Nona – A German Search Engine | Hacker News Reader

Nona – A German Search Engine

nona.de

175 pointsdanskeren96 comments

I couldn't find it anywhere in their data statement, nor in their about page or terms and conditions, where the results were coming from. Finally at the bottom of the SERP there's the indication that results come from Microsoft (Bing). However, all links to Bing seem to happen in the backend, including linking to images. Would like to know how much data they share with Bing, but so far as it's visible from the client side it looks pretty solid. And they have podcast search, which is cute.

Hi. We list every source in our help-section: https://hilfe.nona.de/features/datenquellen

Indexing the web and doing a initial evaluation is hard and costy.

Because of this only a few search engines do so, if we exclude Asia the only starch engines I know which do so are Google, bing and Yandex(?).

All other defer to one of thos search engines (most times bing) while decoupling, post and pre processing queries to a varying degree.

This situation almost screams for web indexing as a service. Would love to see some search engine innovation, especially advanced curation, classification, archival and filtering(!) abilities.

There is index of most sites to download at https://commoncrawl.org/

Brave say they have their own index actually, but also sometimes use other providers (maybe for when their own index didn't return enough results?). https://brave.com/brave-search-beta/

As far as I know Brave brought some Index startup but doesn't actually use it. Instead they use Bing while pretending to have their own index.

Generally I would avoid Brave as the company behind is is quite doggy.

Ironically, Brave bough a German search engine...

Brave search results are quite different from Bing's (they are better) -- where are you getting your info?

Just because it's hard doesn't mean you can't do it.

> if we exclude Asia the only starch engines I know which do so are Google, bing and Yandex(?).

Asia...and Europe? I'm pretty sure that for example Seznam.cz still has its own indexing (https://blog.seznam.cz/en/history-of-web-search/). At the very least there used to be several more indexers as well. A version of one them is even accessible as source code: https://www.ucw.cz/holmes/

There are more.

For example, Syria's Web search engine Shamra has its own index according to its founder Shadi Saleh, whom I have interviewed here: https://irsg.bcs.org/informer/2019/07/syrias-first-web-searc...

Mojeek and Gigablast. Both do not engage in tracking unlike all others. https://blog.mojeek.com/2021/05/no-tracking-search-how-does-...

And there still is the German meta search engine, MetaGer. I already used it more than 20 years ago.

https://en.wikipedia.org/wiki/MetaGer

https://metager.org/

It also has a TOR hidden service:

http://metagerv65pwclop2rsfzg4jwowpavpwd6grhhlvdgsswvo6ii4ak...

I just used this for a few English-language queries and the results... are surprisingly great.

There is also https://fireball.de

https://fireball.de/about

Oh my God, yes.

I barely remember it!

http://web.archive.org/web/19971210215225/http://www.firebal...

Those were the days my friend, we thought they'd never end.

And now let's see, what will happen when in a faraway future, some digital archaeologist explores, whether the early internet really was intended for porn or not...

http://web.archive.org/web/2/http://www.fireball.de/query-fi...

Well, maybe not, but...

http://web.archive.org/web/20000229065924/http://www.firebal....

There you have it, those naughty ancients!

I like that they offer an ad-free experience for 2€ a month. I would like more of the internet to offer this type of alternative.

Google makes like 50$ per month from you and Facebook makes 30$ per month (or something in this ballpark), things are way beyond what most people are willing to pay.

they don’t make nearly as much

Why do you say that? The revenue of Facebook is public (86B in 2020). Is a lot of it not from users, or is the figure incorrect, or do they have many more users than they say they have?

30 cent per user per month, maybe. Not 30 usd

Google Ads cost $~4-8 per lead for our company. Google is raking it in.

How do you then explain Facebook's revenue being 86B in 2020?

Fair point, I stand corrected:-)

The alternative to paying for an ad-free experience is ad-blockers. I very much doubt that they make anywhere near as much from me.

30$/user/year seems to be accurate. Facebook reported 86 billion USD for 2020 [1] and 2.89 billion MAU [2]. When computing the ratio, the geek inside me feels particularly satisfied to reduce billion from the numerator and denominator. :)

So 2€/user/month seems priced well.

