[1] https://news.ycombinator.com/item?id=13284917
Most of the time I run the same search on Google the results are very similar for the top hits. But Google does seem to produce helpful results that are not exactly what I searched for more often.
DDG also skews heavily to US results which is a pain. Any Amazon links are always to .com and not the regional site. Just a general understanding of region specific results seems to be missing, and I end up qualifying searches with the country or town.
I'm a Dutchman living in Germany speaking English at home and at work, so I switch between languages quite a lot: Dutch when I want to know local things (e.g. European laws, or recipes with ingredients that stores here actually carry), German when I need to know something like filing taxes or when trash is being picked up (I don't speak German yet, so I only do this when necessary), and English for everything else. Trying to do that on Google is nearly impossible. It somewhat picks up on the language of your query (certainly better than DDG picks up on that), but for English queries I'll still get a few German results, even after I click the "Change to English?" prompt (and I have to click every time, since I do not store cookies for Google). Like, thanks for this German forum thread when looking for an error message after I already set it to English... In DDG you can just flip a switch.
I've often found it annoyingly hard to get Google products back to a language I understand while traveling (no, the "append `&hl=en` to the URL" trick doesn't always work...)
Is it self sabotaging to not drive a Mercedes (or Audi or Tesla or what you think is best).
Or is it a choice you made to go down a notch in quality to gain something else?
Edit: besides I originally moved because I was really annoyed that Google couldn't respect my searches, so I didn't feel I lost much.
Sadly lately DDG has been copying Googles bad habit of fuzzing my searches to death.
DDG also skews heavily to US results which is one reason I find DDG useful.
I have been largely using DDG for years. What finally drove me away from Google completely is that you cannot get search results that are untainted by localization anymore. Whenever I search for programming related topics or the subjects typically discussed on HN I always get a some confusing local results no matter what language and region settings I use. https://google.com/ncr which did a good job for years doesn't seem to work anymore.
Really the only reason to refer to Google nowadays is if I deliberately search local places, organizations and businesses - this is where Google is unbeaten in my opinion. Even local news and events isn't usually worth the g! switch. These two might be good in the US but here in Germany all search engines are equally bad at it.
It's basically replaced my TI-89 for years, I hate doing calculations with units by hand and it's super error prone to not include them in the calculations.
For everything else, duckduckgo has consistently provided a suitable result to answer my question for 90%+ of stuff, no worse than Google. It's super rare for me to check if Google has better results for normal things. I see them as a different ordering rather than superior ordering, so I think of Google more as an alternative than a fallback. They clearly both interpret whether a site is a good match using different algorithms but I don't know I'd say one is strictly superior.
[1]: https://duckduckgo.com/?q=100+JPY+in+EUR
[2]: https://duckduckgo.com/?q=2%5E32-1
But like you I drop back for unit-aware calculations. For those an alternative to Google is Wolfram Alpha “!wa speed of light / 400nm in terahertz” [3] which also allows for symbolic computation “!wa integral of 2x” [4].
[3]: https://duckduckgo.com/?q=!wa+speed+of+light+%2F+400nm+in+te...
You can use the command-line program units(1) to do this, available on all Unices. (The preinstalled version on macOS is horribly outdated though, so you might have to look for a newer version in Homebrew.)
$ units --verbose
Currency exchange rates from FloatRates (USD base) on 2019-02-20
3070 units, 109 prefixes, 109 nonlinear units
You have: c / 400 nm
You want: terahertz
c / 400 nm = 749.4811449999999695137376 terahertz
c / 400 nm = (1 / 0.001334256380792608193130988) terahertz
sheesh...
Made me feel super dumb for googling "10kg in pounds"
Why not use the Startpage bang? (!s)
If you want to find my stuff, you now have to use DDG or some other search engine. Hopefully we can signal to Google that we are not okay with monopolistic behaviour (that's why I blocked them, not just for being the biggest or monopolist, but for also behaving like it -- see the blog post for details).
