We're not ready to fully support desktop clients on the Streaming API.
Connection counts, permissions issues to protected statuses, OAuth, etc,
still need to be addressed. At the moment, we're trying to move services
over where possible. Desktop client experimentation is fine, but full
support for desktop clients from our end isn't practical yet.
Both Search and Streaming discard all statuses from low-quality users.
Search additionally filters the remaining statuses for relevance and ranking
purposes. This may be hard to see now, unless you cross-reference the
Streaming results, but this divergence will soon accelerate and become more
obvious.
-John Kalucki
http://twitter.com/jkalucki
Services, Twitter Inc.
On Mon, Jan 11, 2010 at 11:25 AM, M. Edward (Ed) Borasky
zzn...@gmail.comwrote:
I just saw the announcement on Twitter-API-Announce about migrating to
streaming. I'm curious about a couple of things:
1. If your application polls for keywords, mentions, is whitelisted
on the Search API, or makes more than perhaps 10 queries per minute,
you should begin your migration to Streaming. Desktop clients should
postpone a migration to Streaming.
I'm making far less than 10 queries per minute - more like 1 or 2. I'm
now seeing a rate limit on the Search API of 150 calls per hour, and
my code adjusts to that. I have something planned that will work best
on the streaming API. It will be a desktop client, though - I don't
have a business case for building servers. Is there any reason for
staying away from streaming just because I'm building a desktop
client?
2. Complete corpus search: Search is focused on result set quality
and there are no guarantees to return all matching tweets. Complete
results are only available on the Streaming API. Search results are
increasingly filtered and reordered for relevance.
I was under the impression from the documentation that tweets going
into the Streaming API were in fact subject to the same Quality
Filtering of users as the search API. That is:
https://twitterapi.pbworks.com/Streaming-API-Documentation
Both the Streaming API and the Search API filter out all statuses
from low-quality users. Unlike Search, Streaming does not filter or
rank statuses for relevance, all statuses are available.
Is there some clarification available on the differences between
what's filtered out of both Streaming and Search and what's filtered
out of Search only?
