[twitter-dev] Re: Introducing the world's most extensive Twitter User Analyzer application
it will be added in the next few days, its in testing right now... Ruth On 5 מרץ, 15:57, Noah noah.cof...@gmail.com wrote: It would be much more useful if I could direct link to specific stats. I found some interesting things about a user and wanted to send them a link to a specific chart, but can't. If you added that and a tweet this stat feature, it would probably do wonders for your traffic. Nice site. On Mar 4, 7:32 am, Ruth yac...@gmail.com wrote: Hi all, Introducinghttp://TwitterAnalyzer.com, the world's most extensive Twitter User Analyzer app. After investing a lot of time and money, Twitter Analyzer is ready to be introduced to Twitter's development community... Twitter Analyzer is analyzing Twitter users with over 50 analyzing statistics represented by amazing graphical charts, and including features like: user usage stats, friends stats, friends density maps, followers growth rate and expectation, friends clustering by bio description or messages, active vs. inactive followers, what friends are writing about you?, who retweets your messages?, and many more... Please be aware that Twitter Analyzer is still in early Beta and bugs are part of development, so be delicate with your criticism. Your Bugs, Feature requests, or Comments are welcomed ... . you can send them to: 1. twitteranaly...@gmail.com, 2. follow us onhttp://twitter.com/ruth_z 3. follow us on twitter.com/tanalyzer (a new account) (we will follow you right back for an easier connection) 4. this discussion group Check it out, I promise that it will be fun... p.s. I would like to thank Alex Payne (@al3x) for helping us (by this discussion group and emails) in the process of writing the application. and cheers to Twitter for releasing a great Api. Thanks in advance, Ruth Zo,http://TwitterAnalyzer.com-הסתר טקסט מצוטט- -הראה טקסט מצוטט-
[twitter-dev] Twitter API Function
Hi All, Can any one explain me how I get the perticuler user block or not for another user by the help of API. Just like we get friendship exists or not. Thanks in advance Pawan Singh
[twitter-dev] Hello Everyone!!
Hi! I am new to this group so I wanted to take a second and introduce myself.
[twitter-dev] Re: friends_timeline.xml?count= not working? (JSON is)
Hi there, That looks like a bug, please open a Google Code issue (http://code.google.com/p/twitter-api/issues/entry ) and we'll take a look. — Matt On Mar 6, 2009, at 01:05 AM, TjL wrote: is 'count' not working for friends timeline if you use XML? I read this: http://apiwiki.twitter.com/REST+API+Documentation#friendstimeline URL: http://twitter.com/statuses/friends_timeline.format Formats: xml, json, rss, atom Method(s): GET API Limit: 1 per request Parameters: {{edit}} count. Optional. Specifies the number of statuses to retrieve. May not be greater than 200. Ex: http://twitter.com/statuses/friends_timeline.xml?count=5 and did this: curl -s --netrc 'http://twitter.com/statuses/friends_timeline.xml?count=50' /tmp/EVERYTHING.xml and got 20, not 50. curl -s --netrc 'http://twitter.com/statuses/friends_timeline.json?count=50' /tmp/EVERYTHING.json seemed to work, up to count=200
[twitter-dev] Re: friends_timeline.xml?count= not working? (JSON is)
I'm getting weird results. Sometimes I'm getting 'count' honored, and sometimes getting 20 regardless of what I ask for. Still checking to make sure it's not pilot error before I open a bug report. Tj On Mar 6, 2009, at 10:10 AM, Matt Sanford m...@twitter.com wrote: Hi there, That looks like a bug, please open a Google Code issue (http://code.google.com/p/twitter-api/issues/entry ) and we'll take a look. — Matt On Mar 6, 2009, at 01:05 AM, TjL wrote: is 'count' not working for friends timeline if you use XML? I read this: http://apiwiki.twitter.com/REST+API+Documentation#friendstimeline URL: http://twitter.com/statuses/friends_timeline.format Formats: xml, json, rss, atom Method(s): GET API Limit: 1 per request Parameters: {{edit}} count. Optional. Specifies the number of statuses to retrieve. May not be greater than 200. Ex: http://twitter.com/statuses/friends_timeline.xml?count=5 and did this: curl -s --netrc 'http://twitter.com/statuses/friends_timeline.xml?count=50' /tmp/EVERYTHING.xml and got 20, not 50. curl -s --netrc 'http://twitter.com/statuses/friends_timeline.json?count=50' /tmp/EVERYTHING.json seemed to work, up to count=200
[twitter-dev] Re: How often do users change their screen names?
