Re: Experimental TESTxx header usage.

[mark_harrison_uk2]

--- In <a
href="/group/xAP_developer/post?postID=bFCRIYcTp6G-TUn-u42WSTb-00W4ga1FWa3Eol9MxQFSLBsXW6592pMgcMQmrSY6DhltyKeuuCXzopBP-XkTxdYC9-o">xAP_developer@xxxxxxx</a>,
"Ian B" <Ian@M...> wrote:
> Hi all
>
> I am working from:
> Specification Release v.1.2 - 29th November, 2002
> Document Release v.1.2-9 (Final) - 1st December, 2002
>
> The word 'test' does not even occur in this document

That's correct. The TEST headers were NOT ratified as part of the
spec.

I think that Kevin's proposal has the "official" status of a
suggestion - it should not be regarded as an official part of xAP...
HOWEVER...

> Quote
> A device should ignore all unknown name, value pairs included in
the header.
> /Quote

Yup - that's what the xAP spec says - The reason for this was twofold:

1: a general agreement that we needed a mechanism in which potential
Specicifaction extensions could be tested.

2: a general agreement that graceful fallback is a worthy design goal
of any system.

> In the embedded world [...] these surprises are not exactly
> welcome.

I don't think you can say that something that is printed in the spec
is a "surprise" :-)

> I validate deliberately strictly and the insertion of new segments
> breaks my code i.e. it will fail and ignore the message.

Do we need a debate on what developers would prefer? A requirement
for a graceful ignore vs. a hard set of header standards.

I, for one, would be grateful if anyone considering using this could
detail what the purpose of their new header items is, and why they
are sufficiently generic so as to require entry in the header rather
than just being in a specific class or schema for the application /
message type in question?

Regards,

Mark