[jdom-interest] Namespace inheritance after cloning

[Elliotte Harold]

Eric VERGNAUD wrote:


> Did it ever occur to you that someone, with a perfect understanding, could
> simply disagree on the purpose of tools and API's and on programming style
> and habits WITHOUT being wrong, but simply different ?

Not, in this case, no. JDOM's approach is the only sensible one. 
Alternatives have been tried, and they've failed. Every single developer 
and every technology for working with XML that has tried to make 
namespaces properties of position rather than part of the element's 
fundamental nature has quickly realized the error of their ways, and 
adopted JDOM's approach. JDOM was not the first to do it, and it won't 
be the last. But there are *no* alternative techniques for 
namespace-aware processing of XML that have achieved any success.

if you want to convince me otherwise, you're going to have to do more 
than talk. You're going to have to demonstrate that the alternative 
approach actually works by producing an API that uses it, getting a wide 
number of people to use it for processing documents that use namespaces, 
and then showing that they are not confused and do not produce buggy code.

You don't have to do it all yourself. JDOM's open source. Fork it and 
try for yourself. But this experiment has been done before and the 
outcome's always been the same. I am confident that you or anyone else 
who actually tried to use an API like this for any significant period of 
time would very quickly see the wisdom of JDOM's approach.

-- 
Elliotte Rusty Harold  elharo at metalab.unc.edu
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim