Why two files?

Currently available are two versions API, but API 1.0 documentation is marked as «deprecated» and March 2013 will be shut down.
Now, you can not search without authentication, and all requests must be performed in the context of the account / token.
All restrictions on the speed of work in the context of the 15-minute window. Each requested method has its limitations. Limitations on the number of connections per minute, normal values ​​are 15 and 180. For example the method «search» limited to 180 connections in 15 minut.V transport protocol will be available only JSON, XML and longer podderzhivaetsya.Twitter informs your current speed and the limitations in HTTP headers X-Rate-the rate Limit.Ogranicheniya reconnecting. Too stubborn customers would ban system once activated by IP for REST API and Streaming API ogranicheniya.Dlya its permanent collection is recommended Streaming API, which allows the flow regime in one connection to get tweets for a specific tematike.Proverte your client to be compatible with API 1.1, since the protocol interaction is different from the API 1.0.Esli you lack speed collection of tweets, you can seek the services of intermediaries paid SERIS recommended by Twitter.
In addition, at the Streaming API:
-In your "pipe" the most can fly 3,000 tweets;
 -If the keyword generated more than 1% of total tweets, they are a random order will recline
 With one-IP can not be more than 10-15 simultaneous connections to Streaming API. Ie do not get to keep a lot of connections to the User Stream'am;
 Also, there are limitations to a specific Application'a.
All these restrictions are driving you to reaching out to the general posrednikami.V it turns out the guard.

fix the link colors in the default theme
fix preference changes not applied
fix old avatars appearing instead of the update ones
more logs to debug some weird cases

Please correct the page cache

Pay attention to the very bottom of the screen, which shows a translucent notification about the fact that it is now loaded.

Text version for clarification:
1) there are several columns: Feed, Mentions, Messages, News, News-it.
2) are on the tab "Tape" or "references" - all right;
3) Go to the "Messages" - there is a notice on loading the next column "News" (I did not even think to look at it, why did she loaded!?)
UPD: preloading begins if at least for a pixel has been scrolled to the current column. (Zachem!? If I'm not going to scroll further)
4) leaf through the allegedly downloaded the "News" - see the notice of the load of the same "News", which I am. (Reloading! OMG!) Should begin loading the column "News-it" (standing by comparison.)
5) flipping on "News-it" - again watch the download of the same worksheet columns on CAUTERETS am.
but that's not all ...
6) to leaf through the back twice already downloaded the "News" ... and what do we see? Right! next download the same list ... Well that "News-it" is not loaded.

Now simulate the situation:
I read the rightmost leaf. It occurs Mansion, or I suddenly wanted to write a reply ... many situations it may be because of what I need to return to the Home tab. made his case, going back again to the far right tab. How many downloads will occur for such movements?
Well if by vayfay or 3g, and if you use a normal 2g \ edge .. how many have to wait?

Or:
I read the list, accidentally pulled the application with the "back" instead of "house." Running back plume, I again have to wait for it to start one or the other list - or rather, everything!

Why not implement the most common cache of these sheets??! 11 Without this pre-load! Better yet, make it optional!

Solving this tank Plume You will have the best functionality not only for common tasks, but also the best he zaimeet implementation speaker lists. And get rid of unnecessary extravagance as an Internet traffic, and battery.

when you install the widget process flies

To what address to send error report?

Well, wait!