perl-unicode

Re: In-Band Information Considered Harmful

1998-10-28 20:12:06
Chip Salzenberg writes:
There are some rules of consistency of markup.  One should define
what the any "editing" operation is doing to markup.

I'm not going to even think about designing markup-rule-enforcement
into the metadata infrastructure of Perl's core.

You see, you thought about dead data only indeed!

No.  I simply thought you were going for something a lot more
ambitious than simple insert- and delete- behavior flags, which
I can certainly support.

As another thought this level of operations should not even care about
the "structure" part of the markup.  Consider a table

   cell11   cell12

   cell21   cell22

which is encoded in the markup as a tree with leaves

     __ cell11
    /     
   /\__ cell12
  /       
  \  __ cell21
   \/
    \__ cell22
         
However, the markup for this (with in-band notations) is

  "<TABLE>cell11<END-LEAF>cell12<END-BRANCH1>cell21<END-LEAF>cell22</TABLE>"

The core does not need to know the difference between different
separators, <END-LEAF> and <END-BRANCH1>.  It need only to know that
they are separators for a bigger structure "TABLE" with appropriate
rules for deletion (and possibly insertion, for out-of-band approach).

Ilya