User name changes are fairly rare, but a real concern nonetheless. A cache with an expiry of a day should be sufficient to guarantee most of your cache hits are successful and valid. Doug @dougw On Thu, Mar 5, 2009 at 1:26 PM, TjL luo...@gmail.com wrote: On Thu, Mar 5, 2009 at 11:45 AM, Nick Arnett nick.arn...@gmail.com wrote: Question for the folks at Twitter - any stats on how often people change their screen names? In another thread, we were talking about the problem of resolving IDs to names... I'm refreshing my user data for lots of users every few days, in large part to catch screen name changes. I could start keeping track of the changes, but I have not done so yet. Excellent question, I was just wondering that myself. Intuition suggests that users would rarely change their screen names, especially if they are active. Do you have any data to support this? Anecdotally, I've seen a few-but-rare name changes in the people I follow on Twitter. Come to think of it, an API call that would give us names changed since a certain date would be very useful for avoiding the need to check everybody. Even better, return friend or follower names changed since a date. Seems like the former would be easier to provide than the latter. It'd be nice if Twitter.com would redirect names (i.e. if you go to http://twitter.com/foo it would tell you/direct you to http://twitter.com/bar) for awhile too, but that's another issue and possibly more hassle than it's worth. TjL -- Doug Williams do...@igudo.com http://www.igudo.com
[twitter-dev] Re: friends_timeline.xml?count= not working? (JSON is)
BAH! It was indeed pilot error. Sorry for the noise. That's what I get for coding at 4am. TjL On Fri, Mar 6, 2009 at 10:19 AM, TjL luo...@gmail.com wrote: I'm getting weird results. Sometimes I'm getting 'count' honored, and sometimes getting 20 regardless of what I ask for. Still checking to make sure it's not pilot error before I open a bug report. Tj On Mar 6, 2009, at 10:10 AM, Matt Sanford m...@twitter.com wrote: Hi there, That looks like a bug, please open a Google Code issue (http://code.google.com/p/twitter-api/issues/entry) and we'll take a look. — Matt On Mar 6, 2009, at 01:05 AM, TjL wrote: is 'count' not working for friends timeline if you use XML? I read this: http://apiwiki.twitter.com/REST+API+Documentation#friendstimeline URL: http://twitter.com/statuses/friends_timeline.format Formats: xml, json, rss, atom Method(s): GET API Limit: 1 per request Parameters: {{edit}} count. Optional. Specifies the number of statuses to retrieve. May not be greater than 200. Ex: http://twitter.com/statuses/friends_timeline.xml?count=5 and did this: curl -s --netrc 'http://twitter.com/statuses/friends_timeline.xml?count=50' /tmp/EVERYTHING.xml and got 20, not 50. curl -s --netrc 'http://twitter.com/statuses/friends_timeline.json?count=50' /tmp/EVERYTHING.json seemed to work, up to count=200
[twitter-dev] OAuth Feature: oauth_access_type added
Hi there, This is mostly for the people in the OAuth closed beta, but that is rapidly coming to an end so other may want to read this as well. One of the major changes requested was the ability for one application to have both read and read+write users [1]. This was a fundamental shift in the security model but last night I deployed the end of it so it's now working. When sending a user to the authorize URL (/oauth/authorize) you can now include a parameter named oauth_access_type with a value of read or write, depending on which you need. If your application needs to change the access type for a user you can send them back again. You will probably want to make sure your app works correctly when people re-authorize this way, since you need to replace the tokens you have. We discussed a 3-button layout but decided that OAuth is confusing enough without moving choices onto the user. We also worked on a way for users to change the access type of a token but in the end every UI was confusing. Re-approval allows your app to handle the state change rather than sendinf them to the connections tab with instructions. Thanks; — Matt Sanford [1] - http://code.google.com/p/twitter-api/issues/detail?id=302
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
I agree with him. Search trends are not available in xml format. I will appreciate, if twitter can provide search trends in xml and so that i can update my social search engine ExploreWWW.com with search trends in real time. Thanks Burhan On Thu, Mar 5, 2009 at 11:09 PM, TjL luo...@gmail.com wrote: Specifically 1) There are WAY to many trending topic bots which fill search results with useless clutter 2) I'd love to see a trending topics list that does NOT include hash tags, you know, to find out what ordinary people are talking about :-) I know this is the wrong place for it (sorry) but I'm not sure where else to go. TjL -- Sincerely, Burhan Tanweer www.explorewww.com expl...@explorewww.com
