The Design of Software (CLOSED)

A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.

The "Design of Software" discussion group has been merged with the main Joel on Software discussion group.

The archives will remain online indefinitely.

A new experimental "similarity-search" engine. Feedback?

We just put online an experimental implementation of a similarity search algorithm we've been designing.

You can try it out at eser.org: http://www.eser.org/usa/en

It's a bit slow, but we wanted to get something out.

Thoughts?
eser.org Send private email
Thursday, June 19, 2008
 
 
Not enough info to build up any 'thoughts'.

An "about"/"What's this?" link/page/text would be helpful, especially for the differences of the four 'search modes' (if that's what they are supposed to be - just guessing here, which illustrates my point ;)
HeO Send private email
Friday, June 20, 2008
 
 
Very interesting approach, and Im glad to see other people trying to attack search as Im convinced this Google world we are depend on is a flawed system in many ways. I really like the fact you have this simple concept of similarity that allows people to limit their searches. I know you can do allot of that in Google via their advanced search area. But to me, this is the fundamental flaw in most engines. Most people cannot consume the millions of selections returned in most queries, and to me, there is no use returning raw data like this, if its not usable. So, seems like you have a concept that starts to address that better.

I have a concept Im working on as well right now. Its based on the idea that web site owners need to have more control over how their sites are indexed, and what they choose to "categorize" their pages as. Right now, its it seems more of an authortarian search system where Google tells us all what we are worth, and how people get to find us. Whay Im working on is a concept that allows sites to not only control who links to them, but also control the people that come to them, who they in turn should go to for information.

I like the idea too that content is not tied to parsed keywords extracted from pages, but a sublevel index that content/website owners control and use to route people to their site (much like your basic hyperlink). It would essentially allow any website on the Web the ability to act like a mini-search engine, as well, and host scripts that link to other resources. The only way system like this can work is to free them from proprietary technologies and commercial sites that suck traffic using their protocol and system of choice (ie digg, facebook, yahoo, etc). I think if more people could adopt a more open mind to whats possible in search, like you are doing, we can eventually build a better web and free us from the "Google chains" that bind us.

Anyway....keep up your own efforts as in time, newer and better ways of search will be better for the Web as a whole.
Ranger
Sunday, June 22, 2008
 
 
Looks interesting. Would be nice to have results sorted by relevance..
Nir Send private email
Friday, June 27, 2008
 
 
"I like the idea too that content is not tied to parsed keywords extracted from pages, but a sublevel index that content/website owners control and use to route people to their site (much like your basic hyperlink)."

That will benefit only search engine spammers who will take control of users search results.  Many useful sites do no search optimization at all (I dont optimize my personal site that has many useful articles because I am too busy).
dd
Saturday, June 28, 2008
 
 
You have positioned the Blinking Caret in the middle of the text box. This is confusing, I was clicking the box and wondering why the caret was not appearing as I was looking at the left of the box.
dd
Saturday, June 28, 2008
 
 

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
 
Powered by FogBugz