ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Donald <>
Subject Re: JSR for Ant (was RE: Ant 1.9 & Task Packaging)
Date Wed, 23 Jan 2002 08:21:41 GMT
On Wed, 23 Jan 2002 05:51, Tim Dawson wrote:
> I don't know all of the pros/cons, but JSR for Ant seems fairly obvious to
> me -- the JSR process is being used to standardize all manner of 3rd party
> Java tools/servers, why not something as common as a build system? This
> JSR/Jakarta tie-in seems to work fairly well for Tomcat, and the taglibs
> project has welcomed with open arms JSR-052 (standard taglib) with many of
> the committers (myself included) becoming members of the expert committee,
> as well as hosting the RI. So why not do this for Ant?

JSRs can be useful but Ant is no where near close enough to being ready for 
that. If it had a more mature featureset then maybe ... however it is 
difficult to see the need of having multiple containers. And without the need 
for multiple containes, making it a JSR would only be a politicial move. Some 
other JSRs have still gone this path but that is mainly so that the vendors 
who supply the product/implementation can say they "conform to a standard". 
It is a nice way of bamboozalling customers but I don't think ant needs to do 
that. Surely we can stand on technical merits alone.

JSRs can be good or bad things to go into. It all comes down to the skill of 
the spec lead. One JSR I was involved with was meant to be a simple 2 week 
job but the spec lead botched it and it is for all intents and purposes dead 
a year ago. Another JSR I am still involved with has a brilliant spec lead 
and the EG members are way mort smarter than I. I am surprised to see the 
rapid progression made even when different EG members seem to be aiming for 
completely different things.

Anyhow even if a JSR could be established and pushed through the JCP, Im not 
sure there would be any one who has the manner and skill to make a spec lead. 
Also the requirements are relativly high - especially if no one is getting 
paid to do the mind numbing edditing and TCK generation and so forth.

> One of the stated goals of Ant2 is to be useful in multiple contexts, e.g.
> command line, IDE, autobuild tools, etc. A JSR process would get direct
> input from 3rd party IDE vendors to ensure that Ant is adopted across those
> tools.

I have been contacted off list by quite a few different vendors. Even some 
that do not make tools but wanted to embed ant into other environments. I am 
sure that many of these vendors would be very very happy to give you their 
comments if you were to provide them with documents describing proposals for 
features pertinent to them - namely the embeddor interface and the TaskInfo 

Theres also quite a few opensource IDEs that we could very easily talk to if 
anybody wanted. No JSR needed for that.



"Science is like sex: sometimes something useful comes out, 
but that is not the reason we are doing it" -- Richard Feynman

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message