Ask Slashdot: What Features Would You Like In a Search Engine?

New submitter nicolas.slusarenko writes Nowadays, there is one dominant search engine in the world among few alternatives. I have the impression that the majority of users think that it is the best possible service that could be made. I am sure that we could have a better search engine. During my spare time I been developing Trokam, an online search engine. I am building this service with the features that I would like to find in a service: respectful of user rights, ad-free, built upon open source software, and with auditable results. Well, those are mine. What features would you like in a search engine?

Comments

privacy?

[tommeke100]

Next to working well, maybe the assurance that not all your search queries were logged and sold to third parties or used for advertisement?

Re:privacy?

[smittyoneeach]

So, how much are we willing to pay as a subscription to make up the lost revenue?
Or do we just expect the vendor to eat the opportunity cost?

Re:privacy?

[rudy_wayne]

Here's a radical idea. How about if the search engine returns results that are ACTUALLY WHAT I'M SEARCHING FOR.

One day I was trying to find a particular video clip. No matter how i tried to refine my search query, all I got was completely irrelevant bullshit. Now, the video I was looking for was somewhat old and obscure and so its entirely possible that it doesn't exist anywhere on the Internet. That's fine. I can accept that.

But, if that's the case, then my search should return zero results. Not thousands of irrelevant results.

Re:privacy?

[TWX]

It has gotten a lot worse, hasn't it?

I want a search engine to identify when someone is attempting to manipulate it and to counter that. I don't want Google Bombs like "miserable failure" regardless of how I feel about the actual politics, to make the results useless. I'm not so childish as to expect an echo-chamber everywhere I look.

This means no more companies whose entire existence is to try to improve someone's search rankings.

As to data being collected, I'm actually okay with the top 80% of searches in a given day being used for advertising revenue, assuming no geographic data beyond nation, and no personally-identifiable data is collected. That's how a search engine would make money, by selling ads based on what people want to know about. If Ford has a press-release about the new Focus, and people search for that, I'm okay with ads related to the Focus or to Fords coming up. I just don't want more than "this term is being asked for this many times on this day" to be reported.

Re:privacy?

[Jane Q. Public]

If there were millions of users, how much would it COST to keep it operational? Figure that out, then add some reasonable profit percentage. Win-win.

Do you think you could run an operation like that for around $3.5M a year, given 1M users? Great. Charge $5 per year per user. I'd pay it.

Re:privacy?

[Rei]

I just want the search engine to stop changing what I'm searching for. I don't want to have to quote every word like I have to do with Google to make sure that the word is actually in the page, and by "the word", I mean "the word I type, not a word that Google things may be similar to the one I typed". It's worst when you're searching for foreign words, product names, acronyms, or whatnot and Google tries to treat them as if they're English words and declines them or chooses synonyms.

"Did you mean X?" is fine. Even "Searching for X (see original results here)", if you're very confident that the person made a common spelling error or whatnot. But just going in and swapping out words as if this is expected behavior? Terrible. At least let me disable it if you want to do that...

Beyond all this: I do like how one can do simple commonn operations on Google - math, conversions, etc. The more of these the better IMHO, so long as they have a standardized format - be they tracking numbers, flight lookups, whatever. It's okay in my book to be a bit Wolfram-y.

Keep the interface plain, simple, the sort of thing that'll work on any browser, from a modern Chrome to a simple text-only browser. Only use javascript where it's not essential for the site to work. Here's an example of something that would be a good use of javascript: if you need to track clicks, like Google does, do it through javascript rather than by having a link redirect like Google does. I hate how I can't just right click and copy link on Google without getting some massive Google redirect link.

Just my thoughts. :)

Re:privacy?

[smittyoneeach]

The cost would seem proportional to the users.
As far as running the business, just let me sharpen my AWS knives a little more. . .
Of course, in our post-capitalistic era, just having a good business plan is necessary but insufficient to protect you from Big Oligarchy and Big Government, assuming you could pry them apart with a crowbar. If you offer a viable alternative, look for "something really, really unfortunate" to arrive with an innocuous name like "Net Neutrality", or something.

Re:privacy?

[dwywit]

Try google verbatim. Saves having to put quotes on every word.

What was wrong with '+' as an operator, anyway?

Re:privacy?

[jenningsthecat]

I just want the search engine to stop changing what I'm searching for.

This, exactly. Google's ideas regarding 'synonyms' for my search terms would be laughable if they didn't waste so much of my time. Also, these days when I do an 'allintext' search it almost always turns up far more results than did the same query without the 'allintext' operator. Now just how in the fuck does that happen?

I would pay two or three hundred dollars a year for access to a search engine with Google's reach and power, but without all the ad-oriented bloat, the lowest-common-denominator attempts at hand-holding, and the Microsoft Clippy-isms. You know - something that's more suitable for real research and for getting a job done than for figuring out where to have dinner or what meaningless bullshit the Kardashians and other such social parasites are up to. And while they're at it, they need to include a way of searching for exactly what I type, including case, punctuation and special characters. And if my search turns up zero results, that's fine. I'd far rather have that than be insulted by Google's insistence that it must have something I'm interested in.

I'm not so naive as to believe that anyone else can replicate Google's massive search capabilities. So I really wish Google would provide a search interface for those of us who have both a good idea of what we're looking for and a clue about how to do research. It would cost them next to nothing, they could charge for it, and they'd be doing the world a favour.

Hell, right now I'd settle for Google circa ten years ago - it was way better than it is now.

Re:privacy?

[Kohath]

They could sell ads without tracking people.

Re:privacy?

[skegg]

Interpreting as:

not a word that Google thinks may be similar to the one I typed

Interpret instead as:

not a word that Google things may be similar to the one I typed

Re:privacy?

[Kohath]

Equal to what? Why should it be equal? A company like Yahoo would be very happy to have a search that was half as profitable as Google's.

Re:privacy?

[Jane Q. Public]

If you offer a viable alternative, look for "something really, really unfortunate" to arrive with an innocuous name like "Net Neutrality", or something.

Since you brought up Net Neutrality: can you give me one example of a small, innovative startup in the ISP business in the last 10 years?

No? Why not?

I'll tell you why not: it is not a competitive market. It's a de facto oligopoly, with barriers to entry that are far to high for the little guy to vault.

That's where government's legitimate role begins: by regulating or breaking up oligopolies and monopolies, so that there can be competition in the marketplace.

Don't expect market forces to fix a problem when there isn't a real competitive market in the first place. That's not very realistic.

And I'll answer my own question: the only "new" broadband ISP to make a name for itself in many years now is Google, and they can hardly be called a "little guy". They are one of the few groups that has the money and muscle to elbow its way in to a business that is very hostile to outsiders.

Re:privacy?

[fahrbot-bot]

Try google verbatim. Saves having to put quotes on every word.

What was wrong with '+' as an operator, anyway?

According to Google Drops Plus Sign from Search Operators

It has to do with limiting confusion about the search engine’s social network, Google+.

To Baio, “it seems obvious that they’re paving the way for Google+ profile searches. When Google+ launched ... they coined their own format for mentioning people – adding a plus to the beginning of a name... The fate of the ‘+’ symbol was clear: protect a 12-year-old convention loved by power users, or bring Google+ profile searching to the mainstream? It was doomed from the start.”

Re:privacy?

[w_dragon]

Actually a search engine is one of the few things where the cost depends less on use and more on the amount of the Internet you would like to index. It takes a lot of storage and processing power to create an easily-searchable index of the Internet.

Re:privacy?

[smittyoneeach]

Well, there are incubators that support some shiny new thing, e.g. http://en.wikipedia.org/wiki/Instagram but they tend to Do One Little Thing Well, a la Unix, and then sell their soul for rock-n-roll.
But we think about IT here. There are also business in the Real World http://www.groundedcoffeeshop.com/ but both of these examples underscore your broader point, if I may radically restate it: economic activity of consequence is a top-down, not a bottom-up affair these days.

Re:privacy?

[bloodhawk]

There are other ways to monetize your product. If the engine is actually good and popular then they can monetize licensing for mobile devices, selling into the corporate space with indexing and appliances for the enterprise, definitely not as profitable as advertising and given greed is number one priority for companies like google, and just about every company for that matter, it isn't likely to happen.

Re:privacy?

[smittyoneeach]

You really gotta walk back that excessive, exuberant, irrational optimism, Sparky.

Re:privacy?

[rtb61]

Seek government funding as an independent service to the community to ensure equal access to all business by all the people. Attempt to create a level playing field, obviously the corporate greed driven solution is not working quite that well. Many companies, in fact likely the majority would be better off setting up an organisation to fund independent equal search with a fixed set of rules to set ranking results, likely in partnership with government.

Designing a search engine. The two obvious things are optional categories (too many categories can be problematic and difficult to use) and regional search with various levels of region, country, state, city, suburb. Next up registered users with, the user being able to block sites they have no interest in, from showing up in their results (and hopefully this information collated to push those places to page one hundred thousand and one of search results).

The flip side of having fully tracked and monitored search is better individual control of search but that needs to be a choice of the user. In that regard also a thumbs up for a good search result, to promote those sites, taking into account the search query and category and region (as well as of course negative responses). Taking on board properly registered volunteer users willing to contribute some time to the community at large by helping to evaluate search results and being one of the quality control methods to ensure fair and good search results. Some level of randomness in search output where, the other factors are equal to ensure more players get a shot at the first three pages and their location on those pages.

