Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The next version of DuckDuckGo

next.duckduckgo.com

Post reply on HN
Print view View original

201–210 of 331 posts

Re: The next version of DuckDuckGo

#201

Post by davidcelis »

ThePhysicist wrote: post #152

Of course the first thing I did was a vanity search of my name and I have to say I'm really impressed (and a bit shocked) by the quality of the results: All my social media profiles, all my websites, every project I contributed to, even websites where people credited me for my work when they used it in their own projects, all neatly sorted and with great context information. Definitely on par, if not better, than Goo…

Wish I could say the same :\. The top result for "David Celis" is a LinkedIn directory of people with my name. The first result that's actually me is my Twitter account, at number 3. My actual homepage, http://davidcel.is/, doesn't even appear on the first page. Guess I need to become more important?

Re: The next version of DuckDuckGo

#202

Post by dekz »

ronaldx wrote: post #196
pnathan wrote: post #188

Earlier quoted context omitted.

Not yegg, but if you store a database of queries, along with frequencies, you can do a prefix match on what's already in the search box and return that. I'm speculating that DDG does something that. It doesn't take any PII to do that unless the query itself includes PII. E.g., "websites operated by paul nathan" is pretty PII. :)

Saving - and now seemingly publishing - search queries is saving and publishing PII. http://donttrack.us directly acknowledges that search queries can be used to personally identify individuals. I therefore understood from the later statement on that site, "[ddg doesn't] store any personal information at all", that this would include search queries: ddg almost literally advertises itself on the fact that it doesn't s…

But that is in conjunction with a browser fingerprint right?

I don't follow that DDG not being able to save anonymous queries disregards their stance on privacy?

Surely for the best experience you'd want DDG to do this anonymously.

Re: The next version of DuckDuckGo

#203

Post by germano »

ronaldx wrote: post #196
pnathan wrote: post #188

Earlier quoted context omitted.

Not yegg, but if you store a database of queries, along with frequencies, you can do a prefix match on what's already in the search box and return that. I'm speculating that DDG does something that. It doesn't take any PII to do that unless the query itself includes PII. E.g., "websites operated by paul nathan" is pretty PII. :)

Saving - and now seemingly publishing - search queries is saving and publishing PII. http://donttrack.us directly acknowledges that search queries can be used to personally identify individuals. I therefore understood from the later statement on that site, "[ddg doesn't] store any personal information at all", that this would include search queries: ddg almost literally advertises itself on the fact that it doesn't s…

By the grandparent's description, the difference would be that DDG is not saving the ip address and browser information associated with the search query. If they're only saving the frequency of each search query and nothing else, then there would be no way to see for a particular user/browser/ip address what search queries have been performed because that dimension to the data simply doesn't exist.

Re: The next version of DuckDuckGo

#204

Post by crazypyro »

ronaldx wrote: post #196
pnathan wrote: post #188

Earlier quoted context omitted.

Not yegg, but if you store a database of queries, along with frequencies, you can do a prefix match on what's already in the search box and return that. I'm speculating that DDG does something that. It doesn't take any PII to do that unless the query itself includes PII. E.g., "websites operated by paul nathan" is pretty PII. :)

Saving - and now seemingly publishing - search queries is saving and publishing PII. http://donttrack.us directly acknowledges that search queries can be used to personally identify individuals. I therefore understood from the later statement on that site, "[ddg doesn't] store any personal information at all", that this would include search queries: ddg almost literally advertises itself on the fact that it doesn't s…

I think the difference here is that there is no association of the search term with a specific search. If they see a term, they simply add 1 to the counter. Of course, that is probably vastly simplifying it. Basically, your search is thrown into a pool with all of the other searches. It possibly never even has an id associated with it. Just a frequency number.

For example, say I start to type "movie t", it'll see that often times after the word "movie", "times" appears at a high frequency so that could be one of the suggestions. I don't see how this introduces a privacy issue unless they are saving the searches by some personal identification number (whether thats an ip, user id, computer, etc), which they specifically deny. Without the specific identifier, there is no way to say it was user A who searched for "movie times" 300 times in a row and not 300 different users searching for "movie times"

Re: The next version of DuckDuckGo

#205

Post by ugexe »

Something that has bothered me in the old version as well as this version is the wiki-summary/product/images/videos header. Often times (especially when im using a VPN) the search results load before whatever needs to be loaded in the header, so I will go to click a result and then the header expands into my click. This means I go to a page I did not intend. It needs to be a fixed size or its size needs to be determined before listing search results.

Re: The next version of DuckDuckGo

#206

Post by nemoniac »

I also use DuckDuckGo as my primary search engine and have had very positive experience using it.

But the single worst thing about ddg is that it's still not possible to deactivate the "did you mean?" functionality. I rarely make typing errors so I spend most keystrokes on ddg putting quotes around things that I did mean.

Re: The next version of DuckDuckGo

#207

Post by ronaldx »

andrewguenther wrote: post #180
ronaldx wrote: post #164

Earlier quoted context omitted.

Understood. Can you clarify how you have the data to provide an auto-suggest?

It looks like it is based on sets of bigrams and trigrams pulled from pages which they have indexed.

Quick example: "Is my password... cats6plus" does not appear to be pulled from any page(?)

Re: The next version of DuckDuckGo

#208

Post by timr »

highwind wrote: post #128
johnnyg wrote: post #15

DDG has been my primary search engine for nearly 2 years. I love you guys for both your service and your mission. You've struggled in the past with local searches and map inferences to non-address searches (think "Godwin Park Houston"). I threw those at next.ddg and the result back were very good. I don't see having to !g a query for these in future. I'll edit as I use the beta in the next few days but first impressi…

Try this query: https://next.duckduckgo.com/?q=whole+foods+falls+church+va

http://www.bing.com/search?q=whole+foods+falls+church+va&go=...

Re: The next version of DuckDuckGo

#209

Post by Spittie »

rml wrote: post #197
fsck--off wrote: post #61

I use DuckDuckGo as my primary search engine and have had a positive experience using it. This redesign incorporates two of the worst design trends today: very low contrast text and gratuitously, obnoxiously large fixed headers. I dislike fixed headers on any non-phone/tablet UI because almost all laptop and desktop screens are widescreen. Fixed position elements on the sides of the screen make much more sense, altho…

I also use DDG as my primary, but I prefer the `lite` interface: https://duckduckgo.com/lite

I'm a fan of https://duckduckgo.com/html/ , not completely bare-bone yet fast to load.

But the new interface is nice.

Post reply on HN