commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Stansberry <bes_commons_...@yahoo.com>
Subject Re: [logging] distribution packaging
Date Wed, 09 Mar 2005 07:48:28 GMT
--- Ceki Gülcü <ceki@qos.ch> wrote:
> 
> Brian,
> 
>  From a cursory look at your document, I have to
> *speculate* that the
> changes you describe do not solve the core flaws in
> JCL but merely
> hide them by falling back on java.util.logging.
> However, I am only
> *speculating* as I have not had a chance to study
> your document with
> the care that it deserves.

This is definitely the case with JCL 1.0.5RC1; the
exceptions you noted were (largely) no longer thrown,
but JCL fell back on java.util.logging.

With the packaging changes, the logging was done using
the expected implementation.  I didn't just want to
rely on my noticing the formatting differences in the
log output between the different loggers, so the test
cases also log the classname of the Log wrapper.

> 
> After careful study of JCL, I am convinced that JCL
> is broken beyond
> hope. While its interfaces can be salvaged, its
> implementation must be
> thrown away entirely. While this opinion is not
> popular around here,
> it is based on verifiable facts, not wishful
> thinking that does not
> survive critical scrutiny.
> 

Perhaps coming down to a solution involving
distributing multiple different jars, teaching users
how to correctly deploy them, and then still having
some use cases where JCL's discovery mechanism doesn't
work qualifies as "broken beyond hope".  I'm actually
still somewhat on the fence on this one.

I think there are two issues here: 1) Does changing
packaging actually solve some of the identified
problems?  This issue can and should be resolved
empirically. 2) Is a proposed change so ugly/difficult
to understand/limited in effectiveness that it's
better to not bother and instead focus energy on a
more radical solution?  This is really a value
judgement that IMHO can best be resolved through
discussion within the community.  My goal to this
point has been to help clarify the empirical issues so
that any discussion of the value judgements could
proceed from a shared base of understanding.

Regarding the issues of politics you raise, I don't
really have the historical background to comment other
than to say that referring to your "Think Again"
article as a rant was uncalled for (and actually
detracts from the content of the wiki page). (OK,
someone out there, flame away :-) ).

<snip>
> 
> In late 1999, National Magazine published an article
> about a newly
> discovered Archaeoraptor fossil, calling it "a true
> missing link"
> demonstrating the relation between birds and
> dinosaurs, supposedly
> bringing to conclusion a debate raging since the
> 1860s.
> 
> When XU XING, a Chineese palaeontologist, declared
> that the
> "missing-link" fossil acquired by National
> Geographic was a fake, the
> illustrious magazine rechecked their facts and
> admitted their mistake.
> They had invested considerably in the article and
> had already checked
> their facts.  However, when XU XING's message
> arrived, they did not
> summarily dismiss it or ridicule his findings. They
> rechecked their
> facts. For the details of this fascinating story,
> please refer to [2].
> 

Thanks for this link.  I'd never heard this story. 
I'm  also a bit of a Sinologist, so the background on
the fossil trade in Liaoning is personally
interesting.


Brian


> Recently the ASF celebrated its 10th anniversary.
> IMHO, if the
> foundation is ever to celebrate its 100th
> anniversary, we better
> develop a better tradition for dealing with critical
> input.
> 
> [1]
>
http://wiki.apache.org/jakarta-commons/Commons_20Logging_20FUD
> [2]
>
http://www.bbc.co.uk/science/horizon/2001/dinofooltrans.shtml
> 
> On 2005-03-08 7:35:11, Brian Stansberry wrote:
> 
>  > I was a little surprised myself, which is why I
> wanted to follow
>  > Ceki's good example and publish test cases that
> could easily be
>  > verified (or debunked) by others.
> 
> 
> -- 
> Ceki Gülcü
> 
>    The complete log4j manual:
> http://www.qos.ch/log4j/
> 
> 
> 
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
> commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail:
> commons-dev-help@jakarta.apache.org
> 
> 



	
		
__________________________________ 
Celebrate Yahoo!'s 10th Birthday! 
Yahoo! Netrospective: 100 Moments of the Web 
http://birthday.yahoo.com/netrospective/

---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org


Mime
View raw message