Search 2.0 vs. Traditional Search

ReadWriteWeb writes "Ebrahim Ezzy reviews 5 new third-generation search technologies — and how they compare to the big guns of Google, Yahoo and MSN. These so-called "search 2.0" companies are combining the scalability of existing internet search engines with new and improved relevancy models; they bring into the equation user preferences, collaboration, collective intelligence, a rich user experience, and many other specialized capabilities. The new search engines profiled are Swicki, Rollyo, Clusty, Wink and Lexxe." Note, as the article points out, that the author has developed yet another search engine, called Qube.

Not so stunning results for the "next generation"

Well, I tried out the "Search 2.0" Lexxe, the one that "does what TSE's already do, but more efficiently."

I asked it a simple question. And it responded. Here is the efficient answer that must surely have Google quaking in its boots:

Lexxe (alpha version) has just encountered a system or internal connection problem, due to too many users using it now.

Interesting choice of names

My confident prediction for search 3.0 engines will be TinkyWinky, Dipsi, Lala and Po.

I say!

Well, I'm off to eat Food 2.0 now and after that I'm going to Take A Dump 2.0

Am I the only one who's getting tired of this trend of tagging on 2.0 to everything? It's stupid. Searching is still essentially the same way as before, it's not like a magic robot comes out of the screen or anything.

Re:I say!

All the op said was the "2.0" tag was stupid, not the content. I happen to agree, what the hell does "search 2.0" mean? Or "web 2.0"? If you want to actually discuss a technology or approach then fine, but these terms are so vague they don't actually mean anything at all.

Its just marketing and hype.

Re:I say!

I agree that buzz words like Web 2.0 sound lame, but they can loosely explain ideas that are not fully formed... Just as the "Information Superhighway" got used way too much as the internet was gaining momentum, I think "Web 2.0" will get used in a similar way, to describe an idea where the scope is too great to be explained in concrete terms.

As for throwing "2.0" at the end of everything, it will happen, and it will be inaccurate to the /. crowd, but it will be beneficial for a lot of people not "in the know".

As mush as we might hate to admit it, the business types that will use this term all the time are just as nessiary as the techies in advancing stuff.

Re:I say!

I thought the increasing amount of amateur porn was pr0n 2.0, and that started before Web 2.0. Porn; ahead of the game as always.

I know! I could create a Porn 2.0 website called fuckhr! Or fuckr. You post your own home-made amateur porn movies, and you tag them with tags like "hardcore", "bj", "cumshot", and "anal"! Then anyone could just search for specific tags and find some good (hopefully) amateur porn to fit their specific fetishes.

Y'know, this almost seems like a good idea...

Its all good

As long as "Lesbian Porn" return plenty of varied and releveant hits

Re:Its all good

As long as "Lesbian Porn" return plenty of varied and releveant hits

Those Web 2.0 thingies are very smart. So you enter the query and they immediately ask: "Do you want to include your sister's name in the search?"

no one gives a fuck

Move along nothing to see here. People don't want flash based 2.0 web searchs, they want to use a search engine as a spring board. You hop on and 2 clicks later you hop off to your location. When you start adding an interface beyond basic input and 12 million adverts around it (hello yahoo), you lose the entire point of using a search engine to find what you want quickly.

Keep it clean and keep it simple, that's all you need for a good interface in most cases.

From the makers of Web 2.0...

...comes the new rage that's sweeping the Internet: Search 2.0! Yes, you've enjoyed Search 1.0 for years but now there's the new and improved Search 2.0! It does all the smae things, but different! No more time-consuming Googling for things -- with Search 2.0, you can have your results in about the same time and have them be remarkably similar!

If they think slapping a fancy title on it will spark everyone to transition to their new search products, they should think again. I suspect Google will simply roll out there 2.0 option at some point and kick everyone else's butt.

Buzzwords

Note, as the article points out, that the author has developed yet another search engine, called Qube.

Apparently he's also working on Buzzword 2.0.
From the Qube home page: AdRoll program aims to enable a new medium that allows free, point based advertising in a proactive manner

With synergy! Concordantly!! Vis-a-vis!!!

You know...

I just don't see any of these names becoming verbs.

You have to trust an article...

...written by someone who is actually working on the same technology for a rival company to the ones listed in the article. There's an unbiased piece of reporting for you!

Uh oh ...

..., a rich user experience, ...

Well, right there's one of the warning phrases.

One of the big reasons for google's success is that it doesn't give you a "rich user experience". The main web page is utterly plain and simple. You type in a word or phrase. You get back a page with a lot more text, but its layout is again simple and obvious. Granted, you can click the "advanced search" and see something more complicated. But they've carefully hidden the "rich user interface" behind something that's simple and obvious.

