As a Sanskrit scholar and busy author, I use the IAST and Harvard-Kyoto m17n keyboard maps many times every day. I don't have terribly strong views about what goes in to the MASTER VERSION and what doesn't. There aren't any ISO standards for IAST or H-K, so it's really up to the community of users to work out what's most useful for us.
I would emphasize, though, that it's really easy to lash up one's own locally customized version of the sa-* files. And I think this is a good solution for anyone writing something in a specialist field (glyphically speaking). I've done that routinely in the past, and it's really no problem. Copy the file with a new name, make your preferred additions, and get on with your writing.
The m17n discussion should really take account of the difference between what one particular individual wants and what gets standardized and included in the international m17n library distribution, and goes out to thousands of people. We should be thinking about core and periphery. The main distribution should contain the core. The core distribution can't possibly cater for all possible future individual needs. It should make as good a job as possible of typing IAST and Devanagari, and be less concerned with mathematical and other symbols.
We don't want to fill up the standard files with stuff that only one or two people actually want. On the other hand, if some additions don't over-burden the keyboard (making typos more likely), then there's probably no harm. But really, I would suggest that local customization is the best answer for such cases.
Incidentally, for anyone writing on Sanskrit grammar, it's not only arrows that are necessary, but other symbols like the root or surd sign, √, superscripts, subscripts, etc. Especially if one is writing about Pāṇini. And we don't have upadhmanīya and jihvamūlīya in the m17n H-K map yet, do we? As real parts of the Devanagari syllabary, they should take precedence over symbols, I think.
But then with truly technical writing, one should really begin thinking about a more powerful document processing system, like LaTeX, which offers a huge range of symbols using simple codes, like \surd, without needing to poke about in a font to find them as glyphs.
By the way, I've just updated my Ubuntu distribution to 13.10 beta, and I notice that the m17n distribution doesn't have the updates to sa-harvard-kyoto.mim that Satish, Daniel and I worked out back in August, that added ऽ, ।, ॥, and ॐ. I don't think these are controversial additions; they are necessary for normal writing in Devanagari, and they should stay in the mim file.
... <checks the
Git repository> Ah, no I see, these changes *are* still in the git repo. I guess they just haven't filtered out to the debian distribution yet, is all.
Best,
Dominik
--
Dr Dominik Wujastyk
University of Vienna