MichKa, I think you are missing the point. You say you do not endode different sounds. sRi is in a way doing that. Stricting what Tamil should do by designating it to a particular sound of the symbol by dictating the combinations. other variations are not allowed to this symbol by UC, but it is the way it is functioning in use.
The stricted variation that UC only permits is the bad version of the symbol. Why restict it with one? sRi has many different combinations and it is context sensitive and also usage sensitive. UC picks the bad form. Now, for science,for understanding purpose, I simply wrote, if we want to build a proper humanoid we need to build the mechanical (or biological) speech organs, not create a limited database of sounds/phonemes. Is this something difficult to understand? if not, is it something difficult to understand that Tamil alphabet represents/names the human organs and other alphabets represent sounds or a kind of mix,in some cases. Is it wrong to catergorically stating that we need to prserve this Places of Articulation for the sake of the world, not just for tamil alone? Now, for politics, is it right for UC to stampead the presently powerless? Sinnathurai --- On Fri, 26/11/10, Michael S. Kaplan <[email protected]> wrote: From: Michael S. Kaplan <[email protected]> Subject: [indic] Re: Revisit Tamil sRi definition in Unicode. To: [email protected] Date: Friday, 26 November, 2010, 5:47 John Hudson said: > This is all very interesting as a > cultural phenomenon, but nothing > to do with computer encoding. More than that, it ignores some basics facts about encodings. Like the fact that a language like English has multiple different pronunciations for several different letters yet Unicode does not encode separate characters for them. Or the related fact that languages like Spanish have different pronunciations for some of the same letters as English and yet these "other letters" are not encoded again. There are countless examples of such things, things that Unicode does not do even though they would in theory make NLP easier -- because they would make many other things harder (and there are many petabytes of data that do not have such "features" meaning that the features would not solve the problem. Sinnathurai Srivas has been espousing this very same argument for at least a decade and no amount of explanation of the facts sways him. He often refers to "science" but it is that unique form of science practiced by some of those who are not scientists that one hears about that ignores facts and evidence and truth -- that unoique form of science one can see whose only purpose is to keep yelling the same statement over and over again in the hope that everyone will accept it as truth. Some refer to is as pseudoscience. I admire the persistence, but I do not admire the refusal to understand that even agreeing with him would not change anything because what he wants is out of scope for Unicode. I find myself weary of it, to be honest.... MichKa
