RE: [aspectj-users] AspectJ and JBoss AOP implementation

["Gregor Kiczales" <gregor@xxxxxxxxx>]

Ruud Steeghs said:

> I don't think that learning the implementation (JBoss,
> AspectJ) is the hardest part. Think that really
> appreciating the value of a new methodoloy / paradigm
> is the hardest part. How long did it take you to fully
> understand OOP????


Charles Zhang said:

> Can someone tell a real learning experience comparison? I myself
> haven't used JBOSS AOP too much and am curious about it.


Macneil Shonle said:

> They don't mind learning new vocabulary, but they should 
> probably be given a short introduction to the big ideas
> before tackling something like "Getting Started with AspectJ."

The above points seem to highlight different aspects of the same
issue -- learning any one AOP tool involves learning both AOP and
that tool.

So

 - its nice to have a gradual introduction, because you
   have to learn AOP and Aspect(J | Werkz) | JBoss both.

 - its important to focus on AOP design qua itself, as
   opposed to details of the specific tool. While what
   the tutorial and the books say mostly applies to AOP
   broadly construed, more could perhaps be made of this

- whatever tool you learn first will frame your
  ontology for all of AOP, so it will affect how
  you see and learn future tools. This last point
  makes me wish people learned AspectJ first,
  even if they then moved to AspectWerkz or JBoss,
  because right now, AspectJ is the only one that
  has the crucial issue of orthogonality of
  join point models right. I personally think that
  fact is one of the most important elements of
  the AspectJ designed. The fact that's its little
  noticed speaks to its importance IMHO.

- it may be time for an expose that really A/B/C
  compares the tools, in concrete code, as Wes
  suggests. This might help to draw the parallels
  and issues.  Hmm... maybe I feel a column topic coming!