Date: Sat, 8 May 1999 17:17:34 +1000
Reply-To: Tim Churches <tchur@bigpond.com>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: Tim Churches <tchur@BIGPOND.COM>
Subject: Deadlines and software development (was SAS Institute - Sanity
vs. Insanity)
Content-Type: text/plain; charset="iso-8859-1"
Jack Hamilton <JackHamilton@FIRSTHEALTH.COM> wrote:
> It's hard to say that longer working hours will produce software more
> quickly. In the short run, it will, yes, but at the risk of employee
> burnout. The end result will be lower quality software produced more
> slowly. I have heard of (but not seen) studies in Europe which show that
> reduced work time actually results in more productive workers - not that
> many US employers subscribe to such a heretical notion. SAS Institute
> may be one of the few to do the right thing.
Yes and no. Producing good software (in a reasonable time frame) requires a
mixture of inspiration and perspiration. There is no doubt that exhausted
workers will produce uninspired, buggy software. But a lot of the time
perspiration is required, which usually equates to hours spent staring at
the little screen. You can't seriously tell me that a software development
project can't be completed more quickly if everyone involved works 50 hrs
per week rather than 37.5, particularly if those extra hours are spent
testing (and yes, EVERYONE should be involved in testing, not just the QA
department - they should just supervise it). One wouldn't want to spend
one's whole life working 50 hours per week or more (although I and nearly
all my friends and colleagues seem to have done just that since finishing
uni...), but in order to meet a deadline, working longer hours for a few
months at a time won't hurt anyone. My background is in health care, where
one just doesn't go home until everything is done (and checked, and
reviewed) and every second weekend is spent working or on call. Mind you,
that way madness lays, but short periods of longer hours to meet deadlines
are not harmful, methinks. Could it be that at SI, because of the guaranteed
revenue stream from the SAS licensing model, many managers don't have any
sense of urgency or even a definite deadline for finishing V8? There is
ample evidence that SAS software is extremely well designed (from the
software engineering point of view) and that all development is very
carefully planned and executed and very extensive testing is done at every
stage - see http://www.sas.com/corporate/quality.pdf for details of their
development process. Sure they might encounter a few unexpected roadblocks
but given the carefully planned process and the evolutionary rather than
revolutionary nature of the enhancements made in each new version, I
strongly expect that development is a fairly orderly process which could be
made to go faster with more resources - either longer hours or more staff.
Given that we are all paying license renewal fees in the meantime, I don't
think that it is unreasonable to expect SI to throw more resources at the
problem of finishing V8 so that it is done sooner rather than when they get
around to it, even if there are diminishing returns for each quantum of
extra resources which they devote to the task - that is their problem, not
ours. Hiring more staff or paying some overtime isn't going to send SI to
the wall. Note that I am NOT suggesting that SI lower or compromise their
quality standards in order to get V8 out the door sooner. What I am
suggesting is that they apply more resources to the task of implementing
those quality standards in order to get V8 out the door sooner.
> I know many people at SAS Institute who have gone out of their way to
> provide top-notch support. Some of those people are also developers.
> It's not fair to accuse them of being slackers on the basis of a magazine
> article. Large-scale software development is complex and difficult to
> predict. There's no reason to think that version 7 or version 8 would
> have been out sooner if they'd just worked a little longer.
I agree, the quality of SAS tech support (at least here in Oz) is superb and
the people here are definitely not slackers. Nor do I think that all (or
possibly any) workers at Cary are slackers. But there is increasing evidence
that the management and some workers at Cary may be just a little bit too
protected from the vicissitudes of life in an age of unrestrained
capitalism. If, after careful research, design and planning, top management
at SI concluded that it would take 2.5 years to get V8 to production stage
(via V7 as a beta release - that's fine, as long as they call it a beta),
then one would like to think that as that deadline drew closer, more
resources would be added to the project to ensure that the schedule did not
slip. Somehow, I don't think that is how things are done at SI. Perhaps I am
wrong. Can anyone enlighten us (anonymously if you like)?
> It's no secret that Dr. Goodnight provides heavy financial support to =
> right-wing political candidates (look him up at
<http://www.tray.com/fecinf=
> o/indiv.htm>), but I haven't heard of people being fired for disagreeing =
> with him. It would be foolish of him to do so - bad publicity and =
> wrongful discharge lawsuits could result - and he's not a fool. In any =
> case, it would be impractical to keep track of the political beliefs of =
> SAS Institute's thousands of employees. There must be more to the story
> than your anecdote suggests. =20
No, no, the anecdote was that the employee was taken aside and was advised
not to discuss politics in the canteen in case his pro-Democrat views
offended Dr Jim. He wasn't fired (at least not in the story I heard). I
shouldn't have related it as it is uncorroborated hearsay and probably
scurrilous, but it is amusing nonetheless. Hmm, I wonder if any of the three
letter agencies (including those which used Cyrillic letters) ever used SAS
to track the political beliefs of citizens. PROC JEDGAR perhaps (but I think
he predates even PL/1)?
Cheers,
Tim Churches
| 