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
Chad, In your experience, do trending bots have a disproportionate participation in the search results for trending topics? Have you done any analysis like that? Doug Williams @dougw On Fri, Mar 6, 2009 at 10:43 AM, Burhan TANWEER btanw...@gmail.com wrote: I agree with him. Search trends are not available in xml format. I will appreciate, if twitter can provide search trends in xml and so that i can update my social search engine ExploreWWW.com with search trends in real time. Thanks Burhan On Thu, Mar 5, 2009 at 11:09 PM, TjL luo...@gmail.com wrote: Specifically 1) There are WAY to many trending topic bots which fill search results with useless clutter 2) I'd love to see a trending topics list that does NOT include hash tags, you know, to find out what ordinary people are talking about :-) I know this is the wrong place for it (sorry) but I'm not sure where else to go. TjL -- Sincerely, Burhan Tanweer www.explorewww.com expl...@explorewww.com -- Doug Williams do...@igudo.com http://www.igudo.com
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
On Fri, Mar 6, 2009 at 11:11 AM, Doug Williams do...@igudo.com wrote: In your experience, do trending bots have a disproportionate participation in the search results for trending topics? Have you done any analysis like that? I'm not Chad :-) but if you click on any of the Trending Topics and watch for any length of time you'll see scads of trending topic bots popping up. I think the most I counted at one point was like 12 out of the top 20 results. It's insane. TjL
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
Well, it's kind of a weird feedback loop. Say you are following a trending bot (many many people do, a surprising number to me). As soon as you see a tweet from your favorite trending bot, you click the link and head over to see the results Well, all the other bots are tweeting at about the same time, so as soon as a new trend appears you get a dozen or so trend-bot tweets appearing in the results you just loaded up. I will admit this can be semi-annoying. Disproportionate? I guess it depends on how many results your browser loads by default. Mine is always set to 100, so I can scroll by the bots pretty quickly, but if people are only seeing 25 at a time, they'd have to click Next or Older to get past the bots. Like I said in my blog post, once you are actually searching for a trend, you don't need a dozen things telling you it's a trend again.. you're already there! Some bots are worse offenders than others and just spew all the trends every 5 minutes or retweet people (randomly it seems) that match the trend (not naming names, I'm sure you can figure them out). As a means of driving traffic they are very effective (at least the one I run seems to be). A little over 50% of the traffic to tweetgrid.com/search comes from links posted by my bot. I am not sure if the effectiveness can be attributed to the mere fact that the bot exists, or because it has some useful information attached (e.g. #trend has risen to the #3 trend! link). Very few of the bots seem to talk about the rank of the trend, but mine does, so it has some added value. I think this has helped my bot, and it also means that it gets retweeted quite a bit (another big surprise to me). In all honesty, I started my bot because one of my competitors convinced one of the existing trend bots to link to their site instead of search.twitter.com. I launched my bot in defense. A long, meandering answer to a short question. I am somewhat conflicted on the issue since I run one of these bots, but I will admit I find the greasemonkey script to blow them away quite nice. How's that for a definite maybe? -Chad On Fri, Mar 6, 2009 at 11:11 AM, Doug Williams do...@igudo.com wrote: Chad, In your experience, do trending bots have a disproportionate participation in the search results for trending topics? Have you done any analysis like that? Doug Williams @dougw - Show quoted text - On Fri, Mar 6, 2009 at 10:43 AM, Burhan TANWEER btanw...@gmail.com wrote: I agree with him. Search trends are not available in xml format. I will appreciate, if twitter can provide search trends in xml and so that i can update my social search engine ExploreWWW.com with search trends in real time. Thanks Burhan On Thu, Mar 5, 2009 at 11:09 PM, TjL luo...@gmail.com wrote: Specifically 1) There are WAY to many trending topic bots which fill search results with useless clutter 2) I'd love to see a trending topics list that does NOT include hash tags, you know, to find out what ordinary people are talking about :-) I know this is the wrong place for it (sorry) but I'm not sure where else to go. TjL -- Sincerely, Burhan Tanweer www.explorewww.com expl...@explorewww.com -- Doug Williams do...@igudo.com http://www.igudo.com
