Re: [twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[M. Edward (Ed) Borasky]

I have over 20,000 tweets under my belt. Even whitelisted at 20,000  
API calls per second ...


I'm thinking Twitter could charge a dollar a thousand for a full  
backdump like that - I'd pay $20 to get a full CSV file of them.

--
M. Edward (Ed) Borasky
borasky-research.net/m-edward-ed-borasky/

"A mathematician is a device for turning coffee into theorems." ~ Paul Erd?s


Quoting Raffi Krikorian :


it is on our roadmap to be able to retrieve all your tweets.

just to be precise, however, all your tweets are retrievable, if you know
their IDs -- you can then retrieve them using statuses/show endpoint.  what
we don't maintain, however, is retrieval of more than 3200 tweets via the
timeline calls.  this is a limitation of our timeline and storage
architecture, but it is actively being addressed.

On Sat, Mar 6, 2010 at 11:09 AM, M. Edward (Ed) Borasky   
wrote:



I filed a request for the ability for an authenticated user to be able to
retrieve *all* of his or her tweets a long time ago. I can find the date and
other details on this group if necessary. Someone back then from Twitter
said that or some equivalent was on the road map.

--
M. Edward (Ed) Borasky
borasky-research.net/m-edward-ed-borasky/

"A mathematician is a device for turning coffee into theorems." ~ Paul
Erd?s


Quoting "@seiz" :

 Thanks all for the replies.


Mark: Then the API Documentation is misleading as it says "Clients may
request up to 3,200 statuses via the page…". My assumption was i can
get a maximum result set of 3200 tweets but could change the timeframe
of the result set by using since_id and max_id.

What you say seems to be the case, but then the wording in the docs
should be changed to clearly state:
   Only the newest 3200 tweets can be received via the API. Older
Tweets are still archived at Twitter but not accessible via the
archive.

After all, this is not a Pagination limit at all then! It's simply a
limit of how many tweets are accessible via the API.

Since there might be cases, where archiving tweets might even be
required by law (in corporate enviroments e.g.), accessing old tweets
might be a rare but needed case and i could see the following as a
solution not affecting your performance.

Allow access to an "archive database" of tweets via a new set of API
Methods - e.g.:
http://api.twitter.com/1/statuses/home_timeline_archive.format
And implement a different rate limit to these API Methods – e.g. 4,000
requests per Month
This should suffice most usecases, as downloading old tweets is often
a one time thing in order to catch up. Once othe old archive is
retrieved, we can use the regular API Methods more frequent to keep
our backup archive current and stay within the 3200 tweets boundary.

What do you think?

Stefan.







--
Raffi Krikorian
Twitter Platform Team
http://twitter.com/raffi

Re: [twitter-dev] Re: Introduce yourself!

[M. Edward (Ed) Borasky]

Thanks!! I'll code something up for her if I can find a way to  
translate a Twitter search query into SQLite SQL or regular  
expressions ;-)

--
M. Edward (Ed) Borasky
borasky-research.net/m-edward-ed-borasky/

"A mathematician is a device for turning coffee into theorems." ~ Paul Erd?s


Quoting Mark McBride :


It's a pretty common request of the search team.  I don't think there are
immediate plans to support it due to its technical difficulty.

  ---Mark

http://twitter.com/mccv


On Sat, Mar 6, 2010 at 3:06 PM, M. Edward (Ed) Borasky   
wrote:



I just had an interesting feature request from a user. It seems like
it would be fairly difficult to do without access to the full
database, but what she wants is to be able to limit a search to just
the people she's following. For example, if she saw a link go by from
someone and didn't remember who it was from, she'd like to be able to
search for it.

I can think of fairly easy ways to code it using the API, but I'm
wondering if something like that could be built into the web app.

Re: [twitter-dev] Re: Introduce yourself!

[Mark McBride]

It's a pretty common request of the search team.  I don't think there are
immediate plans to support it due to its technical difficulty.

  ---Mark

http://twitter.com/mccv


On Sat, Mar 6, 2010 at 3:06 PM, M. Edward (Ed) Borasky wrote:

> I just had an interesting feature request from a user. It seems like
> it would be fairly difficult to do without access to the full
> database, but what she wants is to be able to limit a search to just
> the people she's following. For example, if she saw a link go by from
> someone and didn't remember who it was from, she'd like to be able to
> search for it.
>
> I can think of fairly easy ways to code it using the API, but I'm
> wondering if something like that could be built into the web app.
>

Re: [twitter-dev] Re: OAuth newbie question

[Ryan Alford]

The token is a posted parameter.  The secret is part of the key for the
signature.

