[slf4j-dev] Re: Beta 4 and the new method signatures

[Niclas Hedhman]

On Monday 11 July 2005 23:45, Greg Wilkins wrote:
> So the question is - is there such an underlying "thin" API? and is there
> enough need for it to make it worth the complexity of exposing it?

I belong in the camp that Greg's suggestions of a thin + a thick API makes 
more sense than not.

I like;
 *  Simple API
 *  addChild to Logger, which can be extremely powerful in many IoC scenarios,
    and somethat difficult to emulate.
 *  Loggers able to belong to more than one category hierarchy.

Other people like other things and claim they can't live without it. End of 
the day we are "bike shedding", and probably never totally agree on a single 
approach.


Mission Impossible; The Challenge, Jim, if you choose to accept it...

Create the union of the requests, that will make everyone happy :o)


I think Greg's idea was a decent start. I think the Logger creation/lookup 
should also have a thin layer (application servers, IoC patterns...) and a 
convenience layer (similar to what we have, Logger.getLogger equivalent(s) ).

I think we are reaching a point of either;
  1. Do nothing and stick to what we got.
  2. Choose some of the suggestions of extensions to what we got. Candidates
     being trace(), Markers, ...(?)...
  3. Try to accommodate a thin/thick and/or flexible/convenient layered API.


Personally, I am more inclined towards 3. and if it is considered out of the 
scope of SLF4J, then I am happy with either 1 or 2.
If we can get to some form of consensus at this level of granularity, we can 
carry on with more details. 


Cheers
Niclas