Re: [xml-dev] Re: XML element names

[David Carver <d_a_carver@yahoo.com>]

Michael Kay wrote:
>> This describes why prohibition of certain characters in names 
>> (actually, person names as described in this article), can cause 
>> certain people from not being able to use certain applications. I 
>> think, the concepts of this article apply to XML names as well ...
>>     
>
> Actually, I'm not entirely convinced by this. I think that XML names are
> primarily for use within the IT system and do not need to have any direct
> relationship to strings that are meaningful to end users. Mathematicians
> have always been happy working with symbols in a foreign alphabet, and many
> programmers don't seem to mind it either. 
Michael, I would have to disagree with this. EDI had/has cryptic 
characters and if you are an end user trying to figure out why a message 
isn't coming in as you expect, unless you know the spec inside and out 
for the message being sent, it is very time consuming trying to debug 
what is wrong. The same thing hold's true for XML files and elements and 
attribute names. A human eventually has to look at it at some point 
during the development process. Even programmers, and I know many that 
prefer something they can read and decipher, as opposed to the other 
formats that require more intiment knowledge of the spec (i.e. 
<2A3></2A3> what does that actually mean?? as opposed to 
<PurchaseOrderID></PurchaseOrderID>).
> (Perhaps there is a case for allowing names to use letters from any alphabet
> but not allowing different alphabets to be mixed in the same namespace?)
>   
I would more appropriately say, that the entire file needs to be encoded 
in the same character set. Mixing characterset encodings would be bad. 
I've seen it done and it causes all sorts of havoc.