The problem is is that language isn't 100% predictable.
If it were, people would've used your solution before to translate even between languages that aren't related, for example English and Guarani.
The thing here is that when written in Chinese characters, wo3 and ngo4 use the same character. The issue isn't meant to be the different sounds, it's the different grammars.
A potential problem with manual tagging is that Mandarin monolinguals would not know where to put these in their articles, and to have a squad of people go around and tag all their text for them is rediculous. Also, articles would look rediculous when edited because there is the same issue for other Sinitic languages (although at the moment, it's my opinion they don't really need separate Wikipedias except perhaps Hakka and probably Dungan), and there would be spare tags hanging out all over the place.
When there is a simple difference of terminology and perhaps spelling (ie International vs American English), your idea is very practical, but when there are huge syntactical differences as there are between Mandarin and Cantonese, it's not practical at all unfortunately.
Mark
On Sat, 25 Dec 2004 18:56:29 -0500, Stirling Newberry stirling.newberry@xigenics.net wrote:
On Dec 25, 2004, at 6:47 PM, Mark Williamson wrote:
I agree that this is a good idea, and in certain cases it could be done automatically for all occurances of a word (that's how it works on zh.wikipedia right now; you can do exclusions though if you want).
For example:
K'öi4 pei3 sa:m1-pun3 sü1 ngo4. = Ta1 gei3 wo3 san1ben3 shu1. ("He gave me three books.")
Which could be handled by bracketing: Ta1 gei3 [[dialect: wo3 | gd: ]] san1ben3 shu1 [[dialect: | gd: ngo4]].
Which would probably be macro'd since it is common. {{wo3}} and {{ngo4}} would do it as a pair.