[XeTeX] linux libertine and number

[François Patte]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Le 31/12/2010 15:35, Philip Taylor (Webmaster, Ret'd) a écrit :
> 
> 
> Ulrike Fischer wrote:
> 
>> And please don't repeat that you "was not in any
>> way trying to suggest that this is a general or universal
>> solution.". You _did_ sent the first example without any code which
>> restricted the effect of the catcodes changes and without any
>> warnings about side-effects. This was the posting I commented as
>> this is not my idea of an helpful answer.
> 
> As is so often the case, a suggestion (intended to be
> helpful) becomes a protracted argument from which
> the original questioner is effectively excluded.  But
> as I dislike being accused of something on a fallacious
> basis, let me re-cite the original correspondence :
> 
> François  (1) :
> 
>> I would like to know why the numbers are different if typed with or
>> without the $ sign:
>>
>> Using \setmainfont{Linux Libertine O} 7 is smaller than $7$
>>
>> Is it possible to correct this? I want the same size (ie. $7$)
>>
>> Of course I can always write $7$, but is there another way.
> 
> François  (2) :
> 
>> Thanks for this answer. It is not too nice if you have to write:
>>
>> Multiply 5 by 7: $7\times 5 = 35$
> 
> Self (1) :
> 
>> Bonjour, François : is this perhaps the sort of
>> thing you had in mind ?
>>
>> \documentclass {minimal}
>> \usepackage {fontspec}
>> \setmainfont {Linux Libertine O}
>> \begin {document}
>> Multiply 5 by 7: $7\times 5 = 35$
>>
>> \catcode `\5 = \active
>> \catcode `\7 = \active
>> \def 5{\ifmmode \string 5 \else $5$\fi}
>> \def 7{\ifmmode \string 7 \else $7$\fi}
>> Multiply 5 by 7: $7\times 5 = 35$
>>
>> \end {document}
>>
>> Obviously you can \catcode and \def all digits
>> from zero to nine in a simple loop to cope with
>> the general case.
> 
> My answer thus addressed François' problem /as stated/ :
> it did not suggest that it was a solution to such
> problems in general (although I can see that it /might/
> be possible to misinterpret the last paragraph : when
> I spoke of "the general case", I was referring to the
> case of all the digits from zero to nine, not to the
> entire class of problems of which François' was one
> example).
> 
> Yes, I completely agree that "[I] sent the first
> example without any code which restricted the
> effect of the catcodes changes and without any
> warnings about side-effects.", but I also assumed
> that my answer would be interpreted within the context
> of the question being asked, and not within the context
> of such questions in general.
> 
> ** Phil.

Philip, there is some problem with your answer:

Yes I wrote:

>>> Using \setmainfont{Linux Libertine O} 7 is smaller than $7$
>>>
>>> Is it possible to correct this? I want the same size (ie. $7$)

But it is because I did not think at all that it was possible to change
the shape of the digits when latex is in math mode.

I know how to use latex for many things but I am not aware of all
mechanisms and when I tested your answer, I thought that it was a good
solution to my problem, with some restrictions though: I don't know much
about the 'catcode' but it seemed to me that  this way of doing could
affect some other latex commands using digits, and I understood
immediadly what could be the disaster with the example given by Ulrike,
indeed I will need to use tikz  in the same article and I think that if
I had chosen your solution, I could have waste a lot of time before
understanding that the problem was coming from your solution.

I think that many people are like me: using latex and finding some
solutions to their problems but without going deeply into the technical
part of it. Mastering the side effects of a solution like yours is not
obvious and you should have given some information about the possible
dangers.

Ulrike's solution is a very interesting one because it gives some
homogeneity to the text when you have to mix text and mathematics in a
paragraph, I mean, when your text is not fully a mathematical text.

Anyway thanks for these answers.

F.


- --
François Patte
UFR de mathématiques et informatique
Université Paris Descartes
45, rue des Saints Pères
F-75270 Paris Cedex 06
Tél. +33 (0)1 4286 2145
http://www.math-info.univ-paris5.fr/~patte
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAk0d9+8ACgkQdE6C2dhV2JUjQwCgtm4WcChc8dyEATdxJJsUbIkC
hwUAn2nAPy5Md62Pp1UKBJ+/Wpae4pKS
=t3fI
-----END PGP SIGNATURE-----