Hacker News Reader

Matt, it's great that you think enough of us here to give direct feedback. I wish I could give you some example searches too, but most of what has bugged me lately relates to stuff which might brush up against NDA terms, so I don't feel comfy giving specifics.

Overall I still love Google, I must use the search service 50 times a day so complaints about shortcomings are a bit like grumbling about the paintwork on my free new car :-) So I'll try and identify two recurring (and relatively new-feeling) headaches:

1. quotes. When I put it "in quotes" I don't want ti speell-checked, or cleaned, or made case-insensitive, or whatever else. I would rather get not results and experiment with other strings, than think I've got results that turn out not to be exactly what I searched for. It seems to me like punctuation often gets stripped even if it's inside quotes. For some kinds of searches involving bits of source code or so, this can be a drag.

2. Content farming. I know you are constantly struggling against people gaming the system and so forth. I don't blame Google when I get 50 results of generic junk referring to obscure search terms...the "find [niche product] resellers, hints, tips!" types that are totally generic. but what does piss me off is that a few months ago Google offered a button that would let zap such results, and I could label clutter as clutter. Now on a deep search, I often spend several minutes trying to think what terms are common enough to content farms that -excluding them will prune the results sufficiently that the remaining search results are worth checking one by one.

You remember that Bing ad where someone says something random and their geeky partner starts hypnotically chanting associated but unhelpful phrases, implying an overbroad result set? Sorry, but they had a point.

Thanks for the feedback. Punctuation is tricky, because it only adds value for power searchers, but indexing it would swell the index size quite a lot. Otherwise, using double-quotes should do an exact search.

On content farms, we've definitely heard that feedback. One point up for debate is whether to respond with algorithms-only, or whether we should update our quality guidelines to call out low-quality content farms as a type of webspam. Both DuckDuckGo and Blekko seem delighted to remove sites that most people don't like from the search results. Here's a link for DDG for example: http://www.technologyreview.com/blog/mimssbits/25532/ . The question is: would you feel comfortable if Google removed results that a lot of people don't like from our index? And how do you balance the goal of reducing clutter and junk with the goal of being fair and comprehensive?

One potential problem of removing links that people report is that it's possible to game that system. You'll soon have spammers sending a massive amount of requests to have their competitors removed form the index.

Yes!

and if people wanted a fuller result, you can say, "more lower quality links found. Click here to view all" similar to the way that searching in gmail lets you search in trash & spam.

I know there's no simple answers for these things. On punctuation, consider the (admittedly obscure) situation of searching for some command line string: obscure_utility "-unknown" "-switches" - either the - sign excludes stuff you want, or gets stripped inside the quotes.

Farming-wise, I think you should probably keep all those results in your index, even the ones that are composed of nothing more than your top searches separated by random phrases! even if it's not there now, in future it'll be possible to score page content on whether or not it has semantic value and draw inferences about sites or entire domains that are filled with junk. That will be interesting and useful from security, economic, and scientific points of view. In the meantime people will find useful analyses to run against that 'bad' data in your results which would not be practical if you purged too aggressively.

What I had envisioned (which might be a tad ambitious, but bear in mind that you already have 5% of my local CPU for the asking with the desktop tool installed) is per-user search filtering. I may like sculpture but hate politics, so I would always search for 'statue -liberty', you are the other way around so your searches tend more towards 'liberty -statue'; I would very much like to be able to have complex filters on the client side, either locally or on the client-facing parts of your servers - and not just for spam sites.

DDG takes the approach of allowing regex, which is a neat thing for the people who know enough to want it, and it would be interesting if search patterns and/or selective exclusions (as described above) could be stored locally, as either weighting tables or some sort of white/blacklist - always include wikipedia, never include about.com. I'm already running chrome and using a Nexus one, so perhaps some hashing could take place on my computers rather than increasing the load on your servers.