Google's ads are an example of the same. No "rich" ads; just small, unobtrusive chunks of text. Nothing distracting and annoying, so people don't look for ways to turn them off.

I like wikipedia for the same reason. No flash or pizzazz; just simple, plain, easy to use, and informative.

When I see something touted with a phrase like "rich user experience", my natural reaction (after more than a decade of web use) is to shudder and go on to something that's more likely to be useful and informative.

Categories, duh

No one seems to figure out that the next generation search engine will have to get specific first(patent pending), otherwise how does "bush" know where to go in the search results? It ought to show me a page in between results that says, which "bush" are you searching for? and then has 1 sample result from each "bush" related result group(patent pending). Oh, one other thing -- patent pending.

Lexxe is actually pretty good!

It answered my question accurately and succinctly: http://www.lexxe.com/main.cfm?sstring=what+is+the+ best+search+engine?&clickcluster=fmclk&sstringtemp =fmstr [lexxe.com]

Clusty

Ok... I was looking for something yesterday on Google, but couldn't find it.

Tried out the clusty solution, and found what I was looking for very rapidly. TFA is correct it feels like a cross between Google and eBay.

There something to that. I can see Google copying it.

I didn't try the others because they looked like too much hassle. One of the original appeals of Google was the simplicity.

Ah, "collective intelligence"

...Because none of us is as dumb as all of us. (http://despair.com/ [despair.com])

In Search Of...

Third-generation search technologies are designed to combine the scalability of existing internet search engines with new and improved relevancy models; they bring into the equation user preferences, collaboration, collective intelligence, a rich user experience, and many other specialized capabilities that make information more productive.

Yes, but can they tell me where in the hell I left my car keys??

What About Collexis?

I built prototype search software that revolved around a product called Collexis. It has a medical demo [collexis.net] you can mess around with. The beautiful thing is that it uses a taxonomy to fingerprint documents. It also takes in raw text and assigns it a fingerprint and then uses Sleepy Cat to quickly reference many records and match your fingerprint. Unfortunately, it's not built for "open" domains like everything on the web but works best when you have a finite domain and a large number of documents to search.

I feel the author fails to even address the first thing he should have in this article. Why move from "Web 1.0" to "Web 2.0"? This article is not intuitively laid out.

I found an article in Nature [nature.com] to be much more informative than the article linked in this story.

Niche search

while these are clever ideas, and do indeed provide a slightly different spin on the traditional search engine, I believe that they will not have much hope of ever taking a bite out of google or yahoo. I mean, I just went through the process of creating a Swicki, and while the interface is nice... it is a lot of work.

I still think that the niche search engines are more viable not so much as alternatives to google or yahoo, but as an almost adjunct. Like the site I volunteer for, Diysearch.com, yeah it will never replace the majors, and it isn't intended to do, but because its subject-matter focused, the search results and relevancy are that much higher than what you'd get from a google or a yahoo.

I have no idea if subject-matter focus is the most viable route in terms of focusing search results, but Diysearch.com has been around for a decade and its doing quite well.

Pretty interresting

Just tried Lexxe
Q : Who is the president of united states ?
A : Armed forces

"Rich user experience"

they bring into the equation user preferences, collaboration, collective intelligence, a rich user experience, and many other specialized capabilities

The only "rich user experience" I want from my search engine is to experience a set of results rich in accuracy, without any other bullshit. Unfortunately I suspect this guys idea of "rich user experience" is mostly the kind of crap I want to avoid.

Fine Print...

Let's see... here it is... Search 2.0 requires Web 2.0 or higher. A bribe to your local government official may be required for neutral and/or faster access. Most credit cards and first-born children are accepted. If you broke the shrink wrap, you automaticaly agreed to these terms.

Geek Powerhouse?

The "Qube" webpage calls the product a "Geek Powerhouse", which makes me laugh. The "About" page talks about "browserless search", which sounds a lot like adware to me; and in fact, that's exactly what it is -- you have to download a program to your local machine to use the service. The part that makes me laugh is that despite its being a "Geek Powerhouse", it's Windows-only; no Linux, no Mac OS X. "Search 2.0" apparently means "Now with 200% more evil!"

search 3.0

Eventually search will run like buying a car, I will choose site options (no-popups, easy on the easy, 5 ads per page, no audio, member of the Better Business Bureau) then I will search for what I want and it will bring back sites. This will help weed out irrelevant and wasteful results my results can be as clean I want, it won't be based on other people it will be based on codes and rules, web2.0 is a gimmick which will be short lived. p.s. remember when web search involved driving to the library to pick up the latest list of sites you could 'dial up' and then searching the index for somethign cool.

