Re: Beyond ascii

Ray Blaak wrote:

> Note that I was not at all saying that we should force the use of
> English. Only that the character set not be full Unicode. One can still write
> in their own language even in the face of restrictions. My Greek wife does it
> all the time when working with shitty email tools.

Can you clearify a bit? Suppose I want to use Japanese kanji in
identifiers (which is actually common in those "enterprise"
applications), you suggest to create a localized version of the
language that allows it (which is the case before unicode-era)? Or do
you suggest to abandon this habit, which I guess has been adopted
since introduction of the localized office computers, sometime in
early 80's?

I agree that there could be problems when larger character set is
allowed to write the code. However, in the limited subset, it has
already been in practice and doesn't look like a disaster.

For example, even before the Unicode-era, Japanese had the sets of two
identically looking characters--for instance, see U+FF01 to U+FF5E.
(These were in the original JIS character set, and because of
round-trip compatibility they are included in Unicode).

However, programmers just knew these traps and avoided using them.
The problem wasn't surfaced until the general people started to use
computers, for those people used these characters in filenames and
writing down email addresses and so on.

> And the restriction need not be strict-ASCII I suppose. Perhaps there are more
> "code friendly" subsets to use for identifiers.

Yup. The question is, should the language define these subsets, or
can the language trust programmers to use it wisely? Certainly there
are programming languages whose syntax are so ad-hoc and irregular so
that you would expect it to cause a mess, and yet lots of programmers
using it for useful applications.

.

Loading