HN Explorer

Ask HN: A specialized search engine to find domain name addresses

9 points by ThinkBeat ↗ HN
As with most things of I presume this already exists.

A majority of the start of my interaction with my browser is typing in something like "Nietzsche Cafe" because I want to go to their website and look at their menu. www.NietzscheCafe.fi

So what I need is a hit to their Url: "www.NietzscheCafe.fi"

Or I need look at the prices over at "Dostoevsky Used Cars" I search and what I want is the Url: www.DostoevskyHappyMotor.fi

I wish there was a "pre search engine" that could just find the Url. It could be implemented that if it failed to find a hit it would pass it on to Bing/Google etc.

That would cut my traffic to Google quite a bit over time.

So less tracking and less monetization

3 comments

[ 3.0 ms ] story [ 21.1 ms ] thread
danskeren
I love the idea, and strongly consider doing it with my search engine since I don't have access to any auto-suggest API at the moment.

That being said, not sure how I feel about "Dostoevsky Used Cars" displaying "www.DostoevskyHappyMotor.fi". I'm also curious to hear people's thoughts about whether to include all websites or only the most popular ones in each country, and how you would rank these websites. Which domains should appear when somebody searched "go"? You have go.com, go.co, godaddy.com, google.com, google.etc, go.dev, golang.org, and a million others, not to mention you probably want typo tolerance, e.g. for ogs.gg. Maybe you also want to display the most popular websites for the 'go board game' based on your dostoevsky used cars example. What about seo spam websites that are sometimes named after a popular search query?

ThinkBeat
I think the first step is to harvest domains.

Start with the domains that are by far most popular as a seed for the database.

Then I can see two probably not very good idea.

If the user does not get a hit on their query for domain, the search engine would query bing/google/or other and present a list of possible values.

Then the engine would learn.

It could also be a browser plug in that records domains and search terms used.

There is no reason to record who sends it just what the values were.

Given a lot of people running it across different countries it would slowly build an index.

Since this is a privacy violation it would probably not be popular.

I wonder how long of a list of most popular domains I can find.