Mailing-List: contact geronimo-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: geronimo-dev@incubator.apache.org
From: "Danny Angus" <danny@apache.org>
To: <geronimo-dev@incubator.apache.org>
Subject: RE: [JavaMail] concerns
Date: Tue, 19 Aug 2003 12:29:29 +0100
Message-ID: <HKEFKPNPJLANNFPFMDKJMEKKJBAA.danny@apache.org>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
In-Reply-To: <D3DA1CAE-D22E-11D7-870E-0003934D3EA4@ioshq.com>
Importance: Normal

Alex,

> I think that these comments are starting to get me down. I am at part 1
> of a 3-part development cycle:

Jeez I'm sorry about that :-( but the fact remains that I'm going to keep on
about it until people take an interest and we can get a decsion.

As it stands you and I disagree, and no-one else is interested, I'd like to
hear what others think about it.

> 1) Make the API available with complete signatures, so that developers
> can compile against it for their mail code
> 2) Fully implement the API
> 3) Write transports/stores
>
> It's on the published ToDo list, it's updated with where I am, and in
> the release of the initial code-base I said that I was working towards
> goal 3, but doing so via goals 1 and 2 first.

I'm not disputing this, at all.

> Personally, I'm now starting to wish that I hadn't released any code
> until I'd got to 3), which is essentially the one Danny is wanting.

I believe it is the one which is within the scope of the project.
I believe that there are important issues surrounding 2 which need to be
considered before this becomes part of geronimo and which from what I can
see have not been.

> However, in the spirit of release-early, release-often I thought it
> would be a good idea to release part of the code as it made sense to do
> so, in my published work plan and wiki documented to-do list.

Yes fine, but this is not a one man show, it is a community project, I'm not
stopping you from doing this just raising my concerns.
That is how Apache works.


> Further, now that the API is there, someone else (such as Danny :-) can
> start writing the transport/stores if he wants, because the foundations
> are now there for that to happen. However, if not, then it is something
> that I am working towards.

I can already write an implementation of javaMail without requiring an
Apache owned copy.
And my main point it that if this is so, why do we need a copy.

> The majority of the methods in the API are, in fact, complete. There
> are (give or take) about 800 methods, of which there are only 100 or so
> left; just under 12% to go to reach 2).

Right OK, fine, I'm not trying to stop you, just questioning the value of
this.
Do these 100 or so methods all represent the same amount of work as the 800
already written?

> I'm also curious as to why you keep using javaMail when the trademark
> is in fact JavaMail?

I know that, but I keep mistyping it.

d.