The search rules publicly defined and seek community support in establishing and protecting those rules from interference, at a civil suit an criminal prosecution level ie is corrupt SEO interfering with a persons computer network access by purposefully misdirecting their search results?

Re:privacy?

[Jack Griffin]

And, have it do an "untailored" search by default. If I type in a search phrase I get results relevant to where I live. This can be useful in the odd case I want to find a local service provider, but the 99% of the other times I want non-local, fully globalised results, and I can no longer seem to be able to that.

One example, we had a local rugby player Jarryd Hayne who went to the US to gamble on an NFL career. No matter what I search I only get results from Australian media outlets which I don't want. I want to find out the reaction in the US media (if any), but the search always tailors my results to my location which is annoying. This results in my world view being skewed because I'm increasingly only being exposed to local opinions
The Internet is supposed to be a global village. Google ruining that.

Re:privacy?

[Runaway1956]

Have you tried a VPN? If I log into my VPN through the UK, I get tons of results from Google that seem to be UK-centric. If I log in through Dallas, Texas, the results look very much the same as not using a VPN at all. Log in through Denmark, and everything is presented in some language which I presume to be Danish, and I have to click the little UK flag to get English results. I don't know if I can log in through Oz or not - I may have to check that out. There is a LONG list of places that I can log onto the internet through, but I've only tested a few of them.

You might give it a try, if you want regionally / nationally / culturally flavored search results.

Yes, I realize that you've asked for unflavored results, but my idea offers different flavors, which might satisfy your cravings after all!

Re:privacy?

[smittyoneeach]

The whole point of this thread is that you don't go dictating to corporations to take a haircut. Unless you're the government.

Re:privacy?

[smittyoneeach]

The problem with this is simple: the moment you set up a paid search service, you have to restrict access to those people who have paid the subscription fees.

How so? You're going to have all sorts of promotional access.

Re:privacy?

[jtev]

Instagram may be banal, but it is an innovation. When it was started it was a disruptive technology. It may be a bit "weak" of an innovation, but you didn't actually refute the argument.

You're right that these are not ISPs, however they are companies that would be exploited under a non-neutral net. It's not the ISP barrier to entry that is the problem that net neutrality is supposed to address, it's the barrier for those who provide services on the internet, which he did mention.

Re:privacy?

[gzuckier]

Actually a search engine is one of the few things where the cost depends less on use and more on the amount of the Internet you would like to index. It takes a lot of storage and processing power to create an easily-searchable index of the Internet.

Indeed. Economies of scale most definitely doesn't sound like something an internet search engine would experience. Quite the opposite.

Re:privacy?

[gzuckier]

It has gotten a lot worse, hasn't it? I want a search engine to identify when someone is attempting to manipulate it and to counter that. I don't want Google Bombs like "miserable failure" regardless of how I feel about the actual politics, to make the results useless. I'm not so childish as to expect an echo-chamber everywhere I look. This means no more companies whose entire existence is to try to improve someone's search rankings. As to data being collected, I'm actually okay with the top 80% of searches in a given day being used for advertising revenue, assuming no geographic data beyond nation, and no personally-identifiable data is collected. That's how a search engine would make money, by selling ads based on what people want to know about. If Ford has a press-release about the new Focus, and people search for that, I'm okay with ads related to the Focus or to Fords coming up. I just don't want more than "this term is being asked for this many times on this day" to be reported.

amen.
I'm sick of searching for "painful rectal itch" and having the first two pages of results all offering to sell me the best painful rectal itch at the lowest prices.

Re:privacy?

[gzuckier]

I just want the search engine to stop changing what I'm searching for. I don't want to have to quote every word like I have to do with Google to make sure that the word is actually in the page, and by "the word", I mean "the word I type, not a word that Google things may be similar to the one I typed". It's worst when you're searching for foreign words, product names, acronyms, or whatnot and Google tries to treat them as if they're English words and declines them or chooses synonyms.

"Did you mean X?" is fine. Even "Searching for X (see original results here)", if you're very confident that the person made a common spelling error or whatnot. But just going in and swapping out words as if this is expected behavior? Terrible. At least let me disable it if you want to do that...

Beyond all this: I do like how one can do simple commonn operations on Google - math, conversions, etc. The more of these the better IMHO, so long as they have a standardized format - be they tracking numbers, flight lookups, whatever. It's okay in my book to be a bit Wolfram-y.

Keep the interface plain, simple, the sort of thing that'll work on any browser, from a modern Chrome to a simple text-only browser. Only use javascript where it's not essential for the site to work. Here's an example of something that would be a good use of javascript: if you need to track clicks, like Google does, do it through javascript rather than by having a link redirect like Google does. I hate how I can't just right click and copy link on Google without getting some massive Google redirect link.

Just my thoughts. :)

I'm in the other camp on this. I appreciate whatever it is that Google is doing to widen my search to find things that might be what I'm looking for that I don't have a precisely verbatim search term for. That's something Google seems to do well. It's their ability to produce crap completely unrelated to any of the search terms due to some clever hack by the search optimizer scum that annoys me.
and while I"m at it, I"d like to see more stuff that is something a person might be searching for, and fewer results that are some high school kid's twitter feed.

Re:privacy?

[Old Grey Beard]

Justifiably, there are a lot of comments here along the line of "too many results". Apparently, search engines think that lots of results are preferable to few results. In a way that makes sense because if you get the right answer (more properly, a right answer) early on, what do you care about the following million irrelevancies?

The /. mindset seems to be "...if only I could enter the right (possibly nonexistent today) command or switch or whatever, then I'd get the right results early in the results list, and to heck with the rest." Personally I think this "forward" approach is unsolvable. If it were doable I think we'd be a lot closer to it today.

Instead of the "forward" approach I suggest a "back-end" approach: if the best a browser can do is hand you a boatload of possibilities, maybe the browser should supply you with tools to refine those possibilities. One idea along these lines is to consider your list of results as a series of entries in a relational database, and let you refine the list with, say, SQL-like commands. For example:

DELETE ALL ENTRY IN RESULTS IF ENTRY.sequence > 1000;

-- Limits further operations to the first 1000 results. Note this doesn't delete the actual pages (!), just the entries in the "database" of search results.

FOR ALL ENTRY IN RESULTS IF "clinton" IN ENTRY.url AND "Hillary" IN ENTRY.text THEN DELETE ENTRY;

-- Gets rid of all pages containing "Hillary" if the url contains "clinton". Yes, it's a lot of typing but the browser saves all commands and you can create a file that is auto-executed for every search result, if it means that much to you.

FOR ALL ENTRY IN RESULTS IF ENTRY.date > '2014/06/20' and ENTRY.date < '2014/12/21' THEN PASS ELSE DELETE ENTRY;

-- Removes everything except pages made in Summer/Fall of 2014. It must have been a very good year.

And so on.

About half of the /. community is about ready to pounce on this as being unworkable. Commands that require searching lots of url's, even if limited to 1000, can take a long time to execute. At the moment this is true but that doesn't mean things will stay that way. "All" it takes are support hooks in the master database. E.g., "Is 'Hillary' in <url0>,<url1>,<url2>,<url3>,<url4>,etc." Presumably the search index can answer this without any network I/O.

Probably what's needed is a research project that identifies what kind of properties are useful in the "results database". No need to build your own search engine -- just send a search command to Google (under program control) and harvest the results, building the "results database". It's not that hard. Building an SQL-like command parser isn't that difficult either, assuming knowledge of lex and yacc, or the like. It's a bit more difficult to figure out what properties and commands are useful. (If any!)

Summer Of Code, anyone?

Re:privacy?

[david_thornley]

When I make a search that might have local vs. global results, at least 90% of the time I want local. I might search for restaurants, for example, and although it may be theoretically interesting to learn of restaurants in New York City that I'd like, a list of restaurants in the Minneapolis area is far more useful to me.

Re:privacy?

[flacco]

Conventional methods of payment would tie one's searches to one's identity. I would want assurances that it was not technically possible to log my searches.

Re:privacy?

[smittyoneeach]

How would you do that, unless there were a trusted proxy running your searches?

Re:privacy?

[Jack Griffin]

That's fine, but give the user the choice. Google used to have a toggle for local or global, now I'm forced into local only.

Re:privacy?

[Jane Q. Public]

Well, my comment was really meant in the context of ISPs.

Sure, there are small innovative companies. Like Instagram and even Netflix (it didn't start out big). BUT... what about companies that bring those services to the consumer? The ISPs? That's where Net Neutrality really comes in, and they have erected huge barriers to entry for anybody small (or innovative).

Re:privacy?

[smittyoneeach]

The barriers are exactly the issue I think that voters should focus on.

Re:privacy?

[Jane Q. Public]

The cost would seem proportional to the users.

Of course. Did you not see in my sample calculation "$3.5M given 1M users"?

However, the economies do not scale linearly. You make an investment in infrastructure, and it's good up to X users. Then you make another investment, it's good up to X times 10 users. Etc. In practice it's mostly a step function, not a straight line.

Re:privacy?

