« It's the return of the running joke half the people reading it won't get! | Main | Poorly considered shelving units »
Eric: This is a completely good idea, executed well, completely opt-in, and utterly useful with little work. Why do I expect there'll be a 300 comment long argument about it?
(From Oh No! Robot!.com)
Every so often, someone comes out with something that is unremittingly cool. Action Ryan North of Qwantz.com (the home of Daily Dinosaur Comics as well as many other cool things) and Cool Hand T Campbell of Penny and Aggie and about seven thousand other comic strips have decided to build a tool that's amazingly cool, and give it to the world.
That tool is ohnorobot.com. And the idea behind it is so simple, it's staggering it hasn't been done before. Or if it has been done before (I know Bo Lindbergh has been doing indexing for searching, for example) and implemented in a wide scale before, I don't know it. That's all right, though. I'm allowed to not know things.
The idea is this. Comic creators receive a bit of javascript they can put on their web pages. That javascript causes a small icon to appear on archive pages of the comic (right now, it only happens about one out of four times, to help throttle bandwidth until their hosting changes). Any reader who wants to can click on that link, and enter a transcript of that day's comic into a database.
Once I approve the transcript, that text then becomes searchable. It can be searched along with dozens of other comics in a unified search page, or I can direct people to a custom search page just for my comic, or I can add another trivial bit of javascript to my page and get a search form right on the page.
I added John Stark to the service (hey, I never said I wouldn't mention it), and gave it a try. Transcribing given strips is simple. There is absolutely no coding needed -- just hit return between speech bubbles and two returns between panels and submit. I get a chance to approve all transcripts that come through. And then it just plain works.
Very simple. The tools have been available for years now. And we know that the fandoms of comics would love a chance to dig deep into this -- this is the kind of things fandoms love. It's a chance to pitch in, a chance to feel like a part of things. And since you can put your name into the form, the artist will even know it's you!
And, because the transcriptions are so easy, a dedicated fandom can breeze through adding strips to the archives (well, as soon as the bandwidth throttles are taken off. Pray for bobo.) And then, we'll have a resource that lets us search the archives of dozens of strips at once or just our individual strips in particular. It's fast, it's easy, it's entirely opt-in (you have to add the javascript to your own site. If you don't add the javascript, your site will never be a part of it.) Already, some of the big hitter sites like Wigu, Goats and White Ninja Comics are on board. And given how little work is needed on the part of the artists, there's no reason why most or all of the big guns can't jump in. (Unless, of course, they don't want to. "I don't want to" is fair, you know. Don't try to claim it isn't!)
Eventually, this will become a regular stop for me. Especially since there's a number of strips I sometimes need to refer to in a given comic's run, and finding those strips more easily would be a Godsend. On the other hand, I could see sometimes needing to review Valentine strips, and typing Valentine in and seeing what comes up would be an easy way to do that.
Which means, of course, that the engine can also be promotional.
Dude. This is a good idea. And kudos to T and Ryan for doing it.
Posted by Eric Burns at November 9, 2005 10:35 AM
Comments
Comment from: Dave Van Domelen posted at November 9, 2005 11:17 AM
UbT. Nuff said. :)
Comment from: larksilver posted at November 9, 2005 11:18 AM
Wow! First poster twice in a row.
This thing is freakin' awesome. I can now search for that time, when Pintsize.. ooooh, or when was the last time we saw ChooChoo bear, or.. or..
Way cool tool.
Comment from: larksilver posted at November 9, 2005 11:18 AM
BAH. Dave beat me by his brevity. ah, well.
Comment from: Miklon posted at November 9, 2005 11:19 AM
Well... Google was bound to do this sooner or later.
Comment from: Eric Burns posted at November 9, 2005 11:20 AM
UfbT will get added. Eventually. ;)
Gossamer Commons is waiting for a site redesign, too, since the archive URLs are going to be changing.
Comment from: John posted at November 9, 2005 11:30 AM
This is a cool idea, but I see one problem here -- I just transcribed one of Questionable Content's comics, and it's kind of a chunk of work to type all that!
Some comics are less wordy, sure. But I wonder if fans are going to go for doing all that typing over and over for hundreds of comics. I suppose given enough time and enough readers, it'll all get covered.
Comment from: Scarybug posted at November 9, 2005 11:45 AM
Ha! I sent a friend of mine all the transcripts to my comic because he offered to write a search function for it. I'll have to tell him about this to save him the work.
Comment from: larksilver posted at November 9, 2005 11:55 AM
You think transcribing a QC stip is work? Wow. They're not even the wordiest of the wordy out there.
