Why do many search engines seem to ignore operators (e.g. exact phrases, term exclusions, OR, etc.)? Is there a good reason for having a dumb 1997-level search logic that I'm not seeing?

Submitted ⁨⁨2⁩ ⁨days⁩ ago⁩ by ⁨58008@lemmy.world⁩ to ⁨nostupidquestions@lemmy.world⁩

Wouldn’t it cut down on search queries (and thus save resources) if I could search for “this is my phrase” rather than rawdogging it as an unbound series of words, each of which seems to be pulling up results unconnected to the other words in the phrase?

There are only 2 reasons I can think of why a website’s search engine lacks this incredibly basic functionality:

The site wants you to spend more time there, seeing more ads and padding out their engagement stats.
They’re just too stupid to know that these sorts of bare-bones search engines are close to useless, or they just don’t think it’s worth the effort. Apathetic incompetence, basically.

Is there a sound financial or programmatic reason for running a search engine which has all the intelligence of a turnip?

Cheers!

source

Comments

Sort:hotnew top

HuntressHimbo@lemm.ee ⁨2⁩ ⁨days⁩ ago
Because business majors decided a search engines primary job was actually to serve you ads rather than to help you search for things

source
- Valmond@lemmy.world ⁨2⁩ ⁨days⁩ ago
  Yeah, even fucking linkedin can’t make the difference between C, C++ and C#
  
  source
  - joyjoy@lemm.ee ⁨2⁩ ⁨days⁩ ago
    So basically to find a good job, you should learn all 3.
    
    source
    -> View More Comments
hedgehog@ttrpg.network ⁨2⁩ ⁨days⁩ ago
It’s largely the first one, at least according to The Man Who Killed Google Search.

See also the Hackernews discussion and this follow-up article by the same author (with links to an article with Google’s response, summaries of other discussions on the topic, etc.)

source
- Enoril@jlai.lu ⁨1⁩ ⁨day⁩ ago
  Wow, spent the last 30min to read everything. Thanks for sharing this, really interesting articles.
  
  source
xia@lemmy.sdf.org ⁨2⁩ ⁨days⁩ ago
I could FEEL when amazon removed the not and quote functions… now it’s nigh-unusable.

source
- FauxPseudo@lemmy.world ⁨1⁩ ⁨day⁩ ago
  I hate trying to search for specific things on amazon because negative operators don’t work. I’m frequently trying to find products that don’t contain specific words. Like when I wanted a foam mattress cover that wasn’t cooling. I need all the heat I can get when sleeping. But trying to find one that wasn’t marketed as cooling? No such luck. I tried using search engines that honor negatives but no such luck. Amazon has thwarted every attempt to find what I want.
  
  source
  - Modern_medicine_isnt@lemmy.world ⁨1⁩ ⁨day⁩ ago
    I use the browser search to highlight the word I want to ignore on the page so I can quickly scroll through and ignore those items. It sucks that I have to do that, but at least it helps a bit.
    
    source
Thorry84@feddit.nl ⁨2⁩ ⁨days⁩ ago
Because search engines are much more complicated than you seem to think. The reason the operators worked back in the day (probably later than 1997 though), was because the search engines actually searched through the contents of the pages they indexed. They used a lot of tricks to make it work, but basically they were matching keywords directly to the index.

Modern search engines are much more complex than that, using a lot more abstraction and interesting techniques to both index and search. The amount of data being indexed has exploded since then, the number of users has exploded and the way people use the internet has changed. To keep costs down and search times low, search engines needed to change drastically. And because most people using search engines won’t know how to use those features, they didn’t get preserved.

I do wonder what kind of search engines you are talking about though. I assume you mean the big ones like Google and Bing (or sites using those engines) and not like a simple product search on a small webshop. Because as frustrating as using Google and Bing have gotten, they are still amazing tech and not bare-bones at all. The reasons for their failings are only partly in their control and not even really their fault. (Except for the AI thing Google tried, that was 100% their fault and just dumb).

source
- BearOfaTime@lemm.ee ⁨2⁩ ⁨days⁩ ago
  And yet searx respects operators.
  
  source
- fishos@lemmy.world ⁨2⁩ ⁨days⁩ ago
  Bro, this is just a load of shit. Google removed them by choice, not because of some tech need. Better search engines still use them to great effect.
  
