Re: [aspectj-users] Drawbacks of AOP

[Marc Tremblay <marc@xxxxxxxxxxxxx>]

On Thu, 2003-04-17 at 17:38, Arno Schmidmeier wrote:
> On Thursday 17 April 2003 22:54, Marc Tremblay wrote:
> > > > To me this is the kind of thing that should be handled declaratively in
> > > > a deployment descriptor and applied by some sort of runtime container
> > > > (not nesc. EJB, but along those lines).  To me, implementing things
> > > > like transaction management, security, etc. as an aspect leads to
> > > > hidden complexity and thereby can render the code unmaintainable.
> > >
> > > ...
> > >
> > > > I certainly agree with you that these crosscutting concerns don't
> > > > belong in the code itself, but I just think the deployment descriptor
> > > > approach is a better one than using aspects.  Just because we can solve
> > > > these problems with aspects, doesn't mean aspects are the best way to
> > > > solve these problems.  These concerns are the kind that should be
> > > > handled in a container or runtime framework in my opinion.
> > >
> > > That's identical, though. One could write an aspect that parses that
> > > exact same deployment descriptor and accomplish the same thing with an
> > > aspect.
> > >
> > > It's simply a matter of syntax for maintenance.
> >
> > It's identical with respect to the effect to be achieved.  However, it's
> > very different with respect to on the one hand I'm writing aspects to
> > manage these concerns, and on the other I'm relying on some 3rd party
> > container/framework provider to take care of it for me.  With the later I
> > really don't care how they handle these concerns so long as they do.
> >
> > I guess what I'm getting at is that I see the problems of transactions,
> > security, etc. as already being solved well enough that I'm not interested
> > in finding other solutions for them.
> 
> Where is the difference, between an aspect solution and an EJB or "container" 
> solution, when you simply use an aspect library (maybe from a third party)? 
> You simply have to write your "deploymentdescriptor" in an aspect-form and 
> you are done.
> However the aspect approach has some big advantages. 
> 1. You can simply reuse the pointcut definitions for other aspects. (As I 
> pointed out in my speech at AOSD 2002, infrastructure aspects tend to reuse 
> the same pointcuts.) 
> e.g. you could reuse the transaction-pointcuts for auditing or access control.
> 2. If you can not life with the transaction policy, persistence policy, etc. 
> you can simply switch the relevant aspect. You are not bound by a take it or 
> leave it decision as you are by current "container"-architectures.

Yeah, you're right.  Perhaps I'm being a little myopic in my view as in
my time developing transactional systems I've yet to encounter the need
for going beyond the transactional controls provided by various
frameworks/containers to need to tinker with the way in which
transactions are controlled.  Although, I must admit that this does not
mean that need cannot exist.

-- Marc