API Changes for April 1, 2009

7 views
Skip to first unread message

Alex Payne

unread,
Apr 1, 2009, 8:34:34 PM4/1/09
to Twitter Development Talk, twitter-ap...@googlegroups.com
(Not an April Fool, we promise. We don't enjoy "humor".)

* Feature (REST API): We now return the same representation of User
objects throughout the API. This representation contains all of the
attributes we make available via the API.

A bit more about this change:

Previously, these "full" User objects were only available via the
/users/show and /account/verify_credentials methods. If your
application has been making requests to these methods just to get
extra User attributes, you no longer need to do so. We've had many,
many requests for these extra attributes to be available everywhere,
so we hope to see you all making use of them!

Please note that this new extended view of User objects may not appear
for all users immediately. As cache expiry occurs for users in our
system, the extra attributes will show up. Don't be surprised if this
takes multiple days for inactive users.

Please also note that if your application is operating in a highly
bandwidth-constrained environment, you may want to proxy requests to
strip out attributes that aren't relevant to your client. The
additional bytes over the wire should not impact the vast majority of
platforms, in our estimates.

As always, you can keep up with these changes at http://bit.ly/api_changelog.

--
Alex Payne - API Lead, Twitter, Inc.
http://twitter.com/al3x

Jesse Stay

unread,
Apr 1, 2009, 8:54:40 PM4/1/09
to Twitter Development Talk
Does this include the Social Graph API methods?

Jesse

Zac Bowling

unread,
Apr 1, 2009, 8:57:59 PM4/1/09
to twitter-deve...@googlegroups.com
Fantastic news.

Are the direct message's recipient and sender objects updated as well?

Zac Bowling
http://twitter.com/zbowling

zbowling

unread,
Apr 1, 2009, 9:00:27 PM4/1/09
to Twitter Development Talk
Those apis only return an array of user ids, not user objects. I doubt
they would change that since it functionally break that api.

On Apr 1, 5:54 pm, Jesse Stay <jesses...@gmail.com> wrote:
> Does this include the Social Graph API methods?
> Jesse
>

Alex Payne

unread,
Apr 1, 2009, 9:01:56 PM4/1/09
to twitter-deve...@googlegroups.com
No, only methods that previously returned User objects.

Alex Payne

unread,
Apr 1, 2009, 9:02:38 PM4/1/09
to twitter-deve...@googlegroups.com
Egads! No, they aren't, but that's a quick fix. Will have it out
tomorrow, hopefully, Monday at the latest.

Damon P. Cortesi

unread,
Apr 1, 2009, 10:23:22 PM4/1/09
to Twitter Development Talk
I'm not sure if it's related to this push, but I've started to get
several user objects back with no statuses_updates.

This is a somewhat blocking bug for TweetStats as I try to verify they
have tweets while verifying the account. Though I can just try to
enumerate through and will probably have to do an emergency update to
do so.

here's a sample user object I'm getting back:

<?xml version="1.0" encoding="UTF-8"?>
<user>
<id>20176857</id>
<name>Jonny Beasley</name>
<screen_name>ccfcrule</screen_name>
<location></location>
<description></description>
<profile_image_url>http://s3.amazonaws.com/twitter_production/
profile_images/78298843/Me_With_Cap_normal.jpg</profile_image_url>
<url></url>
<protected>false</protected>
<followers_count>25</followers_count>
<status>
<created_at>Sat Mar 28 20:00:26 +0000 2009</created_at>
<id>1408524555</id>
<text>Checking out my TweetStats! http://tweetstats.com/graphs/ccfcrule</text>
<source>&lt;a href=&quot;http://
tweetstats.com&quot;&gt;TweetStats&lt;/a&gt;</source>
<truncated>false</truncated>
<in_reply_to_status_id></in_reply_to_status_id>
<in_reply_to_user_id></in_reply_to_user_id>
<favorited>false</favorited>
<in_reply_to_screen_name></in_reply_to_screen_name>
</status>
</user>

dpc
> As always, you can keep up with these changes athttp://bit.ly/api_changelog.

Damon P. Cortesi

unread,
Apr 1, 2009, 10:33:34 PM4/1/09
to Twitter Development Talk
Sorry, I meant statuses_count.

