My biggest problem with pluggable Coc is that the same exact piece of
code could behave differently depending on how you deploy it. A new
developer, just coming to maintain a particular project, would have to
1st look at the configuration file to see how it behaved, then the
documentation of the CoC plugin.
Personally, I've always hated implicit "magical" behavior when having to
debug or refactor somebody elses code.
On 4/10/11 3:58 AM, Markus KARG wrote:
> Guilherme,
>
> with the target "CoC" im mind, looking at the *average user* of JAX-RS,
> I cannot find a better word than "rocket science": If a *user* would be
> clever enough to implement such an interface, he wouldn't have a need
> for CoC IMHO, since CoC in my experience is most appreciated not by
> *lazy* people but more by the "not-so-skilled" ones (in other words,
> users like CoC because they don't need to understand what's going on or
> what the correct syntax would be like ["it works somehow magically"],
> not because they do understand how to configure but just don't want to
> type the stuff in). That's why I think for *those* people (in my
> experience: the majority of average users) to get the largest benefit of
> our CoC efforts, the need for understanding such a complex interface
> would be experienced as being "rocket science" so they wouldn't use it
> at all. But if people don't use it largely, there is no justification to
> provide a standard for it. So it could be a really useful extension of
> your framework, but I just don't see that it is so wide-spread needed
> that we should define a standard for it. In my opinion, our CoC target
> should be to define that 80% of use cases that people would love to see
> a simple "best CoC guess" built into JAX-RS, not to define an API for
> the other 20% experts that just are too lazy to type. But that is just
> *my* opinion, maybe the other experts think different.
>
> Regards
>
> Markus
>
> *From:*guilherme.silveira_at_gmail.com
> [mailto:guilherme.silveira_at_gmail.com] *On Behalf Of *Guilherme Silveira
> *Sent:* Samstag, 9. April 2011 23:27
> *To:* jsr339-experts_at_jax-rs-spec.java.net
> *Subject:* [jsr339-experts] Re: Annotations CoC [Was: Convention Over
> Configuration]
>
> If its out of scope I can understand. But I disagree about its
> difficulties, or even rocket sciwnce. Extracting simple interfaces
> should be easier to do than agreeing whether a rest consumer should be
> bound to its server interface.
>
> Regards
>
> On 09/04/2011 4:34 PM, "Markus KARG" <markus_at_headcrashing.eu
> <mailto:markus_at_headcrashing.eu>> wrote:
>
> While obviously CoC is something that is really beneficial to the
> user's productivity, I do not think we could find a real
> justification for adding another complex API for a pluggable rule
> set (it would be counter-productive since it is another thing to
> learn and support). There are other items besides CoC of much higher
> and global interest in our stack, so we should keep things simple in
> this area and just think over a few places where CoC makes really
> sense instead of doing rocket science.
>
>
> > -----Original Message-----
> > From: Marek Potociar [mailto:marek.potociar_at_oracle.com
> <mailto:marek.potociar_at_oracle.com>]
> > Sent: Frei...
>
--
Bill Burke
JBoss, a division of Red Hat
http://bill.burkecentral.com