Re: Ada is not a failure.

[psinntp!vitro.com!news@uunet.uu.net (Morris J. Zwick)]

In article <23s4tl$9ni@news.aero.org> doner@aero.org (John Doner) writes:
>In article <1993Jul27.134205.7881@vitro.com> mzwick@vitro.com (Morris J.
Zwick) writes:
>>In article <1993Jul25.065103.19504@hellgate.utah.edu>
>>matwood%peruvian.cs.utah.edu@cs.utah.edu (Mark Atwood) writes:
>>>
>>>True, your LEAP program makes a stab at it, but it took an Ada vendor too
>>>long to realize what nearly every other software vendor knows, the schools
>>>are what will sell you.  (And don't tell me about "finantial realities".
>>>Once you have the thing developed, it's next to free to copy and distribute
>>                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>This argument is specious. How do you expect to recover the costs of
developing
>>the software if you don't ask a high enough price? The real problem with Ada
>>prices is the (at least cuurently) small market for selling the compilers.
>>Economies of scale do not enter the fray, therefore the development costs are
>>recovered over fewer copies of the product.
>
>Really?  Not so, usually.  Development costs are sunk costs.  You invested
>(i.e., gambled) your money and time up front.  Once the product is there, you
>want to make the most you can from it.  You want every dime you can get from

If this were the case, why not price cars at $100 and sell a ton more than
usual? Easy, because you could never sell enough to cover your costs of
development and production. Why on earth would you invest money into something
where the management says "We need $x million to develop this software. Our
plans are to sell it at some price regardless of development costs."? If
someone risks money, they expect a return on their investment. Otherwise, why
invest? The gamble continues until you WIN or LOSE, not when you place the bet
on the table.

Are development costs a "sunk cost"? In a sense that you spent it, sure. In a
financial sense, sunk costs are costs that have no bearing on a financial
decision because of their inapplicability/relevance to the decision.
Development costs are fixed costs of production and are, therefore, very
important to the pricing decision. Development costs ARE NOT sunk costs!!! The
reason why Symantec and Borland sell their compilers for so little is that the
markets for Turbo Pascal and C/C++ are significantly larger that Ada's. Their
fixed costs (ie. the development costs, plant, equipment) are distributed over
a larger number of units. As you hint, in software, the variable costs of
production are very small compared to fixed costs. That does not mean that the
fixed costs are not important!!!

Now, one can argue that the companies should do more to expand the market for
Ada. This would include a combination of broad-based advertising, lower prices,
and "marketing" to educational institutions and other important "barometric"
groups (IEEE, etc.). These things cost money. As an investor, you must balance
the risk against the reward. I can tell you that defense-sector companies are
extremely risk averse. I suspect that Ada vendors, which are on the periphery
of this group, share a good deal of this aversion. Are they right or wrong?
Neither; from their perspective, they are pursuing a level of risk they are
comfortable with.

>A few years ago, Meridian came out with an Ada compiler and development system
>for the Macintosh.  The price was in the thousands.  I laughed.  How could
>anyone be so totally oblivious of the realities of the marketplace into which
>they were selling?  Think C for the Mac, also an excellent development
>environment, was selling for under $100 in the education market and not much

Was Meridian trying to capture Symantec's market, or were they pursuing a niche
market for Mac Ada compilers required for DoD projects, plus others who wanted
to develop in Ada? I suspect the latter. The Meridian product (which I bought
when the price went down) was not in the same league as Think C (nor is it
now). For Meridian to compete, their compiler had to at least have the same
features and quality level of the Symantec compiler, plus have enough marketing
push and development community pull to convince customers to buy the Meridian
compiler. I suspect all of this would have cost a lot more, and Symantec would
not have stood still while Meridian tried to capture its market. Could it be
done? Maybe, but you would be betting a significant sum of money (risk) to find
out. Would you, as an investor, bet your money on this? I sure as Hell
wouldn't!!

>price it low and sell tens of thousands of copies.  Ask Borland or Symantec.
>Eventually, Meridian evidently realized this and their prices came down.  But
>they may have been too late.

I bet if the Meridian compiler cost $50, they wouldn't outsell Symantec. The
market is JUST NOT THERE!! While Greg's rantings can be irritating, he has a
good point; how can you expect the language to succeed by just establishing
some mandate? Ada needs to be sold to people, not forced down their throats. I
have used Ada, and I love it. No mandate required me to come to this decision.
Alas, governments seem to always pursue this tactic. Prohibition, drug laws,
gun laws, even sodomy laws (none of which I'm defending or attacking) were all
attempts by the government to control behavior. Educational campaigns against
drunk driving seem to have done more to stop drunk drivers than any prohibition
could have. Ada needs a market, not a mandate!!

These views are, of course, my own.

         ___________________________________________________________________
        /  Morris J. Zwick	                 Internet: mzwick@vitro.com
__     /   Vitro Corporation	             Voice:    (301) 231-2784
  \   /    14000 Georgia Ave.                ___________________________
   \ /     Silver Spring, MD 20906-2972      |"I don't want the world; |
    *                                        | I just want your half!" |
                                             |  - They Might Be Giants |