On Apr 1, 7:23 pm, "Damon P. Cortesi" <d.lifehac...@gmail.com> wrote:
> I'm not sure if it's related to this push, but I've started to get
> several user objects back with no statuses_updates.
>
> This is a somewhat blocking bug for TweetStats as I try to verify they
> have tweets while verifying the account. Though I can just try to
> enumerate through and will probably have to do an emergency update to
> do so.
>
> here's a sample user object I'm getting back:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <user>
>   <id>20176857</id>
>   <name>Jonny Beasley</name>
>   <screen_name>ccfcrule</screen_name>
>   <location></location>
>   <description></description>
>   <profile_image_url>http://s3.amazonaws.com/twitter_production/
> profile_images/78298843/Me_With_Cap_normal.jpg</profile_image_url>
>   <url></url>
>   <protected>false</protected>
>   <followers_count>25</followers_count>
>   <status>
>     <created_at>Sat Mar 28 20:00:26 +0000 2009</created_at>
>     <id>1408524555</id>
>     <text>Checking out my TweetStats!http://tweetstats.com/graphs/ccfcrule</text>

Damon Clinkscales

unread,
Apr 1, 2009, 10:55:36 PM4/1/09
to twitter-deve...@googlegroups.com
On Wed, Apr 1, 2009 at 9:33 PM, Damon P. Cortesi <d.life...@gmail.com> wrote:

>> I'm not sure if it's related to this push, but I've started to get
>> several user objects back with no statuses_updates.

>
> Sorry, I meant statuses_count.
>

nor favourites_count, nor friends_count...

here's my record, in full (from /users/show/damon.xml)

$ curl http://twitter.com/users/show/damon.xml


<?xml version="1.0" encoding="UTF-8"?>
<user>

<id>756264</id>
<name>Damon Clinkscales</name>
<screen_name>damon</screen_name>
<location>Austin, TX</location>
<description>I'm a shepherd. Software engineer at VitalSource and
leader of Austin On Rails. I also build apps like
http://snaptweet.com and http://doesfollow.com. </description>
<profile_image_url>https://s3.amazonaws.com/twitter_production/profile_images/73617066/me_smile_normal.jpg</profile_image_url>
<url>http://damonc.com</url>
<protected>false</protected>
<followers_count>974</followers_count>
<status>
<created_at>Wed Apr 01 21:16:29 +0000 2009</created_at>
<id>1434142066</id>
<text>RT @ev on April Fools - There is no Twitter Pro</text>
<source>&lt;a
href="http://iconfactory.com/software/twitterrific"&gt;twitterrific&lt;/a&gt;</source>


<truncated>false</truncated>
<in_reply_to_status_id></in_reply_to_status_id>
<in_reply_to_user_id></in_reply_to_user_id>
<favorited>false</favorited>
<in_reply_to_screen_name></in_reply_to_screen_name>
</status>
</user>

April Fools?

-damon

Adrian

unread,
Apr 2, 2009, 1:10:28 AM4/2/09
to Twitter Development Talk
As of right right now: http://twitter.com/users/show/bob.xml
has about twice the amount of information as say:
http://twitter.com/users/show/WeezerOfficial.xml
> As always, you can keep up with these changes athttp://bit.ly/api_changelog.

Chad Etzel

unread,
Apr 2, 2009, 1:15:28 AM4/2/09
to twitter-deve...@googlegroups.com
On Thu, Apr 2, 2009 at 1:10 AM, Adrian <spiri...@gmail.com> wrote:
>
> As of right right now: http://twitter.com/users/show/bob.xml
> has about twice the amount of information as say:
> http://twitter.com/users/show/WeezerOfficial.xml
>

FTA:


"Please note that this new extended view of User objects may not appear
for all users immediately. As cache expiry occurs for users in our
system, the extra attributes will show up. Don't be surprised if this
takes multiple days for inactive users."

-Chad

Alex Payne

unread,
Apr 2, 2009, 1:23:12 AM4/2/09
to twitter-deve...@googlegroups.com
From what method calls?

Damon P. Cortesi

unread,
Apr 2, 2009, 2:09:11 AM4/2/09
to Twitter Development Talk
Simply a user show.

http://twitter.com/users/show/ccfcrule.xml is still returning the same
data as earlier.