Comment from: Zutto posted at November 9, 2005 11:57 AM
Wow! Definitely a totally awesome idea.
Also the gateway to a scaleable research project about webcomic development. :)
Comment from: Danalog posted at November 9, 2005 12:11 PM
He mentioned John Stark! BURN HIM!
Oooh, nifty toys *scampers off*
Comment from: TheNintenGenius posted at November 9, 2005 12:18 PM
I spent quite a bit of yesterday transcribing Girly and Cutewendy comics so I know this is a pretty cool idea. Yeah, it's a bunch of work, but I don't mind, personally.
But yeah, this project does have a ton of potential. It also proves to me once again that Ryan North's a really cool guy.
Comment from: 32_footsteps posted at November 9, 2005 12:59 PM
If I was planning on opting in, I'd start something on a forum for my fans organizing it so that each person just takes a small chunk. It would make it much easier to handle (especially for wordy comics like S*P).
Comment from: qwantz posted at November 9, 2005 1:58 PM
That's sort of the point, 32_footsteps! Everyone is given the option to transcribe, and there's no obligation. Each person can do as much or as little as he wants, and the system is set up so effort isn't duplicated. We handle that stuff for you!
Comment from: 32_footsteps posted at November 9, 2005 2:11 PM
I mean, I'd actually organize it, instead of having five people all focus on one of the most popular single strips and having a bunch almost completely ignored.
Comment from: Arachnid posted at November 9, 2005 2:33 PM
Very cool. As an accessibility advocate, I like it even more, too. It's only a shame it won't be linked from the 'longdesc' attribute of the comic images, or be on the site itself (so you could search with "foo site:comiciread.com"), but we can't have everything now, can we?
Now, if we're really lucky, they'll add an RSS feed, since they already have something that detects when a new page is up. RSS the comics web!
Comment from: Abby L. posted at November 9, 2005 3:07 PM
I would do this, but I'm afraid no one would transcribe my comics anyway.
Comment from: gwalla posted at November 9, 2005 3:42 PM
Dave: Unfettered by Talent is already transcribed and searchable through the Secret Comics Database
CRFH was going to be in the SCD, but the fanbase was surprisingly incapable of transcribing enough for the search to be useful.
Comment from: Clint H posted at November 9, 2005 4:16 PM
Keenspot is supposed to be transcribing comics for some sort of magic Google thingiemadingie.
Note however Keen projects take a while...
Comment from: Bo Lindbergh posted at November 9, 2005 4:36 PM
CRFH was going to be in the SCD, but the fanbase was surprisingly incapable of transcribing enough for the search to be useful.
Well, not quite. There was an earlier search project that managed to cover everything up to early 2003, but it sort of wilted from lack of daily maintenance (the transcribe-each-installment-as-it-appears part).
Comment from: Aerin posted at November 9, 2005 4:41 PM
Hardly any work is needed in the case of Goats; since all the comics are transcribed already (in crazy detail, I might add), it's a very simple matter to just move them over to Oh No Robot, so there's no duplication of effort.
Also, I just transcribed not one, but two spit takes in QC. This thing rules.
Comment from: Bo Lindbergh posted at November 9, 2005 4:53 PM
Incidentally, the approach I chose is very different from Oh No Robot's; I aimed for "insanely fine-grained". If, for instance, you wanted to find all occurrences of Joyce singing without Walky being there, you'd enter Joyce :g & ! Walky : as the query text.
Comment from: David Morgan-Mar posted at November 9, 2005 5:00 PM
Hmm, that is cool. Of course I've had a search facility on my comic for some time now. But I see that they're offering an easy import service of my already transcribed strips, so I'll try that out.
Comment from: Alexis Christoforides posted at November 9, 2005 5:54 PM
Webcomics need more computer scientists!
Comment from: miyaa posted at November 9, 2005 5:55 PM
I don't know, do we really want to know the number of comics that make reference to World of Warcraft in them?
Comment from: Bo Lindbergh posted at November 9, 2005 6:39 PM
I don't know, do we really want to know the number of comics that make reference to World of Warcraft in them?
At the moment, exactly none.
There are sillier searches, such as what strips contain all vowels of the English alphabet?
Comment from: 32_footsteps posted at November 9, 2005 6:49 PM
Here's a better one - can you search to find out how many strips are missing the letter e?
Comment from: gwalla posted at November 9, 2005 7:33 PM
That all-vowels search has pointed out a typoed QC transcript.