Of course, if your websites generate income based on people that came from google, I would not expect you to follow suit. As I mentioned in the post, the idea is that there are lots of resources on the web that are not there to turn a profit, but that are still valuable to people. Google is not listening when we use words, so I took action.
i.e. the only people that will (likely) see it are the ones that are already doing what you want.
how do you determine which IPs are the google webcrawler, though?
I don't use robots.txt because they say that doesn't stop them from including the site in search results: https://support.google.com/webmasters/answer/6062608 I don't know if returning a HTTP 403 error will, but it seems like it's worth a try.
I also looked into banning IP ranges (that would have been my preferred option), but if I remember correctly they were subject to change and it seems overkill to write a scraper for that page that would then have to generate a config file and reload a service.
If you need to make sure it's actually a google bot when a client shows up with the user agent, you can use reverse dns: https://support.google.com/webmasters/answer/80553
However, in the rare cases that I do switch to Startpage (!s), I get "better" results only sometimes. Conversely, on Google I miss the bangs, the capability to scroll through the results using up/down cursor key (and enter to go to a result), the simple way to specify country specific results (I don't care about Moscow, Idaho and its 23000 inhabitants very much, sorry).
https://www.reddit.com/r/duckduckgo/comments/90ypqz/ddg_craw...
A comment from July 2018:
> We get our results from various sources, mostly Bing, Yahoo and Yandex (more info: https://duck.co/help/results/sources ) so if/when they introduce support for SPAs, we should get that too. I don't know what their plans are for this, however.
This search is expected to return results that provide a tutorial on the implementation or usage of the HTTPS protocol. The first three results seem decent enough, then things get weird. What follows on the first two pages or so are (in order) tutorials for: Ubuntu, PARCC, HTTP (relevant), VPN, React, R (language), Windows 10, Shiny (related to R?), Matplotlib, scratch.mit, AutoHotKey, Quickbooks, Node.js, Java, Python and Kubernetes. As far as returning results for HTTPS tutorials DuckDuckGo has seemingly not done very well. Many of these results are related to software so one would expect these sites to have well-executed SEO. It looks to me like DuckDuckGo has confused the "https://" in the URL for an indicator of content related to HTTPS. But does Google's search algorithm do any better?
Try the same search in Google[2] and it seems to have similar problems: if you go several pages deep the results still do not acknowledge that "https" is missing from the actual content of the result. But the fact that Google is used more frequently has allowed relevant results to bubble up to the top because people click on them. Only one result on the first page does not pertain to HTTP, HTTPS or SSL. Google's advantage seems to disappear after the first page and that makes sense because anything beyond the first page of Google search is rarely clicked.
I don't see DuckDuckGo's problem as being one of how it searches but rather its lack of usage. Maybe if we use it and talk it up to people we can work our way towards a powerful search engine that respects privacy. I could put up my own search engine and try to do better but I would be a decade behind DuckDuckGo's name recognition and that much farther from solving the actual problem.
1. https://duckduckgo.com/?q=https+tutorial&t=ffab&atb=v144-1&i...
2. https://www.google.com/search?client=firefox-b-1-d&q=https+t...
Yet I agree it's not better than or equal to Google when it comes to image search. I guess Google is doing a better job on classifying images.
One way to compensate is to be very specific.
Ex: "python programing language string slice" not just "python slice".
Most people have (without consciously thinking about it) learned to treat search terms as labels on a venn diagram with the center being the search results.
To get the most out of DDG search like it's 1999 - be specific and use operators
I hear this complaint from time to time and it just baffles me. I mean one day I forgot how to spell "guillotine", typed in "french beheading thing", and it was on the first page somewhere.
Sure, Tor might seem like overkill, but over time all those single, isolated queries start to build an elaborate dossier on you and the contents of your mind.
[0] https://duck.co/forum/thread/15228/would-you-adopt-a-canary
(All used with Tor and sometimes a VPN)
- "site:news.ycombinator.com rms_returns duckduckgo": At first the DDG results looked better but multiple results had no mention of "rms_returns" at all.
- "archlinux dmesg audit": Both nice.
- "lange nacht der museen berlin": Wow, unexpectedly relevant results on DDG, did they improve their local results?
I think I should indeed try again. Previous times it was full of spam for local results.
