On Fri, Sep 28, 2012 at 9:32 AM, Kenichi Handa
<address@hidden> wrote:
I don't understand the logic. I agree that "nk" is easier
to type than "^N". But, you wrote "ङ्क् almost never appear".
If so, why do we have to add more keys for that?
Sorry - I miswrote (I have long had those transforms in my machine which caused ङ्क् to appear inplace of न्क्)! न्क् almost never appears, while ङ्क् ञ्च् ण्ट् appear quite frequently.
> I add two desiderata:
> a] It would be great if one could specify a rule which says: For consonant
> characters followed by a space, new-line or daNDa-s, omit ्. For example,
> akSha would be अक्ष, while akSh would also be अक्ष - it would be currently
> rendered a,s अक्ष् - this is good for samskrit, but inconvenient for hindI,
> where words are oft pronounced without the terminal a sound. This would
> render the map useless for samskrit typing work (which I do a lot of), but
> that can be solved by retaining the current map as sa-itrans.mim.
It's possible. But, are space, newline, dondas enough? How
about TAB, !, ", ', and etc?
Oh yes - the rule should apply whenever a consonant (the pattern [क-हक़-य़]) is followed by a non-letter character.
On the same note, it would be wonderful if we could add to the daNDa-rules as follows:
Single '.' followed by a whitespace character (we don't need to worry about " ' etc..) should be rendered as । . (Just typing one dot rather than two is much more convenient, as in English. We would retain .. to । rule for the case where the user needs to insert a daNDa in a position before a non-whitespace character.)
If we add the a] feature, we surely need sa-itrans which
doesn't have that feature. By the way, I don't know what
"vedic svara marks" is. Do you want to add it to sa-itrans
or hi-itrans?
I would add it to sa-itrans. (Hindi users will almost never use it.)
Two basic svara marks could be served by the following maps.
.' to ॑,
.- to ॒
--