Hacker News Reader

Comment by WatchDog | Hacker News Reader

I'm glad to hear that they are working on fixing this, but I would be keen to understand why quote searches have been so much worse over the last ~2 years.

View on HN

Me too.

My anecdote was that during the winter Olympics I was looking for a college friend who's name was several characters off of one of the competing athlete's. It didn't matter how much additional information I put in the search. Even quoted, if it included anything that looked like that popular athlete, then that was all that was returned.

When I finally found an old user name she used, then the searches worked. Her correct name and the other information I was searching with were in the page. So it was indexed, the information was simply ignored, because it wasn't news.

Interesting that DDG did not find the page even with the username. I do not know why, but it appeared it was not indexed there.

> When I finally found an old user name she used, then the searches worked. Her correct name and the other information I was searching with were in the page. So it was indexed, the information was simply ignored, because it wasn't news.

This is exactly what people mean when they say quotes "don't work". Elsewhere are google people in this comment section saying they do, and they may be telling the truth after Google applies some relevance pre-filter to the results. Anything that doesn't pass that, though, basically doesn't exist, which can look like quoting not working at all (and effectively is quoting not working as expected).

I once tried to search for a scammer named "internal revenue services" (with an S at the end) and it just wouldn't let me, quotes or not.

I tried it now and it's better, but not perfect. Most pages contain the term (as a misspelling) but a few didn't seem to (one had the word "services" in many places on it, however) and there was still a big Google Maps result and a Wikipedia result at the top for the actual IRS.

quote search often fails on duckduckgo

I hear this a lot, but it I haven't personally noticed any real problems with quoted search (other than the occasional difficulty actually finding the content on the resultant webpages).

I have noticed a general decline in search quality over the last couple of years, but nothing specific to quotes.

I realise I'm not saying anything particularly useful with this comment, but I just thought I'd add another datapoint.

Edit: Thinking more, my biggest issue is when the quoted text occurs in the "recommended similar posts" section of a page (particularly common with reddit). That section gets re-rendered on each view, so it probably won't be there once I click the result.

Quoted search is provably broken for some queries, try a google search for "[::]" (with the double quotes), it has no results. Similarly, try a search for 'linux next hop "[::]"' (with the double quotes), none of the results will contain [::].

Proof: https://archive.ph/AAa6k and https://archive.ph/9WGe7

> none of the results will contain

More and more frequently I was getting this for the actual search terms, in quotes or not, to the point where I would Control+F any words just to find none of them existed on the page. It's the reason I switched to dumber search engines.

I've assumed the "fast path" is to search for "phrases with similar meaning", rather than actual words. But that really destroys technical searches.

If you haven't read our post, I'd encourage doing do. Quotes do work to find the exact terms specified. But control-F won't locate some of the terms we find when fully rendering a doc -- that's why the list explains using developer tools to search if control-F comes up with nothing.

It might be useful to offer people a way to search for content that is rendered in the page, rather than content that is only visible in developer tools.

The content is rendered on the page. For example, say someone has an email sign-up box. When the page renders, the box appears and it might list all the countries in the world, so that you can pick your country from the list. All those countries are rendered, available if you use the box. But if you ctrl-f search, you might not see that text even though it did render. Real case I looked into which prompted the tip of using developer tools.

This isn't evidence of anything changing, google has always ignored punctuation - treating it as whitespace, as mentioned in the article.

Not always, though: https://www.blog.google/products/search/improvements-searchi....

Interesting - looks like they're doing this via a bunch of special-case rules.

To any google engineers reading:

Please add `really-verbatim` mode, indicated by backtick quotes, which also requires strict matching of punctuation.

I'm a Google engineer way too far organisationally to ever have any say in this.

I wonder if that will ever be worth the hardware cost. Back when I did some coursework on information retrieval, it seemed that you get superlinear savings via reducing the cardinality of tokens. So you'd do stemming, remove all punctuation, words that are too frequent ("do", "be", "and", "or", ...)... Basically remove all grammar. You do the same to your search query and the index. This intuitively reduces your compute by at least an order of magnitude, especially for languages with rich grammar (e.g. stemming nouns in Polish reduces the cardinality of tokens by a factor of 7 and verbs by a factor of 162).

No way they'll inflate their indexes even 20% and add complexity into their algorithms for 0.1% queries that won't bring any additional income.

They don't necessarily have to inflate their indexes. Backtick-quoted results ought to be a subset of double-quoted results, so they can use the standard quoted search algorithm, and then filter out imperfect matches from those results.

Google searches ignore punctuation, so it's not even indexed, so there's no way to search for punctuation without inflating the index

Read what I said. They can use the standard index, then filter the results as a last pass.