Ryan

On Sat, Mar 6, 2010 at 5:55 PM, IDOLpeeps  wrote:

> I've overcome the nuances of generating the oauth signature.  It
> shocks me that the API documentation provides no clear indication of
> how to send the tokens along with an API call.  It's not even a PHP-
> specific question.  Simply put: Where do the "oauth_token" and
> "oauth_token_secret" get embedded in API call: As posted parameters?
> If so, with what parameter names?  Can anybody provide guidance?  I
> have seen many people ask this question, yet see no answer.
>
> As far as why one would want to use their own library vs. somebody
> else's, that's a question for the ages.  One specific answer is that
> many of us have created our own application-specific libraries that
> accommodate traditional http authentication and we'd like to keep our
> libraries when we add Oauth.  To do so, it's best to have an answer to
> this question.
>
> Thank you.
>

Re: [twitter-dev] Re: OAuth newbie question

[Raffi Krikorian]

i suggest reading http://oauth.net/core/1.0a/ and looking at
http://hueniverse.com/oauth/ for OAuth guidance.

On Sat, Mar 6, 2010 at 2:55 PM, IDOLpeeps  wrote:

> I've overcome the nuances of generating the oauth signature.  It
> shocks me that the API documentation provides no clear indication of
> how to send the tokens along with an API call.  It's not even a PHP-
> specific question.  Simply put: Where do the "oauth_token" and
> "oauth_token_secret" get embedded in API call: As posted parameters?
> If so, with what parameter names?  Can anybody provide guidance?  I
> have seen many people ask this question, yet see no answer.
>
> As far as why one would want to use their own library vs. somebody
> else's, that's a question for the ages.  One specific answer is that
> many of us have created our own application-specific libraries that
> accommodate traditional http authentication and we'd like to keep our
> libraries when we add Oauth.  To do so, it's best to have an answer to
> this question.
>
> Thank you.
>



-- 
Raffi Krikorian
Twitter Platform Team
http://twitter.com/raffi

[twitter-dev] Re: Introduce yourself!

[M. Edward (Ed) Borasky]

I just had an interesting feature request from a user. It seems like
it would be fairly difficult to do without access to the full
database, but what she wants is to be able to limit a search to just
the people she's following. For example, if she saw a link go by from
someone and didn't remember who it was from, she'd like to be able to
search for it.

I can think of fairly easy ways to code it using the API, but I'm
wondering if something like that could be built into the web app.

[twitter-dev] Using Normalize function in PHP for status updating. I cannot make it to work.

[hgc2002]

Perhaps this is a quite common question, but I can't find it in these
groups. So I ask it again:

The link  explains how to count the
140 characters.

I'm making tests using PHP. The recommendation is to use
Normalize::normalize($string,Normalizer::FORM_C) or I do have
interpreted it wrong (because it does not seems to work)?

My test: update status to "Probando la integración con Twitter desde
Madrid". Have you noted the accent? That is the reason of this test.

I've tried to normalize the whole string and all characters one by
one. The first way does look to normalize only the first character.
The second way seems to "delete" the accented letter and keep the
rest.

What it's going on?

Thanks for your help.

Herman.

[twitter-dev] Re: OAuth newbie question

[IDOLpeeps]

I've overcome the nuances of generating the oauth signature.  It
shocks me that the API documentation provides no clear indication of
how to send the tokens along with an API call.  It's not even a PHP-
specific question.  Simply put: Where do the "oauth_token" and
"oauth_token_secret" get embedded in API call: As posted parameters?
If so, with what parameter names?  Can anybody provide guidance?  I
have seen many people ask this question, yet see no answer.

As far as why one would want to use their own library vs. somebody
else's, that's a question for the ages.  One specific answer is that
many of us have created our own application-specific libraries that
accommodate traditional http authentication and we'd like to keep our
libraries when we add Oauth.  To do so, it's best to have an answer to
this question.

Thank you.

