Re: Vallian (was: How to minimize "words")

Hi Philip

On 24/02/07, Philip Newton <philip.newton@...> wrote:
>
> On 2/24/07, Jeff Rollin <jeff.rollin@...> wrote:
> > My German is too rusty to be able to tell you whether *"The _I gave
> [him] a
> > hat_ man speaks English" would be possible instead of
> >
> > "Der Mann, _dem ich einen Hut gegeben habe_, spricht Englisch"/"The man
> _who
> > I gave a hat to/to who(m) I gave a hat_ speaks English",
>
> I don't think that's possible -- I think the construction you
> mentioned only allows you to refer to the subject of the relative
> clause, not an indirect object or similar.
>
> When I try to make a preposed relative clause out of it, the closest I
> get is "Der von mir einen Hut erhalten habende Mann spricht Englisch",
> which is, of course, not the same in structure, since the main verb
> changed from "give" to "receive".

Ah, so it is possible with a passive; ok.

> Well,  professional typesetter might be able to type a double-acute (as in
> > Hungarian) over "a", but as Hungarian only uses them over "o" and "u"
> (as it
> > lacks German and Finnish's a-umlaut), Unicode/HTML (possibly) and
> OpenOffice
> > (definitely) lack "a-double-acute".
>
> Unicode lets you produce an a-double-acute by using combining
> diacritics: U+0061 U+030B. (Here's what it looks like for you:
> ...a̋...)

Ah-ha. I didn't know that.

Whether that'll look acceptable depends on the fonts you have
> available and on the rendering engine employed.
>
> > (This may be going off-topic, but IMHO Unicode really should have
> included
> > all possible combinations of letter+accent,
>
> Eh?
>
> As far as I know, Unicode's philosophy is not to encode combinations
> of letter + diacritic (in general) unless such precomposed characters
> already exist in legacy encodings, so your suggestion goes against the
> underlying philosophy.
>
> Not to mention that IMO your suggestion is completely unfeasible.
> There are hundreds, if not thousands, of "letters" in Unicode, and
> dozens of combining diacritics in the U+03xx Combining Diacritical
> Marks block alone. Now if you want to be able to apply more than one
> diacritic to a letter, you'd have, say, 80! combinations of
> diacritics, times thousands of base characters -- how many bits did
> you want to use per character, again? (NB 80! is about 7x10^118, or on
> the order of 2^395.)

Ah, no. I meant only one accent per character, but all possible accents -
for example, to get a-double acute and schwa-grave. But since I now know
that you can combine arbitrary characters and arbitrary accents, as you
point out above, I see that's not necessary.

Of course there are writing systems (such as Vietnamese and perhaps some
phone{t.m}ic transcriptions) that need more than one accent per character

Jeff

Reply