I did read what you said. Imagine trying to search for

  [::]

as 8192kjshad09- suggested earlier in the thread. What standard index results are you going to filter? Since "[::]" isn't in the index, you won't have anything to go on. To do your back-tick really-verbatim searches, the index has to be enlarged.

Ah, sorry, I see what you mean.

I work for Google Search. We did look a this, and we'll keep looking to see if we can improve, but it turns out to be a very hard lift.

The post explains that we see some punctuation as spaces so that query is a search for nothing, which is why it fails.

In my experience, quoted search works very well. I do a lot of really, really obscure searches with quoted search and Google finds the pages for me. Maybe other people get different results, but I'm puzzled by all the complaints about quoted search. You're all using double quotes ", opening and closing, right?

Some search engines seem to just completely ignore quotes now, it's very frustrating especially when I know what I want to search and the search engine just isn't letting me.

The irony was that googles original claim to fame was that it was the first engine to start including all terms by default and respecting quotes perfectly. Then they just went back on it. Maybe it’s SEO pressure, or maybe it’s just incompetent product owners who didn’t understand what made their product click.

I think they just assume that they know better what you want than you do.

That you probably don't want the thing you're searching for, but some alternate spelling, even if it's something completely different.

I imagine this probably ends up being the case for the majority of their users, but for advanced users and people searching for specific technical strings, they've cheapened their service.

I work for Google Search. We never changed how we used quotes as a restriction tool. People seem to have had that impression because our snippets changed to not reflect where we were finding the terms. Hopefully our new snippet change will help stop that impression.

If you did not change how the quotes interact with the indexed text for a page, but you did change the mapping from pages to indexed text, then you have changed how you use quotes as a restriction tool.

I'm not sure what you mean by "mapping from pages to index text" but no, we didn't make any change in terms of retrieval. None. Use quotes, we look for pages that have the quoted material and only show those pages. The change we announced this week was about how we display snippets -- descriptions -- of those pages. Now the snippets will show examples of where we found the quoted terms, when sometimes the snippets didn't. But even if they didn't, the quoted terms were on the page.

So in the year 2000 and in the year 2022, for any web page, when that page is crawled the same text would end up in the index? No observable changes were made to the process of computing the text to put into the index from the page in those 22 years?

> we didn't make any change *in terms of retrieval*

Users don't care which part of the system was changed to cause the results to be worse.

It looks like maybe your reply is only about what changes recently shipped and are mentioned in the blog post. A lot of the discussion in this comment section is about changes much older than these changes.

Yes, we didn't make any changes in how quotes retrieve content with the announcement we made. We changed how we show snippets for quoted searches.

Quotes "not working" is because we search for things in quotes that definitely exist (and may even be possible to induce to appear on a search page, by other means) and get "no results, have some trash instead" in response.

It may be true that quotes are working the way we expect, but the data set to which they apply is very obviously restricted before they get a crack at it.

You are right. I think what I misremembered you guys changed was that WORDS became optional unless you wrapped each one of them inside double quotes which became super annoying.

My best guess is its because people pasting things in with quotes are met with no results. Stuff like error messages often quote the only unique bit of the error leaving them with no results while ignoring the quotes would give them the result they want.

Probably the vast majority of the population have no idea that quotes do anything in searches. Maybe they need some checkbox in the tools bar that makes quotes actually work if you know how they work.

And the vast majority of the population is who use Google now, rather than web enthusiasts. It was probably causing people to get no results when they had quotes, and most people wouldn't know why.

how hard would it be for Googs to do the search as requested, see no results, then try without quotes?

they do this for other things that annoy the F out of me. for example, searching for a phone number. it's a specific set of numbers that is unique. if there's no results for that number (or very few), it gives me results for numbers that are similar like same area code, same prefix, different number. nope. that's not useful. in this situation, I'd much rather see that no results are found rather than making me think there's a result but forcing me to look closely that it is what I wanted

Also, how hard would it be for the majority of the population to learn this? They're perfectly capable of using the internet to find illegal drugs, porn and dirt on their dating partners. It's not like using quotes to get verbatim results is rocket science.

I would prefer a "technical user" mode. Heck, they could hide the option until they've deemed me "technical", by looking at my search terms. Anything is better than giving me pages and pages that don't contain the quoted term I'm looking for.

That mode exists, it's labelled "verbatim".

Verbatim is useful, but it's also going to an extreme.

That's really not what I want quotes to do though, I want a definitive "not found". But it might be a good compromise. Their whole jam now seems to be zero UI except the box, no filters no exact match checkboxes etc. Stuff like that would be really nice to have on the main search.

yeah, but then where's the line? if you start adding checkboxes or other switches to control the search, the next thing you know you have something that looks like a gnarly ffmpeg cmd

A gnarly ffmpeg CLI is exactly what we need for a service as complex as search. The people who can use such an interface are a tiny fraction, but they will produce a majority of economic innovation nonetheless.

