I don't think that discovery without algorithms is possible, because the Web is essentially unstructured. Any means of discovery needs a way to organize all of that information, and then present it in a relevant way. People forget that Google was actually good at this.

The problem isn't algorithms per se, but how those algorithms are implemented. Unfortunately, people coming up with alternatives tend to lean too far in the other direction - we have alternative search engines designed to exclude all sites using Javascript, for instance, which cater to people who don't want to interact with any part of the modern web, but we don't have an alternative that does what Google used to do before search became big business and simply attempt to catalog the entire web (including the parts that HN hates) and display relevant results to the end user.

▲

carlosjobim 3 days ago | parent [-]

How about hyperlinks?

▲

krapp 3 days ago | parent [-]

Hyperlinks alone don't work for discovery as a general problem, and on a smaller scale they just push the problem up a level because you still need to discover the hyperlink. The web is just too vast and complex and they don't contain necessary semantic information. This is why the "web portal" failed and gave way to search engines, which were objectively better for the task.

You need another level of abstraction over the web to make discovery work, as you do with any application to make data useful. It's little different than making queries in a database.

▲

carlosjobim 3 days ago | parent [-]

I find that hyperlinks work very well currently. You read something of interest on one page, and discover more information through hyperlinks to other pages. That also works great for books. Almost every book I read mentions other related books within the content, and from there you discover more and more.

▲

krapp 3 days ago | parent [-]

Hyperlinks don't help for discovering that page to begin with, or related material not specifically being linked to from that page.

Given that you're on Hacker News, it's likely your use case for the web isn't typical.

▲

carlosjobim 3 days ago | parent [-]

Well how would you discover Google? Everybody starts somewhere

▲

krapp 3 days ago | parent [-]

It's much easier to discover Google once and let it do the rest of the work than it is to discover an arbitrary page and manually follow an arbitrary chain of hyperlinks across an arbitrary set of other pages hoping they just happen to lead to what you want.

▲

carlosjobim 3 days ago | parent [-]

Yes, but I understood the discussion about discoverability to be in the perspective of the webmaster. And in those cases, if you have published good information, other websites can link to you for discoverability. That's how an old fashioned web can be built: Pages linking to other good pages. I see it all the time.

And let's not forget that if you have high quality content and submit your pages to Google, they will put you very high in the organic search results.

	▲	krapp 3 days ago \| parent [-]
		>Yes, but I understood the discussion about discoverability to be in the perspective of the webmaster. Fair enough, but I think the problem is the same regardless of the perspective. You as a webmaster wanting to be discovered and me as a user wanting to discover you implies an optimal interface between the two to facilitate that discovery. Unfortunately, the web is no longer old fashioned. Most links are being posted to social media platforms or link aggregators like HN and Reddit. Most pages only link to other pages if it helps their SEO. And unfortunately Google no longer ranks content based on quality because they sell rank space and because search is no longer about discovery so much as it is advertising. Which is why I think we do need search engines and algorithms, but we need the kind of search engines and algorithms we had when Google started and before they monetized search and before SEO ate everything when it actually did surface relevant content based on organic links. I'm not arguing that hyperlinks don't solve the problem of discovery, just that they only solve it at a small scale, and we need a larger scale solution as well.