I don't see your acknowledgement as being an acknowledgement of the
topic that Mike was talking about. When you break existing
implementations, even those implementations are of a draft version,
you do create problems for the people who are volunteering to live on
the bleeding edge --- and tend to be the best people to give you
practical feedback. There should be a good reason for this. Perhaps
we have a good reason here, perhaps not. But I heard you say on
Monday that implementations of drafts should not be a consideration
on how to proceed. I disagree --- it should be a consideration. Not
as big a consideration as Mike believes, perhaps, but a consideration
none the less.
eric
--On March 11, 2008 11:20:02 PM -0400 Dave Crocker
<dhc(_at_)dcrocker(_dot_)net>
wrote:
Eric Allman wrote:
Dave seems to be forgetting that early implementations of drafts
are a good way to get practical feedback into the process ---
it's more than a gedanken experiment. Mike seems to forget that
these /are/ drafts, and drafts do (and should) change.
Hi, Eric. Welcome back.
Perhaps you missed my noting:
> Basically, I believe you are confusing the benefit of getting
data from
> prototypes, versus the more extensive damage done by changing an
> installed global service.
which seems to acknowledge explicitly what you are saying I was
forgetting. It was certainly meant to.
Simply extending the logic at work against change, here, serves to
mandate absolutely no changes, as soon as anyone has written a
single line of code. Since I suspect no one really means that, what
is missing is the offering for how this particular change has an
impact that really is as disastrous as is being claimed.
Lest it be forgotten, the motivating text for this sub-thread was:
Michael Thomas wrote:
> Each time you change it, implementations break in a
> showstopper way.
Now perhaps it registers differently with you, but I'd class that
language as dire. And it is about a working group draft that has
just recently undergone massive change. (Just how recent that
change occurred is yet another item that seems to be getting
forgotten.)
Perhaps we can ratchet down the hyperbole and return to the
reasoned consideration we have recently been enjoying in the
working group?
d/
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html