Mr. Baggett, one of our fine and helpful writers, here manages to seriously confuse some issues.

At 09:34 AM 10/22/2003, Bagotronix Tech Support wrote:
> Don't you think if it were easy (or even possible) to make money -- to
> in business -- at those prices, that there would be companies doing it?!?

No.  That is the theory of perfect markets.  Which is just ivory tower
economics BS.  Observe:

An economics graduate student and his professor are walking down the street.
The student sees a $20 bill laying on the ground.  He exclaims "Hey, look!
I found a $20 bill!".  The professor replies "Don't waste your time bending
to pick it up.  It's just an illusion.  If it really existed, it would have
already been picked up by someone else."

Perfect markets cannot exist when a monopoly exists.  And MS is a monopoly.
And so is any program that locks your data into a proprietary file format
which is undocumented.

The issue is far more complicated. First of all, if a hundred people walk down the street looking for $20 bills, no $20 bill that is in plain sight is not going to be picked up because of the phony argument. It's a good joke and a very bad analogy.

MS is indeed a monopoly of sorts, but the fact remains that it is not easy to replace them. To do so would take a huge investment, and the investors would want, of course, to make a reasonable or more than reasonable return on their investment, commensurate with the risk. And so that software would not be cheap. I see only one way around this, which would be for users to organize to cooperatively develop an OS. Not just software engineer users, but *all* users. That's not going to happen soon unless something rather drastic happens. Not impossible, I am in fact working on a larger problem of which this would be a subset, but by no means do I feel assured of success. I'd feel better if I saw others working on the same problem, but I don't.

I read something a while ago that said that MS has a profit margin of 85% on
their office suites.  That's one heck of a margin!  I wish I could make that
margin on hardware sales!  Assuming Office costs $499, they could reduce
their price to $75 and still break even.  Probably further than that,
really, I bet their margin is even more than 85%.

It seems to me that this is a very serious misunderstanding of margin. Margin does not include development cost, nor does it include support. Rather, margin would generally refer to the difference between sale price and unit cost, expressed as a percentage of sale price, which with software is very low. Microsoft's profit, after considering development costs and support costs, is much much less than what Mr. Bagget would imply. If, in fact, their profit were what he thinks, there would be hordes of vultures circling. Yes, they have tremendous profit, just not what he thinks.

I just looked at one of Microsoft's SEC filings. For 3 months ending 9/30/2002, total revenue was about $7.7 billion. "Cost of Revenue" was $2.0 billion. The difference is called "gross profit," and that would be, in this case, 74%. In another more recent period it is entirely possible that the figure could be 85%.

But gross profit and profit are not at all the same thing. If Microsoft could double their sales, theoretically, they would increase their profit by this gross profit amount. But they have development and other costs, and those costs are steep, more than the "cost of revenue," which would be the cost to them of the packages sold, i.e., CDS, Manuals, etc.

For the period mentioned, those other costs were $2.5 billion. That leaves an operating income of about $4 billion.

Now, what profit would be reasonable for them? That would depend, surely, on what they have invested, on the assets devoted to the business. I'm not sure how much the SEC filings can be trusted, but I don't have better data. The total stockholder's equity in the company is $53.5 billion as stated in the filing. Thus their return on investment, if the assets were appropriately valued, was about 7.5%. I'm not salivating. That's healthy, but not excessive.

This does not mean, of course, that they are not predatory monopolists....

> Most of the folks on this list are hardware people.  I've been developing
> selling software for 20 years, and I can tell you IT AIN'T EASY.
> quality software takes a lot of time and a lot of talent, which = $$$.

Dwight, I visited your website.  Your situation is different than what I am
talking about because you make niche market and specialty software.  You
don't have a large market to spread development cost against.  But vendors
of OS, office suites, RAD IDEs, compilers, and some CAD apps do.

Of course, and Mr. Baggett's comment is correct, I almost wrote the same thing in my response to Dwight myself. But I ended up leaving it out.

Nevertheless, the software business remains quite different from normal manufacturing businesses, because the development cost exceeds the unit cost even when one is selling millions of units. I think even the music business is not so lopsided.

I write software, as well as design hardware.  The software I write isn't as
big as an OS or office suite.  It's embedded software.  And bugs are not
tolerated in embedded software.  A machine control system crashing is simply
not tolerable, especially when it may cause injury.  And embedded software
must be fault tolerant.  Every function I write has to be written with an
analysis of how it will behave if given erroneous conditions.  How long do
you think I could stay in business if my embedded software crashed as often
as Windows or some Windows apps (i.e. Protel, Autocad, etc.)?

And, of course, Mr. Baggett is also in a very different business, more like Dwight's business, but possibly less fault-tolerant. And, if so, I expect that Mr. Baggett's software is *very* expensive, compared to any large-market software of similar code complexity, many orders of magnitude more expensive.

  I see none of
the methods of software testing I have read about over the years being
applied to desktop apps.  Typical PHB Software Manager says, "Fault
simulation, resource monitoring, memory leak checking, bounds checking,
regression testing, what's all that fancy stuff?  Never heard of them.  Just
write it, click on each menu item and dialog once to make sure they pop up,
and ship it!"

I know that this is definitly not true with Altium.... But the question here is not whether or not mass-market software is full of bugs, it is. However, it still costs money to find those bugs and fix them, and, until the market demands better software, the companies are disincentived to improve it. Such software would cost more, there is no doubt about it. Would we pay it?

It's true what you say about quality software taking a lot of time and
talent.  But your time investment doesn't vanish after the current version
is delivered.  You have the source code to make improvements on, and add new
features to, for the next version.  So why does software keep going up in
price, when all that IP already exists and doesn't have to be rewritten from
scratch?  And to add insult to injury, many of the software jobs are being
moved to low wage countries, such as India.  Why doesn't the price of the
software go down when that happens?

Good question. But being able to ask a question that may be difficult to answer does not prove anything. It's a common delusion that if you can ask a question that your opponent in a debate cannot answer, your own position is therefore correct.

There is a problem with software as it becomes more and more complex. Yes, there is an accumulation of IP. I think that is already reflected to a large degree in the price, the software would cost substantially more if it had to be invented to do the same job from scratch. But using that IP is one of the things, I suspect, that leads to bugs. Perhaps a call to a routine was prefiltered to remove malformed data, and in the reuse of the routine, the prefiltering is omitted, thus creating an opportunity for a bug to appear "with a proven subroutine" that had been used by millions of people without a problem. A software engineer would be able to discuss this much more intelligently. I just remember reading a Scientific American problem on the difficulties of dealing with huge masses of code, the exponential possibility of bug creation as complexity grows, etc.

The human brain has been under development for a very long time, and faults are often fatal, and yet there are still plenty of them.... The very mechanism that allows for improvement introduces new problems.

Anyway, my own solution to buggy software is to turn the user base into a continual test and information-gathering base. Error-reporting software should be built in and users given a means to opt into easy, painless error-reporting. Windows is doing this now; whenever W2000 runs into a problem, I'm asked if a report may be sent to Microsoft. Protel could do the same thing, make it trivially easy to report a problem. Right now, it is too much trouble for most of us, we just complain later, when we aren't working on a project or are blowing off steam.

We could be part of the solution, instead of just complaining about the problem.

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
* To leave this list visit:
* Contact the list manager:
* Forum Guidelines Rules:
* Browse or Search previous postings:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to