pretty cool and easy to use.
On a related note, in this site's Search Engine Spider Identification forum, dated yesterday (by moi):
Topsy.com / labs.topsy.com / butterfly.topsy.com [webmasterworld.com]
Butterfly/1.0 *and* libwww-perl [webmasterworld.com]
I don't like it. (not a bit)
interesting. I topsied the name of my city, and thre top results were some public events coming soon, some sports scores, and where to get the lowest gas prices in town.
Not relevant results like City Hall or Munipical News.
These are things people are tweeting about.
I can see how this might be alluring... but I can also see how the results might turn many people off.
The results have an almost random quality. I don't think Google, Yahoo, or Microsoft will feel threatened.
- I searched for one of my favorite subjects and someone named thai4evah has apparently found a way to hawk his spam site in the top 8 spots. The 9th is in a foreign language.
I'd say they have some severe growing pains ahead, it's far from ready.
I hope it doesnt' give results like the Mr Men character.
My site (hobby) is #1... in fact is the only entry. Kewl!
But other searches based on my hobby are only commercial sales sites... Topsy does not seem to have any depth to it. Meanwhile, it will take a little time to find out if there's any bottom to this rollout.
It's flagged as malware by our company firewall :(
Since it's based on tweets it's very easy to manipulate. I did some typical searches in my area and found one site had locked up the top several results for one of them just because over the last couple of months they've released a couple promotional videos and tweeted about them. The other searches looked either normal or empty.
If you go into it looking to search tweets it might be useful. If they market themselves as a niche search engine....
Lets leave the flames for some other site please.
Personally, I see the major potential here. The results are very interesting. I think the algo is clearly in a V1 beta state and will take a lot of training. This is a fresh engine algo approach that has never before been tested or tried. Basing results on popularity of "tweets" is ground-up new.
I found topics through Topsy that I had never seen before about a few products I follow. There are going to be some major finds in there for rep mgt and webmasters that have IP worth tracking, contributing too, and nudging in a favorable direction. Any brand, trademark, website, domain, copyrighted work, or other IP needs care and feeding. Following your interests on twitter is step one, and making sure you are in all important conversations about your brand early-and-often is mission critical going forward in the social media world. Every tool we can have in our bag to do that is a major plus.
Hats off to Topsy for putting together another free tool to do just that. It is currently the only engine of this type available. So far, I am finding it more useful than Twitter's search and far better than Twitter's useless "trending topics" display.
Just a basic observation, if a search input is having an icon input like a magnifying glass icon, it ought to have a mouseover cursor that indicates an action. Such a simple thing but in FF at least there is no action. I click buttons more than I hit enter and i'm sure a lot of other people do to - you need to cover the basic UI stuff if you're going to impress anyone with your results.
For several searches I have done, half of the top ten sites were also in the top ten for the same search made at Google... and they were the best sites for that/those topic(s).
For some subjects people are going to ask a question via Twitter, someone else is going to Google the answer, and send that site reference back as the answer via Twitter, so I'd expect some correlation like that to occur.
Hmm. I see that it often shows the URLs as shorts, not as their final destination, if you used a URL shortener in your message. That needs fixing, and pronto.
I am beginning to suspect that at least part of the ODP database was also used to seed this - or does the fact that some queries closely match the list of sites found there, simply reflect human usage of directories to find the information to answer other peoples questions with?
I have not investigated their 301 or 302 redirect handling, but I expect it to be far from optimum, probably abysmal. .... <edit>For new URLs, they seem to handle redirects OK, but for old URLs that change, I have no idea. I'd expect them to expire the data.</edit>
In my niche the set of links produced by Topsy and the set produced by Google are completely different. Topsy will give you a narrow set of links currently being mentioned, or none at all, but can not even begin to challenge the service provided by Google. If you happen to be interested in what's being mentioned right now, then Topsy does provide a unique and valuable service.
> I am beginning to suspect that at least part of the
> ODP database was also used to seed this - or does the fact that some
no - this is not that type of engine. This is indexing pages that are pointed to by tweet urls (including shortened urls). That makes it a "hot topic" engine. It has nothing to do with the odp.