Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Math The Internet Science

New Company Seeks to Bring Semantic Context To Numbers 264

Posted by ScuttleMonkey from the quick-create-another-standards-body dept.
A new company, True#, is seeking to bring extensive semantic context to numbers to give them obvious meanings just as certain words have obvious meanings to most readers. "Most of us can probably recognize 3.14159 and the conceptual baggage it carries, but how many of us would recognize 58.44? (That's a mole of sodium chloride, in grams, for the curious.) And the response that would work for words — look it up — doesn't work so conveniently for numbers. Only one of the top-10 hits in Google refers to salt, and Bing fails entirely (though it does offer 'Women's Sexy Mini Skirts by VENUS'). Clearly, we haven't figured out how to make the Web work for numbers in the same way it does for words."
This discussion has been archived. No new comments can be posted.

New Company Seeks to Bring Semantic Context To Numbers More

New Company Seeks to Bring Semantic Context To Numbers

Comments Filter:

  • Web searches aren't THAT bad... (Score:5, Funny)

    by Monkeedude1212 ( 1560403 ) on Monday August 10, 2009 @01:31PM (#29014161) Journal

    1337 returns EXACTLY what I expected.

  • Works for me (Score:3, Funny)

    by winkydink ( 650484 ) * <sv.dude@gmail.com> on Monday August 10, 2009 @01:31PM (#29014165) Homepage Journal

    I get one return for NaCl on Bing and nothing about miniskirts.

    Cue the conspiracy theorists.

  • When I search for infinity... (Score:5, Funny)

    by Anonymusing ( 1450747 ) on Monday August 10, 2009 @01:32PM (#29014181)

    ...it returns "number of years it will take before True# turns a profit."

    I'm seriously confused how many companies will jump at this -- and why someone like Google won't just do it for free? Couldn't you use Google Base for something like this?

    • Speaking of infinity, query the number "11 trillion" ... in my view, both Google and Bing get it exactly right!

      Which begs the question, what's the point of this True# service when the major search engines already appear to do a good job with numbers.

      Ron

      • INCORRECT USAGE (Score:2, Insightful)

        by Anonymous Coward

        "Begs the question" is not a synonym for "raises the question".

        http://en.wikipedia.org/wiki/Begs_the_question [wikipedia.org]

        • Re:INCORRECT USAGE (Score:5, Funny)

          by pjt33 ( 739471 ) on Monday August 10, 2009 @02:50PM (#29015277)

          I'm sure you've heard the one about the linguist who was walking across campus with his girlfriend when they saw six descriptivists beating up a prescriptivist. She turned to him in horror and asked, "Aren't you going to help?"

          "No," he replied, "I think six is enough."

          • Re: (Score:3, Interesting)

            by idontgno ( 624372 )

            "'I don't know what you mean by "glory,"' Alice said. Humpty Dumpty smiled contemptuously. 'Of course you don't- till I tell you. I meant "there's a nice knock-down argument for you!"' 'But 'glory' doesn't mean "a nice knock-down argument,"' Alice objected. 'When I use a word,' Humpty Dumpty said, in rather a scornful tone, 'it means just what I choose it to mean- neither more nor less.'

            Welcome to Slashdot, Mr. Dumpty

  • why (Score:5, Insightful)

    by stoolpigeon ( 454276 ) * <bittercode@gmail> on Monday August 10, 2009 @01:32PM (#29014191) Homepage Journal

    In all seriousness - this is not a rhetorical question. Usually I want this information in the inverse order, not just having a number with no context. What is the value in searching in that direction is their some widespread need I don't know about?

    • 42 (Score:2, Funny)

      by edalytical ( 671270 )

      Need I say more?

    • Re:why (Score:5, Insightful)

      by winkydink ( 650484 ) * <sv.dude@gmail.com> on Monday August 10, 2009 @01:43PM (#29014373) Homepage Journal

      Shhhh... Since when has making sense had anything to do with raising money from VC's? Especially when you can use the word crowdsourcing in your pitch.

    • Reverse Engineering and Better Search (Score:5, Interesting)

      by schnablebg ( 678930 ) on Monday August 10, 2009 @01:53PM (#29014531)

      I have had this need when reverse engineering and debugging algorithms in software. There are magic numbers in the formulas and I have no idea what they mean.

