Playing around with the Mac program DEVONthink, which describes itself, among other ways, as a "freeform database." This seems, at first glance, to mean that it functions as a huge data receptacle, one into which you can dump all kinds of raw material, which then gets retrieved not via the normal array of database queries but rather through a sophisticated search interface.

From the documentation:

"With its intelligent organising capabilities, DEVONthink is the number one choice for the 'hunters & collectors' type of people. They tend to store every bit of whatever they get hold of, from text files [to] images, MP3s and Quicktime movies to web pages and bookmarks, and some of them MIGHT even organise it somehow. But most of them won't bother with kinky things like 'groups.'

"And that's all what DEVONthink ... is all about: storing and organising things. So, H&Cs will throw everything they have into the database ... Then, when they look for 'something they are sure they MUST have somewhere,' it's time for DEVONthink to play out its cards: the advanced search functions and the AI-based 'see also' and 'keywords' buttons. ... If the document the H&C is looking for is not [found]. it might at least be similar to one [that is]. One click, and DEVONthink shows a list of all other documents that are similar to the selected one."

I've been seeding DEVONthink with the data from my index card file, and it's been an interesting experience so far. Browsing the cards using this program is different from browsing them in Access, although in ways that are hard to quantify precisely. At this early stage I am prepared to risk falling into the old Mac vs. PC dichotomy by saying that DEVONthink feels "fuzzier" or "more organic" than Access, whereas Access feels more "precise" but also more rigid, and certainly less associational. At this point it's worth it for me to keep both databases up-to-date and active, but I'll let you know if one eclipses the other.

It's worth it to mention that the DEVONthink strategy of relieving the user of the need to categorize by providing a powerful, intuitive, effective search mechanism seems pretty akin to the idea behind the Google Desktop Search hard-drive search utility. Note the way that Google rep Marissa Mayer talks about search vs. directory in the conversation blogged at John Batelle's Searchblog:

"In 1995 ... you could find what you were looking for by browsing a directory like Yahoo. But over time as the web scaled that model didn't scale. It broke, which is why search (became the metaphor for finding things on the web). We are seeing the same thing happening now on personal computers (which have far more storage than even five years ago)."

Labels: databases, indexing

Thursday, October 28, 2004
6:58 PM
0 comments

Had a long, caffeinated discussion about databases this morning with CJO. What makes an effective database, what databases could be used for, etc.

This afternoon I find myself wondering: why is it that the database is not widely recognized as a form for artistic expression? There are certainly times when I feel like the index card file will end up being the best piece of creative output I will ever produce. And databases, in general, are oriented towards fragmentation, discontinuity, heterogeneity, montage, collage, and systematics—major governing principles of contemporary art.

Guy Davenport writes "A work of art is a form that articulates forces, making them intelligible." The database is literally designed to be such a form. So where are the artists striving out in that direction? Milorad Pavic has produced at least one book that points the way...

I suspect there are hypertexts that qualify—almost all hypertexts have, as their back-end, arrays of lexia that could basically be thought of as existing in a sort of database form—although I'm hard-pressed to think of many hypertext works that actually function like databases. Where are the creative works that open within, say, Access, rather than running as independent programs with their own interfaces (such as those generated by a hypertext authoring tool like StorySpace)?

Labels: databases, hypertext, indexing

Saturday, August 21, 2004
2:29 PM
0 comments

I'm enjoying the comments on the Mac thread below, and I thank everybody for their input.

I'm still pretty curious about DevonThink, as the notion of making a "personal encyclopedia" in an open database system is highly appealing. This is a notion that has been of special interest recently, because I have begun the process of digitizing my entire index card file.

This was a project that I've long wanted to undertake, but I've always balked at the sheer amount of data entry that the project will require. I'm only doing it now because the July 3rd hard drive crash forced my hand—in that crash, I lost the Word document that served as the cross-referenced index to the card file, and I figured that if I had to go through the long process of recreating a cross-reference anyway, I might as well put in the extra effort to make a full-text version in a proper database. The advantages of having a full-text digital version are obvious: being able to use the processing power of a computer to filter and shuffle the thousands of cards I have on file will be super-fun, and if I get a version up-and-running on a laptop the entire card file will essentially become portable.

Not to mention reproducible, and thus able to be given away / traded / shared. In Caterina's July 5 post on DevonThink, she talks about how appealing it would be to browse through someone else's database; she goes so far as to suggest that it's a pleasure she'd be willing to pay for / subscribe to. The pleasure of weblogs is, to some degree, the pleasure of reading through someone else's notes, but if you read someone's weblog regularly, it mostly works as a linear process, whereas a database works with multiple points of entry and multiple avenues of potential investigation. The beauty of a weblog is that the entries are arranged in a chronological superstructure (of the sort that I've written about before); the beauty of a database is that the entries are equidistant from one another, and can be endlessly rearranged into different configurations requested by their user. (Half the fun of ITunes is shuffling around the songs in the Library.)

So why aren't more people putting up their notes as Access files for one another to download? Or offering them on the subscription model (each month, receive a "booster pack" of new notes.) Is this happening in a different subculture (say, the subculture of people who trade recipes)? If not, why not?

"And today the book is already, as the present mode of scholarly production demonstrates, an outdated mediation between two different filing systems. For everything that matters is to be found in the card box of the researcher who wrote it, and the scholar studying it assimilates it into his own card index."

—Walter Benjamin

Related: Umberto Eco's fears about a future in which "[w]e could end up with competing encyclopaedias, some of them completely wild." Eco doesn't like it, but (as I've said before) such a future doesn't seem bad to me at all.

Labels: databases, indexing

Monday, July 26, 2004
7:28 PM
0 comments