On 8/29/06, Eric Redmond <[EMAIL PROTECTED]> wrote:
Here are some that I have heard in the past: * Lack of good documentation. * Community unwilling to help me with my problems. * Not "industry supported" or "mainstream" enough. * I don't like conforming to the Maven project layout. * My project is too complex to switch. * There are not enough plugins available. * We already have a large investement in tool X. * I have to build native/non-Java code. Any more reasons? Care to expand these ideas?
Things that keep getting in my way: - IDE support!! (or lack thereof), apparently the m2eclipse team can't seem to get a decent Maven embedder to build on, and the embedder seems to run differently than m2. Getting consistent builds from both an IDE (without using external tools) and automated build tool is problematic - plugins aren't mature enough, too many bugs, not enough documentation, or it's hard to find the documentation because you're not sure if the problem is in maven core or a maven plugin, so you wander all over the place looking for answers - M2 does not seem ready for what I would call casual users, those who want to just download, setup some basic config and go. If I may make a poor and overdone analogy, but it's more like Linux than Mac--more figuring out stuff on your own with lots of config of disconnected parts instead of "it just works" all together and integrated. - seems like archetypes should be used more and easier to use, but they're not. There seems to be a lack of good working example projects (beyond the trivial), though maybe I'm just not looking in the right places - for some corporate developers surrounded by firewalls, proxies, locked-down environments, bureaucracy, inane policies, access controls all over the place, non-standard configurations, lots of integration with legacy stuff, etc., you just can't quite get enough set up easily before you give up and decide to try again in 6 months :( M2 seems to work well for simple and straightforward projects where the basic POM is mostly sufficient. For those very large and complicated projects there seems to be enough win to actually dive in and spend the time getting to know M2 intimately so the ROI is worth it. Most of the projects I work on are somewhere in between, just complicated enough that they don't work out of the box, but not quite complicated enough to justify spending the hours it takes to figure out how to get it to work in our particular environment. I understand it's open source and I don't like complaining when I'm not able to step in and contribute to the solutions, so I end up just not using it much even though I want to. Over time, I'm sure I'll use it more as I learn it bit by bit or the functionality gets more robust. Logan