      Additionally, if something like this was rolled into a more generalized search algorithm, it could be used the other way around. Google could know, for example, that a paper with the number 58.44 a lot of times is probably about NaCl even if it is not mentioned explicitly.

      • Re: (Score:2)

        by JMZero ( 449047 )

        I see plenty of magic numbers in code - but most of them are just arbitrary failures, or sentinels, or tweaked constants. A search engine wouldn't have anything to say about those.

        The few magic numbers I could think of that are meaningful - like for example c704dd7b (used in CRC) - are long and widely used enough to already come up with fine results in Google.

      • Re: (Score:3, Funny)

        by NonSequor ( 230139 )

        I have had this need when reverse engineering and debugging algorithms in software. There are magic numbers in the formulas and I have no idea what they mean.

        Additionally, if something like this was rolled into a more generalized search algorithm, it could be used the other way around. Google could know, for example, that a paper with the number 58.44 a lot of times is probably about NaCl even if it is not mentioned explicitly.

        Sorry, you're out of luck there. Magic numbers work because they're actually magical. Sorry to burst your bubble if you still believe in things like algorithms and the tooth fairy.

        Did you honestly believe that computers worked by interpreting a series of logical instructions input by a mortal human being? Trust me. When you're ready to understand you won't need to search.

    • mole (Score:4, Informative)

      by AliasMarlowe ( 1042386 ) on Monday August 10, 2009 @02:00PM (#29014651) Journal
      If the search term includes "mole" as well as "58.44", the first few pages of google results are almost all for stoichiometry of NaCl. Nuff said, Google works.

    • Re:why (Score:4, Insightful)

      by 93 Escort Wagon ( 326346 ) on Monday August 10, 2009 @02:23PM (#29014933)

      In all seriousness - this is not a rhetorical question. Usually I want this information in the inverse order, not just having a number with no context. What is the value in searching in that direction is their some widespread need I don't know about?

      I suspect the primary use would be using an internet-connected device to cheat on multiple-choice tests.

      • Re:why (Score:4, Insightful)

        by Obfuscant ( 592200 ) on Monday August 10, 2009 @03:03PM (#29015455)
        I suspect the primary use would be using an internet-connected device to cheat on multiple-choice tests.

        So you are saying that when someone sees the question: "What is the molecular weight of NaCl? 1) 34.99 2) 43.33 3) 58.44 4) 3" they'll google each number to see what it says instead of googling the question itself?

    • Re:why (Score:5, Interesting)

      by steelfood ( 895457 ) on Monday August 10, 2009 @04:39PM (#29016571)

      You're right. Numbers are the abstract representation of a specific concept. Words are a specific representation of an abstract concept.

      It doesn't make sense to search the other way. Sure, you can search for, say, 58.44 and get relatively few correct responses. However, what if you searched for 10? Or 1024? Then what's supposed to come up?

      Numbers don't have any a priori meaning. They require context. Otherwise, they're just a meaningless abstraction. Context makes 1 and 12 different, or the same (inches in a foot). It doesn't make sense to compare 1 mol of NaCl with 12 light years. As human beings, we can imply context without explicitly stating it. But that doesn't mean we don't need context.

      It's like the difference between mathematics and physics. Pure mathematics is not useful without an application to a physical problem (not to say that there's no purpose to development of pure mathematics).

  • Errrm (Score:5, Insightful)

    by MyLongNickName ( 822545 ) on Monday August 10, 2009 @01:32PM (#29014197) Journal

    So, a search doesn't bring up what one person would expect and that means the search engine failed? Sometimes the problem with logical fallacies is that they are so big as to defy categorization.

  • Wrong conclusion (Score:5, Funny)

    by godrik ( 1287354 ) on Monday August 10, 2009 @01:34PM (#29014229)

    Only one of the top-10 hits in Google refers to salt, and Bing fails entirely (though it does offer "Women's Sexy Mini Skirts by VENUS").

    Bing seems far superior to my hormon^W^Wme.

  • Sig Figs (Score:5, Informative)

    by Anonymous Coward on Monday August 10, 2009 @01:36PM (#29014251)

    Actually, the accepted weight is 58.443 thats why Bing didn't show any NaCl results.

