Delivered-To: cgu@qos.ch
X-Mailer: exmh version 2.0.3
To: cgu@qos.ch
Subject: Re: Fwd: JSR47 vs. log4j
Date: Wed, 13 Jun 2001 21:40:35 -0400
From: "Daniel F. Savarese" <dfs NOSPAM savarese.org>


>>I also feel, that rather than reinventing the wheel the JSR47 group should
>>incorporate the log4j API rather than recreating it.  The standards put forth
>>by the JSR group will eventually replace any similarly functioning external A
>PI
>>due to its incorporation into the JDK.  For the number of people who already
>>use log4j, such as myself, it feels that Sun by way of including this new API
>>in the JDK is forcing users of log4j to switch.  I understand that there are

This is just business as usual at Sun.  The position may sound extreme,
but the continued marginalization of third party Java solutions is
destructive to the Java development community.  It could almost be
justified if the stuff that made its way into the core or a standard
was actually better than what was already out there, but most of the
time it's not and blatantly ignores the lessons learned by those who
have studied the problem for far longer and had the benefit of many
years of deployment and user feedback.  A one year JSR working group
operating in a vacuum generally produces results that are inferior to a
solution with 4+ years of evolution.

Disclaimer: These are general comments that I do not intend to specifically
apply to JSR-47.  I have not reviewed JSR-47 or compared it to log4j, so
I cannot comment on its quality.  Nonetheless, the fact that none of the
developers of the most popular Java logging API were invited to serve as
a working group expert is a gross oversight.  I can say the same thing
about several other JSRs.

I hate to be such a complainer, but it really makes you feel helpless.
The only recourse we have is to constantly poll the JCP site and provide
as much feedback as possible when a JSR draft is released, which is usually
too late to make any significant changes.  Here's an article providing a
general criticism about the Java Community Process:
  http://www.zdnet.com/eweek/stories/general/0,11011,2666270,00.html

daniel