Guess what: https://www.google.com/advanced_search

I work for Google Search and yes, that's something we are considering. No promise it will happen, but we do understand it might help people better understand there's no matches for what they wanted (though we do say that now, when it happens).

If there's no results, you have plenty of room for a "No results. Try without quotes?" bright blue link.

We don't need a checkbox because if you use quotes, we will match only content in the quotes, as our post explains. If there's no content at all that matches, only in that case will we show a message making it clear there's no match and that we've conducted a search without quotes.

It soumds like they are only modifying the presentation of results, rather than the quality of the results. Useful, yes, but it's not necessarily what people are looking for.

Ultimately they just want to show you ads for you to accidentally click on

I work for Google Search. Nothing changed in how they worked over the past two years. I think it's just that our snippets weren't reflecting where we were finding the quoted terms, causing people to think they were ignored. Hopefully this change will now help avoid that impression.

It's just not true. Some years back (I think around 2017) I was showing to my colleagues how the phrase

    "The magic of Google is gone"

in quotes brings a ton of irrelevant results, none of which contained the phrase or sometimes even any of the words (and no, it wasn't the suggested "results without quotes"). I also posted about it on an online forum which Google could at one point find as the only result, but now it's gone even though that page is still there, on the same forum. Today this phrase yields zero results and suggests some without quotes.

I could occasionally see irrelevant results for quoted searches until very recently, but I presume it's been fixed now.

Quotes searches bring back pages we can see have the quoted terms. That's how it is intended to work, and that's how does work. I've looked into many of these cases, we find the quoted terms there. In this example, we don't see any pages with that quoted phrase. If there is a page with it out on the web, like the forum post you mentioned, it might be that we no longer have the page in our index -- so we wouldn't match it. Do you recall the URL? It would help me look into that more.

I have a different example of Google not finding a page with "" that ought to be found. Assume I want to find https://board.s9y.org/viewtopic.php?t=18685, which is on a ~20 year old phpBB forum that is certainly indexed. The post is from 2012 after all. And I take one sentence from there and search for it, with quotes:

"Hast du besondere s9y eventplugins installiert?"

This leads to https://www.google.de/search?q=%22Hast+du+besondere+s9y+even..., which so far has only one search result, to http://sw-guide.de/2006-11/serendipity-plugins-seitenleisten.... Which by the way does not contain the target sentence, but that might be because the search moved over into non-verbatim mode.

If I search without the quotes I get more results, but I think not the target thread.

Admittedly, this is not "" not working by showing pages that do not contain the target keyword. But it is "" failing by the search not properly searching its own index.

What's going on there? Why is the index of a rather stale 20 year old forum not complete? The site is indexed in general.

I don't see that particular page appearing to be indexed by us. If we don't have it indexed, we can't match it. Ideally, we'd index all the pages out on the web. But the web is really big.

You're still trying to gaslight HN users about this. Just ADMIT YOU and GOOGLE ARE WRONG.

Thread after thread of hundreds of people telling you your wrong and then you going "actually the community is wrong". Month after month, year after year, you pull this crap.

Admit it, Google search is worse now than it was before, and partly because they are abjectly failing at handling quotes.

it might be that we no longer have the page in our index -- so we wouldn't match it

what the ... ? Why do you delete stuff from your index? Even if it's down there's a good chance it'll be archived ...

It's not that it gets specifically deleted. We just don't have infinite space. There's always new pages coming along. Like trillions of pages out there. So we might not get back to revisiting a page and it falls out of our index while other content is added.

How do you folks think about the fact that sites provide privileged access to Google’s crawler but routinely restrict access/paywall the same content to a normal user? Among chronic offenders are LinkedIn Pinterest, Instagram, various media sites, and etc. are you folks well aware that this is occurring, and are there any plans to discourage or punish that sort of behavior?

Regular unquoted search has declined in the eyes of many people. Quoted search worked and works fine (if you take into account the caveats about tokenisation etc). They are improving the snippets. It is all in the fine article.

Doing some form of lossy tokenization would leave you in a pickle like this, but that's pure speculation on my part.

Maybe they're getting gamed somehow? With this change, it should be easier to see if that's the case.

They're absolutely getting gamed (by sticking clickbait keywords into the meta description or invisible content or even just in plain text at the bottom of the page - you'd think they'd have learned their lesson on Youtube but alas). This doesn't change that.

remember when sites would load keywords up with the same font color as the background, or in divs with display:none or visibility:hidden type of shenanigans? they supposedly penalized sites for that. so you're saying that keyword stuffing is back?

Reply on news.ycombinator.com