An Open Letter to Heidi Tandy
Nov. 24th, 2003 11:42 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
Dear Heidi,
FA is a very large archive. I'm not sure if there is an archive dedicated to one fandom that's as big as FA, in fact. Organizing and searching through that thing is complicated, just because of the sheer size of it. The focus right now is what's on what's recent; it's easier to find stuff that's brand new. That makes sense, but it's kind of piling everything on top of each other as it comes in, isn't it.
This isn't a new problem. Every time people start collecting information they hit the same wall. Back when the fandom had three fics it was no big deal, you could have them in any order you wanted. You could say, hmm, I think I want to look at that one, and everyone knew which one you meant. But the more you get, he more you need to organize and, dare I say it, catalogue it. Alphabetical just doesn't cut it; ask any student who's ventured into a library. What if your library collection went with the alphabetical system? You'd never be able to look up a call number and go look at see what's around it, there would be no browsing. There's no relationship between Adams and Adamson except that they sit close to each other alphabetically. One could write about nuclear physics and the other about pornographic scuplture. You wouldn't be able to use the system as a tool to help you find more of what you want, you see. You'd be at the mercy of the alphabet; where is that book? Where can I find that author? Meh, not useful, I tell you. Not when you've got serious volume, like many fanfiction archives do.
I have a modest proposal. Picture it: fanfiction, organized according to...the Dewey Decimal System.
I mean, we'd have to tweak it a bit, but...imagine! You could search your digital bookshelf and find all similar fics together. You could highlight the new ones, or you could have a 'just in' section. Also organized according to Dewey, of course. But we could called the Ivy Decimal System, of course.124: Harry/DracoCome on, it would be beautiful! A masterpiece of organization! Subdivided by geography, style, length, whatever. And if someone comes up with a new pairing you just make a new heading, this is the beauty of the DDC! You could search it by subject, by pairing, by author...think of the possibilities! Your Snape/Remus sap fics grouped together, apart from Snape/Remus sappy fics set in America with a Mary Sue! It's how fics want to be arranged, can you hear them crying?
124.5 Harry/Draco, angst
124.534: Harry/Draco, "how they got together", angst
124.534973: Harry/Draco, "how they got together, angst, set in America
124.55, Harry/Draco, parody
124.55973 Harry/Draco, parody, set in America
130: Harry/Ron
130.5: Harry/Ron, angst
130.5973: Harry/Ron, angst, set in America
130.9999973 Harry/Ron, with Mary Sue, set in America
I hope you will consider my suggestion.
Sincerely,
Ivy Blossom, rabid aspiring librarian
Edited to add: This open letter is not a challenge or a criticism, nor is it a response to anyone turning down this idea. I was just thinking about this after cataloguing class and I wanted to tell Heidi about it, but she was not online. So I wrote her an open letter instead, because I thought some of the librarians and proto-librarians on my friends list might enjoy my inanity. That is all.
no subject
Date: 2003-11-24 08:45 pm (UTC)no subject
Date: 2003-11-24 08:54 pm (UTC)no subject
Date: 2003-11-24 08:46 pm (UTC)As an aspiring librarian myself, I just have to say...
*HUMPS*
That is all.
no subject
Date: 2003-11-24 08:48 pm (UTC)(no subject)
From:(no subject)
From:no subject
Date: 2003-11-24 08:46 pm (UTC)Ivy Decimal System, the wave of the fic future.
no subject
Date: 2003-11-24 09:00 pm (UTC)I just can't manage to do anything and leave it separate from anything else I'm doing...did I mention that I wrote a paper about how the library of congress catalogues books about slash fanfiction? My prof must think I'm insane.
(no subject)
From:no subject
no subject
Date: 2003-11-24 09:01 pm (UTC)(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:no subject
Date: 2003-11-24 08:49 pm (UTC)no subject
Date: 2003-11-24 08:53 pm (UTC)no subject
Date: 2003-11-24 08:54 pm (UTC)no subject
Date: 2003-11-24 09:01 pm (UTC)no subject
Date: 2003-11-24 09:03 pm (UTC)(no subject)
From:(no subject)
From:(no subject)
From:no subject
Date: 2003-11-24 09:04 pm (UTC)And hey, if we have DDC, a nice button that says, "find me fics near this fic."