[1] https://investor.fb.com/investor-news/press-release-details/... [2] https://www.statista.com/statistics/346167/facebook-global-d...

wait wtf why wont they just charge it then

ffs

i thought its 30 per month and thats why they dont do it, now i am even more upset for all the spying..

oh wait, thats 2.8b mau but most of them are worth nothing, look at west europe and north america, and only subset of them, its like 200 million people that have to pay

    >>> (89000000000/200000000) / 12
    37.083333333333336

Most people would not even pay a cent

i think the problem is that the people that will pay a cent are the people they are making 40$ per month from, and the rest of the MAUs are probably almost worthless

so if they start charging, they will only lose

I wonder how many users are unbanked and can't even pay a cent.

because data has some intangible value not reflected in the quarterly financials.

I wonder how much is enterprise revenue albeit the sites are used as personal services.

> Google makes like 50$ per month from you

On average...

Like many people out there I have never clicked on an ad and I definitely never closed a purchase by following an ad link. I never allowed for tighter integration, I avoid other Google services (Gmail, Maps, Youtube) other than Search itself.

Users like me are a burden on google's servers. On average if I was charged 20$ a month for anonomised technical search services then I am fairly confident that they would be making more money than if I just ad-block everything like I currently do.

> On average if I was charged 20$ a month for anonomised technical search services then I am fairly confident that they would be making more money than if I just ad-block everything like I currently do.

You should check out Gigablast (https://www.gigablast.com/). I came across their site yesterday, as one of a very few search engines who maintain their own index of the internet rather than relying on Google or Bing. It's delightfully antique, and surprisingly good for all but the most arcane searches.

Great sugestion, just tried some obscure Electrical Engineering queries and it looks pretty good. Thanks!

Interestingly, this approach seems to be gaining more traction in Germany specifically. One of the larger newspapers, Die Zeit, applies it for the free tier of its website. There’s a big annoying modal when you go to zeit.de if you want to see that approach in practice. Hilariously the choice I make in that modal never seems to stick. Maybe there’s a separate GDPR-compliant toggle somewhere to allow them to track that I want to be tracked ha!

Zeit, Spiegel, Golem… Those are the ones I frequently see, but there are probably more.

What is the relationship between Nona and Chronoto [0]?

Chronoto sells luxury watches, is the same owner, registered at the same address and has the same commercial register entry.

[0] https://www.chronoto.de/impressum/

---

> (1) I agree that Nona Search Technologies GmbH ("Chronoto") may use my e-mail address for advertising Chronoto, including advertising in connection with the sale (e.g. information on the value of the watch and the sales process), the purchase (e.g. interesting offers of watches) and ownership (e.g. offers of insurance and financing, as well as maintenance and repair).

As it says in the imprint: Chronoto is a showcase for the search (and crawler) technology behind Nona, which we are going to offer next year.

Chronoto doesn't sell watches, it's just a (meta-)search engine.

So assuming I want to build a car search website where I get payed via referrals from let's say a predefined dozen of dealers, I could subscribe to this offering from Nona Search Technologies GmbH where it would crawl their sites and give me their new offerings in a predefined format?

It says that Chronoto is a search engine for luxury watches using Nona's search tech.

My ideal search engine would be 2008-2012 Google results but filtering out anything that even remotely smells of blog spam containing affiliate links.

Do you know how to do ranking of that king without pinging anyone to setup your analytics tool and have previous history or searches?

Did 2008-2012 Google have that? I think they mostly used backlink count.

Curious how much would you be prepared to pay per month for your ideal search engine?

For me (student) 5€/month is the no-brainer price for a useful service I use semi-daily. Beyond that it starts to hurt.

I would prefer a netflix model run by the gov. You pay 10 Euro a month when having an internet connection. This is distributed based on your activity. Sites you use often are being paid more to run their infra. At the same time those sites are not allowed to run ads or affiliate crap. Also profits should be capped and taxed with 99% above a threshold. \

Looks like they are using Bing / Microsoft for their search

They list their data sources here: https://hilfe.nona.de/features/datenquellen

Bing for Web+Image (and I assume News+Video) Search, ListenNotes for Podcast, OpenWeatherMap for their Instant Answer weather results, etc.

I like their UI, and as far as I can tell then their icons are custom made, at least I can't find them anywhere.