    • Bing and/or google will also not show results you expect for words if you don't actually put in enough context to make the search relevant for you. I mean seriously, who's going to look up 58.44 or even 58.443 all by itself? What were they trying to find? If you want all instances of the number 58.44, then that's all you type. Otherwise you need to add some qualifier, otherwise you're going to get whatever Bing or Google determines is most likely a match to what you want. Simply add "weight" or "mole"

    • Actually, the accepted weight is 58.443 thats why Bing didn't show any NaCl results.

      http://www.wolframalpha.com/input/?i=1+mole+of+sodium+chloride [wolframalpha.com]
      Wolphram Alpha calls it 58.4

      so much for significant figures.

  • First you need a semantic context (Score:5, Informative)

    by davidwr ( 791652 ) on Monday August 10, 2009 @01:36PM (#29014255) Homepage Journal

    Search [google.com]58.44 and chemistry and you'll find what you are looking for a lot faster.

    This will be much more useful if it allows for approximate numbers and widely-used but inaccurate numbers. "1.4 math" should return 7/5, sqrt(2), and a bunch of other things. "3.142857 and math" should return "22/7" and "approximate value of pi" and probably a lot more.

    • Without units, it's pretty pointless. For example, 0.129 lbs is the weight of a mole of sodium. The possibility space of three-digit numbers is only 1000 entries, and saying that 129 is the weight of a mole of sodium in millipounds is another possible response.

      So, basically, without units asking what a number means is pretty dumb.

      • Where? (Score:3, Funny)

        by davidwr ( 791652 )

        The weight of a mole of sodium varies by location. In most of the universe the weight of anything is almost zero.

      • Re: (Score:2)

        by treeves ( 963993 )
        Strictly speaking you are correct, but as someone with a fair amount of chemistry experience, I'll say that the weight of a mole of X is always expressed as g or kg, even in the non-SI standard USA.
    • I used 58.44 and molecular weight and all top ten hits told me what molecule I was dealing with. Same with 58.44 g/mol. You are right, numbers require context or 58.44 could be anything, and most of the time a generic search without units the context would be wrong. I am sure that I have used 58.44 outside the context of chemistry much more often than I have in it even though I spent many years in a molecular biology lab.

  • Seems to work just fine (Score:5, Informative)

    by gt6062b ( 1548011 ) on Monday August 10, 2009 @01:36PM (#29014269)
    I'll bite.

    http://www.google.com/search?hl=en&q=mole+of+sodium+chloride+in+grams [google.com] - seems to work just fine searching for "mole of sodium chloride in grams" and also works without the "in grams".

    http://www.bing.com/search?q=mole+of+sodium+chloride+in+grams [bing.com] - works for Bing too.

    http://www.bing.com/search?q=sodium+chloride+molecular+weight [bing.com] - also works.

    http://www.bing.com/search?q=58.44+science [bing.com] - "58.44 science" 6th one down. Better results from google.

    Why would anyone just type in a number and expect it to know that you want the molecular weight of NaCl? If you add a little bit of context to your search, it magically works.

    • Hell, even if you had the number, and needed to figure out what it was, WHO SEARCHES WITH NO CONTEXT?

      Type that in and the word "chemistry", (the overwhelmingly vague content area you might have encountered that number in), and 58.44 shows up as sodium chloride with no problems.

      Is this for the same people who type "cocks" into a search engine while looking for birds, and are shocked at the results? I was under the assumption that most people understood that search engines don't read minds..

    • Re: (Score:3, Insightful)

      by Rival ( 14861 )

      I'll bite.

      http://www.google.com/search?hl=en&q=mole+of+sodium+chloride+in+grams [google.com] - seems to work just fine searching for "mole of sodium chloride in grams" and also works without the "in grams".

      http://www.bing.com/search?q=mole+of+sodium+chloride+in+grams [bing.com] - works for Bing too.
      http://www.bing.com/search?q=sodium+chloride+molecular+weight [bing.com] - also works.
      http://www.bing.com/search?q=58.44+science [bing.com] - "58.44 science" 6th one down. Better results from google.

      Why would anyone just type in a number and expect it to know that you want the molecular weight of NaCl? If you add a little bit of context to your search, it magically works.

