Bromite: Add Searx search engine

[ghost]

Add searx.me search engine

[csagan5]

I have declined this request in the past as I thought that the engine would use other random users' engine instances to work. That is not the case, I understand, but rather there are public instances.

The main searx.me instance is one of these; who is running the instance? Are we really going to add a search engine without knowing who runs it?

These are my concerns, that I would like to be cleared before adding the engine.

[ghost]

Searx is fully open source metasearch engine https://github.com/asciimoo/searx
Searx.me ("the default one") is owned/maintenanced by asciimoo https://github.com/asciimoo (creator of searx). Searx.me don't log anything. Check this https://github.com/asciimoo/searx/issues/641 . It's way more private than duckduckgo or startpage. It's also listed on https://www.privacytools.io/#search

[csagan5]

Searx.me ("the default one") is owned/maintenanced by asciimoo

Source? Where can you see this?

[ghost]

https://github.com/asciimoo/searx/issues/641

[csagan5]

Yeah, okay. I do not trust a search engine run by a single person, sorry :) all the others mentioned have a company/organisation behind it.

It would be different if for example it were an organisation running the instance. If you can find a public instance run by a reputable organisation then I might look at this again.

For this reason I will decline this feature request, sorry. But feel free to keep the discussion up on this issue.

[ghost]

It's open source and for example the DuckDuckGo isn't. And if you use another instance it's not as private than using searx.me because searx.me it's more widely used.

[csagan5]

It's open source and for example the DuckDuckGo isn't.

Open source does not mean anything in this case. You cannot "open source" a server instance. What I am talking about is accountability not which programs make up the search engine.
All the search engines added to Bromite have a reputable company/organisation behind them (with the exception of Google that is there for practical reasons, accountable but not privacy-friendly). If any of the others have a privacy mishap they will likely close down so they have an incentive to not do so. An individual might have his/her account hacked and other motives.

As you can see here: https://directory.fsf.org/wiki/Searx

There is no specific public instance being endorsed.

To reiterate: if you find an instance run by a reputable organisation, then the point about accountability would be addressed.

[dimqua]

An individual might have his/her account hacked and other motives.

Isn't the same true for an organisations too?

instance run by a reputable organisation

• search.disroot.org running by Disroot;
• framabee.org running by Framasoft.

[csagan5]

Isn't the same true for an organisations too?

No, a reputable organisation is still better than a reputable individual. I am setting the bar for Bromite this way, sorry. This project is about the opposite of trusting indidivuals/companies so if we really have to trust something, may it be a reputable FOSS organisation not an individual.

search.disroot.org running by Disroot;

They are not very big so the service might be flaky, but I will contact them to ask if they agree to have the search engine added in Bromite.

[ghost]

Any news?

[csagan5]

Any news?

Yes, I recently received a reply. Disroot is run by two volunteers, they do not have a problem with me listing their engine but I still don't think such an organisation is good enough to be listed there (and I am not going to list all Searx instances either).

[ghost]

Yes, I recently received a reply. Disroot is run by two volunteers, they do not have a problem with me listing their engine but I still don't think such an organisation is good enough to be listed there (and I am not going to list all Searx instances either).

Ok, thanks for answering.

[Primokorn]

@csagan5 https://searx.laquadrature.net/
La Quadrature du Net is a well known french organization.
Who they are

[csagan5]

@Primokorn I have contacted them yesterday, waiting for a reply.

Today I browsed a bit through the various tabs of their Searx instance, some of them are broken and contain errors. I think this is a sign of a not well maintained instance, or just the upstream is broken as well.

The bigger problem here is: why doesn't the search engine appear as an available choice after making a search? This works for other search engines I tried.

I suggest addressing this upstream and then you can add any Searx instance you like in Chromium/Bromite.

[ghost]

it would be very useful for me if it were possible to use Searx in bromite 😬

[csagan5]

@speednroll as expressed in last comment you should ask upstream to Searx developers why it cannot be auto-detected in Chromium; once that is fixed, you can use it also in Bromite.

