SV: [jdom-interest] Announcing JDOM b10-rc1

[Per Norrman]

Hi, 

som feedback (late in the game ...)

I have seen the Parent/Child stuff in CVS but never actually looked at 
the changes and consequences in detail. Until now.

To me, the Parent/Child/Container/Content/whatever stuff is really of no

practical use. Over the years (four, isn't it?) I've acquired a certain
JDOM
coding style, which is now suddenly subliminally altered. There was once
a vision,
an interface-less vision, which is now blurred. Funny thing, when I
study my 
collective JDOM code, I never seem to treat a Document node the same as
an 
Element node! In fact 99.99% (or close) of my processing is concerned
with either an 
Element, an Attribute or a Text node. I think there isn't enough
commonality 
to justify a Parent abstraction, at least not in the practical cases.

I really think these changes are going to make more harm than good,
considering 
the (long) history of the API.

As an experiment, I took the beta-10 code and applied it to an older
project, started 
with beta-7 and where beta-9 worked OK. This code does some really
intricate navigation inside a rather large, validated, document.
Element#getParent() not being an Element 
really blowed that piece of code away!! 

As for the method names, getChildren/getChildElements, etc: this issue
is of minor
importance, considering the modern IDEs (as long as the behaviour is
the same).
For the record, I would prefer getChildElements (or even
childElements--the JavaBeans
naming standard is overused).

I think this is my point: Stick with the initial vision, follow through
with the 
consequences. You are never going to achieve the perfect API**. Skip
Parent/Child!

Two other issues, perhaps on a smaller scale:

*) Did you consider the change I proposed to XMLOutputter?
http://www.servlets.com/archive/servlet/ReadMsg?msgId=447565&listName=jd
om-interest


*) Why can't I *get* the current Format from XMLOutputter?

**
Or maybe this is the "Quest for the Holy API".  We are are all going to
be arrested
on a Scottish moor for believing in an ephemeral computational vision,
thus being a 
threat to the IT society as a virtual impossibility ....

/pmn


> -----Ursprungligt meddelande-----
> Från: jdom-interest-admin at jdom.org 
> [mailto:jdom-interest-admin at jdom.org] För Jason Hunter
> Skickat: den 6 februari 2004 10:09
> Till: jdom-interest at jdom.org
> Ämne: [jdom-interest] Announcing JDOM b10-rc1
> 
> 
> Good news!
> 
> After today's work I'm proud to report there are NO known API changes 
> required before our 1.0 release.  What we have now is my best effort 
> at a 1.0 release API.
> 
> We may still have bugs or inconsistencies, but the TODO is wiped clean
> of everything I personally deem important enough to stand in 
> the way of 
> a long, long, long awaited 1.0 release.
> 
> My goal is to get us to a 1.0 release in the next 6 weeks. Here's my 
> plan.
> 
> I've posted on jdom.org a Beta10 Release Candidate #1.  It's built 
> from the very latest CVS source.  You can get it at
> http://jdom.org/dist/source/jdom-b10-rc1.zip.
> 
> I'd like people who are using b9 to download this and see how things 
> work.  Please double check everything's still prim and proper and 
> performant.  The code passes all the jdom-test cases, but they don't 
> have great coverage (although they did help me find a bug earlier 
> tonight).  Subclassers: I'm especially interested in hearing your 
> feedback since we restricted a lot of visibilities.
> 
> Then I'd like people who've been involved in the API design to 
> thoroughly scour the new Javadocs (online also at 
> http://jdom.org/docs/apidocs/).  Do all the new structures make sense?
> Is everything consistent?
> 
> Once we're satisfied this b10-rc1 code and interface is correct (in 
> perhaps one to two weeks), we'll ship the formal Beta10.  After that, 
> we'll get started on the 1.0 Release Candidates.  The first 1.0-rc1 
> will be Beta10 minus the deprecated methods.  (We need to ship 1.0 
> without any deprecations.)  We'll try to spread the word to everyone
> that 1.0 is 
> coming and make sure people test their code against it.  We'll have a 
> month to iron out any issues.
> 
> Throughout this process, API corrections will be more important than 
> code corrections.  I'm happy to ship a 1.0.1 with code fixes, but I 
> want the 1.0 API as solid as possible.  We've done a bunch of
> changes since 
> b9, so there's a lot to review.
> 
> Sound good?  When we ship 1.0, I'll collect all those beers.  :-)
> 
> -jh-
> 
> _______________________________________________
> To control your jdom-interest membership: 
> http://lists.denveronline.net/mailman/options/jdom-interest/yo
uraddr at yourhost.com