Re: [twitter-dev] OAuth newbie question

[Raffi Krikorian]

i'm not fluent enough in PHP to provide guidance, however, i must ask,
what's the aversion to using a 3rd party wrapper?

there are nuances involved in oauth signature generation (search this forum
looking for signature base strings, problems with quotes, etc.), and why not
save yourself that pain, accelerate yourself over that learning curve, and
use a library that is tried in true?

On Sat, Mar 6, 2010 at 1:06 PM, IDOLpeeps  wrote:

> What do you do once you get the oauth_token and oauth_token_secret?
>
> Do the token and secret get passed along with the rest of the API call
> as posted parameters?
>
> Can someone please provide an example of how to form a CURL request in
> php using the token and secret?  I can not find documentation anywhere
> that explicitly explains what to do once the token and secret are
> obtained.  All the documentation I can find provides for use for
> somebody else's php wrapper.  I don't want to use a third party
> wrapper.
>
> Thank you.
>



-- 
Raffi Krikorian
Twitter Platform Team
http://twitter.com/raffi

Re: [twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[Raffi Krikorian]

it is on our roadmap to be able to retrieve all your tweets.

just to be precise, however, all your tweets are retrievable, if you know
their IDs -- you can then retrieve them using statuses/show endpoint.  what
we don't maintain, however, is retrieval of more than 3200 tweets via the
timeline calls.  this is a limitation of our timeline and storage
architecture, but it is actively being addressed.

On Sat, Mar 6, 2010 at 11:09 AM, M. Edward (Ed) Borasky wrote:

> I filed a request for the ability for an authenticated user to be able to
> retrieve *all* of his or her tweets a long time ago. I can find the date and
> other details on this group if necessary. Someone back then from Twitter
> said that or some equivalent was on the road map.
>
> --
> M. Edward (Ed) Borasky
> borasky-research.net/m-edward-ed-borasky/
>
> "A mathematician is a device for turning coffee into theorems." ~ Paul
> Erd?s
>
>
> Quoting "@seiz" :
>
>  Thanks all for the replies.
>>
>> Mark: Then the API Documentation is misleading as it says "Clients may
>> request up to 3,200 statuses via the page…". My assumption was i can
>> get a maximum result set of 3200 tweets but could change the timeframe
>> of the result set by using since_id and max_id.
>>
>> What you say seems to be the case, but then the wording in the docs
>> should be changed to clearly state:
>>Only the newest 3200 tweets can be received via the API. Older
>> Tweets are still archived at Twitter but not accessible via the
>> archive.
>>
>> After all, this is not a Pagination limit at all then! It's simply a
>> limit of how many tweets are accessible via the API.
>>
>> Since there might be cases, where archiving tweets might even be
>> required by law (in corporate enviroments e.g.), accessing old tweets
>> might be a rare but needed case and i could see the following as a
>> solution not affecting your performance.
>>
>> Allow access to an "archive database" of tweets via a new set of API
>> Methods - e.g.:
>> http://api.twitter.com/1/statuses/home_timeline_archive.format
>> And implement a different rate limit to these API Methods – e.g. 4,000
>> requests per Month
>> This should suffice most usecases, as downloading old tweets is often
>> a one time thing in order to catch up. Once othe old archive is
>> retrieved, we can use the regular API Methods more frequent to keep
>> our backup archive current and stay within the 3200 tweets boundary.
>>
>> What do you think?
>>
>> Stefan.
>>
>>
>


-- 
Raffi Krikorian
Twitter Platform Team
http://twitter.com/raffi

[twitter-dev] OAuth newbie question

[IDOLpeeps]

What do you do once you get the oauth_token and oauth_token_secret?

Do the token and secret get passed along with the rest of the API call
as posted parameters?

Can someone please provide an example of how to form a CURL request in
php using the token and secret?  I can not find documentation anywhere
that explicitly explains what to do once the token and secret are
obtained.  All the documentation I can find provides for use for
somebody else's php wrapper.  I don't want to use a third party
wrapper.

Thank you.

Re: [twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[Cameron Kaiser]

> Ah, i see. I thought only the first post would be moderated (and only
> with that i got a notification about the moderation).

The group is configured to auto-promote after a certain number of accepted
posts. The number is internally determined but is in the ballpark of 20 or 30.
I will make this clearer on the group home page.

