Subject: OSS meets JavaScript?
Date: Fri, 1 Sep 2000 11:05:02 -0400

Several recent threads and some offline conversations have made me
aware of an interesting conflict.  OSS developers like to work on fun
stuff.  Companies like to work on stuff that generates market share.

These are not the same.

The particular point that raises this is Tim O'Reilly wondering at why
JavaScript is so clearly taking off but seems to have little uptake in
the OSS community.  The answer is that JavaScript is taking off for
the simple reason that for an important class of projects it is the
only game in town.  But portable JavaScript is a real pain to develop
so people don't want to do that.  Tim's response was that he won't
speak on the technical issues, but it really is important.  Well fine,
but importance is a poor motivator.

A person that I happen to think provides a good model for this type of
issue is Larry Wall.  When he wants people to do something that he
thinks is a good idea, and they aren't doing it, he goes out of his way
to figure out what obstacles they have and remove them.  He wrote patch
to convince people to stay current on rn.  He wrote a2p and s2p to get
people to learn Perl.  He sold Perl in a box so the techies could point
at a box and a receipt for the benefit of management.

In short instead of saying why from his point of view it is important
that people do what he wants, he tries to solve their issues.  In this
case the problem is that portable JavaScript is a pain to develop.  So
if people want to see better uptake, it is important to improve that
situation.  One solution is to standardize the browsers.  Well a lot
of work has gone into that, and it has not happened.  Many reasons lie
behind this, not the least of which is that it is in Microsoft's
interest to break compatibility.  So we need another solution.  Well
let us see what Linus Torvalds has to say on how to make writing
portable code easy (flame alert):

Well he is talking C, not JavaScript.  But the principle applies.  The
potential project that nobody really wants to do, which would really
make a difference, is a standard portable web library.  An API which
has been ported to major browsers, and which will be maintained, that
people can call and then program to.  Designing that will, of course,
be very hard.  But I think it is doable.  To get the full effect you
may need kludges like having Apache autodetect browsers and hand off
different stylesheets.  Gosh, darn, a reason not to use IIS. :-)

Of course the first thing that I did when faced with JavaScript was
figure out how to offload that task onto coworkers. :-)  However if
Tim really wants to see OSS embrace the web as a platform, this would
be a very good first step to advocate.