[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [xmlblaster] Poll for future xmlBlaster direction

Hi all,

Marcel Ruff wrote:

for the next steps of xmlBlaster evolution we need some feedback.
First some basic issues.

Please vote on the different topics if you have an opinion about
it or if your environment dictates it.

The Xmas 1.0 xmlBlaster release is JDK 1.3 compatible on server side.
It will only be patched for bug fixes and will remain JDK 1.3 compatible.
Typical version numberings will be 1.01, 1.02.

The new development will be on the main trunk and will lead to release 1.1

o JDK server side:
  Which JDK should we support on server side, do we need support for

  - JDK 1.3
  - JDK 1.4
  - JDK 1.5

  We would like to drop JDK 1.3 as JDK 1.4 offers SSL, logging,
  scalable socket IO and many more goodies.
  If we can drop JDK 1.4 support we can use the nice JDK 1.5
  features like built in JMX, concurrency and templates.

I would suggest to only keep JDK 1.5 support for release 1.1. This for three reasons:
- we now have a stable release 1.0 which is compatible with 1.2-1.5 (so users who are unhappy with JDK 1.5 will still have todays release)
- upgrading to higher JDK happens fast (probably faster than our releases)
- easier to mantain, particularly for testing and releasing.

o JDK client side If we change to the Java logging API we need for client side at least JDK 1.4 as well (Note we have a tiny J2ME and Applet based Java client lib as well which won't use any advanced java constructs).

  - JDK 1.2
  - JDK 1.3
  - JDK 1.4
  - JDK 1.5

I think the first thing to do here is to cleanly split the client- from the serverside packages. Only then we can consider having different jdks for client- and server side. When deciding which jdk to support clientside we shall look at the utility packages used on both sides (I will refere to them as 'common utilities'). As Heinrich stated it may be important to still support JDK1.3, but if support of JDK1.3 in the 'common utilities' would imply limitations on the serverside, then I would suggest to drop 1.3 on clientside too (here again there is rel. 1.0 making a good job).

o Logging API If we change to JDK 1.4 we should/could use JDK1.4's java.util.logging framework. This will be the standard in future as it is delivered with the JDK. To be more open there are Apache's logging interfaces http://jakarta.apache.org/commons/logging/ Which logging should we use?

  - Keep existing xmlBlaster logging interface?
  - Jakarta commons logging Interfaces?
  - JDK 1.4 java.util.logging?
  - Other approach (http://java-source.net/open-source/logging ...)?

Mmm, I am not sure on this issue. I really would like to use java.util.logging since it is where it should belong but I would first be sure that wrapping/redirecting other logging is strightforward. Somebody who can share experiences on that ?

  Using JDK 1.5 will enable for us simple JMX support
  Is this reason enough to go to JDK 1.5 or should we
  use third party JMX libraries, which do you recommend?

That again is a further aspect making it intresting to focus on JDK 1.5


Please post your opinions,


Michele Laghi
mailto:laghi at swissinfo.org
tel. +46 8 7492952 / mob. +46 70 4103964
spype phone: mlaghi
http://edis.name http://www.xmlBlaster.org