      You're missing the point. The purpose is to provide explanation for numbers which have no context. Presumably, if you're looking at some equation or source code which uses an unrecognized constant, or if a calculation returns a surprising result, one might be able to use such a search to find more information.

      For example, let's say you're poking around with math and discover that 0.5^0.5 == 0.25^0.25 =~ 0.70710678118654752440084436210485. Is that irrational number significant somehow? What if you were o

    • HELP!! I'm stuck in an infinite loop!
      When I google "58.44 science" it just brings me back to this page!

    • Re: (Score:3, Interesting)

      by Deanalator ( 806515 )

      Sometimes, in your calculations (in physics class especially for me), you come across what seem to be magical numbers, and later realize it's something like e^5 or something weird like that. It has also been especially startling to me when certain numbers that appear are perfectly clean square or cube roots etc.

      That's the strange thing about irrational numbers, there are only a few unique ones that are interesting. All the others are built from rational numbers, or other more interesting irrational number

      • 9.8 works just fine. But then again on Earth its the law.
        6.67300 works as well.
        So does 98.6.

        Although -240, 365, 32 didn't work.

      • If you search for the magic number and also add whatever context you DO know, the result should appear. If I'm doing digital signal processing, I'd look for the number and DSP, or acoustics or something like that. If it were optics, or whatever else, just add a word that applies. You would never find a magic number thrown in the middle of something and not be able to figure out what realm of knowledge it fits into. Especially something like source code - you probably know what the program does, so just

  • Hitchhiker's Guide? (Score:5, Funny)

    by R2.0 ( 532027 ) on Monday August 10, 2009 @01:36PM (#29014275)

    So, put in the (numerical) answer and it gives you the question?

    Thank God Douglass Adams didn't know about this.

    • forgive me but I don't remember the actual quote but it goes something like this:

      "Some say that if the answer and the question where known at the same time that the univers would be destroyed and replaced with one far stranger. Some say this has already happened."

  • okey dokey (Score:4, Insightful)

    by Nyall ( 646782 ) on Monday August 10, 2009 @01:37PM (#29014277) Homepage

    Sounds like a solution looking for a problem.

  • Is this a joke? (Score:2, Funny)

    by Anonymous Coward
    How often do you have a number without units? Putting in "58.44g" returns NaCl, sodium chloride, or molecular weight in every one of the first page results.

  • This is really stupid. (Score:5, Insightful)

    by StellarFury ( 1058280 ) on Monday August 10, 2009 @01:42PM (#29014343)

    Why in the hell would I want to search for a number with no context? Who thinks that way? Everyone remembers the concept, not the number.

    You say "3.14" and people know it as pi. But if you said "pi," people would say "3.14." This example is only interesting because it's widespread.

    Nobody would start with "58.44" and say "Hmmm, what does that symbolize?" No. They need to know the molecular weight of sodium chloride, and so they'll search Google for "molecular weight sodium chloride" and turn up the number 58.44. We're not computers, we know semantic context, and need numbers. Not the other way around.

    Though I guess this sort of thing might be useful for some sort of numerical AI, who has numbers but no semantic context. Time to don the tinfoil hats, fellows.

  • Whats next? Astrological web?

  • 420 comes up correctly! :) (Score:4, Funny)

    by Ron Bennett ( 14590 ) on Monday August 10, 2009 @01:43PM (#29014371) Homepage

    For the number 420, Wikipedia's Cannabis information page comes up #1 in both google and bing.

  • 35484.32384 (Score:2, Funny)

    by Anonymous Coward

    3494.13
    332
    8494.354
    2324.234

    Hahaha, I kill me sometimes.

    Get it? 332 ... 3494.13 -> 8494.354 ?

    Man, that is too funny!

  • 1. Eleventeen
    2. 867-5309
    3. 451
    4. 1999
    5. a gazillion
    6. THIS MANY (holding up three fingers)
    7. infinity minus one
    8. approximately
    9. 9/11 (may already be taken)
    10. Top ten

  • So Bing apparently works and is superior to Google.

  • Sounds like Marchex (Score:3, Interesting)

    by Animats ( 122034 ) on Monday August 10, 2009 @01:47PM (#29014437) Homepage

    Marchex [marchex.com] is a "domaining" company. They're the people behind those "What you need, when you need it" ad-heavy landing pages. [90210.net] They tried to buy up all the 5-digit number domains. (So did others; nobody got all of them.) This sounds like a similar idea, only less profitable.

  • Thank goodness my numbers are safe (Score:5, Funny)

    by UnknowingFool ( 672806 ) on Monday August 10, 2009 @01:48PM (#29014449)

    When I searched "1234" on google and bing, the top results are about that Feist song. Thank goodness it doesn't mention anything about it being my root admin password and my luggage combination--hey! Where did my bag go? It was just here, and why is there a sudden spike in my internet tra#%^W&*s%!$AF{:

    ---[CONNECTION LOST]---

    • Re: (Score:2)

      by sconeu ( 64226 )

      When I searched "1234" on google and bing, the top results are about that Feist song. Thank goodness it doesn't mention anything about it being my root admin password and my luggage combination

      Well, duh... that's because your luggage combo is 12345 [imdb.com].

    • When I searched "1234" on google and bing, the top results are about that Feist song. Thank goodness it doesn't mention anything about it being my root admin password and my luggage combination--hey! Where did my bag go? It was just here, and why is there a sudden spike in my internet tra#%^W&*s%!$AF{:

      ---[CONNECTION LOST]---

      That's why i use "1235" - it throws the hackers and thieves off.

  • Google Works Fine (Score:4, Informative)

    by shma ( 863063 ) on Monday August 10, 2009 @01:49PM (#29014465)
    Their example fails because they chose a number that has no significance on its own without including a unit of measurement. If you search 58.44 grams [google.ca], instead of just the number, you get plenty of relevant results. And look at what happens if you take a famous unitless number from chemistry and do a google search [google.ca]. Again, plenty of good results.You can try it with the speed of light as well. A search for 3x10^8 yields nothing, but 3x10^8 m/s gives you the Wikipedia page for Speed Of Light. And as far as I can tell, Google gives you good results for useful numbers in Mathematics like the golden ratio [google.ca]. So I don't see what the problem is.

  • wrong, there are no problems (Score:3, Interesting)

    by iggymanz ( 596061 ) on Monday August 10, 2009 @01:49PM (#29014475)

    Put in "58.44 moles" and you'll have proper answers in the top ten, putting in units makes all the difference for number searches.

    And as for dimensionless numbers, 3.14149 gives wikipedia article for pi, 2.71828 gives wikipedia for e as top answer, even "square root -1" gives i.

    Absolutely false the premise or conclusion of this article. Searching for numbers gives useful information

  • Seems that we'd need a "measurement" element, with a quantity and a unit. <measurement quantity="1" units="m/s"> Then your browser can render the m/s as it wants to.For instance units of a circle (degrees or radians) can be rendered as the number or graphically.

    Of course, this opens a can of issues in expressing equasions, because X could be a quantity too. Expressing <term><<measurement quantity="1" units="m"><over><measurement quantity="1" units="s"></term> would pro
  • Using numerics in search phrase construction just seems like it would screw up the results regardless. It would be similar to constructing search phrases using multiple foreign languages at the same time. Even when searching scientific journals, I use common nomenclature rather than specific values.
  • I find it ironic that the 'net is good with words and not with number, as numbers are what it's based on.

    • Re: (Score:3, Insightful)

      by Pollardito ( 781263 )
      The fact that computers are based on numbers is part of the reason it's hard to be good at this, since numbers are used to represent anything and everything. Numbers sometimes aren't really even "amounts" they can be ASCII codes or addresses, so if you give me 58.44 how do I know that's not just ":.," or room #44 on floor #58 (or room #68 on floor #88 from hex). Numbers are meaningless without context. Even the phone numbers that we use every day would be unrecognizable without specific formatting and/or
  • Sounds to me like the Underpants Gnomes have found their niche on the internets.

  • Google obviously fails. A search for "12345" had exactly ZERO hits to Scroob's luggage OR Druidia's Air Shield.

    What kind of fly-by-night company are you running there, guys?!

  • I typed "12345" into Google, and Google did not know that was the combination to my luggage.

