Software requirements

From: Rob Seaman <seaman_at_noao.edu>
Date: Wed, 21 Dec 2005 11:01:41 -0700

On Dec 21, 2005, at 3:06 AM, Poul-Henning Kamp wrote:

> It must be wonderful to live in a world where software can just be
> ignored or marginalized at whim, I really envy you.

Sarcasm, right?

Software IS a marginal activity - I say this as a programmer. There
is no point to software that doesn't address the external needs of
some class of users. Time standards represent just one set of
constraints placed on a wide range of software. No one is suggesting
software requirements be ignored, but those requirements must be
responsive to real world issues, and software professionals
themselves (like astronomers) are only one community that must inform
the standards process.

The world rotates. Its rotation is slowing. I won't belabor the
implications, but these are facts of nature that must be accommodated
in some fashion. Two options are currently being debated - leap
seconds or leap hours. The former establishes a relatively rapid
cadence of events with a small enough amplitude to allow each
instance to be ignored for many purposes (software included). The
latter permits a slow cadence (for the near term) that permits the
existence of the requirement to be ignored for long periods of time
for many, but not all, purposes. The amplitude, however, is much
larger and many more users (and their software) would have to be made
DUT1 aware. When the inevitable leap occurs, its larger amplitude
will represent a larger challenge for programmers and civilians
alike. The rare occurrence of events makes mistakes more likely.

Everybody on this mailing list obviously welcomes the opportunity to
discuss which option is better. Or perhaps we can establish a
consensus around some other proposal. This is more likely to happen
if the proposal is the product of all the folks on the list, not just
the lurkers. In any event, software requirements will better be met
by designing, implementing, documenting and maintaining a coherent
set of interfaces for transporting time signals of all types. It's
easier to focus on technical requirements when politics isn't allowed
to dominate the standards process.

Rob Seaman
National Optical Astronomy Observatory
Received on Wed Dec 21 2005 - 10:02:40 PST