- "python remove from list": DDG returns the stack overflow answer (mentioning `del`). If I wanted the docs, I'd put in the !py3 bang, which uses the python 3 documentation search.
- "site:news.ycombinator.com rms_returns duckduckgo": yeah, DDG is fuzzing the search terms annoyingly, though I've heard that complaint about google as well? I've had more luck adding a plus, "site:news.ycombinator.com +rms_returns duckduckgo"
This is the exact explanation of why I moved away from Google.
My searching trends toward very specific terms. >2 years ago, Google started ignoring my boolean operands. It became increasingly difficult to craft a boolean search that Google would actually respect.
I finally bailed 6 months ago. By that time, using a minus sign often returned nothing but the results I was trying to avoid.
They use Bing Ads, but I’m not sure if their actual search results are just proxied Bing results.
> We do use results from Bing. Our zero-click info and some other stuff we do, however, uses results from all over.[0]
[0]: https://web.archive.org/web/20180910181021/https://duck.co/f...
A critical feature though is that “g!” to Google will never return AMP pages; as long as that’s true I will always start from DDG.
I wonder whether that's due to users using mobile more and more, where typos are much more frequent. It is very annoying indeed, and coaxing it with quotes and/or plusses doesn't seem to yield consistent results.
> A critical feature though is that “g!” to Google will never return AMP pages;
Why is that?
It returns results from the major search engines (and can be customized for more) while attempting to maintain privacy as much as possible. I'm not sure how effective it actually is from a privacy standpoint, but being able to see aggregate results in a single, clean UI has been a big productivity boost for me. I've been running a self-hosted instance for a few months now and haven't had major complaints.
Bangs are recommended almost every time ddg comes up, and I see people like my brother using them, so it seems like nobody else minds and the choice of symbol will never change. Am I alone in thinking it's one of the worst possible prefix characters?
Try using your right pinky to press shift when typing a bang.
If you want to just use one hand, try holding shift with pinky and hitting 1 with your middle finger. Slightly less stretching.
Maybe it relates to hand pain and stuff
You could use right shift, also, though that’d be tricky for my brain
Also, while it makes you dependent on the reddit search, there's the "!r" search shortcut.
[1] https://help.duckduckgo.com/duckduckgo-help-pages/results/so...
I use DDG as my primary search engine, but the quality of search results and the rate of improvement leave much to be desired. So I tend to go to !s (search in startpage) or !g (search on Google). DDG would have to improve by leaps and bounds if it has to be the only search engine that I have to use (or rather, the only search engine who’s direct results I have to use).
Apart from poorer search results, instant answers aren’t great either. DDG has currency conversions as well as stock quotes as instant answers. But both these are so highly hit or miss that I end up going to some other site anyway. For currency conversion, depending on where you put the currency symbol or code (as a prefix or suffix) or if you don’t leave a space between the currency and the numerical value, you may or may not get instant answers. For stock quotes, there are many symbols where, if you just type the symbol, you wouldn’t get instant answers. You’d have to type “<stock symbol> stock”. For some stocks just typing the stock symbol is enough.
That being said DuckDuckGo is definitely getting much better nowadays. Still end up going back to Google after a few days though.
I actually disagree. Googles optimizing function is to get you to visit sites with their ads. Because DuckDuckGo doesn’t have an ad network on every website, their optimization function is to get you to come back to search (where their ads are).
I think people have also missed Google slowly manipulating search results to maximize profit. Their motives are not aligned with people looking for information (at least not any more)
For example, DDG is really good at documentation lookups like Java or Python docs. Google is much better at complicated queries (I always struggle for an example, had one recently where DDG was just lost and Google gave the perfect top result, and remember thinking "I should write this down" but I don't think I did...), so I use DDG by default and if I can't find what I'm looking for, I fall back to Google. I don't have to fall back for ~90% of my queries, so that's a significant improvement over using Google all the time.
DDG is horrible at images, and Google Images beats Bing Images. But again, 90% of my image queries can be answered by Bing, so in only 10% of the cases I fall back to Google for images.
I've kind of found the opposite.
Because Google knows I'm a developer when I search for libraries etc. they're usually at the top.
