Telling developers from users
Olivier thinks the way the Yahoo API limits queries by IP is a bit strange, which reminded me of what I realized while going to sleep last night: it’s actually very cool. With the Google API, if you distribute an app or build a web app, you have to tell your users to go sign up for a developers key and have them plug it in your app, but with Yahoo your app ID (which, nicely, you choose) tells Yahoo what app is calling, and the IP address tells them what user. Build something cool, and it might generate hundreds of thousands of queries per day, but they are “charged” to the users, not to the developer. (Plus, it’s nice that there’s some good to my dialup existence: if I mess up building something, and go over the limit, I just need to reconnect.)
Maybe I’m missing something but limiting it by IP won’t change anything for webapps right? On a webapp, the server itself will typically be doing the querying (unless it’s using a Java applet/ActiveX to offload the work to the client’s browser) so won’t the webserver be limited to 5000 queries no matter how many end users it’s serving to?
I can see how this would be good for applications distributed and run from the end user’s PC though.
Well, I’m maybe a little too fond of the JavaScript/Ajax koolaid: for a server-side web app, yer screwed, stuck with just 5000 queries. It’s only if you make the client do the querying himself that it helps you. Probably I should have said ”browser app” rather than ”web app.”
I considered the xmlhttp (”ajax”–talk about a buzzword) and Flash based methods but I believe both of those can’t be used across domains (cross-site scripting limitations in most browsers and in the Flash plugin). So an xmlhttp or Flash based client interface would still have to go through an intermediary proxy/script within your domain/server to be able to talk to Yahoo or other external domains.
Grr. I’d save myself a lot of trouble and half-done things, if I could keep that stuck in my head. Too much time spent in the blessed
chromefreedom of extensions, I forget what you can’t get away with as lowlycontent.Limiting by IP sounds interesting at first - expecting that I’m seeing a wider use of open proxies in net use.
I can just see those scrapers now, bulking up the net with scraped results from Yahoo!, generated by programs calling the results from open proxies…
If I were clever enough to use such apps and profit from them, I’d say it was a dream come true. :)