[Jane Q. Public]

The barriers are exactly the issue I think that voters should focus on.

And I repeat: that's where Net Neutrality comes in. It serves to keep those barriers low, in a "market" (ISP) that is not competitive. Otherwise you'd get "tiered" service which keeps those small players small.

Re:privacy?

[smittyoneeach]

The barriers are competition at the last mile and cartel behavior from providers. Having the ultimate cartel, the government, manage the rules is not the same as actual marketplace competition, I don't care what the FCC pencil-necks say.

Re:privacy?

[smittyoneeach]

Not in dispute. As noted above, I'd missed "ISP" as well. I think I was exhausted when I replied initially.

Re:privacy?

[Jane Q. Public]

Having the ultimate cartel, the government, manage the rules is not the same as actual marketplace competition

Of course. But when you don't have competition anyway, nor any good way to create any, regulation is a better-then-the-alternative last resort.

Re:privacy?

[smittyoneeach]

Recreational legislation is an evil exceeding premature optimization.

Best feature

[Goglu]

Confidentiality

Re:Best feature

[Anonymous Coward]

Google and Bing now keep track of which search results you click on. When you hover your mouse over a search result they use JavaScript to show you what the final destination URL will be in the browser status bar, but when you click the link it takes you to a Google or Bing URL to record your selection and then it redirects you to the final destination. Very sneaky, they didn't used to do that. Plus many people log into gmail and leave it logged in all the time, so Google can attach your search queries and result selections to your profile. They are engaging in a war on privacy.

Re:Best feature

[ScentCone]

They are engaging in a war on privacy.

Or, they're trying to make a viable business out of operating mammoth server farms so you can get stuff you don't have to pay for.

Re:Best feature

[Impy the Impiuos Imp]

It isn't just Google or Bing that does this.

Re:Best feature

[ScentCone]

They didn't need to record my search queries or selections for the first 15+ years of operation. Why now?

More competition and plummeting ad revenue.

Simple

[JMJimmy]

Search for what I type in, now what you think I want. I'm so sick of having to change every search to "verbatim" because my search terms are being ignored. I'd switch to someone else but they seem to be carbon copies.

Re:Simple

[CarlosM7]

Also, don't remove search terms, and I want all of them "anded", not "ored."

Re:Simple

[Anonymous Coward]

Add to that:
  * "Search instead" link. My typed search becomes a secondary alternative. And by the way google, a search which return no or few results generally carries more information than your rubbish interpolated mess of results.
  * Switching to a different language depending on my location. When I go to France, I don't necessarily want French results. And if I go to Germany or Poland, I will not automagically speak German or Polish. Duh!

Re:Simple

[swillden]

Search for what I type in, not what you think I want

I want a search engine that searches for what I want, not what I type, and not even what I think I want.

Re:Simple

[JMJimmy]

Search for what I type in, not what you think I want

I want a search engine that searches for what I want, not what I type, and not even what I think I want.

I can't tell if that's sarcasm, stupidity, or trolling... you actually want a search engine that when you search it ignores your input (ie: gives top results saying "missing: stormtrooper" from a search specifically for stormtrooper information), ignores what you tell it not to give you (ie: -books will still give Google Books results) and generally give you crappy results (ie: try searching for terms which happen to coincide with a syndicated news story - it's impossible to filter out every news outlet that picks it up)?

Re:Simple

[swillden]

Why would I want crappy results? I want it to give me what I want, which by definition isn't "crappy".

Re:Simple

[JMJimmy]

Why would I want crappy results? I want it to give me what I want, which by definition isn't "crappy".

And you think a system built by man can divine what you and everyone else wants at the moment you type it in? That'll be the day. Until then, assume I know what I want and not your system.

Re:Simple

[turkeyfish]

Evidently, this is akin to the SCOTUS definition of obscenity. I can't tell what I want, but I'll know it when I see it.

Re:Simple

[swillden]

Why would I want crappy results? I want it to give me what I want, which by definition isn't "crappy".

And you think a system built by man can divine what you and everyone else wants at the moment you type it in? That'll be the day. Until then, assume I know what I want and not your system.

I think systems built by man that knows a sufficient amount about me, my interests and my needs can. We're not there yet, certainly, but the question was what I want... and that's it.

Put it this way: Suppose you had a really bright personal assistant who knew pretty much everything about you and could see what you are doing at any given time, and suppose this assistant also had the ability to instantly find any data on the web. I want a search engine that can give me the answers that assistant could.

Re:Simple

[Bill Dog]

It shouldn't be about what you want, or what you think you want, but what you need. I want (need?) a search engine that will give me that.

Re:Simple

[Oligonicella]

False analogy. There's a huge difference between a personal assistant, who by definition *I* know personally, and a faceless business entity who I know not at all (read adversarial entity) scraping 'enough' information about me to presume it knows me sufficiently to second guess what I want and give me that instead of what I requested. Truthfully, why on earth would I trust such an entity?

That's the problem with hypotheticals. They don't reflect the reality we live in.

Re:Simple

[swillden]

False analogy. There's a huge difference between a personal assistant, who by definition *I* know personally, and a faceless business entity who I know not at all (read adversarial entity) scraping 'enough' information about me to presume it knows me sufficiently to second guess what I want and give me that instead of what I requested.

Not really.

I'd say there's a good argument that all of the information I give Google actually exceeds what a personal assistant would know about me. The real difference (thus far) lies in the assistant's ability to understand human context which Google's systems lack. But that's merely a problem to be solved.

Note, BTW, that I'm not saying everyone should want what I want, or be comfortable giving any search engine enough information to be such an ideal assistant. That's a personal decision. I'm comfortable with it... but I'm not yet getting the search results I want.

Re:Simple

[stoborrobots]

Reminds me of this old chestnut...

Re:Simple

[JMJimmy]

So you want a magic ball

Re:Simple

[JMJimmy]

That's my entire point... any algorithm might be able to show you what you want/need *sometimes* but they can never know what you want/need *now*. Just because you search for, lets say, "batman vs superman" several times does not mean that in this very moment when you search "super bats attack man" that you're looking for batman vs superman.

Just like a personal assistant won't know that today you're feeling like a spiced latte instead of your usual black coffee, neither will any predictive algorithm. If you keep asking your personal assistant for a spiced latte and they keep bringing you black coffee they'd be fired pretty quickly. On the other hand if you ask your assistant for the TSP report and they bring it to you when you really wanted the TPS report you can't really fault them for that now can you?

Better protection against SEO.

[SeaFox]

What made Google so great when it was still relatively new was the results were more relevant, i.e. they weren't just a bunch of advertisements. With the rise SEO that is less the case now, and looking for something on Google for me now means adding "-buy -purchase -price -shop" automatically.

Re:Better protection against SEO.

[JMJimmy]

What made Google so great when it was still relatively new was the results were more relevant, i.e. they weren't just a bunch of advertisements. With the rise SEO that is less the case now, and looking for something on Google for me now means adding "-buy -purchase -price -shop" automatically.

Google seems to ignore '-term' when it comes to certain sites/terms.

Re:Better protection against SEO.

[rudy_wayne]

What made Google so great when it was still relatively new was the results were more relevant, i.e. they weren't just a bunch of advertisements. With the rise SEO that is less the case now, and looking for something on Google for me now means adding "-buy -purchase -price -shop" automatically.

Unfortunately, that's what happens when a search engine is run by a company that depends on advertising for 96% of its revenue. Google is not in the search business. They are not a technology company, no matter how many data centers and driverless cars they have.

They are an ADVERTISING company and their business model depends on getting you to click on as many ads and "sponsored links" as possible, using whatever deceptive methods necessary.

Re:Better protection against SEO.

[SuricouRaven]

You seem to forget just how useless search engines were before Google. I remember using some, and they were utter rubbish. Google shot to success in their early days because they had the better algorithms: If you searched with them you'd probably find what you wanted on page one, not page twenty-five.

Re:Better protection against SEO.

[JMJimmy]

They were only bad if you didn't know how to use them. I found them to be infinitely better than what we have now - even Google of 3-4 years ago was better than now.

Re:Better protection against SEO.

[afgam28]

If anything, being an advertising company provides an incentive to downrank ad-like results. Why would anyone buy an ad if their shopping site already appears at the top of the organic results?

Re:Better protection against SEO.

[Ksevio]

But they are also in the search business since they can't receive money from their advertising customers if they don't receive screenspace from their searching customers. Just because their money comes mostly from advertising, does not make them an advertising company.

Re:Better protection against SEO.

[Anonymous Coward]

They are an ADVERTISING company and their business model depends on getting you to click on as many ads and "sponsored links" as possible, using whatever deceptive methods necessary.

No, they are a SURVEILLANCE company that just happens to engage in a bit of advertising to support their primary business of stalking people.

Re:Better protection against SEO.

[Jack Griffin]

They are an ADVERTISING company

Mod Up!
I'm sick of Google being referred to as a technology company. They are an advertising company that uses technology. But at heart, advertising is their core business.

Re:Better protection against SEO.

[jones_supa]

I am not seeing any advertisements on Google search as well.

Pressing Enter

[snowgirl]

First of all, I would make it so you can press the Enter key and it conducts your search. Forcing people to either tab or navigate their mouse to the button makes it a little annoying.

Re:Pressing Enter

[hercludes]

First of all, I would make it so you can press the Enter key and it conducts your search. Forcing people to either tab or navigate their mouse to the button makes it a little annoying.

Isn't this standard for anything though?

Completely Open Source

[ModernGeek]

I'd like to see a completely open search engine that allowed people to download the search indexes freely so that they may create their own in-house appliances for search without the need for going through some proprietary site that may or may not be available in the next ten years or even months.

A site that promises to deliver you your privacy is not enough, because they could really be doing anything. Google promised us our privacy, and changed and deleted their old privacy policies even though they said that they'd always keep all copies of a privacy policy on archive. They went back on the word "never" and have continued to discontinue online services that people have become accustomed to with little to no notice.

A reasonably sized search index that is extensible based on what one is searching for would be great. Localizing URL suggestions, wikipedia caches, and other toolbar-suggestion searches in a networked work environment would all have benefits; the applications are almost endless. Freeing the shackles of search from a few could do so much for innovation, privacy, and security.

Re:Completely Open Source

[wvmarle]

There are two issues with this.

Search indexes are very expensive to make - lots of data to download and analyse to come anywhere near reasonable coverage of what's out there. Someone has to pay for it.

The amounts of data involved are huge. By the time you're done downloading such an index, even assuming you've got sufficient storage at hand, it's horribly outdated.

There is a reason there are no recent small search start-ups: you have to be pretty big to even consider this. When Google started, the Internet was a fraction of the size it's now, and even then Google's founders could use the massive computing resources of their university. Google's index nowadays is so big that they can not search it entirely themselves: different geographical locations tend to give different search results for the exact same query, as you're searching different subsets of the database.

Exact searches

[Anonymous Coward]

Try searching for someone named "Beiber". Google might find him, but he'll drown in a million entries for some singer named "Bieber". But I did not search for Bieber.

There are many cases like this, where something rare has a name similiar to something more popular. Don't assume I mistyped! I rarely do. But if I mistype, I can search again. But I can't deal with a search engine that blatantly assumes I'm dyslectic.

And finally, let me search for source code snippets without turning up tons of irrelevant stuff. Spaces in an exact search is not separators - if there is no match, just say so. Don't assume I might want something completely different.

Re:Exact searches

[Jack Griffin]

This pattern was one of my main frustration with MS. They increasingly adjusted the interface for stupider and stupider end users, which is ok for that type of user, but FFS at least have the "I've used a computer before and am comfortable fiddling with settings" option for the rest of us. There's a lot of people that know what a computer is and don't need to be treated like imbeciles.

Just search..

[Anonymous Coward]

..and don't mess with my query. If I search for "the best saerch engine", give me the pages that have that string, no questions asked and nothing else. Oh, and while at it, make one that does regex.

Be able to PERMANENTLY disable instant search

[schwit1]

And stop providing results that fail to have ALL of the search terms.

Re:Be able to PERMANENTLY disable instant search

[thegarbz]

I don't agree with this. Certainly results with all search terms should be promoted at the top, but once those run out the rest should be displayed. That allows you to easily find information when you don't know exactly what you're looking for. The term may be misspelt (on the website, not necessarily by user input), or there may be a different term for the same thing and the resulting website gets picked up on the rest of your search omitting the one word that is "wrong".

Re: Be able to PERMANENTLY disable instant search

[BlueTrin]

I disagree with you, also a simple Google search would allow you to learn that you can obtain this behaviour by prefixing the obligatory words with '+'.

Re: Be able to PERMANENTLY disable instant search

[RedWizzard]

I disagree with you, also a simple Google search would allow you to learn that you can obtain this behaviour by prefixing the obligatory words with '+'.

Not anymore. Google now returns results which don't always have all the obligatory words in the page, though it does tell you if some are missing.

Re:Be able to PERMANENTLY disable instant search

[__aabppq7737]

this AND/+ thatfor those wanting to find exact matches.

Re: Be able to PERMANENTLY disable instant search

[shione]

I think the new way around this is to enclose them with ".

Searching

[Livius]

and nothing else.

Stop adding 'features' to things that don't need them!

Re:Searching

[thegarbz]

Why? As long as it doesn't affect searching why not add features?

Why is it such a problem that if I google "Where is my phone" that it brings up the option to call it along with search results for that search? Why is it a problem that when I type EUR-AUD I get a currency conversion screen along with results to articles and currency converters about the Euro?

Re:Searching

[Kjella]

and nothing else.

Stop adding 'features' to things that don't need them!

YMMV, but that's one of the reasons I really like google. For example converting units, what's 53F in C again? I could get a thousand hits that could give me the formula or a conversion table or whatnot but just "searching" for it saves me a step or two. I often use it instead of the built-in calculator just because it's already up. I suppose it could go overboard with Clippy-isms but I haven't felt that has been the case.

Re:Searching

[Altrag]

The trouble is, it does affect searching. When it works its helpful (but not required.) When it doesn't work its horribly frustrating.

To take your example, what if a popular singer releases a song called "What's your phone #?" Suddenly Google is "helpfully" readjusting your "Where is my phone" query and you end up the top 14 pages of results being about a song you're not interested in. Similarly, if a movie is released with a main character named "Euraud."

Now, if you're lucky, the search engine will provide useful ways to tell it to stop trying to outsmart you, and the onus is now on you to figure out all of the necessary hoops and jump through them in order to get a simple search to just bloody work.

And if they don't even provide hoops (or the hoops get too overridden by advertisers and SEO?) Now you're stuck trying to manually search through dozens or hundreds of irrelevant results and just hoping that the information you're actually looking for is obvious from the first 3 lines of blurb.

Re:Searching

[thegarbz]

Your scenario only applies if Google provides unlimited additional information.

In the case of "Where's my phone" If I'm searching for my phone it will be the top query. If I'm not searching for my phone the singer's song will be the next query down. Additional features take up little room in the search results and don't affect the normal results much at all which is to say regardless if one of Google's additional features is displayed or not the number of result from the actual search shown on the screen is the same. There is no need to outsmart, just to scroll one click further down.

But really I'm not sure I follow the logic of your argument. If a singer releases a song called "Where's my phone?" are you suggesting it shouldn't be the top 14 results if it's popular? Regardless of popularity search results do not take precedence over added features, so the end result of seaching for "Where's my phone" would be the additional feature on the top, followed by 14 results from the singer, followed by the iTunes link which would be almost identical if the feature wasn't present.

It looks to me as you're not complaining about "features" but rather about the algorithm deciding the results which is something else entirely.

True boolean search, ability to vote on results

[cowtamer]

In addition to Google-like relevance (which is a must if you are going to survive in this field), it would be nice to have:

1) Boolean search (cat or feline) and not (catwoman or cartoon or dog))
2) Date range which works (e.g., I want to search for websites talking about Enron BEFORE the scandal).
3) If I see a result that's obviously relevant, I'd like to be able to down vote it..

Re:True boolean search, ability to vote on results

[ADRA]

#3 Who has more time to manipulate an open source web engine index? A do gooder looking to relfect bad SEO in a search result, or an SEO not looking to pump their own numbers floating their own crap to the top (through whatever carrot/stick-like measures implemented).

Google had some extremely bad queries some years back due to every SEO on earth trying to game the system. The only reasonable solutions on the top of my head are:
1. some sort of real-id-type verification system that requires actual investment in an acocunt being considered having weight (even then, it makes compromized systems a lot more valuable to hackers)
2. Devise a systemic pattern manipulation which is known to specifically target and down-rate results (like results that only get linked through blog comments or reuse heavily from their referenced pages for instance) -- Google seems to do some form of this
3. Individualized / group based blackack-lists -- Pain in the butt to curate, causes false positives to be burried (forever?), relies on people with associations comparable to their own (eh, ban all GBLT/alternate religions / pre-xyz sites / etc..) and of course having individualized search curation is a butt ton of extra data that needs to be floating around on servers waiting for your specific user ID to hit said server. I think that is may have been one reason blocked sites died in Google. Its just a pain in the ass to distribute the user's search preferences to every possible hosting node (or having slower responses due to limited numbers of nodes being able to respond to them).

Re:True boolean search, ability to vote on results

[Jack Griffin]

#3 Who has more time to manipulate an open source web engine index? A do gooder looking to relfect bad SEO in a search result, or an SEO not looking to pump their own numbers floating their own crap to the top

So include a system to detect this. I'm sure it would be fairly trivial be pick out patterns and have some sort of human audit of major changes. Wikipedia already does something similar.

Re:True boolean search, ability to vote on results

[Altrag]

Wikipedia has a few million pages, all within their domain and (ultimately) under their control, the majority of which have minor or no controversy involved.

Google has many billions of pages that they have no control over at all, almost all of which are controversial (even if the controversy is "I want to be above my competitor in the results.) Human auditing would be immensely impractical, and crowd-sourced human auditing would be even worse due to the fact that the people most invested in breaking the system are generally the ones who can afford to put the most time/money towards doing so.

So they're relegated to using automated systems. And honestly, Google in particular has been really good at coming up with algorithms to beat overly aggressive SEO. Sure they sometimes fall behind a bit (its an ongoing battle after all,) but overall not bad.

