Supporting en@shaw, the Shavian alphabet

I have been asked whether it would be possible to transliterate the English version of Ubuntu into the Shavian alphabet (the locale being "en@shaw", presumably, since "shaw" is the ISO 15924 code for the alphabet). I have some tools which will transliterate .po files automatically, using a pronunciation dictionary, as in this screenshot: http://marnanel.org/shavian/epiphany1.png . My question is whether Rosetta is an appropriate venue for a transliteration which will admittedly have only a few dozen users and which can mostly be programmatically derived from the original text (although in a few cases it will need hand-tweaking, as with the case of homophones) or whether it would be better just to build packages in my PPA. Is there a limit on how trivial a locale may be?

Some background:
The Shavian alphabet is an alternative alphabet to English designed in the 1960s. Essentially it is just another alphabet for writing English, but more regularly and in fewer characters than the standard Latin alphabet. For example, this paragraph in Shavian looks like this:

𐑞 ·𐑖𐑱𐑝𐑾𐑯 𐑨𐑤𐑓𐑩𐑚𐑧𐑑 𐑦𐑟 𐑨𐑯 𐑪𐑤𐑑𐑻𐑯𐑩𐑑𐑦𐑝 𐑨𐑤𐑓𐑩𐑚𐑧𐑑 𐑑 𐑦𐑙𐑜𐑤𐑦𐑖 𐑛𐑦𐑟𐑲𐑯𐑛 𐑦𐑯 𐑞 1960𐑟. 𐑦𐑕𐑧𐑯𐑖𐑩𐑤𐑰 𐑦𐑑 𐑦𐑟 𐑡𐑩𐑕𐑑 𐑩𐑯𐑩𐑞𐑻 𐑨𐑤𐑓𐑩𐑚𐑧𐑑 𐑓𐑹 𐑮𐑲𐑑𐑦𐑙 𐑦𐑙𐑜𐑤𐑦𐑖, 𐑚𐑩𐑑 𐑥𐑹 𐑮𐑧𐑜𐑘𐑩𐑤𐑻 𐑞𐑨𐑯 𐑞 𐑕𐑑𐑨𐑯𐑛𐑻𐑛 𐑤𐑨𐑑𐑩𐑯 𐑨𐑤𐑓𐑩𐑚𐑧𐑑. 𐑓𐑹 𐑦𐑜𐑟𐑨𐑥𐑐𐑩𐑤, 𐑞𐑦𐑕 𐑐𐑨𐑮𐑩𐑜𐑮𐑨𐑓 𐑦𐑯 ·𐑖𐑱𐑝𐑾𐑯 𐑤𐑫𐑒𐑕 𐑤𐑲𐑒 𐑞𐑦𐑕.

You may need a font such as http://marnanel.org/shavian/fonts/AndroclesUnicode.ttf to view this sample. Alternatively, the screenshot given above shows the alphabet well.