> This whole discussion is interesting to me.
As someone who has led major software development efforts, and who has
written commercial software and owned a small company (me) devoted to
same, I can offer a couple of observations that may be helpful.
Software developers are not a continuous commodity, like water or
grain. They come in discrete units, each with a specific set of skills
and preferences. To keep a good software engineer, one must pay them
well--the good ones are in demand from a variety of industries. In my
world, the typical billed rate of a highly experienced software
engineer is about $400,000 a year (including overhead and profit).
When a company has one software developer, the decision to add one and
have two is a big one. It doubles their costs. If they cannot afford
those costs and still support the price points they believe the market
will pay for their products, they have to make priority decisions
about how to use the one person they do have.
Software developers, in my experience, cannot spend 15 minutes on
Product A, 20 minutes on Product B, and two hours on the upcoming
Product C, and all before lunch. When I was actively developing
software, going back to support a prior product required a couple of
days (or weeks, in somoe cases) of re-immersion to become familiar
with the code and environment. Market pressures required each product
to have a more advanced operator experience than the prior product,
which required learning and developing a new set of development tools
and libraries. Thus, I had to let the list of chores on a prior
product build up to the point where it was reasonable from a cost
perspective to go back into the code and make upgrades. Until that
time, I would advise my customers on workarounds and hope they
remained patient. Losing money is not an option--that will quickly
lead to nobody getting any service at all.
Thus, I can well understand that Ten Tec or any other boutique
manufacturer, who can probably only afford one or two software
engineers, probably live at the ragged edge of what they can do,
hoping that the folks who are being told to wait can tolerate it while
they make the best use of their limited resources.
The big companies like Microsoft hire programmers by the thousands,
and still produce software that is buggy and unpredictable. The
support they provide is little if at all connected to the complaints
of any one customer. On balance, I'll take the small company.
One option is to give up on commercial software altogether, and let
the customer provide it. That's what you get with, say, the Flexradio
5000. Short of that, the more software is in a product, the more these
issues will present themselves.
In my world, software has a MUCH smaller potential market than does
ham radio software. We think in dozens. Yet the agencies who buy this
software still expect it to be tailored to their specific needs yet
have the reliability and maturity of a products with an installed base
I'm quite happy not to be in the software development world any more,
and I'm focusing back on my domain expertise and helping agencies
write system engineering documentation and software specs and form
appropriate expectations from those still willing to be in that biz.
It's not any easier, but at least I'm not usually the one getting
Richard W. Denney, Jr. PE|Iteris, Inc. |
Associate Vice President |107 Carpenter Dr. Ste 230 | 703.925.3819
firstname.lastname@example.org |Sterling, VA 20164 |Fax 703.471.1757
TenTec mailing list