jsr339-experts@jax-rs-spec.java.net

[jsr339-experts] Re: UriBuilder, forward slashes, path and segments

From: Markus KARG <markus_at_headcrashing.eu>
Date: Mon, 4 Jun 2012 19:59:31 +0200

Bill,

when you're finished with being pissed-off, swearing and self-portrayal
please let me know, so we can come back to the technical discussion. You
misinterpreted my words and just read what you liked or assumed to read.

I never complained that anybody here is political or not having enough
insight. I solely talked about the fact that *depite* you all being in a
daily discussion *with* users, you *do* have a technical vendor view. A
vendor *cannot* change this (I am a vendor myself, so I do know). This
cannot be changed unless you are not a vendor anymore (and I once switched
sides and was astonished how the world *actually* looks from the other side
of the table). The problem is not missing qualification. The problem is *too
much* technical qualification. JAX-RS Users are much simpler than you all
think. No, I do not talk about those heaving read your book and having
booked your consulting services. I talk about all the others that DID NOT.
*That* I wanted to say. I do not see any insulting here but just telling
social facts about silent masses of not-so-clever JAX-RS users. One cannot
drop his knowledge, and in this case, having knowledge is a key problem:
With too much knowledge, one just cannot see the normal user's actual
problem with the slash. The french king once said "If the people doesn't
have bread, they shall eat cake!" and he did not ever come to the conclusion
that it was ignorant -- he just didn't expect how less (here: knowledge
about URL templating rules) one can have.

About the slash: If you do not want to encode the slash, either use path()
or buildFromEncoded(). Neither of these will encode. But build() *must*
encode, as build() provides CONTENT and CONTENT MUST BE ENCODED ALWAYS. If
you want a slash, the sole reason can be that you DO NOT WANT CONTENT. So do
not use build().

I cannot see what you think is arrogant or ignorant with that.

Regards
Markus



> -----Original Message-----
> From: Bill Burke [mailto:bburke_at_redhat.com]
> Sent: Montag, 4. Juni 2012 18:52
> To: jsr339-experts_at_jax-rs-spec.java.net
> Subject: [jsr339-experts] Re: UriBuilder, forward slashes, path and
> segments
>
>
>
> On 6/3/12 4:26 AM, Markus KARG wrote:
> > As I stated months back already, it seems some members should look
> the
> > API more through the eyes of Java programmers and less through the
> > eyes of implementation vendors or web gurus. Many of the past
> > discussions would not have happened if people would try to make
> things
> > as most simple for the users of the spec, instead of making most of
> their own position.
> >
>
> I find your comments extremely insulting, arrogant, and well, extremely
> ignorant.
>
> Many vendor experts on this specification, specificaly, the most active
> ones (Marek & Co. (Oracle/Jersey), Sergey (Apache CXF) and myself (Red
> Hat/Resteasy)) are not sales guys, nor is our primary job description
> sitting on JSR committees. The 3 "vendors" i mentioned are the primary
> developers of their perspective implementations. Implementations that
> are open source. We answer questions on dev/user lists/forms. We
> schedule and fix the majority of bugs and feature requests asked for by
> users. We interact with users *ON A DAILY BASIS*.
>
> For myself, I've done a number of presentations on JAX-RS, written a
> book on it, and have done a handful of short-term consulting gigs on
> JAX-RS for customers. I've also written a few services and
> applications on top of JAX-RS/Resteasy.
>
> I've been on a few different JSRs over the years, and this
> specification, by far, has the most qualified experts of any JSR I've
> ever been on. Mainly because of all the characteristics I mentioned
> above, but I'll add another. It is by far, the *least* political spec
> I've ever worked on. I'll go as far as saying that there is *no*
> politics in this group and I really commend Marek for this.
>
> So, do you think that we might have *some insight* on things? Any?
>
> As far as the '/' goes:
>
> * I've had the need to *not* encode '/' included in path parameters to
> UriBuilder in a few of the services I've written on top of JAX-RS.
> Specifically when using it to build from uri-templates included in
> links.
> * I've had users complain when we didn't automatically encode '/'
> path-parameters in our client proxy framework. I had different users
> complain when we changed it to automatically encode '/'.
>
> So, between build() and buildFromEncoded(), I'm hoping one of them will
> automatically encode '/' and the other won't.
>
> Moral of the story?
>
> I've HAD EXPERIENCE WITH THE ISSUE. So instead of just blowing me off
> because I'm a "vendor", maybe consider actually reading my concerns
> instead of attacking my integrity.
>
>
>
> --
> Bill Burke
> JBoss, a division of Red Hat
> http://bill.burkecentral.com