Re: [Bug-apl] Request for enhancement

[Juergen Sauermann]

Hi,

there are a few problems related to underlined and overlined characters and numbers.

In Unicode, these characters are represented as two characters. Right now, one APL
character means one Unicode. If we would replace one APL characters by two Unicodes
then quadratic character matrices like those used by ⎕FX and ⎕CR would become
non-quadratic if they contain underlined characters. That would probably break a number
of existing programs.

A work-around could be to put the underlined characters into ⎕AV. But that would not work
with the current 256-byte ⎕AV and we would have to increase the size of ⎕AV. That, however,
would probably break even more existing programs.

Finally, as I observed yesterday, it looks like different terminals (or display channels)
seem to handle the ⎕UCS 773 differently. For example, my xterm overlines the previous
character. If I then copy-and-paste the overlined character from xterm to my email client
 (thunderbird) then it looks still OK. However, after sending the email and reading it
 on the bug-apl web-page it looks wrong. On that web-page it was not the previous but the
next character that was overlined. Building upon such an unreliable Unicode feature
is therefore asking for trouble.

Putting all that together I would say that the problems and incompatibilities that
overlined or underlined characters in GNU APL would create are far bigger than
their benefits.

/// Jürgen

On 07/23/2015 08:41 PM, Kacper Gutowski wrote:

On Thu, Jul 23, 2015 at 6:55 PM, Juergen Sauermann
<address@hidden> wrote:

GNU APL supports the entire Unicode character set to the extent that your
platform supports (= is able to display) it.

On Thu, Jul 23, 2015 at 5:59 PM, Fred Weigel <address@hidden> wrote:

The current definition of letter includes (...)
0.3(overbar) could be expanded to 0.33.. (as many as needed to fill
the precision).

If I understand correctly, the request was about changes in syntax,
not just about being able to use these characters in code in character
literals or to manipulate them.

Overline E2 80 BE (UTF-8), complementary to _ (underline)

I would like to note that symbol ‘complementary to’ _ underline, in
APL world is the high minus which is rendered using ¯ (U+00AF macron).
That's why it's allowed in names.  There is no specific high minus
character in Unicode, nor I see any reason to believe that U+203E
overline would be a better fit for it.  But perhaps there is some
merit in adding it as acceptable alternative just like ∊ (U+220A) and
∈ (U+2208) which are both accepted by GNU APL for epsilon.

Supporting underlined alphabetic characters in names by allowing
combining low line in them would be interesting.  Unicode has no other
way to represent them, and lower-case Latin characters are not the
same after all.  But I'm not sure whether it's a good idea.

As for 0.3̅, it's faster to write ÷3 and in most fonts I know it's
also much more readable.

-k