube doubles if the URL isn't exact. Here's a tip: grab the alphanumeric ID of the video from its URL and search mefi using only that. For example, searching for 9Iq5yCoHp4o pulls up this post, but a search for http://​youtu.be/​9Iq5yCoHp4o and http://​www.youtube.com/​embed/​9Iq5yCoHp4o fail, even though they all point to the same video.
posted by desjardins to Etiquette/Policy at 4:52 PM (19 comments total)

Thanks, that's a good tip.
posted by pb (staff) at 4:56 PM on March 25, 2013

*smacks forehead*

Brilliant! Thanks. :)
posted by zarq at 5:06 PM on March 25, 2013

I'll add it to the FAQ!
posted by jessamyn (staff) at 5:09 PM on March 25, 2013

Wait. This is a pony... give-away? A call-out offering useful advice in a helpful tone? A non-angry non-rant? Is it April Fool's Day already?
posted by GenjiandProust at 5:13 PM on March 25, 2013 [7 favorites]

I already know the answer to this question but what are the chances of that getting done automatically by the double-checking mechanism?
posted by shakespeherian at 5:14 PM on March 25, 2013 [1 favorite]

We'll look into it.
posted by pb (staff) at 5:19 PM on March 25, 2013

Sometimes with search, less is way more.
posted by Toekneesan at 5:24 PM on March 25, 2013 [1 favorite]

I honestly thought this question had been asked before.
posted by cjorgensen at 6:15 PM on March 25, 2013 [2 favorites]

That's a great idea. There are probably some other reductions to a canonical reference that could be helpful for other widely used sites, but this one is the killer app.
posted by unSane at 6:41 PM on March 25, 2013

Agreed, unSane. YouTube seems to have the cruftiest URLs... there are almost limitless ways you can reach the same video. Anyhow, looks like this has come up on StackOverflow (and I'm sure elsewhere) more than once.
posted by Conrad Cornelius o'Donald o'Dell at 9:21 PM on March 25, 2013

I've been saying that for years.
posted by Rhomboid at 10:54 PM on March 25, 2013

Is this a Double post?
posted by Eideteker at 4:53 AM on March 26, 2013

Yes, this post topic sounds similar to me, too. I've been searching for YouTube video IDs for a while, to avoid double-posting. I thought the reason for not auto-checking YT IDs was because so many videos get re-posted, both on YouTube and now moreso on Vimeo, plus on other streaming video sites. But I didn't see anything under the YouTube tag that resembles this topic, so maybe that was a discussion in some MetaTalk thread about doubles... or perhaps I'm making it all up.
posted by filthy light thief at 7:29 AM on March 26, 2013

We have two different places on the site where searching for YouTube URLs come into play. The site search, and the double post checker that runs when you create a new MetaFilter post.

The double link checker already strips out the youtube ID from youtube URLs and searches for that instead of the full URL. So if you're about to post something, it should catch any YouTube-related URL. (As filthy light thief points out, it won't find the same video embedded at a different URL, cross posted to a different video site, etc.) But I think that's a good spot to isolate the ID, and we've been doing that for quite a while.

The site search is a different story. We don't know intent when people are using the general search. Are they looking for a duplicate? Are they looking for a precise URL? We can't know for sure. So that means the person searching needs to do a little more work. We have YouTube ID searching in place where we know the intent is to find similar URLs.

For the most part, our site search is about finding keywords. So I think adding a checkbox for "don't disambiguate youtube URLs" wouldn't be very handy most of the time. It would be extremely useful once in a while. But because we have this feature built in seamlessly while you're posting, I don't think it's as critical to have it in the general site search.
posted by pb (staff) at 8:31 AM on March 26, 2013

The double link checker already strips out the youtube ID from youtube URLs and searches for that instead of the full URL.

Bug report, then, because the checker did not find this video in a past post from a recently deleted double.

http://www.youtube.com/watch?feature=player_embedded&v=4PcL6-mjRNk#at=80
posted by desjardins at 8:54 AM on March 26, 2013

Thanks, the double link checker was being thrown by the # timestamp. It should be accounting for those now.

We probably aren't catching every conceivable YouTube URL style for individual videos that exist, so just let us know if you spot one that isn't working like this.
posted by pb (staff) at 10:03 AM on March 26, 2013

pb, thanks for that breakdown, and the quick duplicate check fix. I often use the site search because if I'm putting together a post less than 24 hours since my last post, I want to check if the material is already there. I realize in this, I'm probably far from the usual user of the site search feature.
posted by filthy light thief at 11:18 AM on March 26, 2013

The double link checker searches the entire history of MeFi, not just the last 24 hours. The tag duplicate checker does have a more limited timeframe, but links are forever.
posted by pb (staff) at 11:21 AM on March 26, 2013

Oh, now I see what you're saying. You can't get to the New Post form within the 24 hours of making your post. Yeah, that's a hassle, sorry about that.
posted by pb (staff) at 11:22 AM on March 26, 2013