You should treat them as one source of activity with one user agent. Martin
Dne po 1. led 2018 9:00 uživatel Tal Barda <412tal...@gmail.com> napsal: > Thanks for your reply. > > I have a question regarding this part: > "The easiest and first thing you should do is make sure your application > sends an unique User-Agent header with its HTTP requests". > > Should I treat my HTTP requests as individual requests per player? > I mean, I wonder whether I should write in the User-Agent just the > application name and my email, or add to it also the specific username of > the request's sender. Eventually, each player sends the requests from its > own device, and I don't know if I should treat all of them like one source > of activity, or separated sources, so if there is an over extensive > activity from a specific user in my app you will be able to block it, and > not the whole app's activity. > > Thanks again > > > > > בתאריך 1 בינו׳ 2018 4:12, "Martin Urbanec" <martin.urba...@wikimedia.cz> > כתב: > > Hello Tal, >> >> here is how I understand the fair use rule. You are free to use the >> Wikipedia API as long as you care about other users of the API and do not >> want to get free profit of the service (by creating live mirrors that add >> adds to Wikipedia only etc). That includes identifying yourself (using >> descriptive User-Agent as Matmarex said), using most efecient algorythms >> that you can figure out, making one request at time, using maxlag in >> noninteractive tasks etc. etc. >> >> That's the reason why the information published aren't perfectly >> describing like you may issue 30 requests per minute and not more. There is >> simply *no way* how can be such numbers figured out. >> >> Thank you for caring about this and for your effort you put into the >> developing process! >> >> Happy new year, >> >> Martin Urbanec >> >> po 1. 1. 2018 v 2:47 odesílatel Bartosz Dziewoński <matma....@gmail.com> >> napsal: >> >>> The easiest and first thing you should do is make sure your application >>> sends an unique User-Agent header with its HTTP requests. In case your >>> application would generate more traffic than Wikimedia are willing to >>> support, this will let the Operations team contact you, or (in the worst >>> and unlikely case) easily block your requests without affecting anyone else. >>> >>> https://www.mediawiki.org/wiki/API:Etiquette#User-Agent_header >>> >>> Thank you for caring about this! :) >>> >>> On Sun, Dec 31, 2017 at 9:28 PM, Tal Barda <412tal...@gmail.com> wrote: >>> >>>> Hello! >>>> I hope this is the correct place to ask this question. >>>> My name is Tal, and I'm developing my own version of the Wiki Game >>>> <https://en.wikipedia.org/wiki/Wikipedia:Wiki_Game> (or Wiki Race, >>>> some may say) for Android devices. There is a great multiplayer >>>> version <https://thewikigame.com/>of it for the iPhone and Web, but I >>>> think there should be a worthy version to the Android. >>>> My version is still in progress, and I put a lot of effort into it, >>>> including the gameplay experience (single-player and multiplayer) and the >>>> design. >>>> _________________________ >>>> >>>> I'm writing this email because I'd like to consult with some wiki >>>> experts about one of the main things which on my mind, and it's my usage of >>>> the Mediawiki API. I'd like to launch the game and upload it to the Play >>>> Store in a few months, but I can't keep working on it while knowing that I >>>> might put my efforts in danger if I will do so, because of unfair usage of >>>> the Wiki API. >>>> >>>> Well, the game is based on generating two random articles: the first >>>> article (the starting point) and the target article (the end point) which >>>> the user should reach to using only the hyperlinks inside the articles. >>>> >>>> The thing is, that unlike the currently exist Wiki Game >>>> <https://thewikigame.com/> version which is played by many users >>>> together at the same time (by joining "game rooms"), or using pre-generated >>>> articles, - I'm generating the articles live for each user individually. >>>> Each user can play in single-player mode and generate his own pair of >>>> articles, his own game, a thing which is gonna keep the game infinite for >>>> all of the players. >>>> >>>> My generator generates two articles before each game and the user has >>>> 15 seconds to decide whether he would like to start the game or not. After >>>> the 15 seconds are up the generator is generating a new pair of articles. >>>> While the time is running, the user may refresh the articles manually and >>>> ask to generate a new pair. >>>> >>>> So why am I worried about my app's usage of the API? >>>> Well, not all of the target articles are reachable through the >>>> hyperlinks in the mobile version of Wikipedia, because the templates view >>>> in the mobile version is very limited and there are articles without >>>> references from other articles, and I don't want the users to get stuck. >>>> >>>> In order to overcome this , and make sure that the target article is >>>> actually reachable, I need to make about 5 requests to the REST API each >>>> articles generation (one request after another, not together). The >>>> algorithm was made after a very long time of research by me, and I think >>>> its efficient for the purpose. >>>> >>>> So, I assume that the average player is gonna play about an hour in a >>>> day, so he is gonna refresh his pair about 15 times, which are 75 requests >>>> from the API in a given day for a user. Let's say (hopefully and >>>> hypothetically) the game goes pretty viral - well, there are gonna be a lot >>>> of requests to the API at the same time, and this makes me unsure about my >>>> way of action. >>>> >>>> I don't mind sending my random articles generator code (as long it is >>>> not copied :P ) for a review, and work together with somebody to make the >>>> game better and fair, and I'm also willing to donate a part of the game >>>> incomes (if there will be any) to the Wikimedia Foundation, but there is >>>> one thing I'm not gonna do, and it's launching the game before I'm making >>>> sure I do a fair use of the API. I know Wikipedia is taking care millions >>>> and billions of requests everyday, and still - I can't move on without >>>> knowing everything is fine. >>>> >>>> This information <https://www.mediawiki.org/wiki/API:Etiquette> about >>>> the usage limits is not very explicit, and it doesn't put the line between >>>> good and bad, and maybe I'm using it badly. I will accept every answer >>>> respectfully. >>>> >>>> Thanks in advance, >>>> Tal >>>> >>>> _______________________________________________ >>>> Mediawiki-api mailing list >>>> Mediawiki-api@lists.wikimedia.org >>>> https://lists.wikimedia.org/mailman/listinfo/mediawiki-api >>>> >>>> >>> >>> >>> -- >>> Matma Rex >>> _______________________________________________ >>> Mediawiki-api mailing list >>> Mediawiki-api@lists.wikimedia.org >>> https://lists.wikimedia.org/mailman/listinfo/mediawiki-api >>> >> >> _______________________________________________ >> Mediawiki-api mailing list >> Mediawiki-api@lists.wikimedia.org >> https://lists.wikimedia.org/mailman/listinfo/mediawiki-api >> >> _______________________________________________ > Mediawiki-api mailing list > Mediawiki-api@lists.wikimedia.org > https://lists.wikimedia.org/mailman/listinfo/mediawiki-api >
_______________________________________________ Mediawiki-api mailing list Mediawiki-api@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mediawiki-api