  You just posted a very long rambling justification for injecting ads and other shit into the results instead of giving you what you asked for.
  
  source
- homesweethomeMrL@lemmy.world ⁨2⁩ ⁨days⁩ ago
  Have you read the article “the man who killed google search”? Google seems to have gone out of their way to screw it up and have roadmapped only more screwage.
  
  source
- conciselyverbose@sh.itjust.works ⁨2⁩ ⁨days⁩ ago
  Here’s the thing though. You absolutely could still use operators reasonably well even if the results are fuzzier.
  
  You just use them to control how you leverage the algorithm. AND feeds the algorithm the two sides and filters to results that appear on both. OR joins the two result sets. “Filetype” filters the result set for results that are the relevant file type. Etc.
  
  If they’re not that common they’re not going to have meaningful costs, especially when most power users don’t use them for most of their searches.
  
  source
  - Thorry84@feddit.nl ⁨2⁩ ⁨days⁩ ago
    Most search engines have filters for stuff like filetype, limiting the search to a specific site and filters for time and location (when applicable).
    
    Like I said, search engines are way more complicated than one might think. Doing the kinds of things you mention would be hard and only very few people would need something like that. There are tools out there that do meta searching for analysis though, so you can use search results as data in your analysis. Most of those are highly specialized and often paid, but when you need them it’s worth the price.
    
    Remember companies like Google invest millions (if not billions) into their search engine and have huge teams working on them. Anytime someone says: “Why don’t they just…” the answer is probably very long and complicated.
    
    source
    -> View More Comments
RedStrider@lemmy.world ⁨2⁩ ⁨days⁩ ago
I’m convinced there’s an AI in Google search now that reinterprets what you put. It never seems to give me what I search, only what it thinks I mean.

source
- AndrewZabar@lemmy.world ⁨2⁩ ⁨days⁩ ago
  I said so-fucking-long to Google long ago and switched to DuckDuckGo. If I ever get really nowhere and think maybe googs might have a result for me, then on the Duck you just use a !g before your terms and it facilitates the search thru Google but without their ads.
  
  Their focus shifted long ago from being the best to just figuring out new ways to get more out of users, no matter how deceitful and manipulative they need to be.
  
  source
rem26_art@fedia.io ⁨2⁩ ⁨days⁩ ago
I know at least duckduckgo says on their syntax page that they're aware that operator usage isn't perfect on their site. Seems to come from the fact that they pull results from multiple sources

source
- homesweethomeMrL@lemmy.world ⁨2⁩ ⁨days⁩ ago
  DDG, because they’re so heavily reliant on bing, has gone to absolute shit recently. I’m just about to quit them. Every link in the top ten goes to some M$N / microsoft-adjacent garbage.
  
  source
  - Bluefruit@lemmy.world ⁨2⁩ ⁨days⁩ ago
    Oh so its not just me who thinks its gotten worse then. I was starting to go a little crazy man. It seems like its slowly gottwn worse over time. I wonder if its due to the amount of ai content messing up seo ratings or something
    
    source
    -> View More Comments
  - konalt@lemmy.world ⁨2⁩ ⁨days⁩ ago
    Duckduckgo is awful at searching for specific memes I need. Sometimes I only remember the text and a rough description of the image, but DDG seems only to consider images with Impact TOP TEXT BOTTOM TEXT captions to be memes. I switch to Google and I find the one I want instantly. If there was a way to have the image results of Google with the web results of DDG my life would be complete
    
    source
    -> View More Comments
  - Fubarberry@sopuli.xyz ⁨2⁩ ⁨days⁩ ago
    Yeah, after some years of DDG getting worse I finally dropped them. I switched to Brave and it’s surprisingly decent for an independent search engine. If you search for something that they don’t have good results for they’ll ask you if you want to get anonymous results from google as well, which means I don’t usually have to switch search engines for harder results.
    