The other reason besides spam is that lately I find myself wanting to do specialized searches, but I don't know how to specify the bounds of the search space. For example, I'm interested in law. But a lot of legal terms are in popular currency, so even if I search for "theft +legal" I may get tons of results for cheap car alarms or something. It would be fantastic if I could get a large set of results by specifying a large number of domain-specific terms - say, "tortfeasor privity precedent appellate" - and then hash and save that result set as a 'search space'. So then I could do more specific searches and know that my results would mostly come from websites devoted to the subject, with few that mention it only incidentally.

In actual fact, the legal resources searchable via Scholar and Books are fantastic. I just picked law as an example of where you might want to temporarily limit your search set because most people can appreciate the difference between writing specifically about legal topics vs things that just mention the subject in passing. If you want to learn how to write a good disclaimer, "legal disclaimer" is not a good start because every 3rd landing page on the net includes that phrase as boilerplate. If users could save and reuse result sets, we'd get more actionable results, you'd (maybe) get lower server loads but more importantly, every successful hit (where the user doesn't search again or try another result for several minutes) is an implicit vote for the relevance of the result to the set, and thus a valid input to a semantic classifier.

I don't want ti speell-checked

/facepalm

Wow.

So, this whole thing started out as a brief rant that's been in my head for a couple of months. Also, I wanted to get rid of those tabs that had been on the far left side of Firefox for ages. I'm not even sure why I posted it to HN; I just did it and then intended to head out the door afterward, except that all of a sudden my web server became unresponsive.

Anyway: the ability to get results from the past year is great, and I was totally unaware of it. I'll update my post shortly with a note about that. Having something like that sticking right out on the left-hand side would be even better. :-)

I think I often see people complaining about search quality (here, and in the other HN thread I linked to from about a year back, as examples), but if you ask them about the specifics of the search, they don't remember. I think that's a problem that needs to be solved somehow. Although it's at least partly laziness on the user's part, I think Google could view this as a huge amount of potentially beneficial information that they're missing out on. Trying to improve search results without knowing what problems people are having with them is challenging, at least.

It would be nifty if there were some kind of "these were not the search results I was looking for" quick form linked from certain types of search results.

People in Google's search quality team really do like to read posts like this as long as someone mentions the searches they were doing. Most feedback posts don't give us enough specifics to go back and figure out what went wrong and how to make things better.

We do have a "Give us feedback" link at the bottom of each search results page which does some neat AJAX-y things. But even then, we get a lot of "Hi, I'm trying to find my great-great-great-grandfather. He last name was Smith. Thanks!" sort of submissions. Also submissions like "My computer keeps humming. I opened it up and cleaned out the dust, but it still hums. Do I need to clean out the dust again?" And some submissions that say "I would like to rank #1 for all my keywords. How do I do that?"

I've wondered whether a Chrome extension or something similar would give higher quality data for spam or bug reports. It might be worth thinking about offering something like that.

I wonder if it would help to give some feedback to the ones that you actually took into account?

I think I've tried more than a couple of times to give some feedback about the search results. Seeing that nothing happened, and getting nothing but an automated response, I've stopped doing that.

Reporting flashing ads to the ad team worked better, they were removed later the same day.

One of the problems with replying to every piece of feedback is that a link on Google's search page--even a link that many people don't notice--gets a huge volume of feedback. And the resulting feedback can be pretty unhelpful ("I lost my cat. I think you should build a search for lost cats.") So we don't have the resources to reply to all of those. Personally, I would like if we offered a lower-volume feedback method with a higher hit rate of good feedback. Hacker News is acting as the prequalifier in this instance, but I find that people annoyed enough to blog often provide good feedback too.

Overall I agree with you, but even if we had that channel, many times the feedback would be: "Yup, that's a bad set of search results. We'll try to come up with a way to make it better, but it might be a while. Getting an algorithm to do better on this search will be hard."

Why not ask permission to transmit the last few searches and a snapshot of the results. I mean, I usually only hit "give feedback" when i'm not finding the stuff I want. You'll probably be able to discern a lot from that.

The "give feedback" form does automatically populate the feedback form with what you were searching for when you clicked to give feedback. We haven't tried transmitting several searches because people would probably complain, and it's not clear how much previous searches would help. But when we revamped the feedback form, that's one of the big wins we got from making it dynamic instead of a generic form.

