[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: strings draft

    > From: tb@xxxxxxxxxx (Thomas Bushnell, BSG)

    > Tom Lord <lord@xxxxxxx> writes:

    > > No.  The c.l.s. idea is to make mandatory the case mappings of the
    > > standard procedures for the portable character set only.
    > > Specifically, 'a..z' must map to 'A..Z' (and vice versa) and that's
    > > all that's required.

    > This gets the Turkish I Problem wrong then.

Not quite, no.

    > In Turkish, i-with-dot capitalizes to I-with-dot.  If you map i->I,
    > then you get this wrong, for Turkish.

Did you actually read the c.l.s. post?

    > >     > I do not object to a Scheme that chooses to do this.  But your draft
    > >     > seems to exclude a Scheme that does *best* Unicode practice, which is
    > >     > not to shove any details under the rug.

    > > No, the draft says that for _linguistic_ processing, the standard
    > > procedures for case mapping are unsuitable.   We need a SRFI to
    > > introduce _new_ procedures implementing "Unicode best practice" for
    > > linguistic processing.

    > Sure.  But since we can easily make a draft that doesn't have this
    > problem at all (by dropping case-mapping as a char-by-char thing
    > entirely) we can fix it.

Not really.  Whether "char at a time" or "string at a time", `list' is
not the (linguistically) ci equivalent of `LIST' in -TR locales.