Google's primary failures in the past few years have been their not-nearly-as-accurate attempts to outsmart their users' search terms. "Having" "to" "double" "quote" "every" "word" and still seeing your top result with a "Missing terms:" disclaimer is immensely frustrating.

Re:True boolean search, ability to vote on results

[monkeyzoo]

YESSS. Complex boolean searching!
This was the best thing about Alta Vista and why I still used it sometimes even after Google's relevancy came along and generally got your the best results intuitively.

Javascript

[Anonymous Coward]

I would like a feature that makes it possible to perform a search query without Javascript enabled, so I guess my needs are fulfilled by all other search engines than Trokam.

Egg Zact

[Anonymous Coward]

I want the old Google exact searches you used to be able to do +"exact keywords" so you can filter out all the sloppy, useless results Google has now. Since 2008, Google has been in a downward spiral. Any other search engine could have stepped up. Why didn't Bing or one of them fill the void that Google left when its search results got useless and sloppy?

Egg zact (hey, good name for a startup) searches are useful if you're looking up exact error codes. Sloppy searches don't work for that purpose.

Who needs links to websites anyway?

[sparkydevil]

I'm wondering what you mean by "Search Engine"? Do you mean a way to sort and rank websites? That's only part of what Google does. You may want to identify what is missing from Google before following the models of the past.

Re:Who needs links to websites anyway?

[Altrag]

Wow. That was a terrible article. Their example is searching for a person's name and then claiming the results didn't give them the information they wanted.

Except they don't tell you what they were wanting! They mention songs, stories, history.. well you could add those terms if that's what you're looking for and get results that are a lot more specific to what you want.

Bringing up the Wikipedia page is probably the _best_ possible option given how vague the query is. Sure the first 3 lines that get turned into the blurb may or may not say anything useful, but everyone knows what Wikipedia is and how it works. If you're just looking for random information about a topic, that's a great place to go as it will have at least a little bit of all of those categories mentioned above and you can go redo your search with more specific terms once you've decided what you're interested in.

And OK sure you can pull up countless examples of when somebody somewhere made a troll edit to the Wikipedia page.. but what on earth makes the author think that the (usually completely unreviewed and uncurated) articles from "way way down the list" will more accurate in any sense?

Same with the news stories. How is Google supposed to know if you're interested in the most recent news or the most controversial news over time?

If you don't even know what you're looking for, how the hell do you expect Google (or any other search engine) to know? A "story" is fine, but if all I'm looking for are concert dates, I just want a damned list of dates and cities I don't want a whole bloody "story," whatever that even means given the lack of context.

Deja vu all over again...

[kackle]

Amen, brother. Similarly, I switched from Lycos a decade+ ago because they dropped Boolean searching (some of us are power users!). I used Yahoo! next, but it was painful on dial-up with all the extra junk on their home page. Then I came across this new, misspelled site called "Google". I loved it; but lately it has been wearing on me as it panders more and more to the masses.

Note to Google: We nerds might be in the minority, but it is WE who direct the non-nerds as to how to set up their digital devices, avoid online trouble, choose their search engines, etc. Don't ruin it for us. I already started to keep one eye open for another search place, because I fear it'll only get worse.

Re:Deja vu all over again...

[JMJimmy]

If you find a good one let me know. I've tried Bing, DuckDuckGo and a smattering of others - they all seem to have the same results or lack any sort of depth which is highly disappointing. My next to try is Blekko - I like the concept of /topic and more of a command line concept.

Re:Deja vu all over again...

[JMJimmy]

oops.. never mind. Blekko is already gone... hrm... maybe Ark? Seems more social than geek though.

Re:Deja vu all over again...

[Mantrid42]

Perhaps it's a ploy to keep us in power, as only we can wield the strongest google-fu.

Accuately read my mind

[PineHall]

To give me search results that accurately are what I meant to type and not what I did type. And a bonus would be if it would accurately know better than me what I am looking for an give me those results. In other words read my mind and correctly anticipate when I am wrong and still give me the correct answer.

best wishes !

[swell]

It wouldn't respond to my request. I had to allow a jquery script. Then it searched but couldn't find 'Benghazi'.

Things have been lost from search. Alta Vista allowed search for 'word1' NEAR 'word2', which proved very useful. Google used to give information about its finds such as date, size, ('cached' is still there, but hidden) and some things so long abandoned that I can't remember them. You know why date is important; size is also important because a very large page containing your terms is probably clickbait. A great sadness for me is that Wolfram Alpha is so wrapped up in fancy scripts that I've never been able to use it with my fairly secure Firefox (oh, it's better today).

Accurate reporting would be nice. I'm looking at a Google result that claims it found "About 54,100 results (0.46 seconds)" when actually there were only 245 unique results.

Location would be nice (maybe a flag icon from that country). An opportunity to vote the relevance of a result up or down and maybe indicate something inappropriate. Wildcards would be incredible. Apple's Spotlight search engine can now search the internet as well as local files- maybe your engine could take advantage of some sinister simpatico surreal symbiosis.

We need a fresh approach after a long period of stagnation. Who knows what clever innovation has been missed?

Re:best wishes !

[Bill Dog]

Google: The IE 6 of Internet search.

Maybe Mozilla could make a search engine and re-light the competition fire and user focus in this space. (Bing just seems to be a clone of the stagnating leader.)

Regular expressions

[John.Banister]

Also, filter out all the pages-of-links search spam
If you want to charge a fee, you could include a link to someone who is better at searching for stuff than I am, or maybe Watson.
Finally, include all the internet that Google hasn't indexed.

Have you seen this list of some of the few alternatives?

Re:Regular expressions

[flopsquad]

This!! For various technical reasons, I don't think a regex-capable, public-facing search engine is feasible right now :(

But I'd beg, borrow, and bite to get some sort of "regex lite" capabilities (we could start by excluding the stuff that's np-hard, like lookbacks).

Re:Regular expressions

[Altrag]

Ugh. Yes this so much!

Also those "answer" sites that are all just word-for-word carbon copies of each other. Even if there's useful answers on the page, we only need one copy of each answer set.

I have no idea how Google could manage that (more from a legal perspective than a technical one) but its certainly freaking annoying from a user's perspective.

Word proximity

[GerryGilmore]

Being able to say "find 'blah' when it is within X words either side of 'bleh'......"

Actually visited your search engine

[stephanruby]

When I open your search engine, I want the focus of my cursor to default your search form.

After I found out that you didn't even have this, which requires no more than one single attribute in html, I didn't have the confidence to go to any further. Usability testing is cheap. The idea that you would forgo any kind of basic usability testing, before asking for feedback from Slashdot users, tells me you don't have the experience, nor the real desire, to make a decent halfway usable search engine.

Culling

[Moof123]

Make it easy for me to specify I am looking for technical information, or looking to buy something, or what have you. All too often I am trying to do a search for technical information, but if that acronym has also been used by Beiber lately I am SOL. I would love it is I could weed out the pop culture hits when I wanted to omit them.

Similarly I would like a search engine that I could easily specify if I also want hits for related words, or just EXACT match, and whether to ignore capitalization or not. It is maddening when an acronym also happens to be a common word and I get flooded with useless crap.

Re:Culling

[snowgirl]

...but if that acronym has also been used by Beiber lately I am SOL. ...

What does the Skilled Occupation List have to do with this?

Some ideas

[grahammm]

1. Return first the results which exactly match the search terms.

2. Do not include results where one or more of the search terms only exists in an advert on the page.

3. Re-introduce a feature which an early search engine (I think it was AltaVista) where you could specify a search term to be 'near' another.

4. (more important in languages other than English) Allow you to specify that any tense, person or case of a search term be matched (eg if searching in French, *aller would match any of 'vais', 'vas', 'va', 'allons', 'allez', 'irai', 'allâmes' etc)

5. Allow you to restrict the results to those where he search terms are actually rendered on the page when you follow the link.

Re:Some ideas

[flopsquad]

I believe google may have (had?) some undocumented "near"-type feature.

Yellow pages, huffington post... both gone!

[EmperorOfCanada]

First I want a toggle that I can never ever ever ever ever see that domain in my search results. So ask.com answers.com experts-exchange.com huffpo and especially Quora you fucking turd pile of shot Quora; I never want to see Quora again in my life.

When I want a little pizza joint or some place that hasn't hired an "SEO" guy all I get are page after page of directories derived from some government database or some crap like that. Their actual page is bottom ranked. I don't want review sites. I don't want anything that was assembled by a machine.

So a simple rule of thumb is de-list any page that offers to "upgrade" someone's listing. Full stop. Also I want a toggle that will remove listings that have any version of "upgrade to our pro service" Literally they could cure cancer but offer to cure it 1 minute faster for 99 cents and I don't want to see that page.

To me right now nearly the entire search results are like going to a dating site and only finding hookers. Some people would argue that they "need" to make money but they don't. There are lots of pages that exist for a specific reason and many of those pages are commercial, as in they offer a specific service such as a pizza places where the page is about their pizza place. Short of the recipes the page is 100% free. But I don't want some shit "Just Eat" website. Maybe they can link to the other page but I really don't want to see it ever again. For instance I loved allrecipes.com. But now it is just upgrade upgrade upgrade upgrade. Some will argue that they should be allowed to make money but quite simply the site existed before some MBA took over and "monetized" the site that's fine, I no longer want it to turn up in my search results. Don't ban it from the entire search engine, just ban it from the search engine when you tick the "No upgrade sites" option.

