RE: restricted use of Dot and DotDot

> 
> Is there any reason why the syntax disallows (e.g.)
>                 ..[whatever]
> but allows
>               (..)[whatever]
> and
>     parent::node()[whatever]
> ?
> 
> That is, why is DotDot an entire step (with no opportunity for
> StepQualifiers), rather than an alternative to 'Axis NodeTest'?
> 
> Ditto all the above for Dot.
> 

At one stage our syntax did allow .[predicate] and ..[predicate] but for
some reason it was reorganised to disallow these constructs, I think as an
accidental side-effect. My view is therefore that this is a bug, but I would
need to check with the parsing experts...

Mike Kay