Hacker News Reader

Comment by nazka | Hacker News Reader

Hmm what do you mean by could? It’s not possible anymore? So far it seems to be still there.

It doesn't work as it once did, Google will still ignore +keyword, It broke with the introduction of Google's social platform Google+

I work for Google Search. It works exactly the same as + did. We don't ignore the terms that are quoted. That was the point ofvthe change we made, that because the quoted terms weren't sometimes in snippets, it might seem like we were ignoring when we weren't. Hopefully this change will help.

> We don't ignore the terms that are quoted.

This has not been my experience. I regularly use quotes and get back pages that don't contain the phrase.

The article suggests that content might not appear because it is in the source of the page, or the page has changed since it was crawled. However, I've grepped the source of the page, and checked Google's cache of the page, and not found the quoted phrase.

I used to report bad results to Google, but it's a complete black box and I have no idea if anyone pays attention to them, so I don't anymore.

If you ever find an example where you think they don't work, please ping me. I've looked into many of these cases, and the material does appear on the page as we've seen it.

On the page: "https://book.simply-logical.space/src/text/1_part_i/1.1.html"

If I take the quote: "At the beginning of this section, we posed the question: can we show that our two definitions of the nearby-relation are equivalent?"

Searching on google lead to 2 dead link, and another website republishing the book as pdf.

Others search engines give me the link of the page I got the quote from, as the first result.

That's not a sign that quote search doesn't work. In fact, the opposite. The pages you get all appear to have that really long quoted phrase. Instead, it's a sign we don't have one particular page with that quote indexed. Certainly better if we did have that page included in our index, of course. I'll pass it on.

> The pages you get all appear to have that really long quoted phrase.

Two of three are dead links, so not really.

While indeed google doesn't look to have indexed this particular page, it looks like others search engine successfuly indexed it.

The "relevant"(where this verbatim string came from) result looks like a pdf dump websites, which is less relevant to what I want.

I also get the dead link, and the 2nd link is now this comment :)

But a false negative (not getting a page that has the phrase) isn't the same as a false positive (getting a page that doesn't have the phrase), which is what the GP is talking about.

Sad that the only way to get Google to “listen” is to back channel requests via some obscure web forum and not through any official google-provided channel.

Please post a general comment on how to submit issues/features to you; created top level stub, but feel free to create your own too:

https://news.ycombinator.com/item?id=32354453

Here's a separate issue I ran into the other day that almost made it impossible to find something Google had indexed. Type in "mumbai comes to norway," and you get a page and a half of results, with the date on almost all of them being 2011. Now give it a time range between 1900 and today (IE, all results with a time stamp should be included). All of the results disappear except for one.

For whatever reason any effort to use the time range can remove _many_ search results that Google knows are from that time range and that should be included. That can make it almost impossible to find old articles, particularly since Google pushes new less relevant hits if you don't restrict the time range.

The only way I was able to find this article the first time was by searching through social media, which sometimes has less crazy search algorithms.

See this Twitter thread I shared (I work for Google Search) on our before/after commands we added to make it easier to do this type of searching. It also explains why it is sometimes difficult for us to determine the date of a document: https://twitter.com/searchliaison/status/1115706765088182272

I don't think this has to do with the issue I mentioned. Google has gives a date for these articles when you do a regular search. As I mentioned, these are mostly around 2011. Again, these dates are coming from Google, not the site. If I do a time range that includes 2011 in it, even a time range that includes the entire existence of the internet (1900 to 2022), all of the sites except one disappear. It's not a matter of Google getting a date wrong, it's a matter of Google not displaying hits that it should be displaying per the data in its own system.

I've also seen cases where the quote seems to be ignored. I don't have a full theory yet but I think it might be when the contained string is too long? Next time I see it I'll send it to you.

I've used quoted searches now and then in the past, as recently as last week or so, and it definitely did not work. There were no usable results, none of the results shown did contain the exact words. Is there a bug somewhere?

If you have an example you're comfortable sharing, I'm happy to take a look with the team (I work for Google Search). It's designed to match content as described in the post.

Search "xaverius franciscus"

Only one of the first six results contains the exact phrase, the others all have other characters between these two words.

Search +"napoloen" And it insists on including napoleon.

Please include an 'expert' mode or something for people who know what they are looking for.

Every result I looked at on the first three pages have those words, and those words being clearly indicated in the snippets. Some of them have punctuation between the words. As the post explains, we see punctuation as spaces -- please do look at the post about this, if you haven't. As for the + operator, that hasn't been supported for ages. Putting the + symbol means nothing. You just need to do "napoloen" for an exact match on that. Which, if you do, causes us to say:

Showing results for "napoleon" Search instead for "napoloen"

So while you said you wanted a quote search, our systems are really really concerned that for this spelling, you're making a terrible mistake. Which, I get as a power user, is annoying. And which, we probably should review and perhaps never substitute a different search.

