[13:57] Saw this and thought you might be interested: http://homepage.ntlworld.com/clivebillson/tube/tube.html
[13:58] oooh
[13:58] thanks
[13:59] drop the tube.html from the end to get at the graphics without the irritating JS
[14:05] aww, shame
[14:05] they only have 1936 on
[14:05] Hmm?
[14:07] the page claims to have maps back to the 1880s
[14:08] http://homepage.ntlworld.com/clive.billson/tubemaps/1889.html ?
[14:09] And http://homepage.ntlworld.com/clive.billson/tubemaps/
[14:10] ah. different.
[14:13] fuck
[14:13] I’ve just found a bug in Epistula
[14:13] ?
[14:14] Epistula executes 1387 queries per page
[14:14] that is a very large number
[14:14] You are not kidding.
[14:15] In fact, it executes one query for every object in Epistula that isn’t a comment
[14:18] I shall go out on a limb and suspect that is not desired behaviour
[14:18] Yeah
[14:25] so what’s up?
[14:27] Ah, a limit problem
[14:29] The front page is $limit items. In order to generate this, Eps gets the last $limit items from each table of things that can be on the front page, sorts them by date, and cuts the array to the last $limit items
[14:29] Sadly, I forgot the limit break on the second clause of this.
[14:30] So it got everything from each table that it could display, then the attachments and comments for each of those.
[14:39] oh hell
[15:23] Right. Down to 190 queries, which is better
[15:28] Most of which are “Do I have a crossreference? Aha, cool. What’s the title of the page I’m crossreferencing to… okay, do I have a permalink for that? No, right…”
[15:30] I’m awfully glad I’m caching all of this, because I suspect that one and a half thousand queries per page-load would probably result in my being killed, slowly, and with sticks.
(I should make Eps have an ‘index’ table so it can grab the last 10 items without having to run four queries, but that’s a Future Thing)