commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jörg Schaible <joerg.schai...@scalaris.com>
Subject Re: [Math] What's the problem with interfaces?
Date Thu, 31 Mar 2011 15:38:56 GMT
Hi Gilles,

Gilles Sadowski wrote:

[snip]

> If I'm not mistaken, this is the answer to the "business requirement".
> And it does not prevent us from evolving the code as often as necessary
> ("math3", "math4", ...).

I think, this is the real point where we disagree. My interpretation of "as 
often as necessary" does not match my expectation of "a major version should 
be stable at least for 12 to 24 months". This is what I consider a "stable 
API".

Looking at the age of CM 2.0 it might be very well the time to move on with 
3.x. But if you do, you should work very hard before releasing final 3.x, 
because the API should not be changed (resp. should not break binary 
compatibility) for this period of time. If you're unsure how the new API 
works out, you can do a minor set of alphas and betas before - we did this 
for lang 3.x also. However, business will always wait for the final release 
and it expects the API to "be stable" then. And this is not even a technical 
problem, but a political.

[snip]

- Jörg

BTW: Gosh, we're so glad that we could finally drop in our company JDK 1.4 
support everywhere with end of last year. That's the business reality with 
global players in sensitive environments.



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org


Mime
View raw message