The other thing that I would kill for is a negative feature option. So any site that uses discus would vanish from my search results. Those scumbags need to burn in hell and I would love any search engine that sent them there.

To me there is a huge opportunity for some new search engine to do to Google what they did to all the others 17 years ago; completely make them irrelevant by brutally ignoring the wishes of the larger websites and completely focusing on the needs of the average user.

No "preferential" treatment of results.

[Jane Q. Public]

Google has been ever-more pissing me off with its "sponsored" results in which I am almost never interested... I have to go further down the page to get the things I want.

Related to this: Google's recent proposal to post "truthy" results before others. Just no. I don't want or need a nanny-search. I'll judge the results for myself.

As far as I am concerned, results "filtered" or sorted according to Google's idea of "truth" is little more than a rather transparent effort toward censorship.

The Best Feature

[JimSadler]

Google!!!!! Do not reinvent the wheel. Google can be a wonderful search engine. How in the heck can a new product provide the variety and depth of search that Google can with their enormous data base and ample hardware?

Give me the option to disregard...

[Godwin O'Hitler]

...all meta shopping sites.
I'm at my wits end when alibaba or ali express or kelkoo or tengo or whatever is in the top five of EVERYTHING I search for. I don't ever want them to be even in the top 1000 unless I explicitly type "Meta shop" or whatever.

Apart from that one filter, just search for what I fucking asked for, not what you think I might have meant.

Clean URL

[Nightlight3]

I find it most annoying with Google last several years that they mangle the URL they send me to so I can't easily change it to the parent or higher level URL.

Ads are not the problem, SEO is.

[barfy]

When I search, I don't mind ads. I mind malevolant ads. But I mind malevolant sites more. One is solveable. Generically the site that has the most to gain through advertising pays the most. And I generally benefit from those. But SEO is more insidious. They are trying to gain my attention from generally more appropriate sites by gaming the system. If the system is evolving correctly the best most appropriate sites get selected automatically. The major search sites recognize these things. Because if they stop providing relevant sites other sites will come along and become more relevant. I am not going to use your site if there are no ads. No ads means less relevant searches for me. The site with the most relevant search wins. Hands down. Everything else is meaningless.

Being able to filter results

[Dixie_Flatline]

So say I'm searching for something with really common words in it. I can't think of anything specifically right now, but this is my most common search failure.

I get back a bunch of results. They have all the words I'm looking for, but they're all about two or three more popular topics. I'd like to be able to select a search result and tell the engine that results like this are incorrect for some semantic reason. Maybe it's a band name and I'm looking for a book titleâI should be able to say I don't want anything related to music. No bands, albums or songs. I'd be willing to tag results with some context to provide hints to the algorithm.

Things like 'windows' tend to mess up results; Google assumes that I either mean the Microsoft kind or the house kind, but sometimes I'm trying to figure out what's wrong with a particular application window. I run into this sort of thing surprisingly often.

Selectable search.

[SuricouRaven]

Specify search areas:
[X] Public internet index.
[ ] Torrent sites only.
[ ] File database of dubious legality.
[ ] Archive of device drivers that actually work.
[ ] The Dark Web, whatever that is.
[ ] Data sheets and manuals.

Category trees and branches

[sleepypsycho]

I would like a search engine that could show me the top categories for my search for results. This might be beyond your goals but it is what I would like

For example I was search for program that would deal card
If the I had the word "deal" in the search I got 1000 shopping sites in a row. But I was not looking for shopping. All my search terms were common and could be used in conjunction. So card and deal together would get business cards or greeting cards. If an engine could provide some sort of organization. I could then drill towards what I am interested.

How the engine would be able to do this I am not sure. Obviously semantic web is one avenue, but the search might recognize lots of links in common or it might look at supplied key word or common terms on the pages

Google had something a bit like it (in labs only I think) but it was buried and not a simple choice about how to present results.

Good results

[Lord Crc]

Without good results, it doesn't really matter about the bells and whistles. I use a search engine to find information, so it better do that extremely well. For example, I just couldn't stand using DuckDuckGo (aka Bing) because of this, and went back to Google. Bing consistently failed to find information the information I wanted, while Google had it on the first page.

So, after your engine returns as good or, ideally, better results than Google, you can start thinking about other features.

One feature I'd really like is to be able to tweak my result set. Something like if I search for "AC DC", I get a bunch of results about the band "AC/DC". That's not really a bad result given the input, but in this case I was after an explanation for the electrical terms.

So I'm thinking some ability to mark one or more of the results I don't want and say "not pages like this", and it would cull those talking about the band, in a weighted manner. Or some other way to help me find the information I want when I search for some ambiguous terms.

GEOIP transparency; User hosted crawlers;

[bug1]

Search engines should shine a light on sites that show different results to different users, maybe its for commercial exploitation (GEOIP blocking), or political propaganda or whatever.

Search engine show allow users to run crawlers in coordinated distributed manner, this helps users have privacy, it adds extra noise to surveillance systems, it might give users deniability as to their intent to access subversive material. It it should help with the first problem.

minus

[pen-helm]

I would like to be able to indicate that a search term should not produce a hit, but should not disqualify the page either.

This is point 1

[Anonymous Coward]

A search engine must not rely on javascript for simple things like submitting a form.

Focus on expert sites

[gman003]

I've started drifting away from using Google/Bing/whatever, in favor of loading a bunch of site-specific search engines into my search bar.

So if I'm looking for, say, a specific Magic card, I don't let Google search the entire net, and find everything that happens to say "elvish mystic", giving me a ton of irrelevant stuff (even searches like "mtg elvish mystic" bring up pages to buy one instead, which usually don't have the info I'm really looking for). Instead I click an extra button to go straight to Wizard's own database page for it.

Repeat that idea for about twelve different gaming wikis, plus Wikipedia for general knowledge, and you'll have the contents of my search bar. If I were into different hobbies, I might have similar search engines for those.

A single search engine that can figure out the context of the search, then go straight to the experts for that context, would be one way to do better than Google.

Re:Focus on expert sites

[Anonymous Coward]

site:www.wizards.com "elvish mystic"

I would like for it to...

[adndgamer]

...read my mind.

Seriously: words suck at describing or conveying a lot of things. And half of these comments are just people who don't know how to use a search engine properly. They can't read your minds [yet] and we are using the equivalent of charades to try and explain to a computer what we're looking for.

Until we can directly share a mental image with a computer, we're going to have to deal with crafting search queries.

Get rid of the crap sites

[E-Rock]

Get rid of the sites that have a single paragraph and then a registration or paywall blocking the rest of the content.

Get rid of the sites that are just copies of other pages with ads.

Or let us easily block a site from appearing in results in the future. Enough users vote a site off, have a human take a look to see if they should remove it for everyone.

Privacy, and symbols

[X0563511]

Privacy goes without saying, of course... but if I put a period or a comma in my search, I damn well meant it to be there. Pay attention to it.

Re:Simple regular expressions.

[X0563511]

It would be interesting to filter the results of a search with an expression - but as you say I think using the expression itself to drive the search is probably not something you'll see happen.

Less

[Snufu]

evil.

Source details

[mango9]

Want to be able to specific more about the source e.g. newspapers only, blogs only, academic only, published in 2007 and so on. Ads I can ignore.

ungoogle

[nospam007]

I'd like a search that actually searches exactly what I type and not what it thinks I might mean.

Including punctuation exactly as I type it instead of ignoring it.

Being able to be more precise

[Misagon]

Search features I depend on:
* Non-English characters. Handle multiple encodings of web pages and URL-encoded characters in search queries.
* site: to search only within a domain. This is often a national domain, such as "site:co.uk" to search only British sites.
* Minus: Begin able to block certain words, or sites.
* Plus: A word prefixed with a plus is required.
* Quotes/hyphen: Searching for exact phrases. "Java class file" is different from "Java File class".

Where current search engines are lacking:
* If there is a period between the words then they do not belong to the same phrase. (A search for "Hello Google" should not return "Say Hello. Google for it." as its top result)
* Use word order in search query to weigh how important a search term is. Rank pages higher wihen those words are closer together.
* Don't correct my spelling by default, assuming that my search query is in US-English. (I am speaking to you Duck-Duck Goo!). I can spell, and I do not always write English. If I misspell then that is my mistake, and sometimes I search for a brand name that was misspelled intentionally.
* When indexing a web page, identify what is the important text on the page and ignore the rest. For instance, on an internet news site, the text in the articles is most important. On a forum text inside the comments. On this forum, articles followed by comments. What people have written in their signatures is not important. Slashboxes are not and ads are definitely not.
It is aggravating when you use Google on a collecting site and you get every other page on that site in every search result because members have listed their collections in their signatures.
If I search for the word "review", I don't want every page on every web store that has a Reviews tab.
Pages on a site often follow a certain pattern - find that pattern to find which text on each page that is the most unique.

