[Top] [All Lists]

Re: Text/enriched ambiguity

1993-10-29 13:07:52
  > example is that there is no matching </colorentry> or
  > </colorentry 0 255 255 0>
      Whoah...   hang on.   WHY is this a problem? 

violation of scope.

a <xxxx> establishes a new state in the document which is required to 
nest within the state established before it.

<x>  <y> ... </y> </x>  but not  <x> <y> ... </x>.

Admittedly some commands would function well if they implied a cancel
as in:

<x>  <setcolor rrr ggg bbb> ... </x> ...<setcolor rrr ggg bbb> <y>...</y>

but there would be no way for a dumb text/enriched parser which didnt 
recognize the setcolor command to parse properly, unless we introduced 
new syntax, and this would impact all present engines, so it may not 
be a good idea:

<x> <<setcolor rrr ggg bbb>>

IMHO the spec should have provided an intrinsic typing in order to 
allow the parser to classify even anonymous commands:

<<nestable_scoped_command>> ... <</..>>
<<<command_with_lengthy_parameters>>> parameters, hide em if dondo<<</..>>>

The purpose being to provide a foundation for a robust parser.
dana s emery <de19(_at_)umail(_dot_)umd(_dot_)edu>

<Prev in Thread] Current Thread [Next in Thread>