On Apr 1, 10:23 pm, Alex Payne <a...@twitter.com> wrote:
> From what method calls?
>
> On Wed, Apr 1, 2009 at 19:23, Damon P. Cortesi <d.lifehac...@gmail.com> wrote:
>
>
>
>
>
> > I'm not sure if it's related to this push, but I've started to get
> > several user objects back with no statuses_updates.
>
> > This is a somewhat blocking bug for TweetStats as I try to verify they
> > have tweets while verifying the account. Though I can just try to
> > enumerate through and will probably have to do an emergency update to
> > do so.
>
> > here's a sample user object I'm getting back:
>
> > <?xml version="1.0" encoding="UTF-8"?>
> > <user>
> >  <id>20176857</id>
> >  <name>Jonny Beasley</name>
> >  <screen_name>ccfcrule</screen_name>
> >  <location></location>
> >  <description></description>
> >  <profile_image_url>http://s3.amazonaws.com/twitter_production/
> > profile_images/78298843/Me_With_Cap_normal.jpg</profile_image_url>
> >  <url></url>
> >  <protected>false</protected>
> >  <followers_count>25</followers_count>
> >  <status>
> >    <created_at>Sat Mar 28 20:00:26 +0000 2009</created_at>
> >    <id>1408524555</id>
> >    <text>Checking out my TweetStats!http://tweetstats.com/graphs/ccfcrule</text>

Damon P. Cortesi

unread,
Apr 2, 2009, 2:10:20 AM4/2/09
to Twitter Development Talk


On Apr 1, 10:15 pm, Chad Etzel <jazzyc...@gmail.com> wrote:
> On Thu, Apr 2, 2009 at 1:10 AM, Adrian <spiritpo...@gmail.com> wrote:
>
> > As of right right now:http://twitter.com/users/show/bob.xml
> > has about twice the amount of information as say:
> >http://twitter.com/users/show/WeezerOfficial.xml
>
> FTA:
> "Please note that this new extended view of User objects may not appear
> for all users immediately. As cache expiry occurs for users in our
> system, the extra attributes will show up. Don't be surprised if this
> takes multiple days for inactive users."
>
> -Chad

Right, but the odd thing is this is data that was showing up normally
before. When this change went into effect, I started getting a lot of
users with no statuses_count showing whereas prior to the change that
was a reliable property of the user object.

Jochem Prins

unread,
Apr 2, 2009, 5:07:48 AM4/2/09
to Twitter Development Talk
I'm having the same issue here. The /users/show method is now
returning the basic user element for some users. Weird.

Jochem

binit

unread,
Apr 2, 2009, 7:04:40 AM4/2/09
to Twitter Development Talk
I am facing the same issue too with the API "http://twitter.com/users/
show/id.xml".
Now the big problem is that in some cases following fields are
missing:
friends_count
created_at
favourites_count
utc_offset
time_zone
profile_background_image_url
statuses_count

Due to this my application buzzom.com has broken down.
Is there an alternate way of getting these fields for a particular
user?
Please suggest.

Regards,
Binit
Message has been deleted

Dossy Shiobara

unread,
Apr 2, 2009, 11:11:26 AM4/2/09
to twitter-deve...@googlegroups.com
On 4/1/09 8:34 PM, Alex Payne wrote:
> * Feature (REST API): We now return the same representation of User
> objects throughout the API. This representation contains all of the
> attributes we make available via the API.

Sweet jeebus! It's Christmas, all over again!

Thanks, Twitter gnomes.

--
Dossy Shiobara | do...@panoptic.com | http://dossy.org/
Panoptic Computer Network | http://panoptic.com/
"He realized the fastest way to change is to laugh at your own
folly -- then you can let go and quickly move on." (p. 70)

Damon P. Cortesi

unread,
Apr 2, 2009, 11:35:44 AM4/2/09
to Twitter Development Talk
On Apr 2, 4:04 am, binit <binit.th...@gmail.com> wrote:
>
> Due to this my application buzzom.com has broken down.
> Is there an alternate way of getting these fields for a particular
> user?
> Please suggest.
>

In some cases, even though /users/show does not have the full user
object, /statuses/user_timeline will. I've been able to use this as a
workaround for some accounts when the fields I need don't exist, but
it doesn't work in every case.

Doug Williams

unread,
Apr 2, 2009, 1:39:29 PM4/2/09
to Twitter Development Talk
Issue 409 [1] was opened to track the loss of fields many of you are
reporting.