Re:Being able to be more precise

[__rze__]

Totally agree with this. :)

I've been using DuckDuckGo for 2y now and can see a lot of improvements, especially in QoR. Some things I don't like, but it covers my needs mostly. Had some problems with things like mentioned in the parent, but worst case I pulled a !g and still got the info I needed without tracking.

There is also another one called www.qwant.com which promises the things that everyone wants, will be trying it out. It is based in the EU, so other laws apply ;-)

Re:Simple (my wish list)

[Kohath]

Why won't this useless search engine tell me the best place to ford the thunderbird river?

2D navigation for search refinement

[iMadeGhostzilla]

Think of Google's search as your type as 1-dimensional suggestion list. I'd like as I type to see around the search bar a matrix of categories: news, videos, documentation, blogs etc. Then as I hover over a category with a mouse I zoom into a matrix of subcategories for that category using the mouse wheel. I zoom out back one level if that's not the branch I'm thinking of.

In addition, I don't want to click until the very end, and maybe not even then. Hovering over a set of results shows me what's at the deeper level, and when I'm looking at a one or a handful of pages that match the criteria as I refine further, it is also shown as a cell. Hovering over it will give me a preview -- from the search engine, not my browser fetching an actual page. Only when I'm certain I want to go there, I'll click.

That would be a search engine of the future. Or, idea #2: make it like google, but when I control-clik on the link for the page it opens a sanitized copy of the page, provided by your server, so I know there are no scripts or malware and crap. And if possible give me that sanitized preview when I hover over the page so if I'm lucky I don't have to click on anything at all.

I know sites wouldn't like it but just saying what I'd like to see that I think is technically possible. Thanks for listening!

Search 'intelligence' is great

[Peter (Professor) Fo]

Suppose I type in a phone number Google understands I live in the UK. The same goes for wanting to buy a mattress. So far so good. That's helpful intelligence.

But if I want to view the poems of Emily Bronte I don't want 100 gazillion results from Amazon.

Just like I use NoScript and AdBlock+ so I want to cut out the shop windows. If I want info from the web then I don't want canned waffle.

Regular expressions

[jlockard]

I'd love to search using regular expressions, failing that, at least a much more precise way of indicating what must and must not be in the returned results.

Stop trying to guess what I'm trying to do

[dackroyd]

It would be very useful to be able to control what the search engine thinks I'm actually searching for. Taken from: http://unqualified-reservation...

A more intriguing question is whether the Graffiti approach can be applied to full-text search. Many modern search engines, notably the hideous, awfully-named Bing, are actually multiple applications under the hood - just like WA. If Bing figures out that you are searching for a product, it will show you one UI. If it figures out that you are searching for a celebrity, it will show you another UI. It may also switch algorithms, data sets, etc, etc. I'm sure Google has all kinds of analogous, if more subtle, meta-algorithms.

While generic full-text search, unlike generic data visualization, remains a viable application and a very useful one, specialized search might (or might not - this is not my area of expertise) be an even more useful one. If the user has an affordance by which to tell the algorithm the purpose or category of her search, the whole problem of guessing which application to direct the query to disappears and is solved perfectly. A whole class of category errors ceases to exist.

My guess is that if there is any "next thing" in search interfaces, it will come not from smarter UIs, but from dumber ones in which the user does more work - the Graffiti effect. If a small quantity of user effort can produce a substantial improvement in user experience (which is a big if), the user will accept the bargain. Hey, it made Jeff Hawkins rich.

Original AltaVista....

[gweihir]

"Near" keyword, logic constructs, all those nice features AltaVista (which was just a hardware-demo) had, and Google never managed. Google is borderline unusable these days and you strongly notice they do not care about good search but only about placing their adds and profiling you.

Regular expressions

[penguinoid]

I'd like to be able to search for regular expressions, or at least boolean. I'd like to be able to specify the context of a word, eg Java (geography), Java (coffee), Java (programming), or similarly the context for the search.

More Robust Search Criteria Formatting

[JimMcc]

1) Exact string matching. As an example, if I search for " 'x.25' " don't give me hits for something with dimensions of 45mm x 25mm.
2) Allow more complex search constructs . For example I'd like to be able to specify the search term " 'x.25' near protocol -handbook ". You can sort of do that with Google's Advanced Search, but it's extra steps and you still don't get terms like 'near' or exact match.
3) Bonus points for boolean constructs such as " (lions or tigers or bears) near woods ".

In short, provide a robust search engine that will support meaningful search terms that can be used for more than shopping for a new TV or figuring out who stars in your favorite reality tv show.

Dear Santa, all I want for Search is...

[Demonoid-Penguin]

a search engine that searches the internet. Not parts of the internet, all of the internet.

I'd also like the search engine to do Boolean and regex.

P.S. I couldn't give a flying fuck if it:- has ads; tries to profile my search queries. I can at least attempt to get avoid those things. But if it does not index the entire internet it's as useful as a range of shoes that consist of one size and one style only.. And no, I don't care if it doesn't come with a free set of steak knives and is 100% dolphin free and kind to puppies, as long as it indexes everything

I don't care if Bill Gates wrote the back end, hell, I'd use it even if it was run by the scumbag behind DuckDuckFuckADuck.

Thanks

Re:Dear Santa, all I want for Search is...

[Z00L00K]

That search engine will be hunted down and convicted of breaking the DMCA.

Re:Dear Santa, all I want for Search is...

[Demonoid-Penguin]

That search engine will be hunted down and convicted of breaking the DMCA.

1. and?
2. so?
3. what?

add results to a blacklist

[gront]

Ability to right-click (or whatever) on a listed result and mark it as "I never want to see this site in any search I run on any topic ever" (useless result) or right-click on a listed result and mark it as "The content of this result isn't relevant to my search, block this page and all others like it from this search so I can find what I'm looking for" (irrelevant result/bad context) and re-run the search.

Re:add results to a blacklist

[Z00L00K]

That's an interesting idea, and then we can get rid of all those useless "experts-exchange" links and similar.

What I also would like to have as a feature is the ability to search also for "special" characters. Google today seems to see "+" and a lot of other characters as spacing characters, but sometimes I really want to search for "a+b" as a term.

Decentralized like Bitcoin or at least Bittorrent

[aNonnyMouseCowered]

An open DECENTRALIZED search system would be way better than a system that's simply built on open source. Hell, for all we know maybe even Google is completely on open source. But the data sets that seed the search engine, without which the algorithms are simply crunching meaningless strings of letters, are kept close to Google's corporate bosom.

What we need is a search engine where everyone that searches can have access to the entire data set if she or he chooses to do so. This is similar to the way the Bitcoin blockchain works. Everybody can choose to have a copy of every Bitcoin transaction ever made, or if they're lazy or don't have the computing resources connect to a full Bitcoin node using something called a light wallet (which downloads only the relevant parts of the Blockchain related to the transactions made using a certain Bitcoin address).

So let there be a basic light version of your engine and a full version. If that's not feasible, maybe you can make an advanced client that processes only parts of the complete data set, but is distributed in such a way that the parts can easily be combined into a complete data set.

Different context switches

[See Attached]

How about a context switches? One for purchasing/acquiring, one for history, one for technical details, One for and people related to an object? When I search now, I have to search between buying opportunities and buying opportunities and more.

Re:Different context switches

[See Attached]

Hey, not for nothing, but does anyone remember Usenet and Dejanews, where you could search up on any technical string and you would find -help- with things, but now, its all about -buying- things.... It was a big part of the value of a search engine to me anyway. Search is all about money which is fine when I want o buy something. i just want a switch to select Buy/Learn/History/people etc...We need a way to filter the search results.

What Google doesn't

[jtgd]

Just look at all the features that Google keeps removing and add those in.

How about a token at the beginning (perhaps "++") that declares that every word must appear? (implied +word +word...)

How about a token at the beginning that declares that every word must appear in that order? (but not adjacent)

One thing I know I'd like to see, regular expressions, is probably prohibitively expensive, but it sure would be nice.

add business listing to addressbook

[davemchine]

I frequently search for a business and google displays the name, address and phone number in a very nice box in my browser. There is no way to move that information to my address book though. It seems like google could have a button to push a cvs file to me.

I know the feature you all REALLY want...

[Yosho]

The "anti-safe search" image search, which filters out all of the things that aren't porn.

Come on, be honest.

Only what is visible on a page.

[magic_user]

I want boolean searching on -only- what is visible on a page. None of that metadata stuff. That alone should bypass all those search conglomerators. I don't mind advertisements on the side, but not mixed with the results. I also want the results based only on what I searched for, no paying for higher rankings. That sounds simple enough...
 

features

[l3v1]

"features that I would like to find in a service: respectful of user rights, ad-free, built upon open source software, and with auditable results"

Well, well, well. For me there's only one single feature of a search engine that makes it a go or a no-go: have a damn good indexing engine that can provide relevant results in a timely manner. Everything else is just a load of crap that I will never care about. If I can't find what I'm looking for, that I couldn't care less how it protects your rights or whether it is open source or not. Oh yeah, about that auditability... forget that. I don't want to find what other people think I should find, I want to find the best match for my queries. That said, good look, develop away, maybe you'll indeed make a better indexer and ranker than Google's and we'll be all better off.

