How Simuze reads a venue agenda
Every morning a small program reads the agenda of a venue. This is what it fetches, what it never touches, and how a venue gets a wrong listing fixed.
Every morning a small program called SimuzeBot visits the website of a venue and reads what is on. It goes to one venue at a time, waits at least two seconds between requests to the same site, and comes back once a day.
What it reads depends on what the site offers. Many venues run WordPress, and WordPress publishes the agenda in a form that a program can read without guessing, so that is one of the first things it tries. If the site has a calendar or a feed it uses that. If the shows are only in the page, it reads the dates from the page itself.
When none of that works, the text of the page goes to Soriku, our own AI, with one question: which of these shows are listed here, and where do they link to. Anything Soriku names has to be on the page and has to link to a page that really exists, or it is thrown out before it reaches Simuze. It does not invent shows and it is used for nothing else.
A show on Simuze says where it came from. It names the venue, links to the venue's own page for the show, and says when we last checked. Tickets always go to the venue or the ticket shop the venue uses, and the show is never presented as something you can buy on Simuze unless the organiser sells it here. The picture is shown straight from the venue's own site, with the venue named as the source. Thirty days after the show we delete it, unless people used it for something, like saying they went.
SimuzeBot respects robots.txt. If a site names it and says no, it reads nothing. A venue can also ask us to stop or to point us to a feed, and if you run the venue you can claim its page and fix the agenda address yourself. When a listing is wrong, tell the venue or tell us, and one of us will correct it.
I know that reading someone's site without being asked sits close to a line. I try to stay on the right side of it by keeping the bot slow and honest, by showing only what the venue itself publishes and sending people back to the venue, and by making it simple to leave. Every page nobody runs has a remove button, and the removal sticks.
The page at simuze.com/bot lists exactly what it fetches. Read it and tell me if something there does not sit right.
Love, Marten
Love, Marten