"Lee Meador" <[EMAIL PROTECTED]> writes:
> If XML is the problem for Maven, why is it not for Ant. Can anyone claim
> that a makefile's syntax is any easier to understand? In addition, make
> isn't procedural or sequential and that didn't, back in the day, generate
> loads of comments.
>
> It's not that you really were arguing the "pro" side of those arguments, But
> IMO the arguments about how 1) XML is a pain and 2) Maven is not procedural
> are not truly problems but for some reason they bother people in the context
> of whatever is really bothering people. (I don't claim to know what that is
> but it seems to exist.)
>
As you have noted, I did not argue about the truthfulness of the
arguments about XML's nature and maven's underlying process. I just
noticed that:
1. people - or at least some of them - I have in my training sessions
generally find it painful to
wirte a pom by hand. I suspect they also find it painful to write
an ant script by hand...
2. they have a hard time understanding all the "magic" behind $> mvn
install when they are used to $> ant all or $> make all. In the
latter case, they can track in the build file the sequence of
build procedures undertaken, while in the former, everything is
built-in in the plugin/lifecycle binding.
What I was considering to solve this "marketability" issues is:
1. wrap maven's pom in some nicer language, maybe something vaguely
looking like rake or scons script. For example, a minimal POM
would look like:
id=toto:tutu:1.0
dependencies=junit:junit:3.8.1,mygroup:mylib:[EMAIL PROTECTED]
Of cours, configuring plugins would clutter things up but you get
the idea. My underlying assumption about this is that maven may be
wrapped as some simple functional language.
2. output the build plan at the start/end of the execution of
maven. Hudson does a nice trick when building a project, it wraps
MavenPluginManager to display all executed mojos. Surely the build
plan features of maven2.1 would give better output.
And if you want my personal feeling, I think that XML really sucks as
a way for human to write something,
whether for ant, maven, xslt or anything else and exists just because
the technology for manipulating easily small languages is not
widespread or widely available.
But of course, all this is my own small point of view. YMMV...
Regards,
--
OQube < software engineering \ génie logiciel >
Arnaud Bailly, Dr.
\web> http://www.oqube.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