But ... we are telling you that happened. And providing a link to override what we did and say yes, I really know what I'm looking for, do it. Which -- if you click on, we'll do.

Thanks. Yes, I understand about the punctuation now.

About 'napoloen', I often search for non-english words that differ only one letter from an English word (English is my fourth language). I wish there was a way to permanently disable "Did you mean?"-results.

A new one... I just got this, searching for: libmad +tutorial

Several results with "Missing: +tutorial"... Why are results included that do not contain the word 'tutorial'?

Again, the + symbol isn't a command and does nothing. If you searched for [libmad +tutorial] we saw that as as search for pages with both or either of two words, libmad OR +tutorial, plus any related words or synonyms, like tutorial.

So with that search, we're trying to show you things we think are relevant, and there aren't a lot of pages relevant with the word +tutorial on them. Tutorial, yes -- but not +tutorial.

In this case, what you really want is this: [libmad "tutorial"] -- that says get pages that have both words or words related to libmad but ONLY pages that actually have the word tutorial on them.

libmad "tutorial". The + does nothing.

So basically always double quote every word. Thanks.

Are you aware of the "Tools -> Verbatim" mode?

Yes I am very aware of this.

Can it be set as default?

Verbatim will still include results with other characters between "franciscus xaverius"than a space. Like a ( or a dot. It includes results containing a phrase that ends with Fransciscus. And the next phrase beginning with Xaverius. This is useless when searching ancestors.

I know the service is free, but it would be really cool to be able to search for any (ASCII) string verbatim.

If can be if you add a custom search engine

https://news.ycombinator.com/item?id=32351006

This implies "we do ignore the terms that are not quoted", which has been a common source of grievance with Google.

Planning your trip ahead, you search for pizza restaurant Chicago. Google ignores 2/3 of your terms and returns local restaurants that serve Vietnamese.

We don't support quoted restriction for local results -- results in a local box -- as the post explains. It also says that's something we're looking at for a further improvement.

What is "a local box"? It sounds like GP was performing a regular web search. Surely you don't mean you ignore quotes just because I put the name of a city as a search term.

No, we don't. Which is why I specifically said local box (it's when we show local results almost always with a map -- the post explains this more). If it's a regular web search, and these are web pages about a local place, quotes work the same as with any web page.

Sounds like you use Google Fit, they will probably optimize your food recommendations based on your health information. ;-)

I don't have relevant queries at hand as you only notice when searching for specific documentation, however terms often weren't in the quoted snippets, but most critically not in the page and nowhere in the cached content as well (assuming the cache reflects what has been indexed). I could simply search for any of the term and not find it at all. Similar case for Bing BTW.

Assuming this is going to improve, were I would submit an example of this not working in the future?

Yes, I'd be happy to review an example if you have it come up in the future.

Google often returns results that just don't contain keywords, not talking about snippets or stop words im talking about the first keyword not appearing in any of the results on the first page but it shows up on page 2.

They seem to be denying this left right and centre on this item. In another thread they said it is due to bugs.

When Google dropped the special meaning of a ‘+’ prefix, the change was not obvious to me (and, I imagine, many other people) because some responses would nevertheless contain the specified term. Consequently, the experience came across as just another degradation of quality.

No, there are bugs, documented them; moved my reply to a top level comment here:

https://news.ycombinator.com/item?id=32354078

Core issue being it feels unnecessarily hard to replicate search results; as documented in comment linked to above.

That's unfortunate, but was never an intended feature and at odds with the whole design of Google search and the reality of web scale data processing and personalization , so can't call it a bug. You want a feature that Google doesn't offer.

I hope Google can bring back the plus operator, as Google+ is now buried.

Quotes do the exact same thing that + did back then.

Why not simply revert to the old, sane option?

money.

I use to be able to filter out websites like amazon from shopping search results with -, no longer works and I hate amazon.

-site:amazon.com works for me.

What about the "-"?

It still works, if you mean the - symbol.

No it doesn't.

https://i.ibb.co/k18PXhW/google-search-2022-08-05.jpg

I think what trying to exclude the .br TLD. Have you tried the "site:" syntax which can include or exclude domains. Loke this:

teste de lingua portuguesa -site:.br

I often do it this way because it used worked in the past, and it works in other search engines.

Well this is a more complex case which includes combining negation with quotation and punctuation marks. I'll bet those factors explain why it didn't work.