    They do seem to have a much lower number of image results though.
    
    source
    -> View More Comments
  - pyre@lemmy.world ⁨2⁩ ⁨days⁩ ago
    every day it looks more and more clear that we’re gonna have no choice but pay for kagi if we want to find anything worth a fuck on the internet
    
    source
    -> View More Comments
  - Flightbird386@lemmy.world ⁨2⁩ ⁨days⁩ ago
    I use -msn , which seems to help somewhat
    
    source
- subtext@lemmy.world ⁨2⁩ ⁨days⁩ ago
  I believe they’ve said that explicit operators are much more expensive to serve than a regular search, so that’s probably why they don’t respect them. Especially a - operator.
  
  source
- skullgiver@popplesburger.hilciferous.nl ⁨2⁩ ⁨days⁩ ago
  DDG doesn’t have their own search index, they’re mostly using Bing and maybe a few fallback providers. They’re entirely reliant on other search engines that they purchase results from, they barely do any actual searching themselves. Things like image search even come from one single source (Bing).
  
  source
BenLeMan@lemmy.world ⁨2⁩ ⁨days⁩ ago
Guys, please. The solution to Google reinterpreting your search queries has been around for years, and it is called VERBATIM SEARCH. (Search options: All results -> Verbatim). Voila, welcome back to 1997.

source
corroded@lemmy.world ⁨2⁩ ⁨days⁩ ago
I’m going to break with what most people are saying and offer the suggestion that search engines are actually doing a decent job. If my mother searches Google for the phrase “Can you please show me a recipe for apple pie?,” she’s probably going to get a recipe for apple pie. If I search google for “c++20” “std::string” “constructors”, after I skip over the ads, I’m most likely going to get a web page that shows me the the constructors for std::string in c++20.

Ad-sponsored pages and AI bullshit aside, most search engines do still give decent results.

source
- Jackcooper@lemmy.world ⁨2⁩ ⁨days⁩ ago
  I think OP is complaining about non Google search engines
  
  source
NeatNit@discuss.tchncs.de ⁨2⁩ ⁨days⁩ ago
I don’t know the answer but I can tell you two things:

It has often been beneficial to me when the search query wasn’t taken literally, it’s not always a bad thing. Many searches are ones where the user doesn’t know exactly what they’re looking for. Granted, that’s definitely not always the case. That said, I don’t remember ever catching it outright ignore stuff like quoted words/phrases.

Regarding “save resources”, Google introduced Instant Search in 2010 which started showing results as you type, thus creating an ungodly amount of extra load on their servers since each user search now created multiple queries. They clearly have no trouble scaling up resources.
source
schnurrito@discuss.tchncs.de ⁨1⁩ ⁨day⁩ ago
For the most part I think they do. I frequently use quoted strings in my search queries (on DDG and Google, I hardly ever use any other search engines) and it usually seems to show me more relevant ones when I do that.

But in general the WWW is now so big that search engines have been having to become more and more complex (and think for themselves instead of taking the queries very literally) in order to be useful at all.

source
thirteene@lemmy.world ⁨1⁩ ⁨day⁩ ago
It’s because websites interpret those characters differently because of how coding requires using the physical qwerty keyboard. Essentially “>” gets used as a compator operator in programming languages, which means that it’s used as a tool to instructs the computer how to do things. When we need to display the symbol, we use “>” as an “escaped character” which basically means treat it as the symbol, not the instruction set. Often search engines will use a very powerful tool called a regular expression which looks like this for phone numbers: ^(\d{3})\s\d{3}-\d{4}

And each character represents something, ^ means start with. \d means digit { means 3 of whatever’s in front of me }. Breaking apart the search parameters is pretty complex and it needs to happen FAST, so at a certain point the developers just throw away things that can be a security concern like special characters like &^|`"'* specially for how to maliciously attack the search engine.

For other characters: www.w3schools.com/html/html_entities.asp

source
Toes@ani.social ⁨2⁩ ⁨days⁩ ago
The tragedy of the commons.

Enough people don’t understand or care to understand, so it gets cut out for the sake of improving the users experience.

source
CorrodedCranium@leminal.space ⁨2⁩ ⁨days⁩ ago
Have you not had much luck using operators?

