Re: More liberal draft formatting standards required

[Julian Reschke <julian.reschke at gmx.de>]

Martin Rex wrote:
> Some more thoughts about this:
>
> As long as the IETF want to continue publishing standards in the one
> single language "English", it should restrict the character sets in
> the texts (and the examples) to ASCII.

Text? Yes. Examples? Contact info? No.

> non-ASCII letters are a royal PITA in so many ways, that they should
> not be used.

Disagreed.

> Finding a Postscript printer driver that prints umlauts is
> extremely difficult (I know, I'm German and I can't get PDFs printing
> properly with Umlauts).  Similarly, quite a lot of screen fonts do
> not contain Umlauts.  And although my software is configured to
> work with iso-8859-1, it is unable to display cyrillic and greek
> characters.  

I have never had problems printing PDFs with umlauts.

> And while I'm German and have keys for umlauts on my keyboard,
> I have serious difficulties to create other funny characters
> from the iso-8859-1 character set (like some skandinavian),
> let alone greek, cyrillic or any symbols from asian languagues.

That's a totally orthogonal issue.

> I can not recognize, name or type any of the symbols from
> asian languages, and I can neither print or display them,
> nor input them on my keyboard.  Which makes it completely
> impossible to search for them or discuss them.

I'm pretty sure that you can display and print at least some of them 
with a recent browser and operating system.

> Really, I see no justification why they should be part of an
> IETF specification, if they're only marginally useful to a 
> (smaller) fraction of the consumers of IETF specs and fairly
> hostile and unusable to the rest.

Non-ASCII characters in I18N examples would be useful to *any* reader. 
Keep in mind that a consumer of an IETF spec will be able to read 
English, thus understand the Latin alphabet, and thus be able to 
understand the difference between, for instance, an "e" and an "é".

> If the IETF want so allow umlauts, it will have to allow all
> of Unicode, and that would become close to a catastrophe.

I would support allowing all of Unicode for contact information (as long 
as an ASCII substitute text is available), and a subset for examples.

>  90.0% of the software *I* use is capable of ISO-8859-1,
>   9.9% is pure-ASCII
>   0.1% is capable to do more than that (not necessarily
>        full unicode)

Any HTML4 user agent supports Unicode (and no, that doesn't mean it will 
have fonts to display all of it).

> Actually, most of the common fixed sized fonts seem to
> contain only ASCII characters (and fixed size fonts happen
> to be the fonts that I use mostly).
> ...

BR, Julian