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
Hi there, We've talked about this among the search folks a few times. We exclude a bunch of bots and things from influencing trends but then they still get displayed. I just opened a ticket for someone to fix that so we can exclude the trend bots using a parameter or search operator. As far as if this is the correct place for search or not, I think it is. If other Twitter API developers disagree please let me know and I'll start a second group. From my perspective keeping up with one is easier for me to manage … and we're planning to merge the APIs in the next version of the API. Thanks; — Matt Sanford On Mar 6, 2009, at 08:25 AM, TjL wrote: On Fri, Mar 6, 2009 at 11:11 AM, Doug Williams do...@igudo.com wrote: In your experience, do trending bots have a disproportionate participation in the search results for trending topics? Have you done any analysis like that? I'm not Chad :-) but if you click on any of the Trending Topics and watch for any length of time you'll see scads of trending topic bots popping up. I think the most I counted at one point was like 12 out of the top 20 results. It's insane. TjL
[twitter-dev] Re: Twitter incorporated into a flash based website
yes, the twitter api http://apiwiki.twitter.com/REST+API+Documentation#update is designed to help 3rd party apps do things like this with ease. Keep in mind there is a 20,000 per hour limit. On Thu, Mar 5, 2009 at 3:40 PM, Abi abi.golestan...@googlemail.com wrote: Hi there, Currently designing a flash based website and trying to create this section where users are able to input messages, and then they are then stored and displayed in another section. I was wondering if it was possible to link twitter to this so that when a user inputs data into our site, it could then automatically update a twitter page? I am pretty new to twitter but was hoping I could incorporate the functionalities in. Hope someone can help!
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
On Fri, Mar 6, 2009 at 8:34 AM, Matt Sanford m...@twitter.com wrote: Hi there, We've talked about this among the search folks a few times. We exclude a bunch of bots and things from influencing trends but then they still get displayed. I just opened a ticket for someone to fix that so we can exclude the trend bots using a parameter or search operator. As far as if this is the correct place for search or not, I think it is. If other Twitter API developers disagree please let me know and I'll start a second group. From my perspective keeping up with one is easier for me to manage … and we're planning to merge the APIs in the next version of the API. It would be terrific if users could self-identify as bots and that data became part of the user profile. Although I'm sure that many people would not bother, we'd at least know that some of them definitively are bots. My bots self-identify in their description, which people seem to appreciate. Hmm. Maybe it would be far easier to simply encourage a hashtag in the description - how about #bot? That's something we could do now, without Twitter having to make any code changes. Thoughts? Nick
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
I am skeptical that bot devs, (outside of the integrious Jazzy Chad), will do anything to encourage segregation, as it would probably lead to a nuking list at some point. I would say this has to be done programatically, with a secret sauce that is known to twitter only. As search is more and more the golden goose apparent, gaming will be enemy number 1. On Fri, Mar 6, 2009 at 8:45 AM, Nick Arnett nick.arn...@gmail.com wrote: On Fri, Mar 6, 2009 at 8:34 AM, Matt Sanford m...@twitter.com wrote: Hi there, We've talked about this among the search folks a few times. We exclude a bunch of bots and things from influencing trends but then they still get displayed. I just opened a ticket for someone to fix that so we can exclude the trend bots using a parameter or search operator. As far as if this is the correct place for search or not, I think it is. If other Twitter API developers disagree please let me know and I'll start a second group. From my perspective keeping up with one is easier for me to manage … and we're planning to merge the APIs in the next version of the API. It would be terrific if users could self-identify as bots and that data became part of the user profile. Although I'm sure that many people would not bother, we'd at least know that some of them definitively are bots. My bots self-identify in their description, which people seem to appreciate. Hmm. Maybe it would be far easier to simply encourage a hashtag in the description - how about #bot? That's something we could do now, without Twitter having to make any code changes. Thoughts? Nick
