New experimental search engine

[Andrew T.]

Hi to anyone reading this group.

I've been playing with a concept for a radicaly different design for a
search engine (I kept finding certain searches near impossible on
DDG/Google/Bing/etc. because of SEO/AI/just plain dumbness).

Anyway, I've called this search engine TagSearch, and the URL for this
search engine is http://search.turnquist.name/ .

The basic idea is that it prioritizes indexed keywords in searches.
Currently full-text is not indexed, but instead using the <meta
keywords> tag as most important, plus <title> and <h?> tags.

But in addition, anyone can create an account and, once logged in, can
add their own weighted tags to a search result, moving the result
higher or lower on their results list. Those tags are also aggregated
into a pseudo-user that affects the weighting in general (not logged
in) searches, essentially a crowd-edited search. With enough users,
inappropriate SEO manipulations of the results should be offset by
others tagging that result down (or up).

URLs can be submitted, but are currently limited to one submission per
hour to make sure nothing goes haywire and unintentionally hammers
someone's server.

There's still a bit of work to do, but I'd be interested in receiving
feedback.

--Andrew

--
Andrew Turnquist, Short Tract, New York, USA (USDA Zone 5)
(Remove numbers and .invalid for email address)
"Do what you can with what you have where you are." -T Roosevelt

[Eli the Bearded]

In comp.infosystems.search, Andrew T. <and...@turnquist.name.invalid> wrote:
> I've been playing with a concept for a radicaly different design for a
> search engine (I kept finding certain searches near impossible on
> DDG/Google/Bing/etc. because of SEO/AI/just plain dumbness).

What sort of searches cause you problems and you think will be solved by
this? (I ask for reasons explained below.)

> Anyway, I've called this search engine TagSearch, and the URL for this
> search engine is http://search.turnquist.name/ .

Firefox tells me "The certificate for search.turnquist.name expired on
12/29/2019."

> The basic idea is that it prioritizes indexed keywords in searches.
> Currently full-text is not indexed, but instead using the <meta
> keywords> tag as most important, plus <title> and <h?> tags.

To me, meta keywords is 100% for SEO, which is why I wonder about how
this improves matters. Title is sometimes good and sometimes junk. The
heading tags I really have never investigated.

> in) searches, essentially a crowd-edited search. With enough users,
> inappropriate SEO manipulations of the results should be offset by
> others tagging that result down (or up).

You might be surprised how much SEO manipulators are willing to do.

> There's still a bit of work to do, but I'd be interested in receiving
> feedback.

It works in Lynx, that's good.

It didn't give me any useful results for the two queries I tried,
one for Perl and one for MongoDB. Both are documenation searches I did
recently. It could well be that your concept is fine, but the
appropriate pages have never been indexed. What sort of queries is this
ready to handle? It will help others evaluate it.

Elijah
------
usually uses DDG these days