ID for language-invariant strings

[Peter Constable]

> From: Tracey, Niall [mailto:niall.tracey at logica.com]
> Sent: Thursday, March 20, 2008 4:01 AM


> As this value is intended for non-linguistic use, I don't see a problem
> with ZXX.

I'm not countering the intended-use argument, which others also made. I'm discounting the content-is-arbitrary argument, which in the case of the particular application scenario I'm dealing with is, I think, invalid.


> It is entirely impossible when defining an arbitrary data field...

I'm not dealing with completely arbitrary data fields.



> We don't need to know anything about the author's language to use the
> string as intended. I'd argue that the data as you have described is
> intrinsically arbitrary -- any linguistic content is extrinsic to the
> system. Hence from a system's point of view, the data is non-linguistic.

For some reason I still find that a hurdle to get over. I guess it's because the names that will likely be used are copies of the name intended for display in some language. But broadening "zxx" to "not applicable" would lessen that hurdle.



Peter