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
I agree, most ppl probably won't abide by any guidelines that they have to 'voluntarily' follow in order to identify themselves at bots. It's pretty darn easy to tell if something is a trend bot or not... especially with the username :) Matt even said they've identified them (uh oh, i'm on some kind of twitter watchlist but who watches the watchlist?) If twitter themselves ever incorporate auto-updating search results like the special election pages, my bot and its links would pretty much be rendered useless D: -Chad On Fri, Mar 6, 2009 at 11:53 AM, Peter Denton petermden...@gmail.com wrote: I am skeptical that bot devs, (outside of the integrious Jazzy Chad), will do anything to encourage segregation, as it would probably lead to a nuking list at some point. I would say this has to be done programatically, with a secret sauce that is known to twitter only. As search is more and more the golden goose apparent, gaming will be enemy number 1. - Show quoted text - On Fri, Mar 6, 2009 at 8:45 AM, Nick Arnett nick.arn...@gmail.com wrote: On Fri, Mar 6, 2009 at 8:34 AM, Matt Sanford m...@twitter.com wrote: Hi there, We've talked about this among the search folks a few times. We exclude a bunch of bots and things from influencing trends but then they still get displayed. I just opened a ticket for someone to fix that so we can exclude the trend bots using a parameter or search operator. As far as if this is the correct place for search or not, I think it is. If other Twitter API developers disagree please let me know and I'll start a second group. From my perspective keeping up with one is easier for me to manage … and we're planning to merge the APIs in the next version of the API. It would be terrific if users could self-identify as bots and that data became part of the user profile. Although I'm sure that many people would not bother, we'd at least know that some of them definitively are bots. My bots self-identify in their description, which people seem to appreciate. Hmm. Maybe it would be far easier to simply encourage a hashtag in the description - how about #bot? That's something we could do now, without Twitter having to make any code changes. Thoughts? Nick
[twitter-dev] What is 140 characters?
Some discussion about this thread popped up on Twitter yesterday: http://groups.google.com/group/twitter-development-talk/browse_thread/ thread/44be91d5ec5850fa Alex states that it's 140 bytes per tweet. So, of course, Loren Brichter and I tried to prove that. With the following results: 1) 140 characters that including ones that include HTML entities: http://twitter.com/gnitset/status/1286202252 At the time of posting, this tweet showed up on the site and in feeds with all 140 characters. After a few hours, the was converted to lt;, increasing the count per character from one to four bytes and decreasing the tweet length from 140 characters to 69. (You can see this truncation at the end of the tweet: the is from lt;) Presumably, this happens as tweets in the memcache are written though to the backing store. I also see a lot of Twitter clients that don't realize how special the lt; and gt; entities are. It took me a LONG time to figure out what was going on here. 2) 140 Unicode _multi-byte_ characters: http://twitter.com/atebits/ status/1286199010 What's curious is that Loren's example with 140 characters uses the Unicode 27A1 glyph. It uses 3 bytes in UTF-8. Why didn't it get truncated? This seems to contradict Alex's statement in the thread mentioned above. As people start to use things like Emoji, tinyarro.ws and generally figure out that Unicode (UTF-8) is a valid type of data on Twitter, our clients should adapt and display more accurate characters remaining counts. I can count bytes instead of characters, but I'm not sure if I should or not. No one likes a truncated tweet: we need an explicit statement on how to count and submit multi-byte characters and entities. -ch
[twitter-dev] Re: Need a list of Friends -- followers/ids.xml isn't enough
