-
Website
http://www.louisgray.com/live/ -
Original page
http://blog.louisgray.com/2009/05/twitter-search-engine-is-very-very.html -
Subscribe
All Comments -
Community
-
-
Top Commenters
-
Popular Threads
- 10 Top New Web Services of 2009 (From My Perspective)
- For All the Gloom Around RSS, Readers Continue to Climb in '09
- Growing Grumblings on Tech News Don't Answer Incentives Problems
- Mozzler Filters Your Social Stream for Links, Highlights Media
- iTunes, Sirius Seem Antiquated After Spotify iPhone Trial
Biz could write a few paragraphs explaining what they’re doing and why it’s important, and put it on Twitter’s blog. Then, @Twitter could tweet this out to its million followers, and re-tweets would mean probably the entire network would receive the news. It’s common sense. It’s common decency.
Instead, they make huge adjustments to things like @replies, tell us *after* they’ve done it, and then sit back and soak up the howls of protest. They tell us they’re paying attention to feedback, but then announce another ludicrous change that nobody asked for.
In a week we’ll probably be told you can now only search the tweets of celebrities. We’ll all complain, so they’ll change it to the suggested user list. But hey, they’re listening.
Search is their single biggest deal. There’s nothing wrong with making it better, or moving other things around so it’s more efficient, but it’s simple common courtesy to announce things like this before you start pulling out the wires. Just a couple of months ago Twitter was riding the wave of praise, while Facebook was getting slated by all and sundry. While they haven’t quite swapped places, I think Twitter needs to be very, very careful the next few weeks. Things can go downhill very fast.
Strangely Twitter only show my tweets dating back to Jan. 2009 when I had been tweeting since 2006.
Like you I went straight to Friendfeed for the information I was looking for, but only found some but not all the tweets. Could this be because not everyone's timeline are on Friendfeed?
Like Jesse Stay says we need to be able to use these tools (Twitter, Friendfeed, etc.) the way we want, to a certain extent, less of what the owners of these sites believe are correct.
However, let them create these sites how they want, that's fine with me. The minute they stop listening to their users though (especially those that have made them) is the minute their users start to leave.
And while the purpose of this post was not so much to highlight the strength of FriendFeed's search over that of Twitter, it did end up having some of that result. At least the data is kept safe somewhere.
I have noticed this for months, but always figured "well, nothing else there is reliable, why should search be?" But that seems out of line with what they think they're going to be able to sell, and it's unfortunate that it's not, because it could be a valuable source of data. The only real hope for Twitter's data (IMO) is to have more companies attach "firehoses" like the one that FriendFeed has.
Twitter is still a valuable source of data - Lots of people post tons of useful tidbits there. Surely the links to these blog articles go there, mine do. And FriendFeed is a huge resource for finding that data. But we will find flaws in their systems too. Just not as huge. And even if FriendFeed would be flawless, they can't index everything the right way for everyone to find everything they want.
Hopefully Twitter doesn't see their way to being "greedy" with their data, and in the process lose it, and its value. The way to prevent this loss is to hook up more hoses that receive the entire Twitter stream. That, in my opinion, is how Twitter should be monetizing. Not building tools to use against the data that they can't keep straight. I think that there will be no lack of companies to buy hoses either, people want tweets.
Data is only of any use if it's shared.
The problem, however, doesn't appear to be explicitly about money. Rather, it's the scale of the engineering challenges they're up against. I find myself wishing I could be a fly on the wall at Twitter -- this is a team that simply must be pulling their hair out to rectify the fruits of early engineering shortcuts regarding integration (search from Summize) and scaling. I've been commenting (apparently) negatively about Twitter lately -- I absolutely adore the service were it not for two things. First, it's unreliable. Second, I want more control to reduce noise. The perfect fix for both of these issues for me has been FriendFeed of late, but I'm still on Twitter and support what their doing. Really looking forward to substance coming out of Twitter HQ in the coming weeks.
Looking forward to improved search asap.
Cheers
Tony
Also from a developer point of view its a pain that the ids used in the search results do not match the id of tweets. It on the fix list but its been there for ages.
But I guess Twitter team is working on it and should release a new version pretty soon ;)
Twitter's biggest failure is that they can't remove indexed tweets on request.
Why Is it important to find Tweets
Particularly when you have that vast
warehouse of info at Friendfeed.
Take it one layer deeper
Why do I need to search how many times
I tweeted Big Lebowski quotes?
Or how many Tweets are about Lighthouses...
Ok I need more coffee
Waiter!
As for FriendFeed vs. Twitter, that's not this article. I've already written that one a few times. The good news is that FriendFeed has a viable, true, searchable index.
Louis, you're the one who keeps bringing up FriendFeed whenever critiquing Twitter, not me. I see them as two complementary, but vastly different, social tools. FriendFeed competes w/ Facebook, not Twitter.
Peace!
One other thing worth noting is that the search database at Twitter is completely separate from the tweet database- since the time that Twitter bought Summize, they have never consolidated the the databases. This results in things like, when you delete a tweet, it still shows up in the search results for some time.
All in all, they've got a ways to go.
For example, on twitter.com, we cannot search profile information. Only user names, first name, last name. As a user, I was trying to find out someone I had followed but forget the name but could not do any search based on the information I remembered.
I see Twitter search like a proof of concept. It has lot of potential but it can't be fully exploited yet.
But one thing that really bugs me is that they still keep deleted tweets in the results. When I delete a tweet, for whatever reason, I want it deleted. (Rather like the stories around yesterday about photos not being deleted) - It just seems they don't re-index things much, if at all. They just sit across 'the firehose', suck it all in and that's it.
But I think some of the comments here are overkill. I use Twitter search regularly as an alerting service for a broad vanity query. It works reasonably well, and I've learned to live with the glitches as part of what comes along with a free service. Could they do a lot better? For sure, and I've blogged about that. But what they have now is nonetheless useful.
That being said, I definitely agree that Twitter needs an angel with experience in scaling. It probably (should have) realized this pretty early on, when the explosion in users was first detected and projections for the future were being developed. It's really just a matter of money ... and that's probably what's slowing things down. I mean, when will Twitter become profitable? For that matter, HOW will they become profitable? At least Google, Microsoft and Apple (by far the least experienced of the three) have deep pockets and well-monetized public channels.
What I find interesting is that despite search potentially being a key asset for Twitter, the better and more useful search applications are being created by third-parties such as Twazzup and Twingly.
The element that's bugging me right now, is how useless Twitter's 'trending topics' feature has become; thanks to spam.
See http://www.tweespeed.com which gives the "instant speed of twitter" and try to estimate the reaching date. Amazing.
http://twitter.zendesk.com/forums/31935/entries...