<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Joseph Pelrine wrote:
<blockquote TYPE=CITE>Well, I can't say much, since I don't consider myself
a Squeaker, but (as
<br>someone who earns a living programming Smalltalk XP-style) I wouldn't
waste
<br>my time trying to do "cool" but unnecessary things (XML/DOM parsers,
B-Tree
<br>implementations etc.). 90% of the structured text we see in text panes
is
<br>method code. If you want color coding, pretty printing etc. for this,
why
<br>don't you just use the built-in parser or the RefactoringBrowser parser
as
<br>a base? That's the way these things were implemented in VisualAge (see
VA
<br>Assist) or VisualWorks (see Travis Grigg's code highlighter).</blockquote>
Your points may well apply within the realm of the development environement
however in more general terms, I beg to differ.
<p>The purpose of the code editing tools is as a bootstrap for the projects
that people are really trying to develop. Their existance and the fact
that 90% of the use of text boxes is in the coding tools is not really
that relevant.
<p>The question is what about the other 10%? 90% of requirements may be
easy, but currently by simple mathematics there are 10% of requirements
are virtually impossible.This is an impasse that could use some chipping
away at.
<p>Therefore what is relevant in XP terms would be a "customer" requiring
the ability to handle large quantities of text with styling, tagging and
editing features. And the simplist thing that could possibly work does
not currently provide the functionality, therefore there is a driver for
some of these "cool" and necessary things! Having got those things developed,
their usefulness in the development environment would be a bonus.
<p>I am such a customer. Of course if I was to do the simplest thing that
could possibly work then I may be up and running far quicker. To do this
I would run a two pass process rendering a complex data model into the
existing simple one which is then rendered to the screen. Then I would
need a complex series of hooks so that editing and edits carried out with
the limited editing tools can be reflected back into the complex underlying
data model. Having done this I would then consider a major refactoring
of the entire framework so that my complex data model can be rendered and
etited directly. Where do we end up... A better framework for text editing
and display.
<p>any takers ?
<p>Keith
<br>
<br>
<br>
<pre></pre>
</html>