Re: [Springframework-developer] Component Centric Web Framework for Spring

[Darren D. <da...@da...>]

On Sun, Mar 26, 2006 at 03:59:21PM -0500, Keith Donald wrote:
> What's most constraining about the Servlet API in your opinion?  Is it the
> API really? Or is something with the implementations of it that concern y=
ou?

I don't think the API itself is badly designed, or that it makes it
difficult to develop against.  Nor do I think the well known servlet
containers are badly implemented - quite the contrary in many cases.

The constraint is *requiring* a servlet container.  The general assumption
is that to run any Java based web framework, you need Tomcat (or similar) at
minimum - a reasonable assumption because it happens to hold true for most
frameworks.  With the servlet container comes dependencies I might not
want, configuration I can't be bothered with, and (re)deployment
issues I'd prefer to avoid.

I agree with Seth that this is something Java tends to do badly - a
necessary evil of the compilation step perhaps, but one that can be avoided.
The client I'm working for now has business users that are raising issues
over the length of time it takes the development team to implement minor
chages to their production web applications.  40 minutes is way too long
apparently.

> What problems do you believe need a scripting solution in the web-tier?

fix re-deployment and permit loose/dynamic typing.


> curious--it'd be interesting to consider integration of a scripting
> technology into the Web Flow engine, and where they might fit.

It's important to ensure that the scripting just works.  If there's a need
to start configuring beans, XML files or whatever just to be able to start
writing code, it's missing half the point.  I know Rob's pointed out that
the scripting is there now but I haven't had a chance to look at it yet and
I don't know what's involved in making it work.

Again, I agree with Seth about how easy we have to make it to start coding.
Anything less than you get now with the Ruby and Python frameworks is
missing the boat.  That means not relying on a separate container to deploy
into, but providing a minimal one that gets your scripts working instantly
(no problem if they want to later deploy to Tomcat or JBoss).
It means not expecting the user to choose and configure a templating
technology, but providing one pre-configured (fine - they can change it
themselves later).  It means not asking them to build an admin interface
around their model, but providing one that adapts to the model as they build
it (ok, let them swap it for a different JMX server later).

So, I was going to now come back to the servlet API and say that to ship a
compliant container would make it difficult to achieve a fast, light, zero
config startup option.  Then I came across this
<http://tjws.sourceforge.net/> which kind of blew my argument away.  (I
haven't tried it, but if it works, it could be perfect).

I'm going to give it a shot and see what I can do with the M2 script
support, a basic object mapper and TJWS. =20

--=20
Darren Davison
Public Key: 0xDD356B0D