Rank original content above harvested content

[bsdpanix]

I posted a request on Google Groups, years ago that Google implement a blacklist of sites, sites that I don't want to see in search results.
I despise results from other search engines, such as ask.com, scribd.com, and a dozen others that harvest other's answers and barrage me with a second layer of ads.
Google did implement "blocked sites" but eventually it removed it. Personally I would think the blocked sites list would be invaluable on a server side "page rank" algorithm. The more users blocking a given site, the lower the sites rank.

Other than the obvious features you've mentioned in your article, simply stop returning other search engines and rank original content higher than harvested content.

Time search

[jbssm]

Seriosuly, I keep trying to use DuckDuckGo as my main browser but I constantly have to switch to Google because of this.

For most subjects, the time of the creation/indexing of the page is not that important, but if you are programing or if you are doing scientific research, you really need to be able to filter the most recent results.

text vs meaning

[ancient_nerd]

There are two different ways I would like to search:

• Text
  • keywords, boolean combinations, verbatim, special characters, etc.
• Meaning
  • stemming of words (e.g. ignore singular v plural)
  • synonyms (e.g. car = automobile)
  • types (e.g. i want an address, an algorithm, a definition, an event)
  • constraints (e.g. i DON'T want to buy one - that should be a tick box in Google anyway!)
  • time (e.g. this year, before the industrial revolution)
  • attributes
  • relationships

This is what the semantic web is supposed to bring us, but we can do much more with what already exists.
A great deal has been achieved in AI and natural language understanding over several decades.

No doubt Google are working on it all right now...

classified by type of search

[jbolden]

1) Exclusions = in a search on X don't give me anything about Y
2) More control on match criteria (i.e. words must appear in sequence vs. can appear anywhere, frequency of word matters, title or keyword bonus up or down), data ranges or data range weighting...
3) I'd like to be able to indicate type of search: news, shopping, academic (i.e. give me papers), physical location...
4) Better handling of non-English words (give me English articles with this Italian phrase vs. give me Italian articles with this Italian phrase)

Oblig

[Hognoxious]

I'd like it to have an integrated init system. Or the other way round. Whichever you prefer, Herr Poettering - you're the boss.

Usability

[dackdel]

Allow me to hit enter to search instead of clicking on the search button. And dont sell my search data. And make it work well

Recognize regions, languages

[MoarSauce123]

One of the biggest annoyances of searching in English is that the results are coming from gazillion places. Sometimes that great, but when I look for lumber yards I really don't care about the ones in England, Australia, or New Zealand. They all might be fine businesses and a pleasure to deal with, but I doubt they deliver to the northeastern US. What also would help bilinguals as myself is to set two (or more) preferred languages. I speak and read two languages fluently and I don't mind and often want search results in both languages. Lastly, making it easier to search within results will be great. I know it is already possible with some search engines, but it is not easily achieved.

maximize control by user

[arctother]

How about: search for exact word or phrase (eg. "within quotes") it is so frustrating that duckduckgo etc don't allow this. No padding results with false positives! Toggle as many options as possible, eg searching for alternate forms or spellings; numbered results; let the user have maximum control over the results.

Feature Request

[Dragon Bait]

I numbered my "features", but they really are in random order.

1. Don't assume that I entered a partial word or that you "know" better than I do what I want to search for. Specifically, if I use the search term "ord" I do not mean "order".

2. Give sufficient context into the results that I know how the page uses the terms. Having the context be part of the links going off the page is of very little value. Specifically, back to the "ord" search above, returning "http://ad.doubleclick.net/jump/%sitename%/blog;pos=%pos%;ord=123456789?" is useless.

3. Only index relevant stuff. See above ad.doubleclick.net example that should never be counted as a hit when searching for "ord".

4. Use https

5. We're addicted to speed. Results need to be returned in a reasonable time frame.

6. If I type in my search results and hit "Enter", take that as hitting the submit button.

7. Renaming the "reset" button "clean" seems like a needless change in terminology.

8. Advertisement that is relevant to the search THAT DOES NOT TRACK ME is tolerable as long as it is clear that it is advertisement. If I type in "tents for sale" I'm kind of asking for advertisements.

9. Don't track me. Don't remember me.

Order by date

[Anonymous Coward]

Google can filter by date, but it would be much easier for me if they could order by date, or at least put a greater weight on timeliness when deciding relevance. (This might have saved them a lot of bother with that whole right to be forgotten mess).

I'll also add another vote to being able to filter out the shopping sites when I'm looking for technical info

What the internet needs is a way to focus searches

[GoodBuddy]

...on knowledge domains.

When Google first came out it was a wonder. It saved me an enormous amount of time (=money) in locating information I needed. But it rapidly deteriorated. I remember when there would be people who rejoiced that X billion new pages to their database. I found that with each massive growth of the site it became harder and harder to find the information that I wanted.

I am retired now, but at once upon a time worked develoing applications in Cold Fusion. It was often easier to location a bit of information by enterin a description of what I wanted in a search engine and finding the answer as opposed to pulling out multiple books and checking the indexes. Data when searching the term Cold Fusion clustered in three areas - the application programming language, articles about nuclear energy, and long dissertations that were basically crackpots babbling. For a while it was possible to narrow the search by doing searches of comp Usenet newsgroups, but Google killed the facility of that when they smashed their Google Newsgroups into the mix, and the New York State attorney general then killed Usenet.

But the idea of knowledge domains impicit in the Usenet heirarchy would be very valuable if it could be applied to the internet. Usenet kept control of which groups could be added to 7 of its top levels and alt was free-for-all. Instead of searching the comp.lang.coldfusion Usenet group it would be good to have a search engine at http://coldfusion.lang.comp./ [coldfusion.lang.comp] Whatever organization controled the site could determine what web sites were worthy of being included in the searchable database. All of the automated spam was a major problem on Usenet. Having control on what sites to cover would go a long way to alleviating this problem.

Of course, I have no power or influence about setting this up. All I know is that as far as I am concerned, the internet is fundamentally broken.

"HOW TO" problem-solving, user-teaching searches.

[JoeTechie]

Not a web-page retrieval search, but a solutions retrieval.
I'd like to start with a query and have the engine then ask me some questions. More than disambiguation, I want it to discern the breadth and depth of my knowledge along the answer line. Then have the engine teach some basic (missing) foundation and fill in my holes up to the topic.
We could dialog back and forth, exploring solutions and arriving at the best ones for me.

Examples:
"How to provide a smart power grid for distributed 2-way customers?" Which would work into monitoring and anticipating power generation/draw, placement and sizing of transformers, capacitors &, conductors, surge and safety, redundancy, weather, balancing, and even pricing models.

"How to minimize crime?" Which would work into size and type of scenario (home, business, city, country), laws. Models of culture, education vs crime. What do you mean by "crime"? Minimum levels of freedom, police, education. Even neonatal nutrition and care and their impact on crime.

Re:Execute JavaScript and CSS

[Altrag]

There's a bit of a practicality issue there. Static pages are very well defined in terms of how they're requested and shown.

Dynamic pages are all over the map. There's a handful of popular frameworks, a larger handful of less popular frameworks, and a huge number of roll-your-own solutions. I mean sure to some degree there should be enough commonalities to make a best effort attempt but its never going to be perfect and frankly, Google's more important to your site than your site is to Google so its not surprising that the onus is on you to make your site work with Google rather than the other way around.

Anti safe search

[GuB-42]

Many search engines have a "safe search" feature to hide NSFW material but no way to do the opposite (NSFW only). This has to change.

New widow/tab for resultant links as default. Plea

[asjk]

nt

switch off personalized search

[Gunstick]

Today's search engine like to optimize results for you. So you get the results based on past searches and clicks.
This puts you into your own personal bubble. You only find stuff you already know because you only get links relevant to your personal knowledge field. But what you actually search is a different view on the world, not your own view.
To escape that, you would need to remove all cookies and even change to a completly different PC or ISP or country.
This should be a user user accessible feature. Sort of "amnesia search".

Search algorithms that are not weighted by ads

[whitroth]

Google is far worse than it used to be, signal-to-noise ratio. Part of the time, it does not appear to respect quoted search terms, even with a "+" in front. I now frequently see part of a word that was quoted, bolded in the non-sponsored results.

It also refuses to allow explicit literal searches: I have an artist friend who uses a period as part of her name - google says "nope, not gonna look for name. lastname, I'll ignore it and look only at name lastname".

Finally, I've found too many times in the last year that one of my search terms isn't even mentioned on the page from the results.

                      mark

All Tips & Tricks

[mundlapati]

https://www.google.com/insides...

You may want to reuse

[mundlapati]

https://github.com/trivio/comm...

anything

[gblankityblank]

(almost anything) would be better than the ad-driven nonsense that passes for an Internet search these days

Why are Disqus scumbags?

[tepples]

So any site that uses discus would vanish from my search results.

Did you mean Disqus? So I guess unlike a lot of other users who have complained in comments to this story about Google's, you want a web search engine to correct your spelling.

Those scumbags need to burn in hell

Could you explain why Disqus are more "scumbags" than other comment section hosts? Or could you explain why all comment section hosts are "scumbags"?

Re:Why are Disqus scumbags?

[EmperorOfCanada]

I misspelled it so that that Disqus PR drones wouldn't find the comment. So I will reply with a question, does Disqus aggregate customers' data and sell any of it?