Ballpoints and pencils facilitate making notes anywhere.
The typewriter sped up writing legibly considerably.
This last point is interesting: being able to churn out fiction can,
it is sometimes argued, allow people to produce lots of lower
quality work than if they have to rework parts. I'm not sure how one
applies that point to programming though, but maybe we are still at
the typewriter stage. Is there any evidence that refactoring
browsers allow people to "leave quality till later" and never get
back to fix it? There are other factors at work though, not just
the tools: time pressure for one.
The 700-lb pen is a flawed analogy, though it offers something of a
good, quick visual. I tend to think that the ease with which people
can now write and self-publish has led to a flood of crap, though it
has also made possible numerous gems that might not otherwise have
been available.
I wonder if there are reasonable analogies between programing tools
and writing tools; I'm inclined to see the Excel macro language as the
equal to the ballpoint pen.
I also wonder about the efficacy of refactoring tools, especially if
one has to think of refactoring as a separate task. If something
isn't easy to do, so easy that it just because part of your unthinking
process, then it will often be pushed off forever. Or else you may
have to alter (stifle?) your normal creative process.
There was a post on some blog someplace (how's that for precision? It
was either a java.net or artima blog, I think) positing that TDD might
stifle creative flow. And I can readily see that.
Like at least a few people here, I tend to first bang out short
scripts to sketch out some possible approaches to a task, refactoring
as I go. Ruby make this very easy, both to try something out and to
move things around; a tool might be a good thing for large
applications, but for most things all I need is gvim.
At some point I may go back and rewrite the application, once I've
discovered it's true purpose, and that's when I am more likely to
write the unit tests first. But since adding unit tests in Ruby is
snake simple I also find myself dropping them in during that first
stage, if code starts appearing less than obvious, so that I don't get
side-tracked looking for goofy bugs.
I may be unfamiliar with available tools (aside from Eclipse, which
makes Java bearable), but I think this style of exploration,
refactoring, and testing is less fluid in Java, and it effects your
thinking. You are more inclined to want to do a bunch of up-front
design, much as when people had to write with quill pens they were
likely more inclined to think a whole bunch about what they wanted to
write. But, in both coding and writing, the physical process itself
changes how you think. You often don't really know what you're
thinking until you write it down.
James