1. http://code.google.com/p/twitter-api/issues/detail?id=409

Doug Williams
Twitter API Support
http://twitter.com/dougw

binit

unread,
Apr 2, 2009, 1:40:47 PM4/2/09
to Twitter Development Talk
Thanks a lot, Damon.
It works!
You saved my day :)

Jesse Stay

unread,
Apr 2, 2009, 2:17:02 PM4/2/09
to twitter-deve...@googlegroups.com
I'm seeing similar from our users on SocialToo.

Jesse

Doug Williams

unread,
Apr 2, 2009, 2:18:02 PM4/2/09
to twitter-deve...@googlegroups.com
The loss of fields (issue 409) is the result of the caching issue described above. You should see these fields reappear as the cache user objects expire.


Doug Williams
Twitter API Support
http://twitter.com/dougw


Jeffrey Greenberg

unread,
Apr 2, 2009, 3:27:23 PM4/2/09
to Twitter Development Talk
Doug,
Grumble: just to say it, this wasn't handled well at all. The fact
that this field disappears whether due to caching or through a coding
error has the same result of completely breaking my app.

How long will it take for this issue to clear up? Days? How many
exactly? and after X days will further requests be populated
correctly?

thx,
jeffrey
http://www.tweettronics.com

Doug Williams

unread,
Apr 2, 2009, 8:36:17 PM4/2/09
to twitter-deve...@googlegroups.com
Jeffery,
This is valid criticism. This bug came as a surprise to us as well. We otherwise would have given developers fair warning. Unfortunately there is no easy fix, and like a bad heart-break, time may be the only answer.

In short, the problem is with the user data cache. To get the extended information into that cache, the user object must either expire or be invalidated through some user initiated update. The expiry on the cache is rather long and you will find that inactive accounts will have abbreviated data for up to 2 weeks.

This is obviously sub-optimal, as Matt would say.


Doug Williams
Twitter API Support
http://twitter.com/dougw


Martin Dufort

unread,
Apr 5, 2009, 9:58:16 AM4/5/09
to Twitter Development Talk
I'm seeing inconsitent user attributes within the *same* request for
the *same* user. One result has full attributes disclosure, and the
other one has not.

I've updated Issue 409 with my results.
Martin

On Apr 2, 8:36 pm, Doug Williams <d...@twitter.com> wrote:
> Jeffery,
> This is valid criticism. This bug came as a surprise to us as well. We
> otherwise would have given developers fair warning. Unfortunately there is
> no easy fix, and like a bad heart-break, time may be the only answer.
>
> In short, the problem is with the user data cache. To get the extended
> information into that cache, the user object must either expire or be
> invalidated through some user initiated update. The expiry on the cache is
> rather long and you will find that inactive accounts will have abbreviated
> data for up to 2 weeks.
>
> This is obviously sub-optimal, as Matt would say.
>
> Doug Williams
> Twitter API Supporthttp://twitter.com/dougw

Marco Kaiser

unread,
Apr 9, 2009, 9:47:10 AM4/9/09
to twitter-deve...@googlegroups.com
I recognize an odd behavior for the "following" property of embedded user object in the friends timeline (XML format). As I understand from the API docs, "following" should indicate whether the authenticating user is following the returned user.

Obviously, all tweets returned on the /statuses/friends_timeline.xml endpoint should come from users I am following (and I manually verified that for some samples), but still some of them have <following>false</following> set. It seems to be at least consistant in a way that the same user always has the same value set for following in a result set - it just is either always wrong or always right.

Is that a known issue?

Thanks,
Marco

2009/4/5 Martin Dufort <martin...@gmail.com>

Marco Kaiser

unread,
Apr 21, 2009, 7:22:44 AM4/21/09
to twitter-deve...@googlegroups.com
Doug? Anyone?

Thanks,
Marco

2009/4/9 Marco Kaiser <kaiser...@gmail.com>

Matt Sanford

unread,
Apr 21, 2009, 10:25:46 AM4/21/09
to twitter-deve...@googlegroups.com
Hi there,

  This sounds like it is related to Google Code issue 474 [1]. Please visit Google Code and click on the little star next to the title to get updates when that issue is updated.

Thanks;
  — Matt Sanford

Marco Kaiser

unread,
Apr 21, 2009, 10:45:29 AM4/21/09
to twitter-deve...@googlegroups.com
Thanks, Matt!

I've added the information from my mail as a comment to that issue.

Marco

2009/4/21 Matt Sanford <ma...@twitter.com>
Reply all
Reply to author
Forward
0 new messages