Clustering?

I haven't seen much talk about clustering search engines, such as vivisimo. They seem a whole lot more 2.0 than the ones quoted in the article. For those unfamiliar, they categorize results from a "flat" search, which can be very handy. For example a search on "UPS" sorts results for uninterruptible power supplies separate from those brown truck people. Very handy...

Sheldon

Bingo!

• First-generation search ranked sites based on page content - examples are early yahoo.com and Alta Vista.
• Second-generation relies on link analysis for ranking - so they take the structure of the Web into account.
• Third-generation search technologies are designed to combine the scalability of existing internet search engines with new and improved relevancy models; they bring into the equation user preferences, collaboration, collective intelligence, a rich user experience, and many other specialized capabilities that make information more productive.

Bingo! [bullshitbingo.net]

Names

What I'm calling Search 2.0 are actually third generation search technologies. To explain the generations: * First-generation search ranked sites based on page content - examples are early yahoo.com and Alta Vista. * Second-generation relies on link analysis for ranking - so they take the structure of the Web into account. Examples are Google and Overture. * Third-generation search technologies are designed to combine the scalability of existing internet search engines with new and improved relevancy models; they bring into the equation user preferences, collaboration, collective intelligence, a rich user experience, and many other specialized capabilities that make information more productive. Examples: Swicki, Rollyo, Clusty, Wink, Lexxe

Okay, I have a few questions about this brave new world...

1. Given collectives like MySpace, Friendster, and the like - I really do not hold much faith in the very idea of collective intelligence. Instead, I think collectives actually devolved into the realm of ignorance, fear, and sheep
2. In the future only the colors brown and red will be allowed.
3. Why is the future always stupidly named by those whose products will not live to see it?
4. Hopefully the future of search will mapped out by throwing out all of the Flash, Shockwave, and other crap in favor of a clean interface and better relevancy. I would trade all of the "rich user experience" for something that gave me clean navigation, relevancy, and throwing out crappy link-sites.

I haven't RTA...

I haven't read the article, I didn't even read the summary all the way through - and I have no excuse other than I am, at this moment, correct spelling and all, very, very drunk - but I'll tell you this about search engines.

I wil put into a search engine an innocuose (sp?) phrase like "toilet ball and cock", and get back all number of results, most of which I have absolutely no interest in what so ever, but because ten thousand, or ten million, people linked to them, I get some really fucked up results when all I was looking for was information on how exactly toilets work.

Search engines, regardless of whether they're Google or some third party engine which does purely semantic correlating, are, for the most part, utterly useless, because they rely on human beings to somehow provide the correct input, and yet, we as human beings, rarely think the same, or put the same data into the same storage style (You say RSS, I say Atom).

The best trick to using search engines is not to rely on what they output, but what they output and how it relates to the next three or four results.

I often search for things using Google, and I generally find what I'm looking for, after much gradual filtering - "Damn, must add -{snarf} to that" - but that is about as far as they're going to get.

Creating search engines which only catalogue blogs, or tech specs, is a handy tool, but we still have to do a lot of the thinking on our end, and I'm glad of this because I don't want some no-nothings to suddenly up and tell me stuff that I already know because they just happened to find it - sometimes two months, at least, after me.

Generation 3 is Search 2.0?

I've heard of innumeracy, but this is ridiculous.

tagjag

No one seems to have mention Chris Pirrillos new project
www.TagJag.com

Is it still unkown in the wider community even after the big launch at gnomedex?

Cheers
Dean
www.collins.net.pr/blog

OMG ColdFusion !!!!!!111one

lexxe.com is implemented in ColdFusion [adobe.com]. ColdFusion used to be a neat solution back in 97 for your first-guestobook(tm) and the sorts. CFML is s server parsed language embedded in HTML. The coldFsuion server was written in Delphi and later with kylix (i guess since it was made available on linux too). They must done quiet some string parsing and when CF hit version 4 it became really slow and did not scale well.

Then someone had the neat idea to emulate the cf syntax by using a java xml parser. compared to the original cf it was pretty fast. So Allaire (back then) decided to move to Java and nearly went bankrupt (might be unrelated).

Macromedia aquired allaire and the coldFusion server has seen some more half-hearted development. ColdFusion was now based on JRun, an EJB Server. JRun had everything but a kitchen sink + the ColdFusion functionality. and because it was based on some "Enterprise" thechnology became the flagship of Macromedias server side technologies. Macromedia wasn't showing much interest in supporting ColdFusion anymore (apart from selling what they had).

now that Macromedia was aquired by Adobe, interest in the product (read: stability, performance , support) is/will be more problematic than ever.

