of APIs and catch-22s

[melannen]

So for the past, oh, year or so, I have been sporadically and half-heartedly working on* a fork of writtenkitten that will let you pull images from more fannishly-oriented sites than flickr, because tbh when you put in "kaiju" or "methos" as your search term it doesn't pull up much from flickr/creative commons.

Anyway, I have gotten as far as rebranding it "wrotegoat", fixing the bug for the first kitten goat, and adding some more user options, but when it comes to actually getting images from Tumblr, I run up against the fact that I have no idea what I'm doing with APIs, and Tumblr makes you "register your application" before you can play with their API.

So if there is anyone here who knows anything about APIs, can you answer me one simple question? To make an app with the tumblr API, I have to register an app. To register an app, I have to have made an app. To make an app I have to register an app. To register an app I have to have made an app. HOW DOES THAT WORK?? Can I register with a dummy URL, keep coding with the files on my hard drive until it at least vaguely works, and then change it later, or something?

...also shouldn't programmers know what an 'infinite loop' is?

*when I know I should be doing other stuff but really really dunwanna

Comments

[stellar_dust]

Trying to recall from what I did at MHC ... I think APIs actually need you to have something running on a webserver for them to interface with. Like, as far as Tumblr is concerned, your computer is just an I/O interface for a human, but for the API to work, your server and Tumblr's server have to be able to interface more intimately than that - software has to talk directly to software, which needs a server, given ::technobabble about the infrastructure of the internet::. Someone else can probably describe that more coherently.

You should be able to set it up on katycat.net, though.


ORRRRRRR you could just click back and forth between wowair.us and icelandair.us until you make a decision, both of which have frankly unbelievably low prices right now (compared to last year anyway).

[melannen]

*sigh*

Well, I can run the writtenkitten javascript code, which uses the Flickr API, from an html file saved to my usb stick, and it works just fine? So it can't be inherently required for an API that the code be running on live on the internet. (I mean this is a really simple use of an API, I'm just pulling URLs for image files out of a block of html, I just need to get it to spit the html into my javascript variable, and there is a tumblr API call that's supposedly designed to do that.)

Possibly writtenkitten is working off USB because it's still authenticating via the writtenkitten website, but in that case presumably I could register a dummy URL on katycat and keep working on the javascript via the USB...? None of the javascript in writtenkitten seems to require contacting the original website - the api key is just a string of hex digits.

(If it has to be running live on the web, I can't mess with it at work, because I can't get into anything fancy from work, plus that's just plain annoying, and if I can't mess with it at work I will never mess with it.)

[synecdochic]

Yes, I haven't looked at Tumblr in specific, but sites generally want to know who you are, how to contact you, and a URL that will be associated with your client/gizmo/etc before they'll give you an API key -- it means they'll know where to yell if your gizmo breaks horribly and starts effectively DDoSing them.

You should be just fine with a dummy URL.

[melannen]

Cool, thank you so much! That was the answer that seemed to make sense but it was surprisingly difficult to get anywhere on the internet to just tell me that in plain language.

[synecdochic]

Yeah, I don't know of any really good "how to use a site's API to do cool things" tutorials out there -- everybody assumes you already know what you're doing!

[melannen]

Yeah, I found plenty of "how to use this particular site's API to do cool stuff", but they all assumed you already knew the basics of APIs.

Well, once I can register, I can mess around with it and figure it out for myself probably - that's the most fun way to learn - it's just those super-basic basics that always trip one up when trying to learn on one's own.