*gets into the act*
no subject
Date: 2003-11-24 09:07 pm (UTC)SOMEONE STOP ME.
(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:no subject
Date: 2003-11-24 09:14 pm (UTC)no subject
Date: 2003-11-24 09:16 pm (UTC)no subject
Date: 2003-11-25 10:34 pm (UTC)no subject
Date: 2003-11-24 09:20 pm (UTC)no subject
Date: 2003-11-24 09:23 pm (UTC)no subject
Date: 2003-11-24 09:26 pm (UTC)no subject
Date: 2003-11-24 09:28 pm (UTC)no subject
Date: 2003-11-24 11:23 pm (UTC)no subject
Date: 2003-11-24 09:32 pm (UTC)no subject
Date: 2003-11-24 11:23 pm (UTC)(no subject)
From:no subject
Date: 2003-11-24 09:36 pm (UTC)no subject
Date: 2003-11-24 11:22 pm (UTC)no subject
Date: 2003-11-24 09:45 pm (UTC)no subject
Date: 2003-11-24 11:22 pm (UTC)*smooches*
no subject
Date: 2003-11-24 09:45 pm (UTC)no subject
Date: 2003-11-24 09:58 pm (UTC)(no subject)
From:(no subject)
From:(no subject)
From:no subject
Date: 2003-11-24 09:50 pm (UTC)But I see you're point. It would be easier. Although, everyone would have to learn the 'Ivy' system, unless the search was kept for a while, which I believe is a good idea. And who is going to read every single fic and place it? What if they fly to America but was originally set in England? Things to ponder.
If someone did the 'Ivy' system, it would be wonderful. I just can't see anyone wanting to take on that task.
no subject
Date: 2003-11-24 11:21 pm (UTC)I've done stupider things. :)
Yes, it is a good idea. But, (and there is a but) you would have to re-organise the whole Dewey system, so why not just start at 1000 (which is where I believe the Dewey system ends). And then, even if you do start at 1000, you have to number every possible couple and feeling and where it's set, etc..
But what would be the purpose of starting at 1000? I mean, most of the Dewey subjects are irrelevant in this context. Basically what we need is a very very specific, and very very simplistic Dewey system with a fairly set number of subject headings. There are a million topics for fics, but only a limited number of ways that fandom tends to group those topics. So you could just pare a lot of that stuff down, but change things like, "subdivide by geog." into subdivide by something else.
But I see you're point. It would be easier. Although, everyone would have to learn the 'Ivy' system, unless the search was kept for a while, which I believe is a good idea.
First, no one would have to learn the system. No one understands Dewey, but the organization works. All people know is that when they go to a specific section to find a book, they find lots of others just like it around it. And that the subjects tend to progress from topic to topic as you walk along the stacks. No one needs to understand more than that, as an end user. If they find a fic they like they can jump out to the catalogue to see what's around it, and browse from there. And who said anything about getting rid of the search? All libraries have searchable catalogues; why get rid of a search?
And who is going to read every single fic and place it? What if they fly to America but was originally set in England? Things to ponder.
You don't have to read them all. The files already have metadata with them. Author, title, pairing, summary, keywords. You'd just need to extrapolate from those.
(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:<333333333333333333
Date: 2003-11-24 09:52 pm (UTC)I love it!!!!!!!!!!!!!!!!!
X
Re: <333333333333333333
Date: 2003-11-24 11:16 pm (UTC)I know how you love to be scared. :)
Re: <333333333333333333
From:no subject
Date: 2003-11-24 10:12 pm (UTC)no subject
Date: 2003-11-24 11:15 pm (UTC)no subject
Date: 2003-11-24 10:34 pm (UTC)Hm. Of course, to do it properly, a controlled vocabulary should be established and agreed upon, otherwise the catalogers will be all over the place, and the users won't have the slightest idea how to search. There are too many instances where a pairing is indicated with a first-name/last-name situation (Harry/Snape), or both (is it Sirius/Snape or Black/Snape?), and then there are some that are almost always first-name/first-name (Harry/Draco). Just to name one sticking point for the end-user.