Yes and a few other data sources for weather, stocks and so on (https://hilfe.nona.de/features/datenquellen). And it's made with Symfony (https://hilfe.nona.de/ueber-nona/technik-tools).

Seems like most new search engines are “bing + some other sources.” Just wondering, does Bing charge for this use?

Personally, in the past month I’ve noticed a significant degradation of results from DDG, and since I don’t know what’s going on I just blame Bing. Because of that, each time we have posts about alternative engines I first check to see if they’re not using Bing. But they all are!

Yes, Bing charges for use of their API. https://www.microsoft.com/en-us/bing/apis/pricing

Regardless, they return far better search results for queries in German than Google does. For example, if you look for something about a changing legislation in the Bundestag, they will have articles published in the past 30 minutes, where as Google won't surface the same articles for a few weeks.

There's only one chance to make a lasting impression as an aspiring Google competitor and there isn't a single one that even made it into my bookmarks. I understand there's enough love for DDG on HN but life is too short for mediocre horizontal search results.

My results are the same or better than Google in DDG as far as I can tell. I haven't missed Google a bit since I switched. So, either the difference for the average user is minimal or there isn't a difference.

It’s a frequent complaint. I wonder if some people just use search differently and the lack of tracking makes it work worse for them on DDG? I’m in the same boat as you (except for their horrible decision to just ignore what you search and show you almost random results when they don’t get enough/any hits) and rarely ever use !g.

Part of that might be that non-english searches in DDG are a horrible experience.

Not in German, IME.

How do you evaluate?

DDG is now even with Google imo. Used to be different for sure. But nowadays, it's a back and forth.

For a lot of people DDG is strictly better: You live anywhere that is not the US and therefore have to deal with a thing called "locality". Sometimes you want search results in language X, sometimes for your location, sometimes you need the "normal" results for Americans.

Google decides all that for you. Badly. There is no setting except - your know, logging into some Google account and finding some setting somewhere for it.

DDG it's literally a switch.

Have you considered that maybe you're just used to Google? There is a learning curve to using a new search engine. The way they answer your queries depends on how you ask.

But at the end, it's about what you care about. Is life too short to have principles?

First thought as a German, why? Second thought, why not? Google killed all others because it was much better. I still use Google because DDG does not work for me, but there might be a market for a German search engine on top of Bing, now that the technology gap isn't as large as it was 20y ago. The benefit of being better for Germans (by whatever definition) could compensate the technology gap.

Compares well against Fireball, another German search engine that also has plenty of results for Donaudampfschiffahrtsgesellschaftskapitän

https://fireball.de/search?q=Donaudampfschiffahrtsgesellscha...

Russian Yandex is okay with it too https://yandex.ru/search/?lr=213&text=Donaudampfschiffahrtsg...

Looking at Ukraine's border... no thanks.

> Donaudampfschiffahrtsgesellschaftskapitän

"Danube steamship company captain" per Google Translate, in case anyone else wonders what that beauty of a word-train means in English.

Split it up into its constituent words, and it becomes easier to make sense of:

Donau - Danube, Dampf - Steam, Schiff - Ship, Fahrt - travel, Gesellschaft - (roughly) Company, Käpitan - Captain

Googling this took me to the electrifyingly long Donaudampfschiffahrtselektrizitätenhauptbetriebswerkbauunterbeamtengesellschaft.

It seemingly won’t even render properly here, at least on my iPhone.

I wonder what would be better? having search engines able to deal with different languages or having different search engines for every possible language, Don't see the point for the latter to be honest, having an universal language makes more sense to me.

Regardless of whether you serve them through the same search box, from a practical standpoint, I think you'd want basically a separate back-end for each language. Stuff like relative term importance, word extraction, ranking, they all benefit from being kept separately.

First we do search. After retrieval comes question answering based on the retrieved snippets. Then comes multi-hop question answering with the need to correlate information between multiple sources. Just watched a video showing how it can be done https://youtu.be/-ethT5YDVmo?t=1256

different search engines for every language.

it looks like a lot, but a typical user will only use a handful of them. and having each tailored to a language or region is better than a universal search engine.

I searched for Donaudampfschiffahrtsgesellschaftskapitän and it had plenty of hits.

I thought you spell Donaudampfschiffahrtsgesellschaftskapitän with 3 "f", i.e. Donaudampfschifffahrtsgesellschaftskapitän

Depends on whether you were taught the old or new rules, and whether you remember them correctly.