On Thu, Mar 5, 2009 at 7:59 AM, Doug Williams do...@igudo.com wrote: The second, and less API intensive method to retrieve a list of all screen names is to page and parse through a user's friends with paginated calls to the statuses/friends method I've been trying this out for the last day or so... unfortunately, it turns out to be quite slow. The problem is that you have to slog through many, many statuses (100 at a time, of course) to get the screen_names. It looks to me as though the most efficient approach will be to use the statuses to get the most active users' names, abandon it when getting news statuses isn't yielding many new names, then use the show call to get the rest if you really want them. Nick
[twitter-dev] Re: Need a list of Friends -- followers/ids.xml isn't enough
Nick, Are you using a caching layer? Initialization of the cache will of course be slow since every user will need to be looked up with a users/show call, but the cache should eventually pay off after the most active users have been entered. Doug @dougw On Fri, Mar 6, 2009 at 12:46 PM, Nick Arnett nick.arn...@gmail.com wrote: On Thu, Mar 5, 2009 at 7:59 AM, Doug Williams do...@igudo.com wrote: The second, and less API intensive method to retrieve a list of all screen names is to page and parse through a user's friends with paginated calls to the statuses/friends method I've been trying this out for the last day or so... unfortunately, it turns out to be quite slow. The problem is that you have to slog through many, many statuses (100 at a time, of course) to get the screen_names. It looks to me as though the most efficient approach will be to use the statuses to get the most active users' names, abandon it when getting news statuses isn't yielding many new names, then use the show call to get the rest if you really want them. Nick -- Doug Williams do...@igudo.com http://www.igudo.com
[twitter-dev] Re: link for to be a follower
Devin, The friendship create method, just like the status create method, requires HTTP POST. Therefore they must be invoked through some sort of server-side script. However, since you are already sending users to Twitter to update their status, have you considered simply sending them to a user's profile (http://twitter.com/user) to follow and unfollow through the Twitter UI? Doug @dougw On Fri, Mar 6, 2009 at 2:35 PM, Peter Denton petermden...@gmail.com wrote: Hi Devin, There is no URL as such. It must be an explicit action, done though the api. http://apiwiki.twitter.com/REST+API+Documentation#create On Fri, Mar 6, 2009 at 11:22 AM, Devin dev...@gmail.com wrote: Hello all, I'm new to twitter and I hope someone can help me with this question. I have a page on my website which display a list of statuses. Each status has two buttons (Follow, Reply) next to it. The link to reply is https://twitter.com/home?in_reply_to=[user_name]in_reply_to_status_id=[status_id]status=%40[user_name]+ But I don't know what link to follow is. It seems like twitter use ajax to add follower. Is there a way to add follower similar way to add a reply. Thank you! -- Doug Williams do...@igudo.com http://www.igudo.com
[twitter-dev] Re: link for to be a follower
Yes, that is my second option and it seems like I'm going with that. On Mar 6, 11:44 am, Doug Williams do...@igudo.com wrote: Devin, The friendship create method, just like the status create method, requires HTTP POST. Therefore they must be invoked through some sort of server-side script. However, since you are already sending users to Twitter to update their status, have you considered simply sending them to a user's profile (http://twitter.com/user) to follow and unfollow through the Twitter UI? Doug @dougw On Fri, Mar 6, 2009 at 2:35 PM, Peter Denton petermden...@gmail.com wrote: Hi Devin, There is no URL as such. It must be an explicit action, done though the api. http://apiwiki.twitter.com/REST+API+Documentation#create On Fri, Mar 6, 2009 at 11:22 AM, Devin dev...@gmail.com wrote: Hello all, I'm new to twitter and I hope someone can help me with this question. I have a page on my website which display a list of statuses. Each status has two buttons (Follow, Reply) next to it. The link to reply is https://twitter.com/home?in_reply_to=[user_name]in_reply_to_status_id=[status_id]status=%40[user_name]+ But I don't know what link to follow is. It seems like twitter use ajax to add follower. Is there a way to add follower similar way to add a reply. Thank you! -- Doug Williams do...@igudo.comhttp://www.igudo.com
[twitter-dev] Survey: usage of since/If-Modified-Since vs since_id
We've recently discovered some performance problems in the use of the since/If-Modified-Since parameter and HTTP header, respectively. What was originally implemented as a way for developers to help us keep our servers happy by requesting only the tweets they need has, in fact, ended up costing as must computationally as gathering up an entire timeline of tweets. To that end, I'd like to find out how many of you really use this parameter or header, and why. If you'd be so kind, take a moment to fill out this brief, three question survey: http://www.surveymonkey.com/s.aspx?sm=z_2b612T_2bkqMCcSLKzXeZLIw_3d_3d Your feedback will determine how we proceed in regards to filtering timelines by timestamp. -- Alex Payne - API Lead, Twitter, Inc. http://twitter.com/al3x
