Ask HN: do you feel Google search result quality has gone down?

"The Modbookish" is an inconsequential six member Ning social network, but doing a Google search on "Modbookish" is interesting. There are the annoying 'Google thinks you are stupid' hits like the site in which there is a sentence ending with "mod" followed by a sentence beginning with "Bookish". The remaining results are sites that link to the Modbookish. However, the Modbookish itself is NOT in the results. As Google ads are on the site, there is no doubt that the Google bot visited it shortly after it went online. Nonetheless, Google has refused to put the site into its web index for a couple months now. Clearly, (and not without cause), it is assuming that the pages that link to the site are link spam. It is probably waiting for some semi-authoritative link to appear before it indexes it.

So, why is this bad? It undermines Google search purpose: to provide the most relevant results. When Google got started, I recall that when Sergey or Larry were asked why Google search was better they often offered the Harvard example. If one enters the search term "Harvard" with no other terms into a web search box, there is really only one reasonable search result: the Harvard University website. However, in the search engines of the time, the university often ended up on page five or six. On Google, it was search result number one. By the same token, if you enter "Modbookish" (no spaces) there is only one reasonable result, and Google doesn't offer it. In contrast, both Duck-Duck-Go and Yahoo list it in the first spot.

Why is this really bad? Many (Most?) websites do not do their own site indexing; they let Google do it for them. Clearly if Google refuses to index a certain important but only occasionally referenced page, that is a problem.

Regarding the query [Linux asynchronous IO] returning older results, here's a tip. Above the search results click the "Show options" link to open up what we call "toolbelt" mode. From there, you can click to show only results from (say) the last year, or in a certain date range.

Toolbelt mode is really handy, e.g. if you search for a product, you can click "Show options" and then click the "Fewer shopping sites" link to get more reviews and manufacturer pages instead of comparison shopping sites.

That is not the entire purpose of SEO. There's good sites out there that don't provide their content in a way that can be indexed by spiders. SEO often solves that. There certainly are plenty of people making bad websites and trying to make them rank, it's Google's job to weed out the useless information.

People still collect links (shameless self promotion: http://internetmindmap.com ), they still have huge blogrolls, there are human powered search engines, a vast amount of directories for every imaginable niche...

Google doesn't need the perfect page of 100 links and I doubt it ever did.

Your mashed potatos example does not make sense. Google gave you generic info for your generic search query. How is that bad?

Now the fact that certain sites dominate a very wide range of search queries, is an interesting point. Personally, I would just add a sidebar or something similar, to be occupied by the "staple" sites, such as wikipedia, about.com etc.

Except I'd never find the obscure stuff that Google helps me find every day. I don't think a lot of links to manuals,mailing lists, etc show up on any of the sources you mention.