Searching for 'Edgar Allen Poe -"The Raven"' does work as intended.

The search I printed was trying to omit some addresses (and I now realize quotes are not necessary and do not help for this specific case), but it did find a result with the term I was trying to avoid as a string.

I'm not an expert, but isn't this as simple as "omitting results that have whatever is found inside the quotes"?

I don't understand why your comment is being downvoted. Downvotes shouldn't be used as a form of simple disagreement. I made a simple question, you gave it a simple answer.

I think that actually raises an interesting question:

Should downvotes be used when a comment states a fact, but the voter knows or believes that fact to be false?

In all fairness, it does work without the quotes.

Surely someone internally would have foreseen that? How in the hell was that greenlit.

They have foreseen it, and the plus sign was replaced by quotation marks.

I thought these meant something else:

+: the term following must be in the text (but may be stemmed)

"": search for exactly this phrase (not stemmed)

So for example: +hiking => a term like hike, hiking, hiker, etc. must be on the page

+"hiking" => the string "hiking" must be on the page

"hiking" trekking outdoor => look for the string "hiking" and preferably return results that contain it, though hits for trekking and outdoor (stemmed) suffice

But apparently I'm wrong, and "" and + don't mean that any longer?

+ hasn't been an operator for years with us now. Quotes mean find the exact term indicated, no stemming, nothing -- just an exact match.

That's about as user unfriendly as it gets. Such things work their way into your muscle memory and to annoy 100% of your power users for some social media play that has nothing to do with search is textbook monopolistic behavior.

Definitely their jump the shark moment for me.

I know people (me included if I'm truly honest with myself) are still peeved about Reader but Google Plus was the Search service's sharkjump for me I think.

Don't teach me how to use your engine perfectly for 20 years then change it for some half baked social thing lmao.

I hope someone got a pisstake massive raise off it so it was at least worth something to someone!

It just happened to me, where I added a "-word" term to narrow it down and the number of results increased somehow. That means that they're doing something behind the scenes that's not as simple as "remove all results that contain "word".

I work for Google Search. Yes, totally agree, it's weird when the count goes up. It's something I hope we can fix. But it's not because we ignored the - symbol. It's because the counts are very fast, rough estimates. And when you go into additional pages, we start to refine them. IE: we actually found more than the count displayed on the first page.

> I work for Google Search

A tangent. You have what I imagine is a hard job and, despite the many threads here about recent apparent quality decreases, the product brings enormous value to its users.

Thanks for taking the time to discuss in this thread, it is very helpful to see your insights.

Thank you. Forum discussions here and in other places are really useful ways to gather feedback. It directly drove the improvement we launched this week.

The "number" of results doesn't mean anything beyond a rough guess. it's an ad promlting the size of the index. You'll never be able to reach nearly all of them.

IIRC not really, at least some times it will still return documents that do not have the exact "+'ed" term in it.

- works, quoting a word was the same as +

- doesn't work for pinterest

When I enter the exact string "bed design ideas" into google images it includes pinterest results, while "bed design ideas -pinterest" excludes them. Is there a specific failure case for you?

-site:pinterest.com

works perfectly, and i use it in my all queries where i search for images.

This should be the default, TBH. Why google thinks that an image search leading to pinterest is going to be helpful is beyond comprehension.

If the original source is still online I indeed prefer it, but occasionally some images of interest only survive on Pinterest, in which case a Pinterest result is better than nothing.

Another thing is probably how images are indexed in the first place. Historically image search could only look at the text surrounding the image (and the alt-text, if provided) to get some sort of tenuous idea what an image was about. Pinterest provides some additional tags for each image, which can help make it more discoverable.

Of course these days Google also runs image recognition (and OCR) on the images it indexes, but I suppose sometimes the tags added by Pinterest still provide some additional context without which the original image wouldn't have been found.

>Why google thinks that an image search leading to pinterest is going to be helpful is beyond comprehension.

Both founder/ex ceo and current ceo came from Google. Pinterest directly deposits >$xxxmil/quarter into google.

I've found that sometimes I have to also exclude specific "*.pinimg.com" (or something similar; don't recall the domain for certain offhand) domains, especially when doing an image search.

I use two Chrome extensions “Unpinterested!” for Google and “Not Pinterested” for Ecosia that both remove Pinterest’s search engine effluent from results.

This is the right answer and I was under the impression that this was in the the Google fu crash course these days.

"-" still works, "+" was bastardized when Google launched Google+ but you can get similar functionality by surrounding your term in quotes, preventing it to be dropped.

“-“ works but I’ve seen it fail so many times that I’m more surprised when it actually filters it properly nowadays.

Reply on news.ycombinator.com