Devanagari & Transliteration

OS X and all modern computers use the international Unicode standard, which maps Devanagari and Latin to totally different codepoint ranges. Switching scripts/languages is done by changing keyboard layouts, rather than by changing fonts as was common in older technology. The Devanagari Qwerty layout provided with OS X lets you produce Devanagari by typing somewhat phonetically. I have not seen a Dvorak version of this, but it could be done using a keyboard layout tool like Ukelele:

http://scripts.sil.org/cms/scripts/page.php?siteid=nrsi&itemid=ukelele

It might also be possible to use this tool to allow a more transcription-like input, but more likely this would require an "input method," like that used for Chinese (and I think Tamil has one too).

If you do find a font of the type you are describing, I'm not sure it would work well in OS X, which is designed for Unicode fonts. But it would be interesting to see such a thing.

These notes may be helpful:

http://homepage.mac.com/thgewecke/TypingDevanagari.html

http://discussions.apple.com/message.jspa?messageID=1200527#1200527

I'm not really sure I understand what you are talking about, but it seems like a really bad idea – on several levels.

So if you press the "kha" key – whatever that is – it would make either "kha" or ख depending on if you are in Devanagari mode or in English mode? But then what happens when you press the "i" key so now it will make "i" or ि.

But the problem is it now needs to read read "khi" or "खि" where the "i" is visually in front of the "kha".

The next problem is that there are many ways to represent Devanagari using the Roman alphabet. Most of them tend to loose something in the translation. For example साड़ी is usually Romanized "sari" but it should really probably be "saaRDii" or some other way that I don't even know.

And what about all those other sounds that we don't have – retroflex and aspirated T and D. And from your user name I'm guessing you're interested in Sanskrit words which are going to be even more difficult since it has nasalized "n" and other things I can't even figure out!

I've always found that the only way to actually spell Hindi (and I'm sure it is the same in Sanskrit) and be sure it is the correct work is to spell it Devanagari. In the end probably easier than trying to get whatever it is that you are talking about actually working.

Other than that, the Devanagari that comes with OSX is pretty darn good. If you are a native English speaker and know the QWERTY keyboard it is pretty easy to learn where everything is. Sorry I don't have any experience with the Dvorak or know of a Devanagari-Dvorak unicode layout. But here is for the QWERTY.

Basic sounds are on the closest equivalent key, aspirated are the same with shift, retroflex are option and the key and aspirated retroflex are shift option and the key.

Vowels are most usually in their matra form, so the short vowel is regular key, long vowel is shift key. Full form of the short vowel is option key and you can guess what shift option does.

Conjuncts are made by typing the first part then the f (halant) and the second part.

Basic nasalization is "m" after the vowel and option-m if you need the moon.

In the end I don't see that it is going to be possible to just change the font and have beautiful Devanagari correctly transliterated to "correct" Roman alphabet.

If you are hoping to go the other way, the rules to get it right would probably be as hard as learning Devanagari in the first place.

There are some sites on the web that I have seen that if you type in the Roman letters – in a very tightly controlled way following the INSCRIPT transliteration rules – you can get a program to generate the correct Devanagari. Not easy to use.

Thanks for the interesting link, Rothrock. Here is one good site I found with some Unicode Sanskrit in both scripts:

http://www.sacred-texts.com/hin/rvsan/rv01001.htm

Most of the other stuff I have seen is in one of several legacy encodings/fonts that probably just hack Windows-1252 , so that downloading special fonts is required and multi-script plain text is impossible.

I think our forum can handle the Unicode version (beginning of Rig Veda):

अग्निमीळे पुरोहितं यज्ञस्य देवं रत्वीजम |
होतारं रत्नधातमम ||

aghnimīḷe purohitaṃ yajñasya devaṃ ṛtvījam |
hotāraṃ ratnadhātamam ||

Interesting, but the transcription doesn't look correct to me. For example: अग्निनीळे should probably be agnimīḷe – the ga is not aspirated. But oddly it seems that ग is consistently translated to "gh" so maybe there is something different about Sanskrit?

And I looked at the next page, where I found अरंक्र्ताः which is transliterated araṃkṛtāḥ, but it seems that ṛ is for ृ not for र and elsewhere I do see it as just r without the dot. Again, maybe the transcription rules for sanskrit are different than hindi…

Unicode input and text is really the way to go. I'm pretty good with it – except some of the more exotic sounds that aren't in modern hindi – so if you need some specific help let me know.

Interesting, but the transcription doesn't look
correct to me.

I checked other sources and confirmed that there seem to be errors in that one. I put the hopefully correct stuff and some other transliterations on a test page at:

http://homepage.mac.com/thgewecke/sanskrittest.html

Aha, I see. I agree that there certainly are uses for both – when done correctly and accurately. Most folks who ask these things are not very precise and expect a "magic" convertor where a work can be Romanized any old way and will automagically be converted to the correct word. (I am still going to maintain that the effort and time to learn how to correctly and consistently spell foreign words in Roman script is not inconsequential. And that, to my mind at least, most folks would be better served by learning a system that was designed to represent that language.:) )

I'm still suspicious of any "font" producer who claims to be creating something like this. Because in the end, the problem isn't a font. As you are learning with your quest for ITRANS – it is a program that can convert one system of notation into another. I am very dubious that such a complex task can be handled by a font. Perhaps this font producer is just simplifying what they are doing to make it understandable to the layman?

Typying Devanagari or Transliteration is most
important to me if done in Dvorak.

I think someone did a Dvorak Devanagari layout a couple years ago. I will try to find it for you. For IAST transliteration in Dvorak, you could use Ukelele to modify the US Extended layout.

I found one font that
allows me to type in Dvorak while in Devanagari and
it is called Avanti.

That would presumably mean working in non-Unicode format, XDVNG or something.

However a Devanagari font producer told me he is in
process of creating one.
I am asking for is not only possible but highly
useful. If I am typing a text that has Devanagari and
requires Transliteration for the general public then
I only need to type in one and have the other.

This would indeed be useful if possible, but I am skeptical too. In Unicode, the font would have to be smart enough to create conjuncts like ग्न (which take 3 characters) when it comes across the Latin sequence gn and vice versa. In XDVNG I suppose the conjunct has one codepoint which could be mapped to gn, but the reverse seems problematical, and of course one is not in Unicode so electronic processing/distribution becomes more difficult.

IAST/Devanagari translation would seem to require more programming than a font allows. If someone could accomplish it with a font alone, that would be ingenious.

Typying Devanagari or Transliteration is most
important to me if done in Dvorak.

PS You can get the Devanagari-Dvorak layout here:

http://idisk.mac.com/thgewecke-Public?view=web

I've never tried it. If it needs modifications, you could probably do them with Ukelele.

My question about this was elsewhere I've seen ṛ be
the transliteration for "ri". So how can it be for
both ri and ra?

The problem is indeed in that particular source. My other source has अरंकृताः araṃkṛtāḥ.

Thanks. I tried to install Ukelele_1.7.dmg but it
wouldn't open. Maybe I should download it again.

Yes, try that. It works fine for me.

Did you try the already-made Dvorak Devanagari layout mentioned in an earlier posting?