Skip to Content.
Sympa Menu

xom-interest - Re: [XOM-interest] Xom and Jaxen dependency

xom-interest AT lists.ibiblio.org

Subject: XOM API for Processing XML with Java

List archive

Chronological Thread  
  • From: "Steve Loughran" <steve.loughran AT gmail.com>
  • To: "David Bullock" <db AT dawnbreaks.net>
  • Cc: xom-interest <xom-interest AT lists.ibiblio.org>
  • Subject: Re: [XOM-interest] Xom and Jaxen dependency
  • Date: Wed, 15 Nov 2006 16:07:49 +0000

On 15/11/06, David Bullock <db AT dawnbreaks.net> wrote:
On 11/16/06, Elliotte Harold <elharo AT metalab.unc.edu> wrote about Maven, a
build tool, in respect of a build issue:

Maven's pretty bad at this stuff.


... and the irony isn't sweet. Java still needs a good build tool.

I'm not using Maven; I'm using Ant1.7RC1+Ivy1.4.1; as of last week Ivy
is in the Apache incubator. While it has lots of issues, and depends
on the maven POMs for a lot of metadata, it does have support for
plug-in conflict managers which let you implement alternate conflict
management algorithms above and beyond Maven's breadth-first-search.

A more fundamental problem with Java is it has a fairly brittle
dependency mechanism, and no official network artifact repository. The
Maven repository with it is metadata is the best we have, and as
anyone who has explored the repository knows, its woefully
inconsistent. While individual projects may upload POM files that are
valid to the extent they include at least artifacts as needed to build
a component, they are flawed in that they often include too many
files, not enough files to run, or just simply the wrong version.

the flaw there is perhaps the notion that individually published
metadata, be it XML or RDF, doesnt necessarily scale to form a
coherent "semantic web". The fact that the maven repository doesnt do
rigorous validation of metadata before publishing the artifacts merely
makes this problem more apparent.

-steve




Archive powered by MHonArc 2.6.24.

Top of Page