For example, if I search for 'express' (the NodeJS package) on Google their website is at the bottom of the first page (just below my local newspaper's website, Paisley Express).
On DuckDuckGo it's 1 from the bottom of the second page.
Of course just a single example but you can see where I'm coming from.
I have 0 reason to not use google. I really don't care if they find out my obsession with electronics and that I have the hots for tall girls.
Out of every company, I'd give Google the nod as most helpful company of my lifetime.
Can someone tell me why I'm wrong?
Otherwise... no, I don't think you are. The people who turn out for threads like these are hardly representative. You shouldn't sweat not passing muster with them too much.
I thought I'd probably change back to the Gman within a few days, but I still use it.
And I only today found out about "!g" (I was literally typing "google.com/search?q=terms" into my address bar everytime).
Averaging about 1 in 5 searches I need to go to Google to get my results. Mostly stuff regarding popularity and currentness, like memes e.g., or (admittingly bad and lazy) natural language queries where Google also has an edge.
Also the ! command is amazing, no more wasted seconds on the homepages of youtube, amazon, ebay (!yt, !a, !e). Now if only we can add arrow key selection to those sites :)
At the risk of being repetitious, I think what I'd like is a completely private search engine that uses only a self-curated list of websites. A year or two of googling and using the good results from that would probably fix me for some time. As usual, switch to google if the results aren't good.
Getting rid of the cruft would be a nice thing.
The only drawback I see that DDG sometimes returns strange results for local searches. Even with Poland switch. For example if I typed some restaurant's name I most likely would get some restaurant in Warsaw with the same or similar name.
* move google analytics to stat counter. Might consider their no cookie option.
* moved email to domain based plus zoho. Keeping web inbox zeroed so I’m not to dependent on web mail or zoho
* did a google takeout so I access to old emails of my acc gets shut down.
* Firefox/safari for home browsers
* duck duck go default SE
* ublock origin
In the other direction:
* using material design for a new side project!
Very buggy and can't take much load, but check it out: https://glorp.co/Search/Hacker%20News
I don't get the people that are claiming worse results from DDG than google - the only thing I can think is that we write our searches differently. But that seems a long shot.
Still growing, thankfully. We make our search traffic public here:
Surprisingly, Google seems to listen to what you actually type.
100% sure any Google alternative is going to be equally "morally bad" if not worse once it gets even 1% of Google's traffic.
2. There's more to it, such as the bangs, keyboard-based result selection, etc.
Also, dns are 1.1.1.1 / 1.1.0.0 Browsers are ff and safari. Email is ProtonMail.
Way better in search and privacy
I like the idea of DDG and I have a strong dislike for Google but DDG is simply not there yet. Having said that it's better than it was five years ago, when it was simply insufferable. So, maybe in five years? ;)
Although I DO get frustrated by DDG's image search. Google's actually takes you to the actual image URL, so it's easier to get a link for sharing. DDG shows all images through their weird proxy thing, such that you can't easily get the raw URL.
It's hard to know for sure what's happening though. Maybe my searches are just very simple and don't need anything complex? Maybe I'm luckier and always have fast internet connections? Maybe it's some other factor I'm not thinking of? Who knows without trying to come up with a more systematic way to compare things.
They have a proxy (just like google) so they don't directly load the images you search for, they make their servers load the images and then send you a proxified version. If you search for obscure terms, they won't have a cached version available so they will have to request it from the actual server, and that makes all images take a while to load. On the other hand Google always keeps the images cached so they can serve them instantly.
Sometimes ddg results are even straight up better. Most of the time I use !g is when looking up location sensitive stuff, e.g. non-english local government stuff etc.
Also, I suggest using !s instead of !g so that you still get Google results without actually being redirected to Google.
OP's only argument (Google is bad so use DDG) is not really convincing tbh
If a company does something you disagree with, but you still support them, because it's inconvenient not to it means you are ok with it.
Sorry for getting politics into it, but I see people complaining about current situation, but also don't want to move their assess to protest, because it is inconvenient to do so. Well duh, if you don't want to put an effort, because it would inconvenience you, then you deserve what you are getting.