Hacker News new | ask | show | jobs
by Waterluvian 2248 days ago
I wanted to love DDG. It promises a lot. And it delivers a lot. I love the shortcut codes.

But after a few months I had to go back to Google. Two main reasons:

1. Image searches could often return literally zero results for simple terms.

2. I'd look up a term and get non-English Wikipedia articles before the English ones, which often wouldn't even show up.

6 comments

IMO it's not about "going back to" anywhere. It's not as if we couldn't use different search engines for different queries.

This all-or-nothing mentality of migrating away from Google (the search engine) seems flawed to me. It's still better for your privacy doing half of your queries using Google, instead of doing all of them.

By the way, no offense intended! We're all free to pick our tools.

I partly agree. I don't want it to be all or nothing. But the way browsers are set up with the "default search", it somwhat is.

Ooooh I would love to be able to custom route searches depending on rules.

DDG ends up working well as the default search for me. The ability to type my query and then decide if I want to use DDG or Google doesn't require me to take my hands of the keyboard. I just add a g! and I'm switched off my default. So it's a great default with an ergonomic way to switch to a different engine.

Google as default requires me to use the browser UX to change the engine, which I believe requires using the mouse or hitting tab or arrow keys. Less ergonomic.

Firefox supports this, go into Settings -> Search and set a keyword for the search engine.

For example, if you set the keyword for DuckDuckGo to "ddg" and Google to "go", then typing "ddg <your query>" into the topbar will search DuckDuckGo, and typing "go <your query>" will search Google.

DuckDuckGo itself supports this. Just start a query with !g to have it routed to Google. For example, "!g search on Google" would bring you to the Google search results page for "search on Google."

DuckDuckGo supports several other commands like this, which they refer to as "bangs."

https://duckduckgo.com/bang

The difference between the Firefox setting described by the GP and the DDG bang commands you mentioned is that the Firefox setting is close to zero latency since it’s handled by the browser. The bang commands need to go to DDG and then come back as a redirect, taking a few seconds more. On the other hand, the DDG bang commands work the same across other browsers too.
A query need not be started with a bang. Just add 'g!' or 'yt!' anywhere in your search terms and it should work fine.
Just as an aside, "bang" has long also been a term for "!" itself - hence the two characters starting shell scripts, "#!", being called "hashbang".
You can set a keyword for any bookmark. Firefox replaces %s with your query. Simply bookmark the search result page of any website, replace the query with %s and set the keyword for the bookmark.

Doesn't have to be search either. For instance, I use it for the nodejs docs, and npm packages too: "https://nodejs.org/api/%s.html" and "https://www.npmjs.com/package/%s". So I can type "node fs" open the fs docs for node, or "npm fs-extra" to look up the docs for "fs-extra" package on npm.

Even easier, Firefox has a right click menu option for form fields it think might be searches to add the search keyword directly.
I've set Firefox its default search engine to DDG. Then if the search results are unsatisfying, I search again but now prepend the search with !g for Google, or !gi for Google Images.
Fwiw, you don't need to prepend it; you can put it anywhere in the query
Oh really? Especially on mobile, that's useful to know, thanks!
> Non-English Wikipedia articles before the English ones

It was only recently that I discovered that each language’s version of Wikipedia is independently edited, rather than being a translated reflection of a canonical source material; and, therefore, that inevitably there will be “better” or “worse” versions of a page (i.e. more/less content, more/less fact-checking, etc.) depending on the language.

This gulf sometimes turns out to be so large, that it’s sometimes more informative to read a foreign-language Wikipedia article through machine-translation, rather than reading the one in your own language!

(I recall, in the recent HN discussion that linked to https://en.wikipedia.org/wiki/Musth, people were pointing out that the German-language article carried a lot more information than the English-language article. That specific discrepancy has probably been since fixed up care of HN readers themselves, but it was eye-opening, especially since the English-language version of the article was phrased in decisive terms like “scientists don’t know X” where the German article instead says “X is caused by Y” citing enough [German-language] studies to thoroughly prove its point.)

I wonder if DDG has just caught onto this trend, and is prioritizing the language of the article that has the most editorial activity. (It would actually be more work to do the opposite, now that I think about it—in raw PageRank terms, there’s always going to be a most-linked-to language-version of a Wikipedia article, and doing nothing means that that version simply floats to the top. Google et al must be doing extra work to group the “same” articles of different language-versions together, assigning them the PageRank of the highest-ranked one in the grouping, while rendering out the link+summary as that of the group-member corresponding to your own language.)

> I wonder if DDG has just caught onto this trend, and is prioritizing the language of the article that has the most editorial activity.

It should prioritize the language of the query. I'm supposing the commenter above searched for something in English and got a non-English result?

Yeah I was looking up some very straightforward things and I'd end up with non-English wiki pages on the first page, and the English wiki page for the same topic was... I guess maybe on later pages.
Are you using a VPN? I've never seen a foreign language wiki article show up, and I've been using DDG for a couple of years.
Oh my. You know what. I might have been on a VPN.
A killer feature on Hackernews would be the ability to just point to a common thread like wiki that people could update; so we don't have to scroll through the comments on how people love DDG, but it just doesn't have good results. So instead of every post having a huge comment section on this restating the entire thing over and over and over, we could just link to the wiki that has the discussion on duck duck go.
It's pretty easy to append "!g" to your ddg search results if they're inadequate.
Yeah, this is so engrained in me now, I find it hard to search without a "second opinion" on devices that default to just Google.
For me the quality of DDG search results seems to fluctuate over time. In the recent times it’s become worse, and I’ve resorted to going directly to !s (startpage) or !g (google). I take many other precautions in my browsing to thwart tracking anyway.
Translation: you prioritized convenience over privacy.

That's your choice to make and I'm not criticizing your choice, but I do think it's a bit off-base to criticize DDG for this. There's a pretty inherent tradeoff: violating privacy is should allow Google to produce better, more personalized search results if they are at all competent.

I would guess that the vast majority of DDG users don't use DDG because they think it produces better search results. We're using DDG because we're concerned about our own privacy and/or the implications of letting a corporation gather very personal data on every person in our society.