I've been building Jive Search. There's already tons of competitors in search (Google, Bing, etc.). There's also DuckDuckGo, StartPage, etc. that claim to be privacy-focused but are just as opaque as Google. The niche I'm targeting is users that understand that for true privacy in search you have to be transparent and open source so that users can view the code and run it on their own if they want.
If you click on their "Sources" link it shows a lot of links but then has this note: "This list is intended as a guide to sources of further information. The inclusion of an item in this list does not necessarily mean that its content was used as the basis for any specific Wolfram|Alpha result."
Basically trying to obfuscate where they are really getting the info.
Been thinking more on this...I think I have an idea how to do it: Wikidata has the USDA ID for a lot (but not all) of items. From there we can get the rest.
I run https://jivesearch.com, so I might be biased. The single biggest thing they can do to improve their product is open source everything. They've instead decided to move away from that by freezing their instant answers with no explanation as to if/when they will re-open source them. Makes me question their commitment to privacy and transparency.
Personally, I love that DuckDuckGo gives me a usable alternative to Google that is privacy focused. It might be better if they were open source, but I'll take what I can get from the industry regarding advertising practices at this point.
I just DDG and have found !bang shortcuts that are broken or 404 for weeks. I've tried reporting the problems on their feedback forms or Reddit, but never got any responses. If it was open source, I could have submitted a pull request to fix them. I'm surprised they don't have automatic heartbeat queries with expected results for every bang shortcut, at least.
Is there a way to go straight to the first result using Jive Search? With DuckDuckGo searching for "example !" does this. (This is the main reason I don't use Searx.)
You can make the ranking configurable. Once we get our own index built we will go back to doing this. The way we used to do it was by having a PostgreSQL function that is customizable once you get everything set up...I'm sure there are other ways to do it as well.
just open a guest window from chrome, and type the same search terms in the two different search engines, and you would see different results. it's not hard to verify.
In fact, DuckDuck Go gets its results from over four hundred sources. These include hundreds of vertical sources delivering niche Instant Answers, DuckDuckBot (our crawler) and crowd-sourced sites (like Wikipedia, stored in our answer indexes). We also of course have more traditional links in the search results, which we also source from a variety of partners, including Oath (formerly Yahoo) and Bing.
We continue to partner with more and more sources to bring you the best Instant Answers from the best sources. Our focus is on synthesizing it all together into a superior search experience.
I suppose that's quite different than 'DDG is a frontend to Bing results'?
It depends. For me, a search engine is about the organic search results and they get all their organic search results from Bing, Oath and Yandex. The 400 sources they talk about are used for Instant Answers, boxes and stuff like that.
> We also of course have more traditional links in the search results, which we also source from a variety of partners, including Oath (formerly Yahoo) and Bing.
Really... You read the last part of that paragraph, but no the first?
>In fact, DuckDuckGo gets its results from over four hundred sources. These include hundreds of vertical sources delivering niche Instant Answers, DuckDuckBot (our crawler) and crowd-sourced sites (like Wikipedia, stored in our answer indexes).
Kinda silly to have a crawler that doesn't index anything.
They have an index but just for their Instant Answers (for which they have 400 sources) and not the organic search results. The last paragraph is what's relevant to the actual search index.
To clarify, we haven't closed anything that was previously open, and all open source Instant Answers will remain open.
For now, we've paused community contributions for new Instant Answers and non-essential changes, until we can find a better way to work with the community and their contributions.
As most business's their internal code is probably a giant ball of mud that runs on caffeine and poor amounts of sleep. It works, but it probably wouldn't be reasonably releasable without refactoring and reworking a lot of the source so that it's componetized in some way.
Personally I just want DDG to make a search API that is free and doesn't require API keys. That way I can setup search in mobile or desktop apps and have it automagically be searching through specific sites like Soundcloud and Youtube and such.
Like if they're reading this, DDG is very unique and if they were to focus upon reducing costs, making efficient software, and better algorithms, DDG could be the service we shove into apps.
Their selling point is trustworthiness, privacy and usefulness. To be useful, they must have quality software/IP. And in their opinion, that should be protected.
If the code was AGPL'd and included contributions for which DDG didn't own the copyright (so DDG would be using the code under the terms of the AGPL), they'd be breaking the licence terms if they failed to disclose any changes.
It wouldn't be proof, but it would be a fairly strong commitment.
I can vouch for that. We recently made a big improvement to our search results at Jive Search and have seen better traffic. People will set it as their default more if it's better.
The !bangs aren't too hard to implement. It's basically just a very large config file. You can see that in our code: https://github.com/jivesearch/jivesearch. I run the project, btw. And, yes, we have that same !bang as well as all of DDG's !bangs.
Beyond what you pointed out, it's hard to really know what's going on with something closed source. Disclosure: I run jivesearch.com (basically, an open source version of DDG). FYI, we've got an open issue on GitHub to add POST requests in addition to a proxy.
