Kiran Jonnalagadda’s Blog

Tuesday, March 9, 2004

My name is キラーン

Translation into Japanese courtesy shradha. Hope I got the characters right (corrected, thanks evan!).

• 

  yellosonja — Mar 9, 2004 10:37:24 AM — # ↩

  cool! could you please tell me how to do mine too?
  • 

    Kiran Jonnalagadda — Mar 9, 2004 11:02:38 AM — # ↩

    Mine name is किरण and your name is सोनिया. I don’t do Japanese unfortunately, but maybe the good lady shradha can help.
    • 

      madhav — Mar 9, 2004 11:25:48 AM — # ↩

      Er..shouldn't that be िकरण and सोिनया ?
      • 

        Kiran Jonnalagadda — Mar 9, 2004 11:35:09 AM — # ↩

        Isn’t that what I typed? िकरण आैर सोिनया। But s/mine/my/ earlier.
        • 

          gromhellscream — Mar 9, 2004 11:44:58 AM — # ↩

          doesn't seem to be you got the 'E' ki maatra wrong in kiran and 'E' wrong again in niya ( you wrote nyia and karin ):)
          • 

            Kiran Jonnalagadda — Mar 9, 2004 2:25:24 PM — # ↩

            Looks right to me. Guess different browsers render differently. Firefox shows it as all ???s.
            • 

              mannu — Mar 9, 2004 2:56:52 PM — # ↩

              Your page needs to have UTF-8 encoding set. That's what LJ uses. See the meta tags in the HTML.
              
              There are 2 reasons why the Japanese text won't come proper:
              
              1) The browser has no clue that the text is UTF-8-encoded or is told otherwise (because of bad META tags in the HTML).
              
              2) The font doesn't exist on the system.
              • 

                Kiran Jonnalagadda — Mar 9, 2004 3:02:38 PM — # ↩

                The Japanese is fine. The Davanagari isn't. Safari on the same OS displays it fine, Firefox doesn't. But the more I use Firefox, the more I dislike it, so I don't really care now. Mozilla & Co. have lost the race to Safari.
              • 

                Kiran Jonnalagadda — Mar 9, 2004 3:29:52 PM — # ↩

                My bad. s/Davanagari/Devanagari/
          • 

            Kiran Jonnalagadda — Mar 9, 2004 3:25:29 PM — # ↩

            Looking at it using IE6 on Windows XP now:
            
            * The Japanese text is all square boxes (missing font?)
            * My first two Devanagari words are rendered correctly
            * Madhav’s version of my name is all squares, while Sonia’s name is incorrectly rendered (dotted circle between na’s matra and na itself).
            * My reply to Madhav’s comment shows all squares.
            
            In Firebird (don’t have Firefox on Windows), Japanese is missing again, my first Devanagari versions are correct, while Madhav’s and my second versions (copy and paste from Madhav’s version, with ‘aur’ inserted) are wrong.
            
            Obviously, Unicode/OpenType support on Windows XP is broken (I’m using a stock install). Safari renders everything correctly except the matra in ‘aur’, which is a few pixels off to the right.
            
            Can any of you guys using other platforms/browsers comment? Sonia says that on Windows 98, my versions are wrong while Madhav’s are right.
          • 

            Kiran Jonnalagadda — Mar 9, 2004 3:48:46 PM — # ↩

            brainz just sent me a screenshot of IE6 on Windows XP on his box, and the rendering is as you describe it. I guess the issue was “fixed” by XP’s service packs.
            
            We should take this up with Google: My versions of Kiran and Sonia, and Madhav’s versions of Kiran and Sonia. From what I can make out, Google says my versions are correct and your browsers are wrong.
            • 

              evan — Mar 9, 2004 3:59:42 PM — # ↩

              I know very little about Devanagari, but: is it one of those languages where you reorder the vowels within syllables visually? Then it's possible that one of you is writing it in the order it sounds, the other in the order it is supposed to appear visually, and whoever's browser (yours, it sounds like) has the better Devanagari support is correctly reordering them to both look right.
              
              Over here (firefox on Linux) your version of Kiran has the character that looks like a single line that curves over on the far left, while Madhav's has it one character in from the left. Which is correct?
              • 

                Kiran Jonnalagadda — Mar 9, 2004 4:15:28 PM — # ↩

                I’m writing by sound, Madhav by appearance. This is what Safari renders it like. The outer words are correct, middle incorrect (the two ticks should originate on top of the vertical bar).
                
                • 

                  evan — Mar 9, 2004 4:39:28 PM — # ↩

                  Then yeah: these browsers (mine included) aren't doing it right. It's a shame, too: Pango looks to have Devanagari support, and it's already a dependency of any GTK2 app, but even though firefox uses GTK2 it uses its own text layout code.
                  • 

                    Kiran Jonnalagadda — Mar 9, 2004 5:08:22 PM — # ↩

                    How about this? Both Safari and Firefox on OSX get it wrong, while GtkHTML/Evolution (on golisoda’s Linux box; this is from his email signature) does it right. brainz sent me screenshots from IE6 and Firefox on XP. IE6 does it right, Firefox wrong (same rendering as Firefox on OSX).
                    
                    The script/language is Kannada, which uses the same characters as Devanagari but different symbols (i.e., loss-less transliteration between scripts is possible).
                    
                    Now it appears that Pango (GtkHTML) is the only renderer that gets everything right. Wish I had a working installation at hand to poke at.
                    • 

                      evan — Mar 9, 2004 5:22:05 PM — # ↩

                      My fonts don't support that, unfortunately. :(
                      
                      I'm not surprised about Pango, though... it's pretty rad.
                      It's a shame that GtkHTML seems pretty impossible to work with (I'm trying to use it in my app and it just crashes if I put it into editing mode...).
                      • 

                        Kiran Jonnalagadda — Mar 9, 2004 5:29:34 PM — # ↩

                        I use the Sampige font available from http://brahmi.sourceforge.net/downloads.html
              • 

                Kiran Jonnalagadda — Mar 9, 2004 4:18:57 PM — # ↩

                Also, Devanagari isn’t a language, but a script like Roman that is shared between Hindi, Marathi and Sanskrit. And yes, vowels are rearranged visually.
              • 

                Kiran Jonnalagadda — Mar 29, 2004 4:39:32 AM — # ↩

                So I just discovered that औ is a character by itself in Unicode, making my आैर spelling wrong. It should have been और instead.
                
                Which also means that OSX is rendering Devanagari correctly. I haven't found any other errors.
                • 

                  evan — Mar 29, 2004 4:51:11 AM — # ↩

                  I don't know enough about Devanagari to comment except to note that Unicode has in many places precomposed versions of characters: for example, there's "roman a" and then "combining character umlaut" and then there's an entirely separate entry "roman a with umlaut". They did it for backward compatibility with other encodings, I think... and also maybe to make rendering easier?
                  
                  So it's possible the precomposed character is one of those.
                  • 

                    Kiran Jonnalagadda — Mar 29, 2004 6:35:03 AM — # ↩

                    Yes, and I was Googling around for ISCII and Unicode and found someone's argument for why Unicode was unusable for Indian languages. He pointed out that Unicode's multiple ways to represent the same glyph makes searching a pain.
                    • 

                      evan — Mar 29, 2004 2:37:56 PM — # ↩

                      I think they're supposed to pick a "normalization form", so that strings compare properly. (For example, if you download a file on OS X that has a Devanagari name, it'll reformat the string to the same way it has done others.)
                      
                      Of course, I get the impression people rarely do this in all cases. (If you have a bunch of UTF-8 input, you have to normalize all of it before you can sort it, which means another O(n*(length of longest string)) pass...)
        • 

          balaji — Mar 17, 2004 4:34:03 PM — # ↩

          what you have written above and that written by madhavan appears as boxes in IE 6.0, Windows XP SP1. what you have written earlier appears correctly (as in Kiran and Soniya).
          
          On Firefox, what you have typed first, is shown as karin and sonyi... the one above appears fine though.
          
          but your name in Japanese appears as boxes in the title bar on both IE and Firefox.
          -balaji
          
      • 

        satyap — Mar 9, 2004 8:19:26 PM — # ↩

        On Windows using Mozilla 1.5, jace is correct. madhav's ki is displaced, also the ni in 'sonja'.
        • 

          madhav — Mar 10, 2004 1:24:58 AM — # ↩

          I'm using Firefox on Windows 2000. Interesting discussion!
• 

  evan — Mar 9, 2004 3:15:19 PM — # ↩

  not quite right: japanese has a few letters that are easily confusable.
  compare:
  リンソ
  that's "ri" (or "li"), "n", and "so". your name ends with the middle one, but you wrote it with the last one: it says "kiraaso", ought to be "kiraan".
  • 

    Kiran Jonnalagadda — Mar 9, 2004 3:50:05 PM — # ↩

    Corrected, thanks!
• 

  shradha — Mar 9, 2004 4:14:23 PM — # ↩

  hey COOL!!:D
  • 

    Kiran Jonnalagadda — Mar 9, 2004 4:20:39 PM — # ↩

    Where are you now ma’am? Bangalore, Delhi, or Japan?
    • 

      shradha — Mar 10, 2004 1:45:46 AM — # ↩

      its my last day in b'lore:(
      will be in orissa tom.
• 

  Kiran Jonnalagadda — Mar 9, 2004 4:43:33 PM — # ↩

  Here's another experiment. The following like should read “Sunil” in Kannada. Does your browser do it correctly? (Safari doesn’t, GtkHTML in Evolution does.)
  
  ಸುನೀಲ್
  • 

    mannu — Mar 9, 2004 5:30:03 PM — # ↩

    I'm on Windows. IE gives square boxes and Firefox gives question marks--both because the font doesn't exist on my system.
    
    The (hex) values for the Unicode characters:
    
    \x0cb8 \x0cc1 \x0ca8 \x0cc0 \x0cb2
    • 

      Kiran Jonnalagadda — Mar 9, 2004 5:34:40 PM — # ↩

      Sampige font for Kannada available here: http://brahmi.sourceforge.net/downloads.html
      • 

        mannu — Mar 9, 2004 6:59:22 PM — # ↩

        Ah, this is a charm. Looks good!
  • 

    ravi — Mar 9, 2004 11:09:35 PM — # ↩

    MS IE 6.0 SP1 on WinXP 5.1 SP1 renders it fine ...
  • 

    madhav — Mar 10, 2004 1:37:23 AM — # ↩

    I can see the Kannada script. I've no idea if that's Sunil or not, though!