The old rules said you only write the 3 same consonants only if they are followed by another consonant. Schiffahrt with 2 f is correct under the old rules, because the f is not followed by another consonant. But it would be Schifffracht (ship cargo) with 3 f because of the last f being followed by a consonant. Or Hellila (light purple) with only 2 l in the middle instead of three (the word is a compound of hell and lila).

The new "simplified" rules from 1996 say you always write the 3 same consonants or vowels. Alternatively you can use a hyphen. Schifffahrt with 3 f is correct under the new rules. Schiff-Fahrt is as well. Schifffracht or Schiff-Fracht. Helllila or Hell-Lila.

While the rules are taught (and following the rules is part of the grade you will get), they are now lawfully binding in German-speaking countries, and indeed there are plenty of people, especially those who went to elementary school before 1996, and even publications that use the old rules still.

I was in school in the middle of it all. I was taught the old rules in elementary school but the new rules technically came into effect while I was in mid/high school. We were given a short introduction to the changes, and were allowed to use whatever rules we wanted. Some teachers would mark you down for spelling mistakes if you mixed new and old, tho.

Wikipedia claims that Donaudampfschiffahrtsgesellschaft is the proper name of a company that only existed until 1991, so the new rules that were introduced in 1996 do not apply. The company never changed its name to the new spelling, so the old spelling should be used to refer to the historical company. The reasoning makes sense to me.

If you're referring to that company specifically, yes, their spelling would not have changed, as it is an Eigenname. The actual name was Erste Donaudampfschiffahrtsgesellschaft (Erste = first). However there were probably a bunch of companies operating steam ships on the danube, which can be reasonably called Donauschifffahrtsgesellschaften ;)

Good quality results on English search, but it logs IP address etc and does not anonymizes users. So I wll stick to DDG till the fix it. Otherwise great initial effort.

Excrept fron ther privacy policy - "Such information may include details of the User’s visit, information about the User’s computer, including IP (Internet Protocol) address, operating system and browser type, the User’s location, and usage information. An individual User will not be identified from or by this information and NONA is entitled to copy, distribute or otherwise use such information without limitation".

Are you talking about Nona? We don't log any IP addresses.

You might want to put the quote in context: The passage is from the area of "email and marketing" or "social media", we are obliged to the passage if we maintain a presence there - so in the case our Instagram accounts for example.

But thanks, we might need to edit the text to make it clearer.

Clever name! Anon backwards.

Hah, you are the first one who got it :)

I dunno, there is something obviously wrong with this engine.

When I search for "saashub", saashub.com isn't on the first page. And that's a website that's been online since 2014 and has more than one million page-views a month... ¯\_(ツ)_/¯.

Bing doesn't show that website either. Some websites are so eager to block spam that they block any automated traffic with the sole exception of Google's crawler.

How can this be a privacy-respecting search engine if it requires me to provide payment details for the ad-free version? On top of that, the ad-free version requires a login and enables them to tie my search queries to my account.

Cool name, good design.. I also like the fact it has podcast search built in. Well done!

Seems to respect privacy. At least it only includes one external js resource, and doesn't leave cookies. And while the interface is in German, it does find results in various languages.

Edit: I must add that I had my ad blocker on...

Yeah, we are using plausible.io as external analytics-service, without any cookies.

Great effort. As a individual developer, How does using Bing results work? Can I signup for their search engine API and modify them based on my secret sauce, without raking up 100$+ bill every month?

The bill will be based on your usage. For Bing API it is $7 per 1,000 queries for full results (news/video/images). So just raw results without any other effort will cost you about $7/mo assuming you do 30 queries/day.

…and you are not allowed to cache results - so you actually have to call the api everytime you do a search.

or you could use a service such as SerpApi.

I love that people keep making alternatives, but I can't shake the feeling that whatever displaces search engines won't be recognizable as a search engine.

(And I have no idea what it's going to be.)

A search engine could replace a another search engine, but it’d need a better take than just being German. Google wasn’t the first search engine after all.

Feels snappy. I could find some people I know which I was not sure if it would find.

English version of the main interface would have been great. I have added it to my Firefox anyway.

Reply on news.ycombinator.com