numeri

numeri

[Eli Zaretskii]

Observe:

    M-: (get-char-code-property ?4 'numeric-value) RET
      => \4

What does the backslash signify in this case?

Re: numeri

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

> Observe:
>
>     M-: (get-char-code-property ?4 'numeric-value) RET
>       => \4
>
> What does the backslash signify in this case?

That it's a symbol, not a number.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: numeri

[Eli Zaretskii]

> From: Andreas Schwab <schwab@suse.de>
> Cc: handa@m17n.org, emacs-devel@gnu.org
> Date: Sat, 29 Nov 2008 17:04:27 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Observe:
> >
> >     M-: (get-char-code-property ?4 'numeric-value) RET
> >       => \4
> >
> > What does the backslash signify in this case?
> 
> That it's a symbol, not a number.

Right, but why backslash?  Why not something like `digit-four'?

Re: numeri

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

> Right, but why backslash?  Why not something like `digit-four'?

I don't understand the question.  What is digit-four in this context?

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: numeri

[Eli Zaretskii]

> From: Andreas Schwab <schwab@suse.de>
> Cc: handa@m17n.org, emacs-devel@gnu.org
> Date: Sat, 29 Nov 2008 18:39:51 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Right, but why backslash?  Why not something like `digit-four'?
> 
> I don't understand the question.  What is digit-four in this context?

I meant to ask why would it be worse idea to have

  (get-char-code-property ?4 'numeric-value)

return a symbol `digit-four' or `digit-4'.

But maybe the answer is because `\4' is ready for the Lisp reader,
while the alternatives are not.

Re: numeri

[Eli Zaretskii]

> Date: Sat, 29 Nov 2008 20:05:18 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org, handa@m17n.org
> 
> > From: Andreas Schwab <schwab@suse.de>
> > Cc: handa@m17n.org, emacs-devel@gnu.org
> > Date: Sat, 29 Nov 2008 18:39:51 +0100
> > 
> > Eli Zaretskii <eliz@gnu.org> writes:
> > 
> > > Right, but why backslash?  Why not something like `digit-four'?
> > 
> > I don't understand the question.  What is digit-four in this context?
> 
> I meant to ask why would it be worse idea to have
> 
>   (get-char-code-property ?4 'numeric-value)
> 
> return a symbol `digit-four' or `digit-4'.
> 
> But maybe the answer is because `\4' is ready for the Lisp reader,
> while the alternatives are not.

Actually, why not make the above return the number 4?  The Unicode
Data File Format says that this property's value is ``an integer or
rational number''.  We could even use 0.2 for characters such as
"VULGAR FRACTION ONE FIFTH".  Isn't that better than `1/5'?

Re: numeri

[David De La Harpe Golden]

Eli Zaretskii wrote:


> 
> Actually, why not make the above return the number 4?  The Unicode
> Data File Format says that this property's value is ``an integer or
> rational number''.  We could even use 0.2 for characters such as
> "VULGAR FRACTION ONE FIFTH".  Isn't that better than `1/5'?
> 
> 

Bear in mind that 1/5 (0.2) isn't actually exactly representable by 
binary floating point, bit like the way 1/3 (0.333(3 recurring)) can't 
be written properly in decimal, so try in emacs lisp:

(format "%.20f" 0.2)

I'd tend to hold off on allowing the vulgar fractions to map to numbers 
without the programmer to do a conversion step and "own" the associated 
potential information loss - at least until there is no loss:  Emacs 
Lisp could one day get CL-like exact rational numbers I guess (though 
bignums would probably be a higher priority...)

Re: numeri

[Eli Zaretskii]

> Date: Sat, 29 Nov 2008 19:01:22 +0000
> From: David De La Harpe Golden <david@harpegolden.net>
> Cc: handa@m17n.org, emacs-devel@gnu.org
> 
> > Actually, why not make the above return the number 4?  The Unicode
> > Data File Format says that this property's value is ``an integer or
> > rational number''.  We could even use 0.2 for characters such as
> > "VULGAR FRACTION ONE FIFTH".  Isn't that better than `1/5'?
> 
> Bear in mind that 1/5 (0.2) isn't actually exactly representable by 
> binary floating point, bit like the way 1/3 (0.333(3 recurring)) can't 
> be written properly in decimal

The important issue, IMO, is whether the value of that property should
be a number whenever possible.  The rest are details that can be
figured out once we make that decision.

Re: numeri

[Stefan Monnier]

>> (get-char-code-property ?4 'numeric-value)
> Actually, why not make the above return the number 4?  The Unicode

Agreed.  It says `numeric-value' so it should return a number.
I understand that 1/5 cannot be represented exactly as a number in Elisp
right now, but the floating point approximation should be sufficient: we
can recover the /15 fraction from it, if really needed, by considering
additional assumptions such as "this is the numeric value of a unicode
char".


        Stefan

Re: numeri

[Kenichi Handa]

In article <jwvy6z2xi4x.fsf-monnier+emacs@gnu.org>, Stefan Monnier <monnier@iro.umontreal.ca> writes:

>>> (get-char-code-property ?4 'numeric-value)
> > Actually, why not make the above return the number 4?  The Unicode

> Agreed.  It says `numeric-value' so it should return a number.
> I understand that 1/5 cannot be represented exactly as a number in Elisp
> right now, but the floating point approximation should be sufficient: we
> can recover the /15 fraction from it, if really needed, by considering
> additional assumptions such as "this is the numeric value of a unicode
> char".

Ok, I've just committed changes to make the above return an
integer or a floating point (if non-nil).

---
Kenichi Handa
handa@ni.aist.go.jp