• matjeh (unregistered)

    This application has failed to start because the application configuration is incorrect. Reinstalling the application may fix this problem.

    Note from Alex: This has been fixed (for me in XP at least). I rebuilt it and reuploaded it.

  • Thygrrr (unregistered)

    And this is ranked #6 ???

    OMG... awesome idea. Can't even barely fathom the others.

  • (cs) in reply to Thygrrr
    Thygrrr:
    And this is ranked #6 ???

    OMG... awesome idea. Can't even barely fathom the others.

    I don't think it is ranked 6th... I believe the top 12 are being published in the order submitted...

  • Pinturic (unregistered) in reply to matjeh

    +1

  • (cs)

    omg that's fantastic hahahaa..

  • RC Pinchey (unregistered) in reply to matjeh
    matjeh:
    This application has failed to start because the application configuration is incorrect. Reinstalling the application may fix this problem.

    Same problem here- any ideas anyone?

    As for the submission- I've been looking forward to this one ever since he posted a snippet of his configuration file on the contest forums. Absolute insane genius (but I still think Massimo's going to win!)

  • snqow (unregistered)

    cool!

    this would be an interesting application for palmtops and alikes

    it also reminds me of a Aasimov story where everyone had some sort of palmtop as a calculator and everyone forgot how to make calculations without it.

  • (cs)

    Very 1337 indeed. A completely WTF design to a WTF task. My favourite!

    Configurable numbers. Very useful for when they decide to redefine the shape of zero. Come to think of it -- this may be the world's only completely globalisable calculator. Redefine the shapes to match your culture's number representation...

  • (cs)

    OCR shape definitions in plain English?

    A calculator you can draw on?

    Really. WTF or not, this submission is COOL.

    It even has an application. I'm thinking of a calculator printed on electronic paper, or something...

    Also, I'm vaguely disappointed by most submissions so far. Sure, they use ridiculously roundabout ways to accomplish a simple function, but it just isn't the same when it's done on purpose. They're missing that hint of despairing ugliness, of code grown organically for years in a vast monolithic corporation, of developer fossils who emulate COBOL's behavior in .NET, of Wizards who write a program part Visual C++, VBScript, and part decompiled FORTRAN.

    They're weird, but more shiny than ugly.

  • (cs)

    On Vista with UAC on I had to run the application as administrator or it couldn't find the instructions file in the same directory :D

    Bit hard to draw the stuff but so much fun :P

  • (cs)

    Wow... just wow.

    Up until this entry, I thought I still had a chance at winning. Now I'm not sure I deserve to win.

  • Corporate Cog (unregistered)

    This solution has gone around the circle past wtf, past over-engineered, arriving back near an acceptable solution (if there were no calculators). Well, maybe not... or may be. I'm certainly impressed.

  • Reuben Harris (unregistered)

    The number-shapes definition file is a work of art! I am so going to copy that approach in a future project...

  • (cs)

    Sadly my handwriting with a mouse is absolutely aweful, either that, or the app doesn't work very well. (I could only get a 1, a 0, a -, a / and a =. All other signs were not recognized)

  • Ozzy (unregistered)

    Awesome!

    Has anyone defined the roman numeral configuration file yet? As stated before... insane genius.

  • Hit (unregistered)

    Anyone notice how he parses out comments?

    Any line with a ' anywhere in it is considered one.

    ....

    craaaaazy indeed.

  • Oh (unregistered) in reply to madjo

    I have the same problem. I guess the OCR is picky.

  • (cs) in reply to madjo
    madjo:
    Sadly my handwriting with a mouse is absolutely aweful, either that, or the app doesn't work very well. (I could only get a 1, a 0, a -, a / and a =. All other signs were not recognized)
    The great thing about it, is that you can configure it to recognize your handwriting. :)
  • anne (unregistered)

    "it looks like a sheet of paper"?

    or... eet looks like sheet?

    (I don't actually think it looks like sheet. I just like saying "sheet".)

  • (cs)

    Is anyone else having as much trouble as I am defining the D'ni numerals?

  • [ICR] (unregistered)

    I once saw a video of a Mac OS widget that did a similar thing. You could just write a sum and it would convert it to text in place and give you the answer. You could also circle numbers and drag them about. I could never find it again though, so I don't know if it was a piece of completed software or a concept.

  • Massimo (unregistered) in reply to RC Pinchey
    RC Pinchey:
    Absolute insane genius (but I still think Massimo's going to win!)

    You're talking about my self-destroying objects?

    I guess you're wrong, 'cause it looks like I didn't even make it to the finals (no email from Alex) :-(

  • SomeCoder (unregistered)

    Wow...

    Some people have WAY too much time on their hands :)

  • Marc (unregistered)

    I especially like how the keypad copies the image to the work area, rather than bypassing the OCR function for keypad'ed characters. It's a built-in user-driven unit test. Brilliant!

  • (cs)

    This is the best one yet, all because of the plain English config file. Brillant!

  • RC Pinchey (unregistered) in reply to Massimo
    Massimo:
    RC Pinchey:
    Absolute insane genius (but I still think Massimo's going to win!)

