Re: XHR LC comments

Bjoern Hoehrmann wrote:
> Being able to send wf-but-ns-illformed documents would not make much
> sense if you couldn't also read them back in, and the odds that some
> wf-but-ns-illformed has been created deliberately for use with XHR
> as opposed to being created due to some mistake, seem too low to make
> supporting this worthwhile.

I do not disagree with that.

But as you said, it's probably a mistake. There are several ways to 
handle this:

1) throw an exception on serialization

2) send the document, and let the recipient detect the problem

3) silent data loss (sending empty body)

In cases 1) and 2), the caller will find out about the problem, in case 
3) s/he might not. That is a problem (keep in mind that there are HTTP 
resources/method handlers that both support an XML request body and an 
empty body, and the difference *does* make a difference).

BR, Julian

Received on Monday, 19 May 2008 07:02:33 UTC