axis1.x->axis2的最重要特征是:
hot deployment(axis2 addresses技术)
hot update
但我个人却对hot deployment/update不感冒,因此我仍然抱着xfire不放.
很多人都认为axis2是垃圾代码产物, 打开axis2的team页:http://ws.apache.org/axis2/team-list.html
,你发现wso2几乎主导了整个axis2设计, wso2据说是一个斯里兰卡的公司, axis2好多都是由学生式代码堆砌的, 可以用非常烂来形容.
http://www.jroller.com/page/fate?entry=axis2_why_bother
有些人被apache的不纯洁感到忧虑, wto2的主导者davanum srinivas本身是一个顾问, 他是否想通过axis2获取更多的顾问费, 不得而知了.
用江南白衣的话来形容axis2, "那是一点都不pojo,不spring!"
最后, 看看denis robert是如何批斗axis2的:
no question about it, stick with xfire. you’ll be
happy about it. my only gripe with xfire is the docs,
which are woefully incomplete. hopefully that will
change with time. for the time being, you have to
plow through the source for any complex service.
but architecturally, it’s really sound.
axis2 is a nightmare. even with xfire’s incomplete
docs, i was able to go through the source to figure
out what i needed. axis2 is such a jumble of code that
doing the same thing would take weeks, not hours.
also, compared to axis2, xfire’s docs are positively
brilliant! not only are axis2’s docs fragmentary
at best, half of it doesn’t correpond to the current
version.
xfire looks like it’s going in the right direction,
and dan diephouse (the lead) seems like he’s on top
of the project.
you also have to take jax-ws into account. whether or
not it’s all it’s cracked up to be is another
discussion, but it nevertheless is the official standard.
the axis2 team have made clear that they have
no intention of supporting it. jax-rpc was horrible,
but it was at least common ground, and was the api
used by most enterprise users. same will end up happening
with jax-ws and jaxb 2. websphere users will
end up using that, and knowing it’s out there will
make interop a lot easier. xfire has taken a “can’t
beat ‘em, join ‘em” approach here.
the way i see it, the axis team dropped the ball on
this one, and the new kid has taken the lead.
it’s the circle of life…
另外, tss的axis2讨论也非常激烈:
http://www.theserverside.com/news/thread.tss?thread_id=40280