    You're talking about my self-destroying objects?

    I guess you're wrong, 'cause it looks like I didn't even make it to the finals (no email from Alex) :-(

    :-O

    But you made my two favourite entries! The self-deleting objects, and the single-function-overwriting-its-own-operator-instruction-in-machine-code thing... how you've not got through is a mystery to me, but you can console yourself with the fact that you're a firm favourite in the office!

  • Duston (unregistered)

    This application could usher in a whole new era of "softcoding!" Not only could the business rules be softcoded in config files, but the actual look and feel for the entire application could be softcoded! Brilliant!

  • Anonymous (unregistered)

    To all having trouble to enter digits: You have to do it one shape per mousedown. (e.g. three: mousedown-semicircle-mouseup-mousedown-anothersemicircle-mouseup) If you draw the entire digit at once, it won't recognize it.

  • (cs) in reply to Anonymous
    Anonymous:
    To all having trouble to enter digits: You have to do it one shape per mousedown. (e.g. three: mousedown-semicircle-mouseup-mousedown-anothersemicircle-mouseup) If you draw the entire digit at once, it won't recognize it.
    I was just going to post that, I figured it out as well.

    Here is an example of me doing all of them: http://www.javaop.com/~ron/images/screenshots/ocr1.png http://www.javaop.com/~ron/images/screenshots/ocr2.png

    Go me!

  • Brandon (unregistered)

    Ok, just wow. I've worked with OCR before and knowing that to write a from scratch OCR-ish application based on a plain English configuration file... I'm blown away. The WTF here is that someone spent this kind of effort on a programming contest.

    Very very impressive nonetheless.

    Btw, I've gotten it to recognize all of the characters, but it's picky about number of strokes, each stroke must match a line in the config rather than the whole shape.

  • (cs) in reply to RC Pinchey
    RC Pinchey:
    Massimo:
    RC Pinchey:
    Absolute insane genius (but I still think Massimo's going to win!)
    You're talking about my self-destroying objects?

    I guess you're wrong, 'cause it looks like I didn't even make it to the finals (no email from Alex) :-(

    :-O

    But you made my two favourite entries! The self-deleting objects, and the single-function-overwriting-its-own-operator-instruction-in-machine-code thing... how you've not got through is a mystery to me, but you can console yourself with the fact that you're a firm favourite in the office!

    bow

    Even if I too think my submissione were great (of course ;-)), I'm not going to make any complain, because it has been made clear from the beginning that the judgning criteria were purely subjective... and I'm quite sure Alex and the others tried their best to be as impartial as possible, even if actually isn't possible to be impartial when judging entries in such a contest; they're an art form (well, sort of...), and everyone of course has his own likes and dislikes.

    From what I've seen until now, they appear to like overly-complicated approaches and terribly convoluted code more than little but crazy programs; which is fine, of course. Besides, some entries were actually quite great :-)

    I'd have really liked to see more small-but-utterly-crazy entries, those that (in my opinion) make you scream "WTF?!?" a lot more than over-complicated ones; I definitely prefer them, as I already stated in another thread, and as it should be clear from what I submitted ;-)

    Addendum (2007-06-06 11:22): Update: as Alex states here, looks like my guessing was right: the most complex solutions were liked the most.

  • (cs)

    Wow, the plain-text grammatical English configuration file is amazingly awesome.

    The vector definition syntax(describing a line implicitly gives it a name "first", "second", etc., and lines are given position by being joined to each other) is actually very readable. I may have to borrow that sometime.

  • Anon (unregistered) in reply to java.lang.NullReferenceException
    java.lang.NullReferenceException:
    Configurable numbers. Very useful for when they decide to redefine the shape of zero. Come to think of it -- this may be the world's only completely globalisable calculator. Redefine the shapes to match your culture's number representation...

    I'm thinking that you would not be able to configure this for roman numerals because he foolishly hardcoded it to use only decimal systems of numbers. The real WTF is that I can't configure it for binary or hexadecimal ;)

  • (cs) in reply to Ozzy
    Ozzy:
    Awesome!

    Has anyone defined the roman numeral configuration file yet? As stated before... insane genius.

    I was working on it, but it looks like I can't tell it to put a line beside another one. Instead, it stacks them. For example, I configured "II" (2) and it puts the lines on top of each other. If I include "never intersects", it makes a really long line (stacked). For "IX", it originally puts the I in the middle of the X. If I put "never intersects" for both slashes, it puts the I on top of the intersect point instead of to the left.

    shrug

  • (cs)

    In case anyone is interested, here are the two examples I mentioned:

    Two is a long vertical line a long vertical line first never intersects second

    Nine is a long vertical line a long backslash line a long slash line first never intersects second first never intersects third second intersects third

    Nice job, BTW. It's mind-boggling.

  • Southern (unregistered)

    This is brillant ...!! Awesome idea.

  • Anon (unregistered) in reply to AbbydonKrafts
    AbbydonKrafts:
    Ozzy:
    Awesome!

    Has anyone defined the roman numeral configuration file yet? As stated before... insane genius.