Long comment short meaning; WTF are they using such an utterly bad performing (and expensive as well) technology to implement a meta search engine. Do they expect any traffic soon? or is this just a playground of some kid ?

i have tried to use lexxe just now and got the following response

Sorry, Lexxe has just experienced Internet connection problem. Please try a few minutes later. Thank you for your cooperation.

When they were disconnected from the "internet" then i would probably not even get an error message, right? right!. So i gues this "search" engin just got slashdottet, hehe. Or the search providers are blocking their requests ?

Ketchup vs Catch up

The difference between ketchup and catch up is that the former can be put on top of a burger whereas the latter may include needing to actually make one. There are a few good ideas raised by new searchers mentioned, but they certainly have a lot of catch up to play. I think the idea of relevant categories to browse in addition to however many results you are displaying on the page is good, and I liked how Clusty did that. The problem is that the main results are very poor. Not to say that google would never give you irrelevant results, all I mean is that I would never be inclined to hit any "I'm Feeling Lucky" button on any of this new searches right now. It's much harder for the new developers to perfect relevancy than it is for google to add in results plopped into categories. I'm sure they could produce a Beta version of it that would last longer than the lifetime of some newer companies. It's great that the new startups would like to be the next google and want to influence the world, but not all things are so replaceable. Hold on though, I need to go register for the Green Party.

Wrong number

Sorry, Google *is* search 2.0. Search 1.0 was the fscking phone book. And I don't think this new stuff deserves a major version number. Perhaps search 2.1?

What I want is weighted search

I want to put in keywords an give them relative importance to each other and then I want to see filtered results instantly..... So let me put in three keywords, hit submit.. get results... okay 65,056 results... now let me make keyword number 2 3 times more important than the rest... preferably using a slider widget with 100 tick marks on it and a dragging marker to indicate the number... and... the results refactor before my eyes without a refresh so i can see how this has impacted the results list...

For instance, I often look for code examples... now I usually want them in a specific language... lets say ruby which should narrow down my results to just pages that talk about ruby, but my second keyword is xml... and it's much more important to me because I'm looking for a xml processing function, not just a mention that xml is handled by ruby and finally I'm looking for examples but it's not as important as xml to my search because It may filter it too much in favor of just tutorials when an api reference may have what I'm lookig for.

So i get a results list and voila... yep lots of tutorials on ruby xml processing... but lots of books in that list... so I give xml a weight of 30 to see if it brings up a more rigorous set of results with more about xml and less about examples, but still on the topic of ruby... well I would do this if it existed.

Okay, ready set CODE!

Slashdot 2.0... can we move on now

Will someone please shoot the writers/editors who keep referring to things as "two-dot-oh". Please. It's like a rolled up ball of bad-naming crud.

What we have is an ugly rehash of the late 80's and early 90's when everyone who wanted to add "new hotness" to their product name called it 2000, the dotbomb when everything cool was e or .com, and the recent "Applefication" of products to i.

Get over the 2.0 already.

Please try out my new search engine

I would appreciate any feedback on my new metasearch engine, Zeedex.com. Zeedex suggests terms to help you narrow your search. The terms are submitted by people. Anyone can contribute lists of terms, edit other people's lists, and leave comments. It's like a wiki.

I wanted to make something that would help people who are new to a topic, or are looking to dig deeper into a topic.

Anyway, it's only been up about two months, and I would be grateful for any suggestions, thoughts, etc. There aren't a lot of lists there yet because the site is still so new, so please test it with either "computer" or "civil rights". And if you like it, please register and add some lists!

Thanks,

Russell Miller

[e-mail]

http://www.zeedex.com/ [zeedex.com]

I want real improvements instead

What I am looking forward to is search engines where I can choose to search specific types of pages, like forum discussions, blogs, news articles, product support pages, etc.

I also want search engines that ignores menus, or other things that aren't part of the main content of the page. Why should every page on slashdot be associated with Apple just because it appears in the sections menu?

Another thing that could use improving is the removing of pages with similar content. There is no need for there to be 50 wikipedia clone pages in the search results.

There are lots of useful improvements that could be done to current search engines, but adding useless gui features aren't one of them.

I don't get it...

Ok, so I skipped TFA and went right to Qube. Download, scan, dissect, install. Great, so I have this hefty MFC application on my system now that allows me to search without opening a browser.

Wait.. huh? No browser?

