Setting Itemdelimiter to multiple character string

[Sivakatirswami]

I am writing a processor to handle input from a raw index dump from Quark
Express which is "massaged" back into a tagged format for re-upload into
Quark, getting Metacard do what the index person usually has to do by hand
over many days/hours of manual work

 Parsing is extremely difficult due to the subtle similar, seeming
infinitely variable combinations of things like

<tab> See also God, nature
, see also Parasakti
: See

Etc. where any instances of a single character of punctuation, or word
space/count etc. can rarely be used as a dependable delimiter. The logic of
which is  quite reasonable to those types of minds that create the Chicago
Manual of Style and its ilk, but which defy any rational algorithm ...

I saw an earlier question which arose but was never answered but which is
quite important for a lot of this kind of text processing work. Simply:

Can we safely build processes that poke multiple chars into the itemdelimter
to do such things as the following:

==============
## Form 3: Text string ending with ", see " and one or more words/phrases
## which are to be italicized...
    
    if x contains ", see "  then
      put ", see " into itemDelimiter
      put "<I>" before item 2 of x
      put "<$>" after  item 2 of
    end if
    
Put another way, barring earthquakes and unknown acts of God in the realm of
cybernetics, is it a safe bet that the itemDelimiter will support this
multiple character string input in the future?

Himalayan Academy Publications

Sivakatirswami
Editor's Assistant/Production Manager
katir at hindu.org 
www.HinduismToday.com, www.HimalayanAcademy.com,
www.Gurudeva.org, www.hindu.org

Read The Master Course Lesson of the Day at
http://www.gurudeva.org/lesson.shtml
[Built daily at Midnite by a Metacard CGI on a Solaris]