[twitter-dev] Re: Formats for specifying annotations
Hi, 1) annotations shouldn't go via the GET method 2) if we to reserve characters then the pipe will be the one 3) allow multiple fields be combined automatically in the POST: annotations=domain|attribute1|value1 annotations=domain|attribute2|value1 4) encoding will be needed anyway, so why not go with the html-encoded POST On May 17, 8:15 pm, M. Edward (Ed) Borasky zn...@borasky- research.net wrote: On Monday, May 17, 2010 11:54:41 am @epc wrote: On May 17, 1:35 pm, Marcel Molina mar...@twitter.com wrote: Hey folks. I'd like to get your advice on the Annotations feature. Initial feedback: - I'd use a shorter variable name - Instead of using annotations once, allow repeated instances, each instance being a new annotation. | We want to allow you to use any arbitrary set of bytes for the type, | attribute names and attribute values of an annotation. I think you need to rethink this a little, at least declare a set of reserved characters. If you reserved : for example, you could reduce this down to: annotations=type:attribute:value (and yes, that'd mean that value could potentially be double-URL encoded/decoded if it has a : in it). So your example would be: twurl /1/statuses/update.xml -d status=Tweet with annotations! annotations=movie:title:Terminator 2annotations=movie:url=http://www.imdb.com/title/tt0103064/; I can't find anything to back up this but I believe that POST data is ordered. If so you could alternately do: annotation=type:attribute annotation_val = value twurl /1/statuses/update.xml \ -d status=Tweet with annotations! annotations=movie:titleannotations_value=Terminator 2annotations=movie:urlannotations_val=http://www.imdb.com/title/ tt0103064/ …which would avoid the need to double URL escape the annotation value. Can types use UTF8? Attributes? Values? Could I post annotations=映画:言語:英語 (or annotations==映画:言語=英語 in your original notation)? (I'm hoping that reads annotations=movie:language:english in Japanese. Apologies if it does not.) -- -ed costello What are the constraints? We're just talking about serializing / marshalling objects here, right? Documents? Trees? Or are we talking about an API for creating / modifying objects / documents / trees? As long as you're constrained to (UTF8) characters, I don't see any reason not to use JSON for objects. Now if *binary* is allowed, well ... seems like the only issue is little-endian vs. big-endian. ;-) -- M. Edward (Ed) Boraskyhttp://borasky-research.net/m-edward-ed-borasky/@znmeb A mathematician is a device for turning coffee into theorems. ~ Paul Erdős
[twitter-dev] Re: Introduce yourself!
Hi, my name's Alex and I'm a software developer in London. I just launched http://topytalk.com - a Twitter talk-oriented timeline On May 9, 3:06 pm, Georgios kapero...@gmail.com wrote: Hey all My name is Georgios (@georgioskap) and I have been developing on the Twitter API for the last 6 months. I just launched Favorious (http:// favorious.com) which is a website that aggregates Twitter favorites. Favorious aims to become the most comprehensive service based on Twitter favorites. You can view your favorites, your tweets that have been favorited by others and the most popular tweets and people overall. You can also explore popular tweets and favorite tweets of other users. Favorious is based on Twitter's REST API and has been developed on Ruby on Rails with MySQL database in the backend. Please give it a go and let me know if you have any feedback. The site development is not over yet as there is more functionality to be added in the future. Cheers Georgios
[twitter-dev] Re: What tools do you use?
Hi there, I'm Alex - the creator of http://topytalk.com - a Twitter talk-oriented timeline My software stack is that : Windows Web Server 2008, IIS 7, .Net/c# 2.5, MemCached, MySql 5.1, TweetSharp library In the future I want to migrate MemCached to Redis and use Linux/Mono for the services part. My choice of .Net/C# is purely because of my corporate development background. I'm considering moving into the Rails land, but not in the immediate future On Apr 27, 11:52 am, glenn gillen gl...@rubypond.com wrote: I use ruby, the twitter-text library, yajl for json processing, and mongodb for storage. -- Glennhttp://glenngillen.com/ -- Subscription settings:http://groups.google.com/group/twitter-development-talk/subscribe?hl=en
[twitter-dev] Help whitelist my ip
Hi, I'm waiting for a response regarding ip/account whitelisting for about a week now. I've first filled in the required form, then after several days emailed to a...@twitter.com, got a reply suggesting to fill the form again, did it 2 days ago. I run http://topytalk.com - a talk-oriented timeline and considering expanding my offering but am not able to do so without prior elevated access to the api. My account is @topytalk. Many thanks
[twitter-dev] read/write vs read-only oauth - usability issue
Hi, creating a consumer tokens for my app as of now implies that the app uses either read-only or read/write access to user accounts. But many apps (mine included - http:topytalk.com) have rich functionality in both read-only mode and read/write mode. the app could earn additional trust if the user was able to choose which kind of access she is comfortable with. I can create a second set of credentials for the app and offer that choice to the user. Question is: how Twitter treats this?