The Trilema Article Database, a toollet.

January 9th, 2018

In its ten+ years of life, trilema has amassed over 8,000 articles, a feat that has terrorized many, and confounded still more. The sheer number and variety of pieces can make recalling what was said where a bit of a bitch, especially if one can't remember an exact distinguishing phrase.

The Trilema Article Database is a humble attempt to meaningfully, if somewhat idiosyncratically1 index Trilema articles, making them searchable beyond Google's arguably useful options.

I first designed TADB to work as a lisp IRC bot that'd idle in #trilema, but opted to put it up on this site as a php script following two issues: 1. nobody's gonna like a bot capable of (eventually) spitting 8,000+ title/url pairs in channel, and 2. I'm a noob2 and lisp is beardy, how about we not wait forever for the echafaudage of what is, after all, an experimental and possibly even not all that desirable item. Should the thing prove useful I imagine it'll see a lot of surgery.

As of this post solely January of 2017, chosen arbitrarily, has been indexed, to allow potential users to chime in on any modifications3 they'd like to see --obviously, any change to the indexer itself will require a re-indexing of all known articles, which prompted the minimal initial run. Anyone interested in taking on some chunk of pieces, say a month, or a year, or even a category, is welcome to join in --I find it's a great exercise in revisiting old pieces and restructuring them in one's own head, a pleasant reward for the time spent. I'll also make sql dumps of the database available to the lordship for the asking, should anyone want to use it in part or total for Mad Science.

Please give it a spin; TADB lives here.

  1. I see no way out of this; while a number of objective criteria may be extracted from a given article and thereby used to search, an index will contain some interpreted data if it is to be used to match readers' memories to extant articles (that is, unless I'm the only one who doesn't think in terms of "Gee, what was that one piece with 587 words 42% of which started with t?" I guess it's possible. InformAssimilate me! ) []
  2. I listed it as #2 so it's less obvious. That makes it less obvious, right? RIGHT?! []
  3. Additional criteria would be especially appreciated. []

2 Responses to “The Trilema Article Database, a toollet.”

  1. Diana Coman says:

    Looks interesting. It took a bit of spinning it to start getting some idea as to how it works tbh so I might be entirely not getting it but if I understand it correctly and the whole thing runs on a manually (or including a manually made part) built set of labels for the indexed articles, why not offer for each part the actual list of relevant labels? Basically some help for the explorer as to the paths that this particular map really knows anything about.

    A search for "poetry" as genre under "Literature" with "yes" for "Discussion present in article" returned all the Disgrace chapters. I can see that as potentially poetry, sure, but then I'm a bit at a loss as to what wouldn't be included as poetry from the category "Cuvinte Sfiinte" in which case, well, I'll go and get the category itself. Pushing this, I guess it would be interesting to see at least how is this exactly more useful than the categories themselves (tbh that's quite often how I end up finding something on Trilema if a direct search doesn't work: by most-likely-makes-sense category and then possibly some approximation of time).

    As a side: at "History" one can currently select both yes and no.

  2. hanbot says:

    I suppose I could change the entry fields to drop-down lists that'd include all known labels (with the exception of "names", which is already enormous and likely to become a monster). I guess that's more of a browser than a searcher, but I can also see the argument for doing both, if time's spent on one. I'll mull this, thanks!

    Re the poetry: the query you described only pulls up those chapters of Disgrace that actually discuss poetry (Lord Byron shows up a number of times, as the main character is attempting to write some operetta based on his life), not the entire set of chapters. I specifically tried not to classify articles as one subject or another, as it'd be quite the feat if more than one or two readers could agree. ;) Moreover, it's supposed to be about what they include, even if it's incidental.

    Nice spot on the history buttons, fixed!

Leave a Reply