Comment from: Bo Lindbergh posted at November 9, 2005 7:47 PM
Here's a better one - can you search to find out how many strips are missing the letter e?
Doesn't look like it. (But the secret comics database can, of course: ! *e*)
Comment from: Robert Hutchinson posted at November 9, 2005 9:45 PM
This is the kind of thing that makes me type for four or five hours.
(I did one, and I might do another tomorrow.)
Comment from: Connor Moran posted at November 9, 2005 10:38 PM
Useful utility, yeah yeah. Promotion opportunity, yeah yeah.
Isn't somebody going to note that the Robot himself is perhaps the greatest thing ever? Lookatim! Combined with the phrase "Oh no! Robot!" I think that there's a level of awesomeness that could kill a man.
Comment from: kirabug posted at November 9, 2005 10:40 PM
Meh, it's a good idea, but as I'm already entering the text of the comics into the "alt" tags for usability while I'm retrofitting a number of the graphics, so that the Wordpress search box can find them anyway, I'm not sure there'd be any significant benefit to having someone transcribe them again.
Comment from: Benor posted at November 9, 2005 11:32 PM
In keeping with the title of this post, I will claim that this idea is useless and horrible. The grenade has been thrown!
But on a more serious note, how does this work with pay comics? I'm not a coding guy (at ALL), so I'm not claiming it would or wouldn't. But if someone transcribed a pay comic's archives, and then the little robot gave them the link in question, would they be stopped from seeing the comic but know that it had the terms in question? Or would they be able to see the comic?
I imagine it's the former, but I'm curious. Oh, and it might be nice to see an advanced search option later on, where you can choose to search within a particular set of comics. Such as choosing only Clan of the Cats and CRFH, and then picking out a term to find their crossover.
Comment from: Godspiel posted at November 10, 2005 12:03 AM
What I'd find just as useful is a search through text descriptions of the visuals. I'm constantly substituting in synonyms for words that actually appeared, so I tend to use a combination of date guessing and rapidly clicking "Next" when I'm trying to find a specific comic in someone's archive.
Being able to type in "Davin, Davin's computer, Choo-Choo Bear, Green Shirt, 'cock'" and being able to get a corresponding comic, as opposed to just filling up my livejournal, would be wonderful.
The fact that the Goats search covers locations and props is fantastic, as is the fact that the text and these are included on the page itself for Web Search Engine Indexing.
Do the major Web Search Engines index alt text?
Comment from: jjacques posted at November 10, 2005 12:11 AM
"This is a cool idea, but I see one problem here -- I just transcribed one of Questionable Content's comics, and it's kind of a chunk of work to type all that!"
As of right this second, there are 513 pending transcriptions for Questionable Content. That's not even counting the hundred or so I approved yesterday. My readers are crazy dedicated.
(Astute readers will note that there are not in fact 513 extant QC strips at this point. We've got a lot of multiple entries in there right now, which I am currently pestering Ryan to find a filtering solution for.)
Comment from: quentin mcalmott posted at November 10, 2005 2:18 AM
I was going through the QC archives, transcribing some, and I found a few things that could be helped:
First, the "image to transcribe only loading 1 of 4 times' is kinda annoying, because I'm not sure if it's not showing because I haven't refreshed enough times, or because there's already a transcription in the archive, so I have to go and check to see. I do understand that it's for bandwith issues, but it makes it hard, at the moment.
If there was a way to check what comics needed to be transcribed, or a list of comics that were transcribed, by webcomic, that would be entirely helpful, because I'd rather spend time doing something that hasn't been submitted yet. Just to be more helpful.
eh, either way, great idea, guys! keep up the good work!
Comment from: qwantz posted at November 10, 2005 7:48 AM
Thanks everyone for these really great suggestions here. The 'transcribe' button appears even when a transcription has been submitted (but not approved) to protect you from having a malicious dude going through and submitting junk - but I may change this, because I wasn't expecting so many duplicates, and people do seem to be generally quite awesome.
The "1 in 4" thing has been scaled back to "1 in 2" and will be dropped entirely soon - I'm sorry about that, but it was the only way to keep the site up during the unexpected initial blast of popularity.
As for pay comics, where you can't see the archive, they would link (probably, assuming the web server is set up as I think it is) to a comic you'd have to subscribe in order to read.
And I really like this idea of RSS feeds! I'll add it to the list of Features I Need To Add Soon.
Comment from: Tim Tylor posted at November 10, 2005 8:40 AM
The 'transcribe' button appears even when a transcription has been submitted (but not approved) to protect you from having a malicious dude going through and submitting junk - but I may change this, because I wasn't expecting so many duplicates, and people do seem to be generally quite awesome.
