RE: Experimental TESTxx header usage.

[Kevin Hawkins]

> -----Original Message-----
> From: Stuart Booth [mailto:<a
href="/group/xAP_developer/post?postID=vbRVfAfM0voWUbIhs2OA6vQ7q5YWmKXGaP7V9HlC2ssTCW53wcy4mVp5FogA2QsC2KrXPSTUATAjQL7lBA">lists@x...</a>]
> Sent: 05 August 2003 18:54
> To: <a
href="/group/xAP_developer/post?postID=YUBYd9aAg6GtAJqopRqQEPoa3XhhfucPkxc71mxGoz2IVY6pbIUYgITa0U2ROlu09oVAUYDSfB_8lxGgTRv8dLYVHnt2kDc">xAP_developer@xxxxxxx</a>
> Subject: Re: [xAP_developer] Experimental TESTxx header usage.
>
> On Tue, 5 Aug 2003 15:15:29 +0100, "Kevin Hawkins"
> <<a
href="/group/xAP_developer/post?postID=z45tFmGY974-ABVFvPznNSqBsNR66YsluB_vYdDLLYZXhR3Io5aHolQFVgmNemnXRq3Y-gkVu_0FopVRll0HtnVX">lists@u...</a>>
wrote:
>
> > The xAP specification committee agreed in the v1.2 spec release,
> >(although we did not publish it), to allow the inclusion of TEST
> headers
> >within xAP messages as a way of trialling suggested features. It
is
> >perfectly permissible to use, for your own purposes - (see below)
the
> name
> >value pairs Test00 to Test09 in the HEADER .
>
> Did you mean Test99?

TBH I can't remember what we agreed - I think it was Test0 to Test9 but it
could also have been Test00 to Test99 - and that is how I originally wrote
the email - that is probably why now I have two numerical digits after the
test. However I also reasoned that we didn't want to see a load of Test
name
value pairs in headers so I knocked it back down again to imply the very
special nature of these - I left the double 00's such that we could expand
it later without breaking code if need be . At the moment let's take it as
I
wrote it and see if that suffices.
>
> > Please note that the spec for
> >xAP v1.2 says that all 'unexpected name value pairs' will be
> gracefully
> >ignored so this should not break anyones implementation. ;-). All
such
> >parameters can safely be ignored as they are not part of the spec.
>
> Oops. You just knew you'd get a reaction from me, didn't you? :-)) Yes
> folks, my framework just says "No!" when it encounters
rubbish. I
> prefer to stop that stuff and fix it rather than casually let it pass,
> but in this case I'll change it.

I have two views on this. Firstly that a xAP development framework should
be
exceptionally thorough on ensuring validity of the content of xAP messages
and hence shouting about such things is good. However a second view says
that an application will be far more resilient if it can ignore stuff that
it can't cope with which also allows us to do things like this. Perhaps a
good compromise would be to allow Test00-Test99 through but still 'flag'
other name value pairs but not necessarily throw the teddy.

>
> >What
> >I should also mention is that for example Test06 as used in my
> applications
> >(UKUSA) is NOT to be confused with Test06 as used by say Stuart
> (KCSOFT). To
> >that end you should regard my usage of Test06 as really
"UKUSA.Test06"
> >different to KCSOFT.Test06. Now whether you personally chose to
> validate
> >this is your choice - you run the risk that someone else's app may
use
> the
> >same name/value pairs for a different purpose if you don't.
>
> I'll add that to my ToDo list for my next batch of xAPFramework
> updates.

Again there is a double edged sword here - if we decide that we wish to
experiment with a communal set of values eg the Test00 to Test04 pairs then
they shouldn't be qualified with the xAP developer name to be regarded as
valid - however higher ones should be. I don't know what best to suggest
here - maybe others have ideas - we could always say all test0x are shared
perhaps and start at Test99 downwards but I don't want to see loads of
superfluous headers appearing. Maybe we need xTest00-09 and Test00-09
pairs.


Kevin