No fixed set of symbols can express the infinite space of ideas. To express an
idea not represented by a known symbol, you must eventually use a compound
symbol.
On paper (or electronic paper), space (or lack of space) is the typical way
humans communicate the idea of "compound" vs "distinct". Note the letters in
this email - the ones grouped together form words, and whitespace allows us to
separate words.
Chinese may have tens of thousands of characters, but there are more ideas in
the world than that. At some point, even in Chinese (or APL), one will have to
group symbols together to express an idea. Changing the spacing will change
the grouping, whether in the eyes of a human or a computer program (hence
newline-delimited sentences even in APL).
I think using non-ASCII symbols in J is a terrible idea. They're hard to type,
they're hard to communicate (transfer over e.g. email), they're unreliable in
rendering, and they are (ab initio) no more suggestive than ASCII. All
"suggestivity" of epsilon or iota or whatever is bred from familiarity, and one
can become just as familiar with ASCII-based symbols to render them suggestive.
For example, now when I see # I can't help but think of a little sieve,
filtering ore from dross.
Moving away from a exotic symbolset to the vanilla, ubiquitous, and reliable
ASCII standard was a major motivating force in the creation of J. That force
has not diminished. When Unicode (including all the weirdo characters we're
proposing to use here) is just as vanilla, ubiquitous, reliable, and standard
as ASCII, then it will be just as good an idea to use Unicode as ASCII.
But we're not there yet.
Having said that, I would be fine with changing J's underpinnings to be strict
UTF8, so that it would permit Unicode identifiers, comments, literals, etc.
This would move J closer to the emerging standards of the internet, but more
importantly, would give each individual user the _choice_ of how he'd like to
write his J scripts. If someone wanted to use iota and epsilon, then fine:
{iota} =: i.
{epsilon} =: e.
Done. And it's still compatible with all existing J processors (including the
ones residing in the J community's mental machinery). Plus we wouldn't get
into an endless, fruitless, enervating debate about which exotic symbolset to
use. We could each use what we prefer, but fundamentally, we'd be using the
only correct symbolset: ASCII.
-Dan
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Roger Hui
Sent: Friday, April 12, 2013 11:10 AM
To: Chat Forum
Subject: Re: [Jchat] [Jprogramming] J Symbols
What I said was:
> FYI, the differently spaced versions of the Chinese sentence do not
> quite
have the same meaning. Certainly not the same effect.
Put in enough whitespace, esp. different amounts of whitespace, and one effect
is that it makes it look like you were drunk when you wrote the text. Put in
enough whitespace, and it has the meaning and effect of punctuation (such as
comma or parens). I know of at least one example learned in grade school where
punctuation changed the meaning of a sentence to its exact opposite.
I haven't thought much about it, but I believe the same thing works in English,
where you put the whitespace between words rather than letters.
On Fri, Apr 12, 2013 at 5:05 AM, Raul Miller <[email protected]> wrote:
> I do not know chinese, but I'll take Roger's word that the whitespace
> still has some significance there.
>
> [Replying in chat, also]
>
> Thanks,
>
> --
> Raul
>
> On Thu, Apr 11, 2013 at 10:17 PM, Skip Cave <[email protected]>
> wrote:
> > Raul,
> >
> > Who said that ASCII English was ideal?
> >
> > Here's the sentence "I do not see why this should be an ideal" in
> Mandarin.
> >
> > 我不明白為什麼這應該是一個理想的
> >
> > And here's the same sentence again in Mandarin, with different
> > spacing,
> but
> > with the same meaning. .
> >
> > 我不明 白為 什 麼這 應該 是 一 個理想的
> >
> > And here's the same sentence again in Mandarin, with even different
> > spacing, yet with the same meaning.
> >
> > 我 不明白 為什 麼這 應該 是 一個 理 想 的
> >
> > So true single-glyph symbolic languages are space-independent, and
> that's a
> > GOOD thing for writing. Your example shows why languages that use
> > multi-glyph words or symbols like English and J and thus are NOT
> > space independent, are a BAD thing for handwriting.
> >
> > When you write your sentence on the board in English, you have to be
> > careful to clearly indicate where the spaces are, or you get what
> > you showed in your first example. With a single-glyph languager like
> > Chinese, the spaces don't matter much.
> >
> > Skip
> >
> > On Thu, Apr 11, 2013 at 7:12 AM, Raul Miller <[email protected]>
> wrote:
> >
> >> On Thu, Apr 11, 2013 at 6:02 AM, Skip Cave
> >> <[email protected]>
> >> wrote:
> >> > Ideally a written version of the language should be space-independent.
> >>
> >> Id ono ts eew hyt hi ssh oul db ea nid e al.
> >>
> >> I do not see why this should be an ideal.
> >>
> >> --
> >> Raul
> >> -------------------------------------------------------------------
> >> --- For information about J forums see
> >> http://www.jsoftware.com/forums.htm
> >>
> >
> >
> >
> > --
> > Skip Cave
> > Cave Consulting LLC
> > Phone: 214-460-4861
> > --------------------------------------------------------------------
> > -- For information about J forums see
> > http://www.jsoftware.com/forums.htm
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
