Hacker News Reader

Comment by noelwelsh | Hacker News Reader

Neither side made their expectations clear at the beginning. I can't find either blameless. I also don't believe that engineering is the most important factor for the kind of product being described. For what the author calls "secrets", yes. Early Google was just so much better than the competition that it won. However, a lot of tech history shows that marketing trumps tech in many cases. Windows wasn't technically superior to OS/2 for example.

View on HN

Honestly, the OP is trying to be a bit ingenious trying to push the example of PageRank into the argument. Google's core advantage is its tech. The PageRank algorithm was so much better than anyone else on the market that it was bound to win.

In this case, there is no real innovative, patent-worthy tech. It's just an idea with some very basic tech any developer worth his salt can put together in months if not weeks.

I see this all the time: developers confuse pure tech companies (like Google or SpaceX) with tech-supported companies.

Oh no, again this PageRank myth.

Google won, because it was fastest and had good links.

Fastest was more important than better links. How did they get faster than their competitors? They bought cheap servers all over to beat the transportation times to the clients. Also their page was smaller, not overloaded, so the results could be presented faster.

Their bot was also more aggressive. A more aggressive bot contributes more to the link quality than the algorithm, because you get deeper and new stuff more timely. People are searching for new stuff.

That's how Google won. Not because of PageRank alone. PageRank was a contributing factor. But renting out cheap servers in every datacenter out there and keeping the page small and fast and dealing with the consequences of cheap servers (HD fails, fallbacks, ...) was more important.

Correct me if I'm wrong, but didn't they start off with custom servers? I read somewhere that Page made a server stand out of Lego...

Speed was certainly a factor. I remember I used to use a desktop application[1] for web searching, and after using a little while always noticed how the results from google were returning about 30 seconds before any of the other results.

It wasn't long before I dropped that application and just used google, I wasn't waiting around for any of the other results anyway.

[1] Copernic - Long since pivoted into desktop search

Eh, I remember the early days of Google, their results really were a LOT better than the existing competition (AltaVista? other?). Whether that was mostly due to the PageRank 'secret' or not, I couldn't say (and we all know at this point Google's relevance algorithms are orders of magnitude more complex than a 'PageRank secret'), but people didn't just start using Google because it was fast, but because they found what they were looking for much better than in existing solutions. In my memory.

That's absolutely correct. AltaVista wasn't slow (until they filled their page with garbage; they were originally as minimal as Google), but you'd have to hunt through pages and pages and pages of results. Google usually found what you were looking for right away. This is what I remember from 1999 - I had been using AltaVista as my primary search engine for several years, and once I found Google I started using it almost exclusively because it was so much better.

Another reason Google won was because their minimal page. People like me, who was managing large corp networks in the latter half of the 90s set Google as the default home page for all machines because yahoo was (is) a hideous pile of shit as a main landing page.

This helped get users used to google as it was the first page on every machine I had control of as it manager...

I think you missed the point. He said that the starting idea was not important, and made examples of actually important starting ideas (page rank). His point is that the execution mattered way more, given that the idea was not really mind blowing.

"Windows wasn't technically superior to OS/2 for example."

Windows required less resources to run and that made it "superior" in mass availability.

Reply on news.ycombinator.com