Re: Save text as unicode

[Brian Stafford <brian stafford uklinux net>]

On Thu,  7 March 12:37 Olaf Frączyk wrote:
> On 2002.03.07 13:10:39 +0100 Carlos Morgado wrote:
>> 
>> On 2002.03.07 11:27:20 +0000 Olaf Frączyk wrote:
>>> Hi,
>>> 
>>> I would like to have a possibility to save text (or plain message) as 
>>> unicode.
>>> It's because I use mostly iso-8859-2, and if I save document in linux, I 
>>> have problems to read it in windows.
>>> 
>> isn't the correct approach to call ms support and ask them for proper
>> iso-8895 support ?
> What do you mean by 8895?

8859?

> Anyway, the only way to do it is unicode. In normal text file you don't have 
> information if it is 8859-1 or 8859-2. The only way is to tell the 
> application how to display it.

Well what exactly is unicode?   AFAIK it takes 4 octets to cover the full 
unicode characetr set.  M$ use the UTF-16 encoding of unicode which is a two 
octet encoding.  It isn't possible just to say this is a unicode file, the 
actual encoding in use is what matters.

> The same is for unix.

Many apps in the *nix world use UTF-8 which encodes the unicode characetr set 
into a variable number of octets, though UTF-8 has the nice feature of being 
compatible with ascii for the characters below 128.

> So call to ms won't help :(

Maybe not, but their choice of UTF-16 is unfortunate, unnecessary and 
incompatible.  Also insane - using UTF-8 means you can preserve existing APIs 
because unicode characetrs fit in a char * - with UTF-16 you need a short * 
for characetrs so that means duplicating *every* API that takes a string.  And 
it meant they had to add non standard extensions to the compilers to select 8 
or 16 bit characetrs in strings.  This seems an awful lot of work to achieve 
nothing but inconvenience, unnecessary inflation of text files and yet another 
source of bugs.

Even if it doesn't help, M$ deserve to be pestered for their stupid decisions.

--
BCS