Re: p:error

We should keep the code as required, otherwise how could a pipeline 
author perform a p:choose on /err:errors/err:error/@code ?

On a side note, the error vocabulary in the spec (appendices D and E) 
are bound to the c: prefix. Moreover, if @code is made required for 
p:error, it should be required as well on E.2 (err:error) and on the 
schemas.

Rui


Norman Walsh wrote:
> I think we're fooling ourselves if we believe that requiring a code
> or description on p:error is going to force authors to write either
> meaningful codes or meaningful descriptions.
> 
> They ought to both be optional.
> 
>                                         Be seeing you,
>                                           norm
>