[Tugindia] Comparison of input methods

Amitabh Trehan amitabhtrehan at softhome.net
Tue Apr 22 15:07:13 CEST 2003

Thanks for carrying out the work. My two bit -->

1) I would just like to venture that your conclusion may be a bit hasty
- needs more work. As Raj Singh has written in his reply to this mail -
the characters you are talking about as missing in devnag are ligatures
i.e. combined characters, and there is a big range of them in devnag.
So, by personal experience, if one knows a bit of phonetics, one can get
any of these characters in devnag.Ileap uses dny for gn( j~na as in
j~nAn in devnag)- this makes no phonetic sense.
2)We need to consider the differences between TeX typesetting (character
based) and ILeap (WYSIWYG). devnag seems to be using a technique of
pre-tagging i.e. . before t to make it .t, others post-tagging Nj. Which
is more linearly parseable? 
 3)I suppose when you talk of ILeap you are talking about the phonetic
English keyboard (they recommend inscript). I suppose one question to be
asked is can the other input methods handle the range of ligatures that
devnag handles?  
4) devnag uses character-vowel complex i.e ka for a full k. I suppose
the same holds true for TeleguTeX and ileap. That means there are many
implicit a's to be added to the characters in the chart while inputting.
5)I suppose the test is how readable is the english text file to a human
6) Are there not ways in which we can handle all the Indian languages
together on the basis of their common phonetics?

As far as I know, ITRANS builds upon devnag, what are the
standardisations done? 

Interestingly, on ftp.dante.de , in the TeleguTeX directory, the file
manual.dvi is actually the devnag manual!!!


On Mon, 2003-04-21 at 22:11, Duvvuri Venu Gopal wrote:
> A comparison of input method used by different programmers to 
> input Indian symbols using English keyboard.

>  From the above it is evident that Mukkavilli Lakshmi and C-DAC 
> followed a systematic pattern, except in a few instances where 
> they differed. Whereas Devnag of Velthuis and the other two 
> systems vary drastically in the T-varg. Velthuis also 
> unnecessarily complicated by introducing two different types of 
> inputs for the letters - A, I, U, ri, kh, gh, ch, jh, th, dh, th, 
> dh, ph, bh. There is no mention about the characters ksha and  gna 
> (as in Jnani) in Velthuis's Devnag.
> It appears from the ITRANS 5.3 help pages this sort of 
> standardization was already done by Avinash Chopde for ITRANS. The 
> problem is I could install it properly and there is no help 
> available.
> Conclusion:
> All the developers of Indic TeX may follow the system developed by 
> either Mukkavilli Lakshmi or C-DAC, as the roman input is more 
> readable than Velthuis's format. In that case it is necessary to 
> change the Velthuis's Devnag preprocesser also for 
> Sanskrit/Marathi/Hindi appropriately.

More information about the Tugindia mailing list