1. my one complaint, is that when a post is old enough that it isn't in my RSS feed anymore I can't find it via feedster, so it's great for recent searching but for older stuff, not so much.

Posted at 8:03PM on Dec 18th 2005 by Sean Bonner

2. Hi Sean,

That's a great comment but let me explain how we work so at least this makes sense.

Feedster indexes content in feeds (rss, atom, rdf) and when we find a new feed for the first time, we start indexing it. Now since back content doesn't appear in feeds, our index is generally from that point in time forward. So, for example, we have everything from scoble indexed from roughly early March 2003 forward. What we also do is provide a backlog loader: http://feedster.com/backlog.php that allows people to manually create a custom feed of everything and then we'll index back to Time 0 (the start fo their feed). We've done a poor job pushing people to do that and we should do better. That I'll definitely agree on.

Now please note that this is very, very different from Technorati which ONLY indexes the past 7 days. Our archive is now north of 58 million documents.

Scott

Posted at 8:03PM on Dec 18th 2005 by Scott Johnson of Feedster

3. Google is killer because it helps me with my horrible spelling problem. If a word is spelled wrong, perform the search and display the results just in case, but give a suggestion or two of what the user might have met.

A lot of the search results I get are in languages I do not speak. Sometimes the characters returned are complete gibberish. A good example is when you enter a query of "google". Man it seems like I am infatuated with that company but I really am not. OK maybe I am. But it would be nice if the results were sorted by my language preference, which my browser will tell you if you ask it.

Posted at 8:03PM on Dec 18th 2005 by Peter T. Brown

4. Feedster would be ( better | perfect) if:

more complex: show "spheres"
Often results come in clusters. Say I search for "John" - I might get a bunch of political sites (John Kerry) and a bunch of Law sites (John Doe). There are probably a few degrees of separation, or cross-links between sites in the same category, but not between the categories. Of course a basic categorisation system could do this, but it might be more useful if it was dynamic.
This could also be displayed as visual "clusters" of sites.
Basically I know then more easily, on a vague search, which results to ignore, and which ones to drill into more...

simple add: keyword hliting. search results come back without the keywords picked out. eg if i search for 3 words, the results don't show if these words are grouped together in a single sentence, or in different parts of a page.

PS I dont have a GMail account yet, but would like one! Does it support Japanese correctly? Yahoo.com, Hotmail etc still have encoding problems. Hope its as good as Opera's M2 in latent searches/labels etc...

Posted at 8:03PM on Dec 18th 2005 by dc

5. Thanks for the note, I didn't realize feedster kept an index of everything, for some reason I thought it was only active feeds - good to know for sure.

I checked out the backlog loader and while it's a cool idea, I don't know how useful it really is. I mean, certainly some people will go through the efforts to put together an RSS feed with everything they have ever blogged, but most people probably won't - or can't - Personally I started blogging in 99 and was entering posts by hand, for maybe a year, after that I switched to Grey Matter and used that for a while before jumping to MT, so the thought of going back through all those archives and reimporting them into something to try to make one giant feed is frightening. There's no way I'd ever do it, so I know a huge chunk of my content will never been in Feedster, which is fine because I know that walking into it.

Don't get me wrong, I like Feedster a lot, and use it regularly. I also like technorati a lot and use it regularly, and several other searches to find stuff. At this point no one has it nailed, and it takes a combo of several services to get all the info I'm looking for on a daily basis.

Does that make any sense?

Posted at 8:03PM on Dec 18th 2005 by Sean Bonner

6. It makes a lot of sense. Most of the sites being put together to help navigate the massively syndicated web are themselves web services. It is easier than ever to combine dynamic feeds from multiple vendors (just be careful on ToS, of course). What is your ideal mix of features? I bet all you have to do is ask and a meta-site will spring up to make your life easier.
--The Other Scott at Feedster

Posted at 8:03PM on Dec 18th 2005 by Scott Rafer

7. now wouldn't that be grand scott...

what i am endeavoring to accomplish here with this series of questions is this dream sharing... like sean, i use a combination of tools that include feedster, technorati, k-collector, waypath, google, gmail, and bloglines just to help track and trace where these blog post conversations are going... and i know that i am still not getting a complete picture...

power to the blog post... i want a UPS like tracking mechanism on each post that will tell me where it has gone, who has touched it, responded to it, quoted it, commented on it, bent, folded, and mutilated it... (-:=

Posted at 8:03PM on Dec 18th 2005 by judith

8. Feedster can't take credit for coming with with the solution to your dream scenario, but we'll participate in delivering it. ScottJ's been part of the open standards effort focused on aggregator syncing via OPML. It will create a synchronization standard between multiple aggregators, so that "mark as read," adding new feeds, etc. works from your aggregator at home, to the one at work, to the one on your mobile phone, etc. All that will take months, of course.

Also, while more vendors needed to be recruited to support the effort, Feedster plans to support the AttentionXML spec that Steve Gillmor and Dave Sifry proposed. By embedding AttentionXML into OPML synchronization, (approximate and anonymous) information on how many people read your postings, how often, and in combination with what other kinds of information should become available.

Do we have REM yet?

Posted at 8:03PM on Dec 18th 2005 by Scott Rafer

9. Half the time I go to Feedster it is either not updated, and many times it is slow/down. I know from Scott they are upgrading things and getting more bandwidth, but Feedster would be my fav blog search *if* they were faster/mre updated. I'm sure they will get there... growing pains!

Posted at 8:03PM on Dec 18th 2005 by Jason McCabe Calacanis

10. I would appreciate very much a better support of not-english feed, like Italian ones, for example. It's not easy as it would to search non english feed, in my opinion, but I appreciate Feedster as the only one feed search spot.

Posted at 8:03PM on Dec 18th 2005 by Luca - Pandemia