16 May 2000
Basler is the President of Aurora Systems,
Guest Editorial: Robert Basler
Insulting OS/2 Software Developers Everywhere
In reading the column "Karen's
Corner" article "OS/2 is Rescued by the Internet" by Karen L. Mansbridge-Wood,
I was appalled at her offensive comments towards OS/2 software developers and by
her obvious lack of understanding of how software development works. Having spent
some time reading her posts in comp.os.os2.advocacy I was familiar with her opinions,
but seeing them leave that venue and pour over into articles intended for public
consumption was disturbing.
In her column, Karen rails against
OS/2 software developers, calling their hard work "substandard" and accuses
them of "exploiting" the OS/2 community with "inflated prices."
That she could say such things clearly demonstrates her lack of understanding of
how commercial software development and simple economics works. That she would insult
people who for the most part develop software solely because they love to, is embarrassing
to the OS/2 community as a whole. We all benefit greatly from the hard work of software
developers and they deserve our praise and support.
Most software initially is developed
by one person who has a cool idea, or who needs something done that can't be done
by existing software. Even in large software houses, this is how many software products
are born. That person then pours their heart and soul into designing, coding, debugging
and testing the software, often for several years, sometimes adding more people
to help finish the task. This takes long hours, sleepless nights, and causes no
small amount of financial stress. Even a simple project by a single developer can
take months to complete. If you have to hire people to help, when a software developer
can make US $5000 per month easily, you can quickly collect a lot of up-front expenses.
Even a single developer working in his basement just squeaking by on living expenses
is probably laying out a couple thousand dollars each month to pay the mortgage
and bills. For many commercial software projects, this is the incentive to ship
To develop version 2.0 and grow your
company, you take the money that you are earning from 1.0, hire developers, get
feedback from your customers as to which features they need most, and get to work.
A year or so down the road you have a shiny new version 2.0 to release to your customers.
Of course, these early versions don't
have all the features you might want. Many of the "higher quality" products
from "major software developers" that Karen is so interested in have had
the benefit of several major revisions. They have grown to become mature, stable
packages, they weren't born that way. While there are certainly exceptions, in general,
large software houses don't produce high quality software simply because they are
large, they produce high quality software because they've had five or ten revisions
to refine it. If they produce a high quality product on the first release, as the
gaming industry has to do, it is because they spend a fortune on teams of people
who perform extensive feature development, testing and debugging.
The problem with a lot of OS/2 software,
is that it rarely gets this chance to grow into a mature product. Having talked
to a number of OS/2 software producers over the years, other than Brad Wardell,
I can't name one that has made a lot of money doing it. Small ISV's are small because
they don't make a lot of money doing what they do. Many of the ones I have talked
to that produce OS/2 products do so because they have reasons other than money,
usually because they love the platform and it is what they like to use.
Why aren't the OS/2 software developers
Karen so despises making lots of money and refining their products and growing into
major software developers like the Windows developers? It is simply numbers. To
survive selling software, you have to cover your development costs. If you want
to grow, you have to make a profit. If you want to sell software cheaply, you have
to sell a lot of it. To sell a lot of software, you need a large customer base,
and you have to be able to reach them.
The commercial OS/2 software developer
has a number of disadvantages compared to software developers in other markets.
The biggest one is that there is not a mass-market of OS/2 users anymore. Because
of this, you can't get shelfspace in the local WalMart, Future Shop or CompUSA or
coverage in mainstream PC magazines. If you want to advertise, there is only one
print magazine I know of dedicated to OS/2 and not everyone reads Fidonet or has
internet access to Warpcast or OS/2 E-Zine or Usenet. How is an OS/2 developer to
reach their potential customers?
OS/2 ISV's are probably more intimately
aware of the state OS/2 market than anyone, they see it in their paycheques. Since
software costs the same amount to develop, regardless of the size of the market
it sells into, you often see ISV's charge higher prices for OS/2 software in an
attempt to cover their costs and stay alive to produce more, better products. This
is unfortunate, but no one has yet come up with a workable alternative. Those that
suggest vendors should fund OS/2 development with profits from other platforms clearly
don't understand what a terrible business strategy that is.
I agree with Karen that IBM's idea
of cross-platform software development is a good one, I just don't believe that
Java and web based applications are ready for prime-time yet. I do however eagerly
await the day when they are, and I suspect most OS/2 ISV's would agree with me.
The "write once, run everywhere" promise of Java and browser based applications
is compelling when companies such as mine spend up to 80% of our development effort
on testing. My company tests our products on Windows 3.1, 3.11, 95, 95b, 98, 98
2nd Ed, NT 4 SP 3, and Windows 2000 - eight separate platforms. Unknown to most
industry outsiders, testing every platform you run on is very necessary. Even between
Windows 95 and 95b where there is little difference to the end user, we invariably
find problems that appear on one platform and not another. If I could bring my applications
to OS/2 (or BeOS or Linux or Mac or SunOS or...) with no additional development
expenses, I would, as I believe would most other ISV's.
Unfortunately, while Java is a great
technology and it has mostly overcome the speed criticisms of the last few years,
its API's are not yet complete enough to do everything the developer wants. Even
if you avoid the use of proprietary extensions on some platforms, its cross-platform
support is still of the "write once, test everywhere" variety which just
isn't adequate for many companies if supporting a small market like OS/2's means
they have to justify additional development expenses. Also, many users are very
unforgiving of the shortcomings of cross-platform strategies like Open32, Odin or
Java and web browser applications. We want, and often require, native applications.
I understand all the problems the
OS/2 developer faces bringing a product to market because I have many of the same
problems. My software sells to the minutest fraction of the Windows user base. I've
run a small software development house for ten years, and while I make a living
at it, I won't be buying a sports car any time soon, despite the price of my products.
I know that the bulk of the OS/2
community is sensible and understands the simple economics of their situation, but
it still bothers me when someone takes a public forum to insult an entire group
of people that they obviously have no appreciation for or understanding of. I for
one am grateful for the OS/2 software I use, even though I have spent quite a bit
of money on it, because it allows me to use OS/2 to get my work done. I applaud
OS/2 developers everywhere for the hard work they put in to bring out software that
I can use. I hope the OS/2 community would agree with me.