Ok, so I give it a shot. We'll skip the part where only Alt-Z seems to work and none of the right click features do. It's beta, that happens. So, anyway, I search for a few things. Can you guess my results? Yup.. Every one of them is a web page. So, uh.. How does this help me? Now I need to pop up a web browser to view the pages that I just searched for. Or, I can use the preview feature in cube which seems to be an embedded version of IE... Wait, isn't that a web browser?

The feature list is nice, but I'm not really seeing anything interesting :

Browserless search - Ok, neat I guess, but I still need a browser to view the results anyways
RSS Feed Reader - I use Firefox for this already
Search History - Between browser history and what some search engines already support, why do I need this?
One Click Search - One click and a keypress maybe, tho it wasn't working for me..
Built in Previewer - Umm.. Doesn't this mean you're opening a web browser? albeit an embedded one?
Adult Filter - Don't believe in em, don't use em.
Search Refiner - Most search engines support this already...
Progressive Results - Again, I believe most search engines do this already, you just need to load the page
Realtime Suggestions - Another prevalent feature
Dictionary - define: works fine for me...
No Spyware/Malware - Notice that Adware is missing.. Not to say that Adroll will be malevolent, but it's there...

So maybe I'm missing something here, but if I'm using a search engine, and the results are all web pages, how does it help me to not have to open a web browser first?

The problem has never been search..

..it's scale. I remember a company that built an index in SQL; the concept they were using worked but thier back-end did not scale. There is also a huge start-up cost when building a modern cralwer, which is why there are so few new search companies that survive.

In other news

In other news, I tested out new Web 2.0 mail applications. They're called Eene, Meene, Miney and Mo!

version numbering

Third generation search engines are called Search 2.0? Ok, who put Sun in charge of vesion numbering?

Legitimate concern 2.0

With all the hype around 2.0 I see a serious issue ensuing. What happens when 2.0 is upgraded? Obviously we can't call it 3.0 or even 2.5 that just wouldn't be "hip." Therefore we need to come up with something even more original and exciting. Be sure to watch for my next new and upgraded post, "Legitimate concern 2.0 2.0"

ViewFour.com is the next big thing - visual search

I think most of the SE's already have search down very well... except how it's displayed.

Check out how to visually see results at ViewFour.com

Search 2.0 is already here

It's called google. For anyone old enough to remember, altavista, et al were very "dumb" searches; there were no heuristics involved other than popularity (the more instances of "string" on a page, the higher that page was ranked). Google changed that by adding page popularity (as well as other rules) to give more relevant results.

So, who's the president of the US?

Thomas Jefferson [lexxe.com], of course!

Don't be quick to dismiss these

Don't be quick to dismiss these as ways to cash in on the "2.0 buzzword bandwagon"! I actually find that Clusty is good at something that Google isn't. Specifically, ever try searching for something where your search term has many meanings? For example, "apple" could refer to the computer, the fruit, or the Beatles' record company. "record" could mean "a world record", an audio recording, a 12" vinyl platter, the act of "recording" something, ect.

To exemplify my point, take a look at the differences between the results of a Google search for "record" vs. a Clusty search for "record". Clusty's results are much, much, much better:

• http://www.google.com/search?hl=en&q=record [google.com]
• http://clusty.com/search?query=record [clusty.com]

Scrapers ? Where are the bots?

OK, 3rd generation ... now how comes that i am running 30 domains, which of some are indexed on these engines and never saw a hit from any of these engines or their bots?

How comes that they find some of my domain names with obscure names that they do not possible index for ....

I tell you how : expanding the 2nd gen means : scraping results. But how comes, that none of the engines had any human rating system visible to the visitor ? Not that I examined all the menupoints, but a human judgement system would include a rating next to every search result... hey even a browser plugin, so I caould click on "scaper" "spam" "porn" when on a site i did not expect to be.....

but hey these are just my ideas of human controlled ?

no try a search : "phentermine" or "viagra" or "penis enlargement" into these engines, and you see the amount of trash that blows into your face, and all the blog spam that pops up .....

OK, yes I also run a home developed search portal for my own entertainment (rss filter, kindof) , so I do not want to be harsh, I am fighting with spam and bw/proc capacity as well. It's just that the review claims many things you do not actually see in these engines, and some seem to me like dogpile.com, that collects msn,yahoo,google into one portal ...

The author REALLY misrepresents history

He says that a first generation of search is exemplified by engines that focused on page content, such as Altavista or Yahoo.com.

In doing so, he ignores well over a decade of pre-Web text indexing products.

What's more, it's nonsense in any case, since Yahoo started as a directory, with a search engine added only later.

Since it's anyway admitted that he's biased in how he defines the market space, I wasn't motivated to read on further. Maybe I'll click on the five sites mentioned, but the review itself is almost certainly worthless.