  • ...that as others have pointed out, pretty much any useful number search can be done with existing search engines. Meanwhile, a Google search for "true#" turns up nothing relevant.

    This is why we don't put funny characters in our company names, kids.

  • Sounds like the Encyclopedia of Integer Sequences (Score:3, Informative)

    by OfficeSupplySamurai ( 1130593 ) on Monday August 10, 2009 @02:05PM (#29014715)
    This sounds a lot like the Online Encyclopedia of Integer Sequences [att.com].

    That site labels and stores integer sequences for easy lookup, and will let you simply search for a subsequence to find the one you're looking for. This proposed site keeps track of numbers instead and incorporates more than the pure math that the sequence encyclopedia limits itself to, but it sounds very similar in concept.
  • That's a mole of sodium chloride, in grams, for the curious.

    Can we at least have some pumpkin seed in that too?

  • "Clearly, we haven't figured out how to make the Web work for numbers in the same way it does for words."

    We haven't figured that out even for our brain!

  • If it is anything like this, then I might be interested: http://xkcd.com/526/ [xkcd.com]

  • The 1990's called, and they want their business models back (if they'll give me my 401K back it's a deal!)

    So, we can either:

    1. Pay this company money to create a link from a number on my web page to their page, where full details live,

    2. or

    3. Define an open XML namespace which we can pull in to describe a number, something like this:

      <sn:number_use fields="chemistry" description="molecular weight of NaCl" href="http://en.wikipedia.org/wiki/NaCl">58.44</sn:number_use>

      and have something we can use free

  • when you Google for 58008 ?

  • Using 3.14159 and 58.44, those are particularly poor examples.

    First, the value of pi can't be written down exactly, in fact the term "pi" is the shortest and best, so that's not a good example.

    And the gm/M of table salt isn't too keen an example either-- that number is going to vary depending on the isotopic composition of the sodium and chlorine.

    So maybe "pi" and "table salt" are already good semantic descriptors.

  • One reason a Google search for 58.44 turns up so much crap is the non-adjustable punctuation filter Google uses. AFAIK, you can't search for the exact phrase '58.44' and have it exclude IP addresses that contain *.58.44.*

  • I've got your semantic context for numbers right here.

  • I've long been looking for a browser plugin that will convert any numbers and units it finds on a Webpage and replaces them with SI equivalents. Come to think of it, you wouldn't need tagging for that: convertible numbers will invariably have a relevant unit symbol next to them.

  • There are a few similar things out there (Score:3, Informative)

    by John Sokol ( 109591 ) on Monday August 10, 2009 @02:50PM (#29015275) Homepage Journal

    http://www.research.att.com/~njas/sequences/ [att.com] This is huge.

    http://www.amazon.com/Dictionary-Real-Numbers-J-Borwein/dp/0534128408 [amazon.com] A Dictionary of Real Numbers (Hardcover)

  • and all these responses. OK, sorta cool, I guess. But I don't understand how someone is going to make money on this.

  • That's because numbers without context are completely meaningless.

    Here's proof by examples:

    29A would scare a lot of people were it to magically appear on their bathroom mirror overnight, but only if written differently. That alone is sufficient proof for me that numerology is stupid, and that superstition is all in your head. You'll never hear a profit talking about 29A.

    52 is at the heart of a lot of not-so-inside jokes. That is, if you first convert it from octal.

    0 is the true/good/success in *nix, not

  • This sounds like one of those gee-whiz attempts to capitalize on current buzzwords.

  • 58.44 is nothing, as is 5.58880653.
    Unless you add the units.
    58.44 grams = 5.58880653 ounces.
    Grade school Science class will teach you this.
    In science a number without its units is nonsense.
    Pi = 3.14159 is a ratio and thus has no units.

    So try your google search with units.
    And low and behold bingo it matches.

    Unless we miss the point.
    Let not start searching for salt.

  • Google had my number (Score:3, Funny)

    by Skapare ( 16644 ) on Monday August 10, 2009 @04:09PM (#29016181) Homepage

    2635622779696759818963956926355997625653382829357706805515232 / 838944787028681613144502774660896402692975681322322888764935

    I have lots of better ones. But they'd probably break Slashdot to post them.

Slashdot Top Deals

Genius is ten percent inspiration and fifty percent capital gains.

Close