Re: Multiple Message Bodies

[Mark Hindess <xpl@xxxxxxxxxxxxxxxxx>]

On 11 December 2005 at 12:46, "Ian Lowe" <ianlowe@xxxxxxx>
wrote:
>
> A couple of good suggestions for tightening up the protocol doc -
>
> Clarify that it's 1 header, 1 message body (that used to be in the
doc, I'm
> sure!)
>
> Clarify that a developer is free to add custom fields, but not repeat
schema
> fields - unless the schema specifically allows it.
>
> We'll also need some of the schemae docs updated to reflect that they
use
> multiple fields (and which ones), however, that can be done over time.

It would also be good to make sure the schema are consistent.  For
instance, Mal mentioned a schema where they use duplicates, but other
schema - for instance, x10.basic - use separate fields - i.e. data1 and
data2.

This suggests that the unwritten policy is that schema developers should
use duplicates where you might have one or more but use named fields
where you know how many you should have?

> I'll get them added to the wiki.

Great! Thanks Ian.

I notice that some implementations don't currently support duplicate
fields.  Mine certainly doesn't (yet).  Handling the duplicate case
is actually relatively easy in Perl but it is somewhat harder in
languages which are more strongly.  It's probably a good idea to
clarify if handling duplicate fields is a requirement.  While probably
not important for clients handling only a couple of schemas (that
might not care about duplicates) it is kind of important that all hub
implementations handle them correctly.

-Mark.