Re: [Springframework-developer] Release plan

[<rod...@in...>]

>Rod, are you already on your way to Boston (you told me that 
you flew today)? ;-) 
I'm leaving for the airport at 1500.

>So we seem to have decided to put out release 0.9 tonight. 
>I'll do the actual release creation at SourceForge, I guess. 
>Please add any stuff that you want to have in it today, 
>especially docs and readmes. 
Unfortunately I won't have time to contribute anything today, 
besides the high-level stuff I've already put in the wiki.

>We will still use com.interface21, announcing a change to 
>org.springframework in the main readme. This will 
undoubtedly be plausible, if we mention the roots in Rod's 
book. IMO, we don't need to hurry in terms of the package 
name change: I propose to switch when we reach 1.0 Release 
Candidate state, not in a 0.9.x release (if there will be any 
further 0.9s before 1.0 RCs at all). 

I like this. I think Spring is mature enough to go 
0.9/1.0RC1. Further 0.9 releases should only be necessary if 
any significant bugs/omissions are apparent.

<jh>
For the first 1.0 RC, our main tasks will be docs and 
examples (besides the package name change). I'll do further 
polishing of the Javadocs (already today, as far as 
possible), and provide short essays on resource/transaction 
management and Spring web MVC vs Struts (as announced, I 
already have drafts). 
</jh>
The stuff you've posted to this list, Hibernate forums etc is 
great. Pulling this together in one place will be an 
excellent start.

I'll try to contribute a fuller discussion of AOP before 1.0.

<jh>
As far as I see, we don't need additional functionality for 
1.0 - JMS and Web Services can wait until 1.1. Let's rather 
target a date as early as possible: 1.0 RC1 at the end of 
July should be feasible, probably releasing 1.0 final at the 
end of August. Is this an acceptable plan for everyone? Any 
differing proposals? 
</jh>
Agreed regarding JMS and web services. Spring does enough 
right now--we need to start selling it. Dates sound good. One 
area I'm interested in is metadata attribute support as an 
option for AOP. Also a special factory that can detect 
whether a target object has metadata tags such as 
transaction/read-only tags and can generate the appropriate 
interceptor chain, based on factories registered in the root 
bean factory. (This should ensure that developers can simply 
use a common entity reference for most tasks.) 

This should simplify AOP usage for typical scenarios quite a 
bit. I think this is important before 1.0. It should be a lot 
simpler than other AOP frameworks I'm familiar with, and 
almost comparable to .NET.

Btw, I'm looking for a volunteer to help with the 
AOP/metadata enhancements, as I have a pretty clear vision of 
how it should work, but I'm not sure I

The AOP Alliance has kicked off again, so there will be 
package changes and some other changes that we'll have to 
incorporate before 1.0 RC1 too.

Btw, what's the current test coverage? (I had a JDBC version 
problem running Clover last night and didn't have time to 
sort it out.)

Regards,
Rod