[twitter-dev] Re: Need a list of Friends -- followers/ids.xml isn't enough
Nick, Have you looked into memcached [1]? Attribute-value pair caching is what it was designed to do. Perfect for the write-through cache that is needed here. It will also handle the pesky details like resolution expiry for you, too. If you would like help, ping me offline, I can get you started. [1] - http://www.danga.com/memcached/ Doug Williams @dougw On Fri, Mar 6, 2009 at 2:49 PM, Nick Arnett nick.arn...@gmail.com wrote: On Fri, Mar 6, 2009 at 11:15 AM, Doug Williams do...@igudo.com wrote: Nick, Are you using a caching layer? Initialization of the cache will of course be slow since every user will need to be looked up with a users/show call, but the cache should eventually pay off after the most active users have been entered. Yes, I'm putting them into a database... it is especially slow because I decided to capture the status data in addition to the user name, so there's a fair bit of overhead. I'm going to code up a light version that just grabs names and see how much faster it is. I'm fairly sure it's nothing people would want to wait for in real time... so as you say, building up the cache is the key. The speed of my current code is also limited by the database, which is CPU-bound. I'm using an old server that will benefit from a lot more memory, which I'm going to go and purchase this afternoon! I haven't offered any hard numbers because of these constraints... and I may be bandwidth-limited some of the time. I'll post some numbers at some point. I'm wondering how many unique screen names I'm getting, on average, per API call (it's less than 100 because there will be multiple statuses for some people) and what the average latency is. Those are things I can't control, so they ultimately will create the upper boundary. By the way, I'm not just looking at this as a problem. It's also an opportunity and I may have a source for the resources to address it. Nick -- Doug Williams do...@igudo.com http://www.igudo.com
[twitter-dev] Re: OT - where's the proper place to talk about search.twitter.com?
On Fri, Mar 6, 2009 at 2:56 PM, Nick Arnett nick.arn...@gmail.com wrote: I created @RoboTweeters this morning. I'll probably start feeding it screen names and ids of the ones I find, since that's quite simple. Nick Just make sure not to feed @RobotTweeters to itself... you may rip a blackhole in the tweet/space continuum!
[twitter-dev] Re: Search API and Feeds ... using a sinceID ... please?
some information inline … On Mar 6, 2009, at 01:25 PM, Scott C. Lemon wrote: I'm working on our site - http://www.TopFollowFriday.com - and am currently using the search API to search for the #followfriday hashtag. All is well, and it's working ... except ... The search feed only returns the last 15 items. There is a since_id, but that is useless as it only appears to work *within* the last 15. Then there is paging ... but I'm unclear exactly what good paging does? If I make a request, and get 15, and then make a request for page 2 ... what exactly does page 2 consist of? I'm not passing anything else by the page, and I'm guessing that you don't store server-side state information ... so page 2 doesn't really mean anything to me ... 1. It could be that page 2 will somehow be exactly the 16th-30th items in the list from when I made my first request ... but I somehow doubt that ... Actually, both the JSON and atom APIs return an attribute called 'next_url' which includes the page parameter as well as max_id so it works as you would expect. 2. I'm thinking that page 2 will maybe be the 16th-30th items in the new list that now includes all of the tweets that came in since my initial query. Bad. See above, about max_id. I'm caught with an API that I'm confused with ... how can I make my queries in a way that capture all of the tweets ... but not have to pound the server with requests? Can you do one of the following? 1. Straighten me out, and explain to me how I'm missing the boat here. I'm wishing that I was missing something, but it just seems this is how the API works. 2. Provide *more* than just 15 items ... 25 ... 