2 ms·

Article is written by armchair pundit with zero internal experience but who thinks he knows how Google works better than the person who ran the company for a de

by sytelus 12y ago

Article is written by armchair pundit with zero internal experience but who thinks he knows how Google works better than the person who ran the company for a decade. Title is obviously a link bait to invite flame wars. I guess he needs to pay his bills by writing these kind of junk. The guy doesn't understand the first thing about experimentation: It's normal to fail 100s of times before you can have blockbuster success. Also people who think search business is just web search have no clue about what search business is. In today's world, if you want to build a search engine its not enough to index the web. You need spell checker, deep links, answers, image search, map search, local business search, video search, controlling presence on mobile. You can't be a one-hit wonder, you need to excel at all of these areas if you want to be search business.

threeseed 12y ago
> want to build a search engine its not enough to index the web Yes it is. People by and large are just searching for sites. Spell checking is easy, deep linking is easy enough to get something basic in place, answers/images/video are hardly mandatory. And business/map search I can get by buying data the same way Google did. Google's search engine is better than Bing for example because its index is larger and its core search algorithm better. Everything else are all "nice to haves".
- sytelus 12y ago
  I can tell you don't have much experience working in search engines them. Search engines are all about tail. Even if you had just as good algos and index as Google, you likely won't have any chance to take market share because you will fail in, for example, local queries which constitutes 10-12% of all queries. Someone entering USPS tracking in search box won't see delivery time. Someone else entering flight number won't see its status. A kid entering 100+200 won't see answer. And so on. All these things matters. When comparing things people don't see what works, they see what doesn't work. Also what you say "easy", for example deep links, are typically significant multi-year large team efforts with large number of open questions and very active academic research.

by sytelus 12y ago

threeseed 12y ago

> want to build a search engine its not enough to index the web Yes it is. People by and large are just searching for sites. Spell checking is easy, deep linking is easy enough to get something basic in place, answers/images/video are hardly mandatory. And business/map search I can get by buying data the same way Google did. Google's search engine is better than Bing for example because its index is larger and its core search algorithm better. Everything else are all "nice to haves".

sytelus 12y ago
I can tell you don't have much experience working in search engines them. Search engines are all about tail. Even if you had just as good algos and index as Google, you likely won't have any chance to take market share because you will fail in, for example, local queries which constitutes 10-12% of all queries. Someone entering USPS tracking in search box won't see delivery time. Someone else entering flight number won't see its status. A kid entering 100+200 won't see answer. And so on. All these things matters. When comparing things people don't see what works, they see what doesn't work. Also what you say "easy", for example deep links, are typically significant multi-year large team efforts with large number of open questions and very active academic research.