-- 
 personal: http://www.cameronkaiser.com/ --
  Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com
-- He is rising from affluence to poverty. -- Mark Twain --

[twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[@seiz]

Ah, i see. I thought only the first post would be moderated (and only
with that i got a notification about the moderation).
Thanks!

Re: [twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[Cameron Kaiser]

> Apologies if my reply comes through multiple times. But google tells
> me my reply was posted but i never see it arrive here.

Just remember that the group is moderated (I removed your duplicate).

-- 
 personal: http://www.cameronkaiser.com/ --
  Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com
-- armadillo, n. the act of providing weapons to a Spanish pickle. 

Re: [twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[M. Edward (Ed) Borasky]

I filed a request for the ability for an authenticated user to be able  
to retrieve *all* of his or her tweets a long time ago. I can find the  
date and other details on this group if necessary. Someone back then  
from Twitter said that or some equivalent was on the road map.

--
M. Edward (Ed) Borasky
borasky-research.net/m-edward-ed-borasky/

"A mathematician is a device for turning coffee into theorems." ~ Paul Erd?s


Quoting "@seiz" :


Thanks all for the replies.

Mark: Then the API Documentation is misleading as it says "Clients may
request up to 3,200 statuses via the page…". My assumption was i can
get a maximum result set of 3200 tweets but could change the timeframe
of the result set by using since_id and max_id.

What you say seems to be the case, but then the wording in the docs
should be changed to clearly state:
Only the newest 3200 tweets can be received via the API. Older
Tweets are still archived at Twitter but not accessible via the
archive.

After all, this is not a Pagination limit at all then! It's simply a
limit of how many tweets are accessible via the API.

Since there might be cases, where archiving tweets might even be
required by law (in corporate enviroments e.g.), accessing old tweets
might be a rare but needed case and i could see the following as a
solution not affecting your performance.

Allow access to an "archive database" of tweets via a new set of API
Methods - e.g.:
http://api.twitter.com/1/statuses/home_timeline_archive.format
And implement a different rate limit to these API Methods – e.g. 4,000
requests per Month
This should suffice most usecases, as downloading old tweets is often
a one time thing in order to catch up. Once othe old archive is
retrieved, we can use the regular API Methods more frequent to keep
our backup archive current and stay within the 3200 tweets boundary.

What do you think?

Stefan.

[twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[@seiz]

Apologies if my reply comes through multiple times. But google tells
me my reply was posted but i never see it arrive here.

[twitter-dev] Re: Can't access old tweets via statuses/user_timeline

[@seiz]

Thanks all for the replies.

Mark: Then the API Documentation is misleading as it says "Clients may
request up to 3,200 statuses via the page…". My assumption was i can
get a maximum result set of 3200 tweets but could change the timeframe
of the result set by using since_id and max_id.

What you say seems to be the case, but then the wording in the docs
should be changed to clearly state:
Only the newest 3200 tweets can be received via the API. Older
Tweets are still archived at Twitter but not accessible via the
archive.

After all, this is not a Pagination limit at all then! It's simply a
limit of how many tweets are accessible via the API.

Since there might be cases, where archiving tweets might even be
required by law (in corporate enviroments e.g.), accessing old tweets
might be a rare but needed case and i could see the following as a
solution not affecting your performance.

Allow access to an "archive database" of tweets via a new set of API
Methods - e.g.:
http://api.twitter.com/1/statuses/home_timeline_archive.format
And implement a different rate limit to these API Methods – e.g. 4,000
requests per Month
This should suffice most usecases, as downloading old tweets is often
a one time thing in order to catch up. Once othe old archive is
retrieved, we can use the regular API Methods more frequent to keep
our backup archive current and stay within the 3200 tweets boundary.

What do you think?

Stefan.

Re: [twitter-dev] Non-Smart phone location

[Raffi Krikorian]

Twitter doesn't do anything (yet) with updates that come in via SMS.



On Mar 6, 2010, at 9:54 AM, Abir  wrote:


Hey Guys,

If I update my status via text, using a non-smartphone in the US, does
twitter pick up the location variables? If so lat/long or just
neighborhood?  Thanks

[twitter-dev] Non-Smart phone location

[Abir]

Hey Guys,

If I update my status via text, using a non-smartphone in the US, does
twitter pick up the location variables? If so lat/long or just
neighborhood?  Thanks