HN Explorer

Ask HN: What do you think a successor to Google Search looks like?

7 points by igammarays ↗ HN
Thought Experiment: let's rethink the search engine. How does a small team of engineers, with tiny compute resources, build a new kind of information-finder?

The old model of centralized crawling/link/text-pattern-matching (a.k.a. Google) seems so 1998. So arbitrary, so open to manipulation (e.g. the Boris Johnson story today). I want something fundamentally different, something that takes into the account the sheer size of the internet today. Something not manipulatable by "SEO".

15 comments

[ 3.0 ms ] story [ 47.2 ms ] thread
tinytrader
What if your search engine was just made up of sites you add to a personalised "pool" of sites. This way your news comes from a consistent source and/or your technical questions are answered by your favourite dev community site, etc. Certain sites could be tagged for relevancy around a certain search keyword.

That's my two cents :)

igammarays
I like this idea. You can also start with a pre-populated pool of curated sites for particular topics. You choose these seed pools based on your trust of the curator, which could be a university research department, for example.
freediver
Sounds like a bad idea. It more describes a news app than a search engine (which should be able to find info about any string of characters you input into it, including obscure searches that you enter now and then)
tinytrader
I hear you - but a question I have would be: How often do you use the rest of the search results Google shows you? 9.9/10 out I'm using results from the first page. Would it be bad to curate the sources of that initial search, and disregard the rest?
freediver
The point I am making is the web is huge and your curated list likely to be small and not contain any information about 'tail of search' kind of query.
hans1729
Uh, you might want to look into the disciple of Information Retrieval.

I'm confused by your question: what exactly do you want? It seems like the SEO-story of Johnson caught you off guard, and now you're like "Google search bad". Then: well, what exactly do you know about search algorithms? What better product do you want to develop? Asking HN before you learned anything about Information Retrieval (because that's what it feels like to me) seems inadequate.

igammarays
I don’t know what I want. I just know that I don’t like the text-matching, machine-learning path that Google has taken. The internet can no longer be distilled into 10 blue links. Maybe.

Thanks for the suggestion about reading up on information retrieval. But maybe the point of this exercise is NOT to use well established methods. I want to rethink the interface and mechanics of search.

ToFab123
I don't think it looks like anything. It will sound like something. Think star trek computer.
pi-squared
Think what problem you are facing with the web right now and how you would solve it. Don't solve a problem for a thousand or a million people because all of them will have different problems. Start with you. Do you not like the interface? Do some css play, mockups on gimp or pen and paper.

What are your queries usually? Take a look at your own Google history, find and try to group what you are looking for and what would you ideally find that Google/other search engines don't satisfy.

I understand your desire to do something "revolutionary". But most things are incremental.

dorait
The current search engine has very limited context (your profile and the sites you visited and the searches you have done).

What if the next incremental improvement is asking you questions about the search to get better context before actually doing the search and giving you a smaller set of usable results?

Just dreaming :)

QueridoGuy
I don’t think it’s easy to make such successor. However, I think the best approach into making a better search engine is to start with meta search engine, and collect analytical results of users decisions. At the same time, that search engine needs to develop its own. If you stay with meta engine, not many is going to find it authentic, and it would be a cluster. However when you start as meta engine then you implement your own engine using it, that’s when users will be comfortable in using
freediver
Re meta not being authentic - would you say the same is true for DuckDuckGo?
QueridoGuy
I guess the less harsh word that I wanted to say is reliable.

DDG has it’s own community. It’s growing, but I don’t think it will catch up with google. There are other factors too.

As for authenticity. I think people interpret DDG as unwanted, unfortunately. Because it was frequently set as home page as a bundle. If you search its name, you will find people frequently asking “is it a virus”.

It’s hard to advertise an engine these days without being viewed as a forced advertising. But there are better tactics like offering incentives.

[dead] ronaluma
creating a successor of google.com is difficult but not impossible considering that it is something it hasn't changed in years and that people are more and more concerned regarding privacy. I've started to build something a couple of years ago with proxycrawl.com as I had to crawl other sites first, then my idea was to relate the pages with other searches from users with a rating system. It's complex but I decided to abandon the project and dedicate to something else