Twitter Annotation Proposal – Image
This is a proposal for a twitter annotation that I hope to be blessed by both the development community and twitter itself.
Image Namespace
Images are something commonly shared throughout twitter. Something that would be fantastically idealistic would be embedding image data into twitter annotations.
Due to constraints, images should be resized to a lower-quality “preview”. Possibly VGA or so? A link to a privately hosted image should also be available.
Using the current preview information supplied by the twitter development team, I imagine that it would go something like this:
"image": { "preview": "data here.. base64 etc.", "preview_encode": "base64", // Or whatever else floats your boat. "bin" or "binary" for raw data. "full_src": "http://example.org/link/to/image.jpg" }
Any additional thoughts on other information that should be included? Maybe a title and description, or is that not needed? Please, leave all of your thoughts and ideas in the comments!
Why SuperTweets won’t kill URL Shorteners
Now, the title of this blog post makes it sound like I’m going to write an essay about why SuperTweets (and the probability of them having a URL metadata for tweets) will not be killing URL Shorteners like j.mp and bit.ly anytime soon.
1) URL Shorteners Keep You Safe
One of the things URL shorteners do now is they keep you safe from malicious websites. You can preview the site you are visiting, and if its determined to be delicious the short URL will either be deleted or blocked or a warning will be shown, letting you know it is no longer safe to visit that URL.
2) URL Shorteners are Branded
All I really have to say here is “Bit.ly Pro” URL Shorteners now have custom branding, so it makes it even easier to send people to your website or promote your brand on twitter by including the link in the text.
3) URL Shorteners are Easy to Remember
When they are used correctly. Services like bit.ly allow you to give a custom name to your short link. This is especially useful in media such as Television or Print: http://j.mp/cnn-transgender is much easier to remember and type than www.cnn.com/2010/LIVING/04/14/transgender.irpt/index.html?hpt=C1
With even just these three simple reasons in mind, it is very clear to me that URL shorteners will not be dying anytime soon, no matter how much metadata you can attach to tweets. I’m not even sure they’re bad for the internet, anymore.
Sneaky Twitter, or My Imagination?
Fairly surprisingly, twitter went down today (for what appeared to be no rhyme or reason), but something much more surprising happened when it did. A tweet I posted (from my own, self-coded twitter client that has no “store tweets until twitter is back up” system) was there on the site when it came back up (viewable in the image above).
I’m not entirely sure if this is what I want to think it is, but could it be that twitter has a secondary service that stores tweets that come in via the API until the system is back up, or something? I didn’t think the update had posted at all, as the API didn’t return any of the proper data (nor an error).
So, is this just a coincidence, or does twitter have a system in store to keep the stream flowing even when its not?
Rumor: Twitter Close to Unveiling Contributions
Do I know for sure? Absolutely not. Do I have inside information? Absolutely not. Was I randomly observant one night and saw something that seemed to push me to think in this direction? Yes.
Back in December, Twitter blogged about its “feature test with businesses” of a new Contribution API they were adding in to Twitter. One that would allow companies to give users permission to tweet on behalf of the main account, and still attribute that post to the user who wrote it.
If you wanted to see this in action, all you would have to do is look at the main Twitter account, where almost every post and retweet is attributed to one of the employees.
They originally announced that this feature would improve usage of applications like @CoTweet and @HootSuite. But if you look at their timelines, you see very little Contribution API dabble – until recently, that is.
Looking through CoTweet’s posts all the way back through December, none of them have contribution metadata – except for the latest two on February 18th and 19th. HootSuite has only one, posted on February 18th (none earlier, and none later). This brings to mind: Twitter must have recently been rolling out (or testing) their Contribution Feature – or are we really supposed to expect this to be coincidence?
And all of this with the Twitter Developer Meetup Scheduled for Monday, March 1st 2009.
All I’m saying is, I think they’ve gotten much closer to rolling out Contributions. Maybe they’ll announce it at this small developer meetup, since everybody already knows about it. Then again, maybe they won’t.
Settings Needed on Twitter
- The ability to block someone from showing up in your timeline, but not block them from following you.
- The ability to allow users to send you DMs, even if they’re not following you (and block them from doing so on an individual basis).
- The ability to see all @replies from your friends – I WANT THIS BACK
That is all.
Twitter PHP Library
Well, being a PHP developer I quickly got excessively tired of looking for the proper Twitter Library with all the features that returned good results, and I got tired of looking for proper libraries and classes that could interface with OAuth.
Then, I found a relatively new (albeit “Beta”) Twitter OAuth Library, but it still required more steps than I really wanted to do just to get user authentication in a program. So I decided to finally piece everything together and write Twitter API Library.
This library basically does everything you would expect it to do, except for Unauthenticated Requests and the Search API… I haven’t gotten around to coding that in yet, as it was not the primary focus of this library.
The library is not thoroughly tested, but theoretically shouldn’t encounter any errors. I would be very grateful to anyone who’ll test it for me. (I can’t work out the bugs without testers, so please, test!).
SimpleTwitter Update
I’ve pushed a simple update to the SimpleTwitter JavaScript. Change log (if its even needed) below.
- Support for Twitter Lists
- Uses the same format as twitter, for example @navarr/shogi
- Support for Turled Profiles
- Using the variable &turl=true will link @replies to turled beta-profiles instead of twitter profiles.
- Surprisingly, this doesn’t break in co-ordination with lists. It just doesn’t display the list.
- Uses Search API
- We moved to the search API for retrieving tweets, as well as pulling a JSON file instead of XML. This MAY BREAK your current implementation (though it should not). If it does, you can use the older script by pulling old.php from the server instead of index.php
Yeah, that’s it. Was thinking about making the default for &turl being true instead of false, but figured that’d probably annoy more people than it was worth (at least until I get turled to be of more use as a twitter replacement).
Twitter gets new Following Block
As seen in the image above, Twitter got a new following block to replace the small button they had previously, mimicking the same look and feed on the following and followers pages.
JavaScript: Changes to the Simple Twitter Feed
Recently, I had resumed working on the Simple Twitter Feed, and decided to make some improvements after reading Dave Winer's "What am I up to?" post.
So, the Simple Twitter Feed has one fix, and four additional optional parameters.
| start | STR - Symbol or Text to look for at the beginning of a post |
| time | BOOL - Whether or not to display the "about x minutes ago" |
| client | BOOL - Whether or not to display the "from x" |
| br | BOOL - Whether or not to place a break between the tweet and the extra data |
The simple fix is that the count parameter would check the last (x) amount of tweets, whereas now it displays up to (x) amount of tweets.