I’ve found them relatively reliable

ahrefs.com/…/google-advanced-search-operators/

duckduckgo.com/duckduckgo-help-pages/…/syntax/

…startpage.com/…/4521473758228-Advanced-Search-on…

source
db2@lemmy.world ⁨2⁩ ⁨days⁩ ago
It’s cheaper for them not to do it and you’ll still search so they don’t care.

source
- j4k3@lemmy.world ⁨2⁩ ⁨days⁩ ago
  No any more. I use an offline open source LLM first quite a bit now because it is better than their junk. It may only be accurate 80% of the time, but that is a far higher percentage than any present search engine.
  
  People complain about web scrapers, but scraping is the only practical alternative for finding info and sources now that the web crawlers are worse than trash.
  
  source
  - brygphilomena@lemmy.world ⁨2⁩ ⁨days⁩ ago
    No. The issue is websites are trash, not the crawlers. SEO has created a weird amalgamation of content, filler, and keywords. It’s why recipe sites have stories with every recipe.
    
    Google very much is responsible for the current web design though.
    
    source
A_A@lemmy.world ⁨2⁩ ⁨days⁩ ago
We need a browser extension that would post filter the garbage provided by search engines … they are treating us like their product, we should treat their garbage at its right value.

source
- chickenf622@sh.itjust.works ⁨2⁩ ⁨days⁩ ago
  UBlacklist does this, but you have to define what URLs are garbage manually.
  
  source
  - A_A@lemmy.world ⁨2⁩ ⁨days⁩ ago
    Thanks 😌 … … but, well …
    i was hoping for an extension filtering based on search terms and not based on URLs … and also doing the necessary boolean logic like requested by the O.P.
    
    source
skullgiver@popplesburger.hilciferous.nl ⁨2⁩ ⁨days⁩ ago
Because you can’t just boolean search the entirety of the web the same way you do a local database. You’re not getting all results every time you search, that’d be insanely inefficient, so doing full filters like with boolean database search won’t work.

That said, based on my experience with Google, negations work just fine, as do double quotes. Last time I checked NEAR even worked pretty well. AND is implied, OR used to work but is probably derived from the rest of the query these days.

People hate it when their query doesn’t return anything. So, whenever you search for something and get very little results, search engines will relax their boundaries to find something that may answer your query.

Search engines in the early 00s had them because they required very specific phrase matches and the experience was horrible. You either got millions or results that didn’t relate to what you were searching for or you got none and had to start over.

At some point, search engines started interpreting what you mean instead of what you type. For most people, searching for “rain” and getting results about “precipitation” is exactly what they want. Using the 90s/00s search term syntax, you’d need to type “~rain” to also get synonyms, which is obviously a terrible user experience that serves only the most pedantic people.

source
kevincox@lemmy.ml ⁨2⁩ ⁨days⁩ ago
There are a few reasons. Some of them are in the users’ interest. Lots of people phrase their search like a question. “How do I turn off the wifi on my blue windows 11 laptop?”

While ignoring stopwords like “the” and “a” has been common for a while there is lots of info here that the user probably doesn’t actually care about. “my” is probably not helping the search, “how” may not either. Also in this case “blue” is almost certainly irrelevant. So by allowing near matches search engines can get good articles even if they don’t contain all of the words.

Secondly search engines often allow stemming and synonym matching. This isn’t really ignoring words but can give the appearance of doing so. For example maybe “windows” gets stemmed to “window” and “laptop” is allowed to match with “notebook”. You may get an article that is talking about a window of opportunity and writing in notebooks and it seems like these words have been ignored. This is generally helpful also often the best result won’t have used the exact same words that you did in the query.

Of course then there are the more negative reasons.

Someone decided that you can’t buy anything if your product search returns no results. So they decided that they will show the “closest matches” even if nothing is anywhere close. This is infuriating and I have stopped using many sites because of it.

If you need to make more searches or view more pages you also see more ads.
source
- BearOfaTime@lemm.ee ⁨2⁩ ⁨days⁩ ago
  And yet when I put quotes around something… It ignores that. Well Google would anyway. Decent search engines don’t.
  
  source
- brbposting@sh.itjust.works ⁨2⁩ ⁨days⁩ ago
  
  my blue windows 11 laptop
  
  lol so true though
  
  source
not_woody_shaw@lemmy.world ⁨2⁩ ⁨days⁩ ago
I start to feel ill whenever I need to use the search in BitBucket.

source