It sounds like so much fun!
no subject
Date: 2003-11-24 11:14 pm (UTC)*defends original parody idea* WHAT'S WRONG WITH ME?
no subject
Date: 2003-11-24 11:15 pm (UTC)no subject
Date: 2003-11-24 11:58 pm (UTC)No wonder they're always hiring.
(no subject)
From:(no subject)
From:(no subject)
From:no subject
Date: 2003-11-24 11:51 pm (UTC)To start with, Dewey is fundamentally hierarchical. It divides all texts up into subjects, then subsubjects, then sub-sub-subjects, and so on. Then it imposes a linear ordering on them. This makes sense from a library's point of view - after all, you've gotta shelve these books somehow, and we can't have complex information topologies in our plain old 3-D space. (Though I just had this really cool vision of of a library with non-Euclidean topology, where stacks of books intersect based on converging subjects...it'd be like an Escher library!)
But information, at its core, does not follow a hierarchical pattern. Books and stories and other snippets of information fall into multiple categories, so you end up with a graph instead of a tree. I remember my first experiences using the library. I was looking for stuff on warships - but I found that the books of interest were spread far and wide. Current reference books like Jane's Fighting Ships were in the reference section. Historical narratives were in the 940.54s, primarily, but were spread throughout the 900s. I found a bunch in, I think, the 380s, which is God-knows-where (I remember it was near the porn section, though).
That's not very conducive to browsing. Ideally, I'd just be able to go to some information nexus and find everything I wanted about warships. But what if someone was looking for historical information on European history? Some of these books would belong in their nexus too. And what if someone wanted military weapons in general? They'd also need a subset.
(contd in next comment)
no subject
Date: 2003-11-24 11:51 pm (UTC)(Incidentally, you might note that your computer's filesystem works exactly like the Dewey Decimal System...you have a hierarchy of directories, starting at the root, where you have folders to put your files. If a file belongs in two different categories, you need to create a Shortcut to it, which is basically what the card catalog does. But the next version of Windows will work much more like the database I just described. You attach attributes to your files, and can then query at will for whatever attributes you like.)
To give you a sneak-peak into the forthcoming FA system, it'll store data on ships, characters, genres, and eras, along with more mundane stuff like date published, hit counts, and title and author. Then you can search by any combination of those. So, you could say "Give me all post-Hogwarts romance fics that feature Harry/Hermione or Ron/Hermione where the main character is Hermione", or "Give me all Ron/Draco fics that take place in their 5th-7th years and involve lots of angst". I think (though I haven't read the spec recently) that you could even do stuff like "Give me all fics featuring Ron or Harry at Hogwarts that are not Ron/Hermione but are romance".
The one thing I don't like about such a relational database is that the attributes themselves are fixed. You can specify any combination of ships/genres/eras/characters, but we have to make the decision that ships, genres, eras, and characters are what are important. If a new attribute came along, say, stylistic devices, we'd have to modify the software by hand to incorporate that. And such a modification would be fairly painful.
As a little on-my-own research project, I'm investigating tuple spaces as a means to solve that. A tuple space is sort of like a relational database, but it doesn't impose the same constraints on records. Basically, tuples can have whatever attributes make sense for them. This makes adding a new attribute fairly easy. But then there are problems in searching, if everybody makes up their own attributes. Sometimes the extra constraints on relational tables really do make things easier.
BTW, I prefer Library-Of-Congress to Dewey. Dewey suffers from having only 10 major subject areas, so when computers came around, they had to shove them in the 001s, which is I believe next to the occult. LOC offers up to 26^2 main subject areas, which is much more appropriate IMHO.
(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From:(no subject)
From: