parallel invention of the wheel? (was: Document Search Summary)

[Wilhelm Sanke]

On Fri, 13 Jun 2003 (Subject: Document Search Summary)
 Dar Scott <dsc at swcp.com> wrote:

(snip)

> We have asked for search.  We have it in abundance.
>
> Dar Scott
>
Just for the record: Indeed, after some waiting, we have it in abundance
now.

There are times in history when the situation is "rife" because
circumstances and discussions demand new solutions, which then may be
developed nearly simultaneously.

Because of the discussion I had added a special button for searching the
Transcript Dictionary to my general text search tool "Topsearch 1" that
was added to the list of Revolution User Contributions. The extra button
seemed to be necessary as the "revdocs" stacks are not "topstacks" by
their nature.
Two days later (please take "two days" not as an accurate description,
but as a metapher for a slight difference in time) we were informed that
a new -although somewhat hidden - search routine had also been added to
Revolution to search the Dictionary, differing in a number of details.

Further following the discussion, I considered - in a post to the list -
to expand the search routines of "Topsearch" to encompass the whole of
the Documentation. The result was "Topsearch 1.2", also to be found
among the User Contributions.

Again two days later (for accuracy see above) we got expanded search
routines for the Documentation with Revolution 2.01 that were so
competently described by Dar Scott. There are a lot of parallels between
the details of the "Topsearch 1.2" solution and the new Revolution
search routines, because both approaches attempted to deal with the same
set of problems. I will not go into details, you can  compare the
routines for yourself if you like.

I mention only one of the common features, because it may be of interest
also outside of search routines, namely when dealing with larger
quantities of text and intending to "hilite"  specific strings in that
text. This can be achieved by a combination of  the replace command
together with the htmltext property that provides the possibility to
colorize (foreground or background of texts) all specific strings even
in huge portions of text in the split of a second.

I do not claim any credits for the approaches of the Revolution team; I
claim credit for my side of the parallel development.

As "Topsearch" is primarily a tool to search the text contents of *any
current topstack* - hence the name "Top"-search - it would be worth
considering if such a general text search routine (that works in a
different way than the "find" routine) should be also integrated as a
tool into the Revolution environment. Another "two days" would be
sufficient to accomplish this.

That would then  make the stack "Topsearch" completely obsolete,
although it could still be adapted to be used with Metacard and its
sparser documentation.

Regards,

Wilhelm Sanke