A rising tide lifts all boats; many alternative search engines have
started appending links to other search engines at the bottom of their
SERPs, to allow users to try their searches elsewhere if they're
unhappy. This could be a good opportunity to promote other
non-mainstream engines with their own indexes. Mojeek and Brave are two
such examples of engines with these links.
Marginalia, Teclis, Searchcode, Mwmbl, and Alexandria would be good
options; all of those except Teclis are FOSS, and all except Mwmbl run
without JS.
--
Seirdy
On Mon Jul 11, 2022 at 6:55 AM CEST, Rohan Kumar wrote:
> A rising tide lifts all boats; many alternative search engines have > started appending links to other search engines at the bottom of their > SERPs, to allow users to try their searches elsewhere if they're > unhappy. This could be a good opportunity to promote other > non-mainstream engines with their own indexes. Mojeek and Brave are two > such examples of engines with these links.>> Marginalia, Teclis, Searchcode, Mwmbl, and Alexandria would be good > options; all of those except Teclis are FOSS, and all except Mwmbl run > without JS.
I'm open to this. I think we should pick one at random and stick a link
to it at the end of the search results (<div class="alert">didn't find
what you were looking for? Try this search on <whatever>). Only FOSS
providers and let's also avoid recommending mwmbl until they have a
javascript-optional frontend.
On Mon, Jul 11, 2022 at 11:50:52AM +0200, Drew DeVault wrote:
>I'm open to this. I think we should pick one at random and stick a link>to it at the end of the search results (<div class="alert">didn't find>what you were looking for? Try this search on <whatever>). Only FOSS>providers and let's also avoid recommending mwmbl until they have a>javascript-optional frontend.
Semi-agreed; I do think that providing multiple links rather than a
single random one would be preferable, as it's a more established
convention and will increase the probability of users discovering a new
engine they like. I'll reach out to Daoud about progressive enhancement
for mwmbl; I do think it's worth keeping an eye on. I especially like
its distributed crawling of centrally-collected pages.
Of the engines I've collected at
https://seirdy.one/posts/2021/03/10/search-engines-with-own-indexes/,
these are all the engines that meet the criteria of FLOSS and usable
without JS:
- Gigablast (repo is a bit out of sync with the live engine, because
people were trying to game the ranking. But it is one of the OG
alternative engines and is worth a mention)
- https://www.alexandria.org/: powered by the Common Crawl!
- https://search.marginalia.nu/
- https://searchcode.com/ (source-available, with eventual-GPL)
- https://wiby.me/, https://wiby.org/ (just went open source!)
- https://searchmysite.net
- https://infinitysearch.co/ (source code:
https://gitlab.com/infinitysearch/infinity-decentralized)
- YaCy (many public instances, one is https://searchlab.eu/en/)
- https://www.yioop.com/
Another one is Lieu: https://lieu.cblgh.org/. It searches the excellent
Merveilles Webring, which I brought up earlier in chat. Also written in
Go; might be worth exploring some cross-pollination.
--
Seirdy
On Mon Jul 11, 2022 at 6:30 PM CEST, Rohan Kumar wrote:
> Semi-agreed; I do think that providing multiple links rather than a > single random one would be preferable, as it's a more established > convention and will increase the probability of users discovering a new > engine they like. I'll reach out to Daoud about progressive enhancement > for mwmbl; I do think it's worth keeping an eye on. I especially like > its distributed crawling of centrally-collected pages.
How about:
<div class="alert">
Didn't find what you were looking for? Try one of these search engines:
<div flex space-around>
<a href="whatever">whatever</a>
(bullet point)
...
</div>
</div>
With perhaps 4 or 5 different ones selected at random.
> - Gigablast (repo is a bit out of sync with the live engine, because> people were trying to game the ranking. But it is one of the OG> alternative engines and is worth a mention)> - https://www.alexandria.org/: powered by the Common Crawl!> - https://search.marginalia.nu/> - https://searchcode.com/ (source-available, with eventual-GPL)> - https://wiby.me/, https://wiby.org/ (just went open source!)> - https://searchmysite.net> - https://infinitysearch.co/ (source code:> https://gitlab.com/infinitysearch/infinity-decentralized)> - YaCy (many public instances, one is https://searchlab.eu/en/)> - https://www.yioop.com/>> Another one is Lieu: https://lieu.cblgh.org/. It searches the excellent > Merveilles Webring, which I brought up earlier in chat. Also written in > Go; might be worth exploring some cross-pollination.
I know that YaCy is not very good, so I would want to hear some kind of
judgement on quality for each of these before agreeing to include them.
NACK on source-available as well, and searchcode is not really useful
here -- should only include relatively general-purpose search engines.
On Mon, Jul 11, 2022 at 07:14:35PM +0200, Drew DeVault wrote:
>On Mon Jul 11, 2022 at 6:30 PM CEST, Rohan Kumar wrote:>> Semi-agreed; I do think that providing multiple links rather than a>> single random one would be preferable, as it's a more established>> convention and will increase the probability of users discovering a new>> engine they like. I'll reach out to Daoud about progressive enhancement>> for mwmbl; I do think it's worth keeping an eye on. I especially like>> its distributed crawling of centrally-collected pages.>>How about:>><div class="alert">> Didn't find what you were looking for? Try one of these search engines:>> <div flex space-around>> <a href="whatever">whatever</a>> (bullet point)> ...> </div>></div>>>With perhaps 4 or 5 different ones selected at random.
Sounds good. Might wanna use role="note" for the alert.
>I know that YaCy is not very good, so I would want to hear some kind of>judgement on quality for each of these before agreeing to include them.>NACK on source-available as well, and searchcode is not really useful>here -- should only include relatively general-purpose search engines.
Yeah, probably not a good idea to include those then.
--
Seirdy