Re: [xml] encoding tutorial draft

[Igor Zlatkovic <igor stud fh-frankfurt de>]

> I am not sure Igor if you do not expect too few. Of course there are
> programmer who never put a thought on the general problem of data
> representation in computing, but from time to time every programmer is faced
> with the problem, that (s)he needs to know about the actual representation
> of data. (This does not relate only to character data but to numbers and of
> course the more complex data types). Not knowing what "encoding" is produces
> so many traps. I think you can expect, that most of the libxml users DO know
> what encoding is and that there are many different ways of representing
> character data, since everybody has once walked into one of these traps or
> the other. 

Sure, you are right. Still, the fact remains, that people have problems 
believing that the conversion is necessary. Statements like "I have found a 
bug, libxml misinterprets my strings when they contain äöü" are too common.

I wasn't trying to introduce any reader in C, but to emphasise the necessity 
of conversion as much as possible.

> And don't forget the people we are talking about are C
> programmers! I emphasize again: C!! Only that they need a small reminder
> perhaps.

C programmers are the targeted audience and every C programmer was a 
beginner once upon a time. Many posts here could be observed which hint that 
the poster knows little beyond the first hello-world lesson. But, these 
beginners in C are also our audience. What better way to learn programming 
than to write software? Libxml is not reserved just for the expirienced.

> More grateful they will be if supplied with a simple way of achieving the
> conversion to utf-8. Maybe the smaller code snippet I made for the FAQ
> should at least be mentioned in the tutorial, since 80% of all appliances of
> conversion will be from iso latin-1 (or some local windows codepage almost
> the same) and about 10% the other way round (just a guess and not counting
> east asian programmers who know the business of encoding quite well :-):
>
> in = "some null terminated iso latin-1 string";
> temp = size = (int)strlen(in)+1; /*terminating null included*/
> out_size = size*2-1; /*terminating null is just one byte*/
> out = malloc((size_t)out_size); 
> if (!out) {
>         if ((ret=isolat1ToUTF8(out, &out_size, in, &temp)) || temp-size) {
>                 free(out);
>                 out=NULL;
>         }
> }

Sure, everything that makes sense to someone is worth mentioning. Each human 
perceives the information differently and only the readers of a tutorial can 
judge if it helped them learn something or not.

The fact remains, that libxml is there to process XML, not to transcode data 
from one format into another.

Ciao
Igor