I have given feedback on search for "BMTC" not returning the official website of BMTC (bmtcinfo.com). (2-3 months back). But the search results haven't changed since then.

I did a quick check, and that website has some issues. Here's a wget on bmtcinfo.com:

$ wget http://bmtcinfo.com/ --2010-09-14 09:01:43-- http://bmtcinfo.com/ Resolving bmtcinfo.com... failed: Name or service not known. wget: unable to resolve host address `bmtcinfo.com'

The url http://bmtcinfo.com/ just doesn't work in a browser or with wget. And trying to fetch the "www" version of the website, it does a 302 redirect to a deeper url. 302 redirects are the ones that don't pass PageRank. The last time we tried to crawl the page, we got an error--probably from the non-www version. I'll see what we can do to find this site better, but returning errors instead of content to Googlebot really hurts that effort.

By the way, we provide a free webmaster console that would let the owner of this domain self-diagnose these issues. The "Fetch as Googlebot" feature lets the owner send Googlebot to fetch a url on their site and see what Googlebot saw, including errors and redirects. If the owner of bmtcinfo.com were to use that tool, they'd probably notice the errors pretty quickly.

Thanks a lot for the reply. I have passed on this info to bmtcinfo.com through their feedback form.

If you get through to them, it looks like a couple DNS servers in their chain have been serving us NXDOMAINs. Point them to sb[12].mooo.com, hosted on afraid.org, and ask them to let Googlebot in. :)

Almost 2 weeks over. No response from them so far :(

People in Google's search quality team really do like to read posts like this as long as someone mentions the searches they were doing.

Okay, here goes. Lately I have been doing a lot of C# searches. Google suggest strips out the hash (#) from the suggest items, so I keep getting faked out by likely C programming matches. Can you please stop stripping off the "#"? Otherwise, I'll need to attach a hot wire to my down arrow to break the habit. Thanks.

I'm pretty sure that several search quality people are reading this thread now, but I'll try to mention this explicitly to people. Thanks for the suggestion.

Normally Google doesn't allows searches for punctuation, because doing so would swell the index size a lot for very little return in terms of helpful searches. But since we're engineers, we do search some punctuation, e.g. underscores so you can search for sprintf_s, and terms like C# and F#. I'll ask whether the # can carry through into Suggest.

You should really automate this feedback process by mining the web for examples where people complain about Google quality AND they mention the search queries themselves :-)

Thanks, nhebb. ntoshev, now you're thinking like a Google employee. :)

You're a Good Egg, Matt.

>My point on this query is that you can use the left-hand search bar to restrict the results to a certain time range (e.g. only results from the last week, month, or year).

I am amazed at how many users still don't know about this feature. It is one of the best features in google search (that was added recently). Google should promote it better to its users.

I find myself resorting to 'Show Search Tools' -> 'Past x' more and more frequently these days. As the web ages, it seems the proportion of old and potentially outdated information in Google's index to new information is increasing.

While Google Instant is technically impressive, the feeling of speed is undone whenever I have to next manually restrict the results to the past year or 6 months or whatever (and sometimes experiment with multiple durations) in order to filter out discussions from 2008 and earlier. Same with using regular search as well.

Hence, might I suggest you guys consider something like inverting the default search, to only include pages created or modified within the past year (or whatever duration you calculate most optimal for returning relevant results)? Or perhaps changing the PageRank weighting of recency, at least for domains that change rapidly. I know this could cause some problems, and I'm not sure the web is quite at the age where this is necessary yet, but I think there's an inflection point coming soon where more searchers will expect recent results in their top 10 rather than years-old pages.

It might also be worth moving the duration filter up one tier of UI so it is both clearly visible on the search results page, only takes one instead of two clicks to activate. The OP isn't the only one who's recently needed searches restricted by age and didn't realize just from the UI that this could be done. A few months ago I made a similar complaint on another forum and was informed of the same thing - use Search Tools -> Past X. But it's not obvious to do that on the current UI.