I suppose one solution would be to let people see the previous submitted transcripts when they click 'transcribe', so they could see whether or not there was a good transcript already entered. If there was a mostly-decent transcript with errors, they could copy-and-paste it into the submission box and make corrections. I've no idea whether that would be possible, though, not knowing a thing about the technicalities of the database.
Comment from: xbishop posted at November 10, 2005 8:46 AM
I know this is off topic, but has anyone else seen today's GPF?
Comment from: Danalog posted at November 10, 2005 9:19 AM
I guess that depends if two negatives make a positive, or in this case, two had me and lost me's make a read me.
Comment from: Benor posted at November 10, 2005 11:51 AM
"As for pay comics, where you can't see the archive, they would link (probably, assuming the web server is set up as I think it is) to a comic you'd have to subscribe in order to read."
That's what I figured. i was just curious.
Comment from: larksilver posted at November 10, 2005 11:55 AM
But hey... if you're a new reader to Narbonic, for instance, wouldn't you be able to still see the text?
Might be nice if you're trying to figure out what the hell they're talking about when they mention giant robotic feet.
Comment from: siwangmu posted at November 10, 2005 1:41 PM
So, I love this transcribing idea, but... format? I mean, what is ideal, here? I have no idea how to transcribe in a way that's optimized for searching. Let's say I do a Penny Arcade where somebody stabs somebody in the eye. What if someone searches poke instead of stab? What if I had only written stabs someone, and they search for eye? These examples are lame, but hopefully someone gets what I mean, here. Is there any set of guidelines for stuff like this, how much detail, etc.? This is not that different from people saying they wish we could search blue shirt and such, I'm just wondering, practically, how should we be making this call?
Comment from: siwangmu posted at November 10, 2005 2:45 PM
Update: when you actually go to transcribe something, it does in fact link to a set of guidelines! Guidelines that directly imply that only text should be included, because "Stuff that advances the story is what we're interested in." Not that I'm saying the stuff on a character's shirt (their example) necessarily advances the story, but... dude... this way you could never search "name kisses name" unless you remembered the stuff they *said*. Is just interesting. However, with guidelines in hand, it's off to look for things to transcribe because it's a very appealing thing to have done I think!
Comment from: gwalla posted at November 10, 2005 4:35 PM
The duplicate transcripts problem can probably be solved by setting it up so that when a transcript is accepted for a given stip/page/whatever, all other pending transcripts for that strip are dumped.
I don't believe that either search engine can search for actions. Synonyms would make it infeasible. Neither is set up to search for props either, although I think the SCD is capable (it's just that nobody's added that info in transcripts because it's kind of tedious). The SCD can seach for characters appearing in-panel, which it doesn't seem like ONR can do (you can search for a name in ONR, but you'll get results for every time that character appears and every time the name is used in dialogue, and you can't limit it to in-panel appearances)
Comment from: Aerin posted at November 10, 2005 7:35 PM
The SCD can seach for characters appearing in-panel, which it doesn't seem like ONR can do (you can search for a name in ONR, but you'll get results for every time that character appears and every time the name is used in dialogue, and you can't limit it to in-panel appearances)In that case, I believe you'd search for "Name:". It wouldn't work in the rare occasion that a character shows up and doesn't speak, but the rest of the time you'd be fine.
Comment from: qwantz posted at November 10, 2005 8:26 PM
Yep, the duplicate transcript problem was taken care of this morning: if you approve a new one for a strip, the others for that strip are dropped. And yeah, the database model for ONR is much simpler. It's a matter of necessity: the average reader (i.e., not the hardest of the hard-core) is not going to take the time to note props, background characters, etc - so we give them a plain text field and a few simple instructions and go with that. I think ONR and SCD are serving two different interests: ONR is for searches (though not as fine grained as who kissed whom with what when) across comics, while SCD is a more-complex database for a smaller set of comics. There's room for both, I think! Though I confess I was unaware of SCD when I built ONR - it was clearly a SECRET.
Comment from: Bo Lindbergh posted at November 11, 2005 12:39 PM
Oh No Robot—the Keenspace of search engines.
Comment from: larksilver posted at November 11, 2005 1:02 PM
Oh No Robot, as I see it, can serve as that "readily available" reference site for someone new to comics. Perhaps new reades won't be quite so intimidated by years.. and years.. and years.. of Schlock, or Sluggy, if they can look up the name of a character the team has obviously seen before.
Post a comment
Thanks for signing in, . Now you can comment. (sign out)
(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)