Forum Moderators: bakedjake
[mooter.com...]
I kinda a like it. It is a very impressive start! (it's one of those clustering/kartoo like engines)
[mooter.com...]
"Users shouldn't have to work hard to find what they want. Mooter sets out to solve the problems of most search engines by understanding the psychology of how users interact with information.
"Mooter analyses the choices you make while searching, then reorders the results based on what you are actually looking for at that moment without you having to go back and rephrase your exact needs."
So instead of giving users long lists of scrambled results, Mooter displays simple, sensible categories of information. As users search, the algorithms shuffle the results in the background, ensuring that more relevant results are displayed.
Mooter is simply the worst search facility I have ever seen.
Really? Where did you look at all the years? There are tons of so called search engines out there which are extremly bad. Compared to them, Mooter is the best thing I could think of.
Want to see a real one?
[vivisimo.com...] or
[kartoo.com...]
It may improve, I hope so. They certainly should not have gone public with it the way it is.
There's only so much that can be achieved in a lab. Put it in the public domain, cop a few hits to the chin, improve and release again.
DEMO OR DIE!
We expected a small article in an Australian local newspaper, and a small steady stream of local users for the first week. Instead we got about 1/4 page in most of the major Australian newspapers; plus it got onto the wire, which ignited the system about a day or two before we expected it. We are also on national radio and cable TV news here, so it has been an interesting week. People seem to like it, and it spread from there. (Maybe journos find a female ceo of tech co unusual for some strange reason? shld that be a new topic?)
Why we don't want to be like traditional engines:
most search engines tackle the non-trivial task of organizing massive amounts of info from the same angle: 'lets decide on behalf of the searcher what is most relevant, based on what the bulk of the general public thinks is relevant.' We decided to rather expend our efforts watching how individual PEOPLE actually search, see where it hurts, fix that. So, you know how it goes: type in search phrase, 3 mill results, read, read, read, scroll, click, back to results, read, read, click, then go back and type in a different phrase to educate the engine more about your needs. So there already are about 10 things you have to do, and that’s assuming the sort of search where you have a pretty clear idea what you after to begin with.
While you are reading, your mind is forming a sort of scaffolding with all the data. Hanging around in fluid nodes in this scaffolding are roughly held together conceptual groups – we don't really like handling more than 3-5 chunks at once. The reason for this is pretty simple - back in our troglodyte days you would die rather quickly if your mind didn’t work like this: you had to focus on essential clumps on data (and clump details if there were too many), and discard anything extraneous (even though the 'extraneous' in today’s world may be critical info). This means sometimes we got it wrong, by putting things in wrong nodes or assuming incorrect relationships between nodes, but it was better than slowing your decisions down and being eaten by something meaner and (possibly) smellier than yourself. That’s why so few people go past results page 2 of traditional engines. And that’s why we group conceptually, rather than try override human hardwiring.
Some of the things MOOTER does are not as readily apparent as the opening starburst of clusters. When you put in a search phrase, mooter analyses all the results. The clusters are not formed by counting up batches of keywords, but by some fairly intricate linguistic analysis of underlying themes. If the user is traditional-engine trained and needs a long list of results, we give a list in ‘all results,’ but have put our DAA algorithms on them. Because mooter understands the themes underlying websites, we are also able to pick up dynamic trends in your searching. So lets say you search for ‘dog,’ but then for the next few minutes all you do is look at sites with Chihuahuas, we are able to figure that out and push Chihuahuas plus dogs up. We also have a few other interesting algorithms, but I won't go into detail on those just yet!
SO in short: Mooter is slow this week because of massive largely unexpected press & traffic (working on that), and also slow because MOOTER is actually thinking (or partially thinking anyway, for AI purists) not just regurgitating.
We would much rather have a dedicated niche of smart users than dominate the planet.
We would much rather have a dedicated niche of smart users than dominate the planet.
Smart users dont need to be told what to do, thats why they like google...i think u should be catering for the dumb users.
Two things
"So lets say you search for ‘dog,’ but then for the next few minutes all you do is look at sites with Chihuahuas, we are able to figure that out and push Chihuahuas "
So if i was smart user and wanted my site to be number 1 or 2 ,3,4,5,6 id make different domains search for that result and boost up my site?
I think thats a disaster plan against spam.
Also I think you sooooooooooooooooooooooooooo forget one BIG thing?
