From: Dave Chapman <dave_at_dchapman.com>
Date: Wed, 24 Mar 2010 00:17:02 +0000

Paul Louden wrote:
> On 3/23/2010 3:13 PM, Mike Giacomelli wrote:
>> Anyway, without a rational the NODO list is kind of bizarre. We
>> should have a
>
> So how do you give an objective rationale why one feature is worth the
> size, and another isn't? Some things will have to be subjective, it's
> just a given.

I agree that giving a rationale for some no-do items is going to be very
hard - especially when Rockbox runs on such a wide range of hardware. A
feature that uses 100KB of RAM is obviously unlikely to be acceptable on
2MB targets, but will have far less impact our 64MB targets.

In my view, binsize shouldn't generally be used as an argument against
new features. Instead, the argument should be that we don't want the
added complication to the code. Often these two go hand in hand.

But having said that, I fully agree with Frank's proposal - the barrier
to entry of the "no-do" list should be high, the reasons transparent,
and the collective opnions of developers on the no-do items should be
sought regularly (devcon seems ideal).

Dave.
Received on 2010-03-24