    I was working on it, but it looks like I can't tell it to put a line beside another one. Instead, it stacks them. For example, I configured "II" (2) and it puts the lines on top of each other. If I include "never intersects", it makes a really long line (stacked). For "IX", it originally puts the I in the middle of the X. If I put "never intersects" for both slashes, it puts the I on top of the intersect point instead of to the left.

    shrug

    I pretty sure you will not ever get it to work. From the original article:

    * ten symbol names ("Zero" through "Nine") * order of symbols ("Zero" preceeds "One" which preceeds "Two" which preceeds "Three" ... which preceeds "Nine") * "Zero" acts as a placeholder symbol

    These limitations would make it impossible to configure it for roman numerals (for instance, roman numerals has no 0) regardless of whether it can recognize two lines next to each other.

  • Def (unregistered) in reply to RC Pinchey
    RC Pinchey:
    matjeh:
    This application has failed to start because the application configuration is incorrect. Reinstalling the application may fix this problem.

    Same problem here- any ideas anyone?

    As for the submission- I've been looking forward to this one ever since he posted a snippet of his configuration file on the contest forums. Absolute insane genius (but I still think Massimo's going to win!)

    This is due to the executable being built with an embedded manifest file which references two different versions of the same DLL. (Usually the CRT runtime DLL.)

    This usually happens when you link a static library built with the RTM version of Visual Studio 2005 into an executable built with VS which has service pack 1 installed. (And vice-versa.)

    The executable will work on machines which have both versions installed and registered (such as the machine the program was built on or any machine which has VS SP1 installed) but will most probably not work on any others.

    It's the new version of DLL hell for a new millenium, basically.

  • (cs)

    Still doesn't work for me... I'm with XP SP2

  • porgo (unregistered)

    I have to say -- from all the entries so far, this is the one which both should and shouldn't take the cake. There is no "a normal wtf" here, but this is so awesome, so crazy, so perfect by its idea that there hardly can be any match for this.

    I wouldn't say this should win the competition. Seems too well written for a wtf. But this is an amazing piece of work.

  • Lummox (unregistered) in reply to Thygrrr
    Thygrrr:
    And this is ranked #6 ???

    Can't even barely fathom the others.

    The WTF in these comments is your use of grammar. :P

  • guy (unregistered) in reply to Duston
    Duston:
    This application could usher in a whole new era of "softcoding!" Not only could the business rules be softcoded in config files, but the actual look and feel for the entire application could be softcoded! Brilliant!

    Isn't that what WPF is?

  • Tom (unregistered) in reply to Hit

    Not quite...

    If a line has a ' in it and it is not preceded by a space, the whole line is a comment. But if a line has a ' in it which is preceded by a space, only the remainder of a line is a space. E.g.

    This is not a comment but 'this is now a comment 'This is a comment too This whole line's a comment

  • (cs) in reply to Welbog
    Welbog:
    Is anyone else having as much trouble as I am defining the D'ni numerals?

    Winner of nerdiest person on this site.

    Or does that make me runner up for knowing what this guy is talking about?

    crap.

    This is just plain awesome.

  • (cs)

    I feel like a piece of shit. I have no talent at all, not even to think of software like this.

  • (cs) in reply to JamesKilton
    JamesKilton:
    Welbog:
    Is anyone else having as much trouble as I am defining the D'ni numerals?

    Winner of nerdiest person on this site.

    Or does that make me runner up for knowing what this guy is talking about?

    crap.

    This is just plain awesome.

    Nope, nerdiest person would actually be able to code them.

    D'nian number system. Ah, the nostalgic memories of Riven.

    I think it's possible given the grammar in the definition file, but it would take a whole lot of statements.

  • ItsMeAgain (unregistered)

    The entry needs a rename.

    It's not really OCR, more like MCR (Mouse Character Recognition).

    If it was real OCR I could write an equation on paper, put the paper on a wooden table...

    (... all non-newbies know where it goes from there)

  • Andrew (unregistered)

    Impressive.

    Too bad it wasn't implemented as a web service. We could have mounted a camera or a scanner to the output tray of a printer. On a wooden table, of course.

  • (cs) in reply to Erick
    Erick:
    JamesKilton:
    Welbog:
    Is anyone else having as much trouble as I am defining the D'ni numerals?

    Winner of nerdiest person on this site.

    Or does that make me runner up for knowing what this guy is talking about?

    crap.

    This is just plain awesome.

    Nope, nerdiest person would actually be able to code them.

    D'nian number system. Ah, the nostalgic memories of Riven.

    I think it's possible given the grammar in the definition file, but it would take a whole lot of statements.

    Weird... Yesterday, I was nostalgic for the game and I pulled out the game and started playing it again.

    So, would it be nerdier if I could get the Narayan numerals working?

  • (cs) in reply to Anon
    Anon:
    These limitations would make it impossible to configure it for roman numerals (for instance, roman numerals has no 0) regardless of whether it can recognize two lines next to each other.

    Obviously true Roman numerals couldn't be done. But, I was experimenting with just 1 through 9.

Leave a comment on “OMGWTF Finalist #06: OMG!OCRCAL”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article