The internet is made up by content. This explains the problem finding content in foreign languagaes because english content is more predominant at the moment.We webmasters make up the content. So accurate results are all dependant on opimization factors put in place by webmasters.
Also who defines good content?
Also i hate the fact that people are trying to outsmart me, i dont think other people like that idea either.
We would much rather have a dedicated niche of smart users than dominate the planet.
Again on this point , i bet you would change your model if these "smart" users do not buy things or click on advertising.
I like google, cos content is king, not what someone else thinks i would like.
AI + search would be an incredibly combo, the linguistic end of the field has yet to be tamed, and is wide open.
Though naturally, that kind of high level processing, etc / will take a bit of time before it's readily apparent to the end user in the results sets that are available.
Will this be the first "fuzzy search engine" that uses fuzzy logic & fuzzy sets in a major way? Imho, that & neural nets are the only way to get the keyword mappings & groupings done, and have a mathematics that will scale to the type of calculations that'll be necessary...
Good luck.
Penfold25 I concur, Webmasters and SE's shld be in a gorgeous symbiotic dance, not antagonists.
'content is king, not what someone else thinks i would like' that’s just the point. Who do you know has honestly ever looked all the way up to result number 107698345 of 200 million results? Where's the choice or ability to see relevant content in looking at the first 10 or 20 sites (which most ppl only do), as decided by the great big SE? Anywhere in the first 200 (at best guess) is probably buried the most relevant site for that human at that point in time, but who even looks to 187? MAYBE the site that would meet yr needs best is at number 1, as decided by the relevance algorithm, but maybe not. So isn't the traditional SE deciding on the top 10 a stronger example of being told what to do? Plus at a recent SE conference we heard constant speculation that up to 65% of search pages may end up being paid. In contrast, our first cluster page represents sometimes hundreds of most relevant pages to yr keyword, but mooter gives you the choice where you steer into that sea of content.
So optimizing for MOOTER will be about simply ensuring general concepts and groups of thought in website content are what your user is looking for and are what the site is about.
Now, I don't know whether you've 'met' GoogleGuy, but I'm thinking something like "MooterMiss" or "MooterMadam", or maybe even "MooterMum"! ;)
We would much rather have a dedicated niche of smart users than dominate the planet.
This could present a problem according to the 22 laws of marketing ("the law of the hype") as I remember ...
Good Luck!
I was very critical early only because people were comparing this new engine with Google which was totally wrong considering that at the time the results seemed innacurate.
Anyway they seem to be flying along now and I wish them all the best.
This is a global SE is it not? Not Australian as such? In other words it retrieves results from all over, not just Aussie? (otherwise it should have the .com.au domain name)
And also I'd like to know in what programming language mooter is programmed in.
Sid
I didn't notice any overlapping issue so perhaps they have fixed it now.
Sid
Sorry about the clutter the scooter thing creates. The overlap is something we have to put some work into. Our UI is challenging because we draw the page dynamically, so we have all sorts of challenges most engines don't deal with. We are still fiddling with UI, logo's etc, but have had to give priority to other things. UI still in a fluid state so would love input from a savvy audience.
We are global search - we did mistakenly assume that we wld get mostly Aussie traffic for the first few months, so our index is currently a bit Aussie-heavy, but we are working on that. We have a bit of a multinational team - our CTO speaks Esperanto and I think has Scottish roots. The Neural Network PHD guy is Greek, I was born in Central Africa etc...
Oddly enough, the relevant listings included pages with lots of links to other relevant pages--yet strangely enough these weren't showing up in the "cluster."
I'd be interested to see how this works with a bigger db.
Also finding the same irrelevant pages linked to from multiple clusters was annoying. Will moot too be spammed into oblivion?
<added>Liesl--you need to have a link at the bottom of the page for page two, page three, etc of search results, not just at the top left corner. At first I thought each cluster was only returning 10 results or less, which didn't make me feel good about the size of your index.
Yahoo, Google, etc have always had such a link at the bottom of their pages, and with good reason. When I scroll to the bottom of the listings, the link to the next page should be right there.</added>
I'm curious where they are getting their data.
Are or were?
Looks like an old crawl well before we discussed putting a star in the Title field - the data for my pages is that old as I have had a symbol on my personal page since and Mooter's data is older.
[mooter.com...]
Not afraid to mention the critics either - nice.
