Re: [Classpathx-discuss] Activation implementation

[Nic Ferrier]

>I do seem to recall some criticism of that API as it happened,
>much along the lines you sketched:  "it would have been more
>sensible..." to make a general purpose MIME classes rather
>expect javax.mail.*, javax.servlet.*, and so on to each have
>one more implementation.  That's what a lot of people asked for
>at the time, but I think the javax.mail.* people were focussed
>on just their (primarily client side) part of the world.

Javamail sucks in many ways. It's a very poorly designed API and I
haven't found that the javamail people are very creative thinking
about changing it.

It's also a bit of a backwater, there just isn't much interest in
it. Many people use alternative APIs because of the difficulty of the
Sun version (and also, historically, performance was terrible).


One of the reasons I want a really good GNU alternative is so that I
can force the Sun Javamail team to take notice of some ideas about how
to get it changed. Maybe get a JSR setup for a major enhancement.


BTW part of the problem with the Javamail API is that it covers some
of the same ground as some of the things in java.net. And we all know
what a mess that is.


Nic