Hacker News Reader

Comment by mahmud | Hacker News Reader

Yep. I never used to use search operators unless I was looking for something really specific. Nowadays I get completely irrelevant results and I am forced to quote strings and explicitly specify term precedence, conditionals and other regexy stuff.

Last night I was searching for the syntax of DEFTYPE when used with various types (i.e. MEMBER, SATISFIES, OR, etc.) and the #1 his for "deftype member" was the personal MySpace page of some guy.

I think they're optimizing for "social" results now.

View on HN

They'd probably do well to have a seperate 'technical search' for searching for things related to technical matters, eg programming languages, physics, chemistry, medicine, engineering, etc. And remove the casual stuff (facebook, myspace, pages that are clearly not about a technical subject, etc) from that index.

It would probably be a highly praised feature to seperate off a second index like that, as specifically searching for programming language concepts and documentation can be difficult (the C# and .NET problem).

Having separate indexes and separate searches for sub-domains of knowledge is an interesting idea. The original idea of PageRank was that each link constituted a 'vote' for a page. Perhaps different epistemic communities on the internet use links to mean different things. So, the meaning of links on technical webpages is slightly different from the meaning of links on social websites. Interesting idea to play around with.

It's also aggravating that quotes don't allow matching of specific symbols within the quotes.

And, yes, the results are different. I'd agree with you they seem enhanced (eg., classified results) and fresher for popular culture, but somewhat worse for domain specific queries.

Which is dangerous, since it seems like the biggest threat to Google has always been the 'vertical' search market. If they don't have a better way to narrow search results to a specific domain, they're going to run into issues from competitors.

Wait until they integrate twitter. It will be 99% real-time crap and 1% historical results... Google should just stay away from real time frenzy, or maybe fork the search engine to avoid disrupting the good results it has.

Can't agree more. Let's hope in the future we can choose to search for tech-specific results.

There's evidently a market for a tech-specific search engine..

Just out of interest, do you find lispdoc.com useful generally for this kind of thing?

Reply on news.ycombinator.com