50 ... or let me specify up to some maximum? I swear that I can write some good rate- limiting code that would automatically adjust the numbers to try and keep it optimal. you can do this, check out the rpp parameter at http://apiwiki.twitter.com/Search-API-Documentation 3. Implement the since_id so that it actually worked properly - not capped by 15 items - so that I could call the API at some reasonable rate and pass along some since_id and get all of the tweets since that tweet. It would even be ok to put a max size on that also ... it works as expect when you paginate. We can't support a call that returns millions of entries (since_id=0) so the max_id/page is the correct way to handle this. 4. Tell me about the items in the bug list that I need to vote for to make this happen ASAP. :-) You guys are awesome ... this has been a fun project, and the first friday was a great success ... I want to clean up my code though, and see if I can get the data that I want while being respectful of the API and rate limits ... :-) @Humancell
[twitter-dev] Re: RESTful API to unshorten URL's from twitter
Only 1 in 10 tweets tends to have a link - so you would need to receive 1000 tweets in an hour (on average) to run out of requests - but I guess that is a possability, we are at very early stages but we will look at upping to maybe 200 an hour and see how it goes. On Mar 5, 9:49 pm, Ed Finkler funkat...@gmail.com wrote: This is a cool feature, Nick. I appreciate you opening it up to the community. One concern I have, as a desktop app developer, is that a user could probably plow through 100 reqs/hr pretty easily if they're following 1000 or more users. Now one might say most people don't follow that many people, so you're an idiot, and they'd be right. I happen to follow that many, though, so I'd notice it 8). Also, people who are essentially edge cases when it comes to # of folks they follow also tend to bitch the most. Like me. Right now I'm doing url un-shortening in the client, but I would look seriously at using your service if the limits were a bit higher. -- Ed Finklerhttp://funkatron.com Twitter:@funkatron AIM: funka7ron ICQ: 3922133 XMPP:funkat...@gmail.com On Mar 4, 3:38 pm, Nick Halstead nickhalst...@gmail.com wrote: Today we launched an API for tweetmeme, for those who havent tried it, we aggregate all the twitter URL's to rank the most popular stories. Well the upside of this is that we have massive database of all the short URL's - and where they resolve to, included in this we also go and grab the page that it points at, and so we fetch the title, category of content, and a few other bits. We have tried to stick very closely to the RESTful + twitter style API The documentation is here -http://www.tweetmeme.com/apidoc.php An example of the url fetcher -http://api.tweetmeme.com/url_info?url=http://is.gd/lznv We also have two methods that let you fetch the most popular + the most recent stories. Would love to get feedback on what other data mining methods we could expose.
[twitter-dev] Re: RESTful API to unshorten URL's from twitter
Right now I prefer that the link is direct - if we wish to track (globally) then we could go via our own shortened URL which we will look into. On Mar 5, 9:34 am, Santosh Panda panda.sant...@gmail.com wrote: Hi Nick, Good stuff! We are ready to launch Tweetmeme popular tweets with Twitblogs. Btw the story URL points to the source of the story, I was under impression that user will be redirected to Tweetmeme to ready the full story? Thanks, Santoshwww.Twitblogs.com Twitter :-http://twitter.com/santoshpanda On Thu, Mar 5, 2009 at 8:30 AM, Nick Halstead nickhalst...@gmail.comwrote: We have every single short url + long url that is posted on twitter, we take the whole firehose, there are outages at times with the firehose so sometimes we still miss bits, but coverage is very high. On Mar 4, 8:51 pm, Nick Arnett nick.arn...@gmail.com wrote: On Wed, Mar 4, 2009 at 12:38 PM, Nick Halstead nickhalst...@gmail.com wrote: We have tried to stick very closely to the RESTful + twitter style API The documentation is here -http://www.tweetmeme.com/apidoc.php An example of the url fetcher - http://api.tweetmeme.com/url_info?url=http://is.gd/lznv We also have two methods that let you fetch the most popular + the most recent stories. Cool... I'm doing the same kind of thing, but instead of trying to do it comprehensively, I'm relying on predictive modeling and social network analysis to minimize the data. I'm able to identify most, if not all, of the popular URLs by making a system that is smart about who to track and how often. How comprehensive is your data? Are you trying to do the entire firehose? Would love to get feedback on what other data mining methods we could expose. By offering the API, you'll make it much easier for people to build on top of it. Maybe the best thing you could do is to make that service as complete and robust as possible. I think the future of things like this are in vertical-ization and personalization. Nick