Re: Interleave (was Re: [relax-ng] RELAX NG telcon 25 October 2001)

[James Clark <jjc@jclark.com>]

Maybe we should consider making

  oneOrMore//interleave

a prohibited path.  We already make oneOrMore//interleave//attribute 
prohibited, so we would be simply generalizing that: it would simplify the 
spec rather than complicate it.  I can't think of any use cases where 
oneOrMore//interleave is useful.

The other two are trickier.  If the third one was changed to:

(a* & (b, c)) | ((a, b) & c*)

would it still be hard? If so, I don't see any solution, since we need to 
allow

  (a, b) & c*

and we need closure.

--On 25 October 2001 17:52 +0900 MURATA Makoto <EB2M-MRT@asahi-net.or.jp> 
wrote:

> James Clark wrote:
>
>> 2. Issue 40. Restriction on interleave.  Any further news from
>> Hosoya-san? Should we allow <interleave> inside <list>?
>
> Hosoya-san reported that the following patterns are hard although
> they satisfy the restrictions in our draft.
>
> 	a, (a & b), b
>
> 	(a & b)*
>
> 	(a & (b, c)) | ((a, b) & c)
>
> He is leaning towards some global restrictions, but does not have any
> concrete ideas.
>
> It appears to be difficult to find something better than our (arguably
> insufficient) restrictions in the near future.
>
> Cheers,
>
> Makoto
>
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>
>
>