...As I see it, these are the problems people have with the build process:
Nice analysis! Let my try to elaborate on it (just my 2 swiss centimes):
1) emotional shock at having to build at all.
Positive IMHO, helps people realize that Cocoon is a tool for *developers*.
2) java environment problems (jvm instead of jdk, JAVA_HOME not set, endorsed libs problem?)
You have the same problems when running Cocoon later on anyway, having to build it first makes no difference here AFAIK.
3) winzip empty libs problem
ok.
4) failure to read INSTALL.txt
People should do their homework, and maybe help us fix the docs if they're unclear (read the "getting started" tracks for example).
5) trouble deciphering the options in build.properties and blocks.properties
Agree on this one, that's where a web application (or a build based on a higher-level config file) could help. But generating local.*.properties files would be enough IMHO.
6) misinterpreting build output (warnings!=failure !)
ok, one for the docs I guess.
the web based builder idea helps with all of those, except postpones the problem with java environment (#2).
So maybe a web app that would explain (once again) the steps needed to run Cocoon and spit out local.*.properties files according to the user's options (by activating groups of related blocks) might be useful. More than that is not very practical IMHO.
I don't mean to sound cynical - if people are willing to build and maintain front-ends to the build system, why not.
But again, I don't think we want Cocoon to be a shrink-wrapped thing that fools people into thinking it is a point-and-click thing, only to have them miserably fail later on. Failing early is certainly better if people do not have the necessary skills/mindset.
Again, just my 2c.
-Bertrand
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