[Technically-Alexander]

It seems as though searX will detect user agent as either Chrome or Chromium now, and will automatically add "/search" to the end of the URL, and it switches from POST form submission to GET form submission. Chrom/e/ium doesn't accept POST requests it seems, and this workaround is required. It is working for me now on Chrome in Windows, Chromium in Linux, and Ungoogled Chromium in both.
More about it here: https://github.com/asciimoo/searx/issues/768
And here: https://github.com/asciimoo/searx/issues/768

I recall not too long ago it wasn't working to auto-detect though, and this post being from early 2017 kind of hints that it wasn't working.
It seems so unstable that any search term from any Chrom/e/ium based browser that uses the term "downloader" anywhere in the search phrase returns "rate limit exceeded" error page, regardless of the amount of requests submitted. I have submitted this issue to their github here:
https://github.com/asciimoo/searx/issues/1422
Between the inconsistencies and other points noted, and the POST/GET and Chrom/e/ium issues make me feel like searx is a bit unstable. For example: https://github.com/asciimoo/searx/issues/1287 -- closed in one day without providing the OP adequate time to reply with more details that were requested. Simply, it is called a feature and closed. Doesn't inspire confidence when inconsistency with the search engine interacting with Chrom/e/ium is already known and established.
Hopefully my more accurate bug report (compared to past reports on the Rate Limit Exceeded feature) will bring attention to this and lead to stabilization of the search engine in searX
It would be nice to see it in Bromite, but given all the points above, I can't blame you for leaving it out, not being reliable enough... yet.

[csagan5]

@Technically-Alexander thanks for the information you have dug out; it is not added because there is no company/organisation behind it (searx.me) and thus different (privacy) legislation apply to individuals.

Regardless, there are plenty of decentralised instances and it should work like any other browser/portal by supporting OpenSearch.

[smnthermes]

Disroot has now six volunteers: https://disroot.org/en/blog/welcome-the-new-team-members

[Nemris]

@csagan The instance at https://search.privacytools.io is ran by PrivacyTools.io. Is it a sufficiently reputable organization?

[csagan5]

@DMSalesman as explained in https://github.com/bromite/bromite/wiki/SearchEngines you need to ask Searx developers to fix the OpenSearch support (or fix it on your own and submit a patch upstream); I am not going to add any Searx instance in particular, they all are poorly maintained or give some error (even the one you proposed).

[pixelprofile]

Maybe allow people to add their own search engines if need be but put a disclaimer saying that it's not recommended?

Also by your logic, people shouldn't trust Bromite as it's primarily maintained by one individual. Unless the binaries can be easily decompiled to ensure that the code used to compile them is the same as the source code posted to the GitHub. Most people will not compile manually, and even if they do, most of those people probably won't go through and audit much of the code for potential privacy issues (not saying that your program has privacy issues, but when applying the logic you used in a general context, it quickly falls apart).

[csagan5]

@pixelprofile read the comment above yours and the linked wiki page. You can already add your own search engine, it is literally explained there.

Also by your logic, people shouldn't trust Bromite as it's primarily maintained by one individual. Unless the binaries can be easily decompiled to ensure that the code used to compile them is the same as the source code posted to the GitHub. Most people will not compile manually, and even if they do, most of those people probably won't go through and audit much of the code for potential privacy issues (not saying that your program has privacy issues, but when applying the logic you used in a general context, it quickly falls apart).

That is correct, you are responsible for verifying the software you run and under no circumstance you have ever been exempted from this. Your comparison does not hold because Bromite is not a search engine and thus there is no server-side data to collect. In the case of a search engine you need to trust that data is not being collected/sold but the data must pass through the server of the search engine for search to work; in the case of a browser no data must pass through a server (of the browser company/provider) in order for the browsing to work.

[brightersun99]

searx is total shit