Interesting comments about this subject.
Here is a new perspective on why moving to windows will be a
difficult choice to make. How many hams are using their "old"
computer for logging? Most of the DOS based logging programs
run very well on a 386-SX25 with a few megabytes of RAM.
Q: How do you make a good 486 machine run like a good 286 machine?
A: Install windows on it.
Now, if I decided to port TR Log to windows, I would have two
programs: TR Log for DOS and TR Log for Windows. The amount of
work involved to port over to windows is something I have started
to understand in more detail (using Delphi, which uses the same
basic Pascal compiler I am currently using). Simply put, the changes
are significant enough that I would have two packages with very
few common elements.
TR Log continues to evolve with significant updates monthly. In
order to update both packages, I would now have doubled the amount
of work involved.
Or, I could just say that TR Log won't support DOS any more.
That would solve the double work problem. However, there are
simply too many people who are using computers in DOS mode who
don't want to use windows.
If I could create enough demand to do this full time, I would
do it in a minute, but as someone already pointed out, we are
underpaid and this is impossible.
Note that none of the above points mention the technical problems
associated with generating CW in this environment. This is "hard"
because I don't know how to do it. Maybe someday I will learn
(acutally, W5UN has a method that would work just fine), but it
doesn't solve the other problems.
It isn't clear that there are any advantages to making these DOS
based programs a true windows application anyway... You can run
them under a DOS window and generate CW (if you have things setup
correctly). A mouse just doesn't look too useful to me for
contesting...
In closing, I would like to copy a message or two someone else wrote that
does a better job explaining some of this than I can:
>OK, at the risk of flames/mailbombing, here goes:
>
>Why is the "Rolls-Royce/Cadillac/Mercedes" (your choice of car/country) of
>contesting software still not ported to a multitasking environment like
>Win95 and/or OS/2 or Linux? Any thought for the 000's (maybe millions) of
>poor 3rd World users like yours truly who do not have access to a
>PacketCluster and must rely on a single machine for a multitude of tasks
>like running CT, connecting to a WebCluster etc.etc.?
>
>In any case, why the fixation with DOS, which will disappear soon? If a
>multitasking OS of choice, why not Win95 or Linux? Living in this day &
>age why must CT stick to the 8088 era? At $ 80 a throw surely one can
>expect to get CT in the OS of choice, or am I wrong?
>
>73 de Tariq/AP2TJ.
I'll offer an opinion or two. I write software for a living, and I've
used several contest logging programs for about 8 years.
1) Writing great software is hard work. Supporting people with widely
varying skill levels is even harder work.
2) The amateur radio community isn't large enough to make any real money
(enough to make a living, or enough to entince someone away from a "real"
software job). You just can't sell enough copies to make it worthwhile. If
you're going to bust your butt working a lot of strange hours, it's easier
explaining it to your XYL if you're paying the bills and creating a future
for your family. Working strange and long hours year after year to give the
stuff away cheap is hard to explain to your wife, kids, dog, and even
sometimes your mirror.
3) The common denominator is PC DOS. Even the business community is largely
still mostly on PC DOS with Windows 3.1. Granted all new machines have
32-bit multitasking operating systems on them, and that trend will continue,
but PC DOS compatibility is still extremely important.
If, hypothetically, you were to write some great 32-bit-only ham software,
there would be a very large group asking to have it revised to run on their
system (the 1 megabyte 80286 that runs CT or TRLOG just fine). And even
though it's arguably easier to write for the newer systems, those people who
want it to work on 16-bit machines would feel that they should pay less for
it. They figure YOU should spend time revising your program to work on
their machine rather than them buying a current machine. And when they do
upgrade, they'll want the 32-bit version. For free. After all, they bought
the 16-bit version.
3) To send good smooth morse code, you probably need to mess with the PC
clock. I'm not an expert in the techniques, but I think it is difficult to
provide smooth speed changes (from 22 to 23 to 24 WPM) and "accurate" morse
code with a clock as coarse-grained as the standard-rate PC time-of-day
clock. So I suspect CT and its peers run the clock at a higher speeds.
Most multitasking OSs won't let you mess with the hardware unless you
install a device driver, and talking Joe Ham through the details of device
driver installation isn't something I'd like to spend much time doing minutes
before a contest.
4) But a major reason is that to do a good contesting program you need to
really love contesting, and if you really love contesting you probably would
want to enter contests, not take all the phone calls from guys who decide a
half a day before the contest that they'd like to learn how to set up their
computer. Sometimes this happens because stations are thrown
together at the last minute from a crew that meets for the first time on the
day.
And I do this all the time. But would I like to be spending my
pre-contest hours working on my station or answering questions about IRQs
from one of my customers?
5) Hams can be a bit fast and loose with license agreements.. For every
paid copy of a contesting program out there I would suspect that there are
several others that the users have ahhh "deferred" payment for. And of
course the author will get support calls from all users, whether or not
they've bothered to register their copy.
I heard once that the author of a very popular contesting program took tens
of calls EVERY DAY during contest season from people who thought they needed
his help. That would be fun for a while to talk with your friends, but I
suspect it would get pretty old after a few years. I know I and my wife
would tire of it.
