Re: [xml] re:[xml][no subject]

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

> I read your page ( http://xmlsoft.org/encoding.html) and all the
> documentation: well the page doesn't give any useful information. I have
> used your search engine with the keys "iso-8859 problem" but in the
answers
> there is only one example which is wong.

You were looking for a problem, you named it a problem, but there was
nothing that indicates a problem. The right search words would be "default
character set" or "internal text representation", or similar.

* The ISO-8859-1 character set is very widely used and that for some time
now. It is very unlikely that a software of this sort which has been in
production for several years can have fundamental errors in processing this
character set.
* The fact that the type 'xmlChar' is an 'unsigned char' and not a 'char' is
good enough an indication that something is different to traditional strings
used in C. In your code sample you even casted a string to xmlChar* without
asking yourself why.
* The presence of a macro named 'BAD_CAST' and its usage in all tutorials to
cast a constant string to xmlChar type is another strong evidence for
something different to a regular usage of character arrays.

The above points represent a very strong indication for the fact that the
internal storage of text is different than what is regularily used in C.
This is nothing but a sound reason.

I am not saying that you shouldn't ask questions, quite the contrary. I am
not performing any critics on you. I am just very puzzled about why did you,
like so many others, see a problem? Knowing that would be very helpful in
creating a documentation which leads the people in the right way. That is
the whole point I am stressing here.

> I know that you are a very busy man and then you have no time to answer to
> everybody and write some more useful documentation, but you should
remember
> that we are busy too, and if a person does a question is not because he
> hasn' spent time to read the documentation (everybody have read it) but
> because he has a problem he can't solve and maybe he needs help.

It has nothing to do with being busy and not having time to write useful
documentation. Any software developer is
* the worst possible tester for programs she wrote herself,
* the worst possible person to document what she developed.
When dealing with what you made yourself, you are always strongly biased
unless you have been trained not to be. It is very difficult to make a
documentation which is considered useful by everyone. Noone can produce such
thing alone, based solely on own perception.

Be Daniel's documentation good or bad, the fact is that it does contain
information about UTF-8 being the only acceptable encoding in a call to any
libxml function which accepts text. If you could make clear what exactly
prevented you from finding that bit, you would do a great contribution to
the design of a more useful documentation.

> In the tomcat user list I spend some hours a day to answer questions
> (.......often the same questions) but people need help and then I give
them
> it.

Then you should be the first to understand that answering the same questions
over and over again is a nerve-killing business, eh? :-)

Ciao
Igor