I do not mean this as a flame. It's not a bad question. But maintaining an
application through all the various PC hardware, PC software, and radio
control permutations can become a life's work. If you're going to take on
that sort of crusade, there ought to be a good living in it. And there just
isn't one there with the kinds of numbers contesters represent. Cheap
software is cheap because you sell many thousands of copies. If there are
relatively few purchasers, expect to pay thousands of dollars for a
professionally written and supported application.
73 de Dick, K6KR
k6kr@contesting.com
>On Friday, 12/5 Tony KE6YNR wrote:
>But... The going rate for annual software
>maintenance for 'big' workstations (Sun, HP, etc.)is about 12% of the
>cost of the product; _not_ 44%. The 12% cost has the customer paying
>for bug fixes as well as feature enhancements.
>
>In line with software upgrades for personal computers, 40 to 50% would be
>an appropriate price to upgrade a major revision. Say from version 9.x.x to
>version 10.x.x. A mini upgrade could happen if significant features were
>added without the major rev incrementing.
>From dieven@msn.com Fri Dec 5 10:40:13 1997
From: "Dick Dievendorff, K6KR" <dieven@msn.com>
To: "Tony" <ke6ynr@garlic.com>
Cc: <ct-user@contesting.com>
Subject: Re: [ct-user] CT BBS
X-Priority: 3
X-MSMail-Priority: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
Content-Length: 4873
Tony:
The "rules of thumb" that apply to large-scale commercial software do not apply
to specialized software with such a small distribution as CT or NA or TR.
Some of these companies are selling hardware and software, and the cost-recovery
formula for software is all over the map.
The company I used to work for sold hardware and software. For many years, the
software was "thrown in for free" with the hardware rental or purchase. And that
included on-site custom software development. This made it hard for independent
software companies to get started.
The company I now work for sells a lot of PC software. The consumer price is
very low. But we sell millions of copies, and it is a very competitive market.
As the volume goes up, the development cost becomes a smaller percentage of the
total cost picture, and costs of marketing, distribution, materials (CD-ROM
versus diskette) and books in the box, support, and advertising become very
significant. There is a very focussed effort to control costs, particularly
per-unit costs, in order to keep the prices low enough to keep the volume up so
that you can afford a significant R&D and marketing effort (which doesn't go up
as the number of units goes up).
My wife is a court reporter. She has a specialized PC DOS application that
helps her translate machine shorthand into editable text. This is a limited
market, perhaps a few thousand customers at most. She paid about $3000 plus
about $500 a year for "technical support", which includes software maintenance
and upgrades. She has an 800 number she can call at any time, and it's staffed
by people that can get answers. The software comes on 3 diskettes, and that
includes several large dictionaries and lexicons, so the software effort isn't
really all that huge as these things go. And this is a good business deal for
my wife. She looked at several competitive packages, and she wasn't
particularly price sensitive. She uses this software every day to generate her
income. It's the small numbers of demanding users that make this a reasonable
figure. The company is reasonably responsive to requests for new features. She
wants this company to stay in business in order to support this software.
I think the reason we expect low-priced software is that our expectations are
set by the prices for mass market consumer software like games, spreadsheets,
and word processors. Take a look into the market for the prices of software
tailored to specialized fields with good support. The prices are quite high.
And they're often a great business deal for the customer.
I spend a fair amount of money and time on antennas, radios, computers, internet
providers, telephone lines, and software for other applications. Why not a
hundred annually for great contesting software for a hobby that's important
enough to me to dedicate many hours several weekends a year? My perception of
the contesting experience is very much dependent on the capability and quality
of that software. At 36 hours into a contest, do I worry about saving a few
tens of dollars on logging software? Some people go on pretty expensive trips
to rarish locations for contests. Wouldn't a few hundred dollars spent on
contesting software be a drop in the bucket?
The public domain versions of any of the popular logging programs suffice for a
casual entrant. Those of us that want the best software available should expect
to pay enough to make it worthwhile for a few groups of people to make a career
out of contest logging software development and support. I want a field with
healthy competition and choice. And there's nothing like the profit motive to
stimulate that competition.
Using a percentage of a small amount is erroneous. It costs a certain
per-customer amount to open the envelopes, deposit the checks or run the credit
card numbers, and answer the phone or e-mail for technical support.
This is why software manufacturers can sell software so cheaply to PC vendors.
If it is sold with the machine, the software developer deals with one very large
customer ready to move a lot of units. The software company doesn't have a
per-unit transaction or an individual customer support task to deal with. Some
end users require a great deal of support for what seems absolutely trivial.
73 de Dick, K6KR
--------------
End of quote from K6KR.
So - for anyone still reading this very long message, I would say the
people who are insane enough to do this are already doing it. From
time to time someone with a callsign I don't know from contests
announces they are going to port CT over to windows. We are still
waiting.
I predict either K8CC or myself will eventually do this, but only
after another 5 years or so and everyone except K3ZO is using a
32 bit operating system in their computer. Perhaps it will be
some future operating system that hasn't even been announced
yet. I have taken the first step in this process and some of my
post contest or log checking software will become windows applications
soon.
73 Tree N6TR
tree@contesting.com
--
CQ-Contest on WWW: http://www.contesting.com/_cq-contest/
Administrative requests: cq-contest-REQUEST@contesting.com
|