Re: Fantastic Ada promotional piece from Rational (long)

[mfeldman@seas.gwu.edu (Michael Feldman)]

In article <1994Sep8.225002.12999@ocsystems.com>,
Kevin D. Heatwole <kdh@ocsystems.com> wrote:

>For Workstations, software vendors typically charge a one-time fee for
>the product, maybe a required one-year maintenance/upgrade fee, and then 
>optional yearly maintenance fees.  The yearly maintenance fee is generally 
>10-15% of the "undiscounted" purchase price if it includes maintenance
>only (e.g., phone support, bug fixing, etc.).  If the fee includes free
>upgrades, then the fee is usually around 20%.  Ada compilers are usually
>a "high" maintenance item.

Yes, I understand this, and for teams working on deadline-driven projects,
it makes sense for them to pay for, and have access to, good support.
See below.

>Regardless, I think you might be a little rough on the Ada compiler vendors
>whose pricing policy is one-time charge with an annual maintenance fee.
>This pricing model is used by many software vendors (at OC Systems, we
>pay annual maintenance fees on just about all the commercial software
>we have purchased for our workstations).

Remember, I was talking abut academic pricing. See below.

>Borland is pricing software for the massive PC market.   Also, Borland's
>compilers were much easier to implement than Ada compilers.  I even implemented
>a Pascal compiler when I attended school in the early 80's.  Even GNAT
>has taken several millions of dollars and several years to implement, and
>it still has a way to go yet.  So don't expect Borland's pricing to
>be the same as a commercial Ada compiler vendor just because they both
>sell compilers.

Many Ada vendors have told me they were pricing for margin, not volume.
These companies (the ones with the Ada/PC compilers) convinced themselves
that Ada would always be a low-volume thing, and thereby guaranteed
that Ada would be a low-volume thing. I don't think anyone seriously
tested the elasticity; look how hard it is even to find out that
one of those compilers _exists_. When it comes to publicizing their
PC products outside the Ada community, they keep secrets better
than the CIA (well, that's no criterion these days...:-))

>As for "academic" pricing, as far as I can tell, there isn't enough of
>a market for Ada compilers from academic organizations that are willing 
>to pay, to make it worthwhile for most Ada compiler vendors (excluding
>the Ada compiler vendors that have targeted the PC marketplace with PC-like
>pricing models - they may still have a chance to make a profit).  

My argument has been that universities should not be seen as contributing
to profit, but rather to long-term growth, by making sure that all those
students demand Ada on the job. I am not asking - have NEVER asked -
that companies lose money on us, only that they not _make_ money on us.
I don't know why I have to keep saying this.

>Therefore, most vendors seem to have figure now that they might as well 
>give it away (with minimal support) because this really doesn't cost them 
>much.  Besides, the more users using your compiler, the more solid the 
>compiler will become in the long run.  

I'm glad the point has finally sunk in; we've tried to make it
for ten years.

>This wasn't true several years back, because
>most vendors probably felt they could make a profit "selling" into the
>academic market. 

That was their big mistake. We told them, but it took 10 years (almost)
to sink in. The universities voted with their feet.

I have NO problem with "distribution fees" of - say - a few hundred
bucks - to cover the vendor's avoidable costs of a compiler copy.
I know no educators who'd object to this. You imply that "minimal
support" is a new idea - it is _exactly_ what we asked for all
along. Compilers in a teaching situation do NOT require vendor
support, once the thing is stably installed. Ask any of us how
often we call our vendors. This is - always has been - a red herring.
>
>Another minor point that might be hindering many Ada vendors from pricing
>software cheaper is that the US government continues to be one of the
>major sources of income for Ada vendors.  Often, the government will 
>require a vendor to give the government "most favorite customer" pricing
>(especially, if you are listed on the GSA schedule).  I think this means
>that the vendor can't sell the product cheaper to customers outside
>the government.  

I understand this. It still goes to the high-margin/low-volume
argument. Why should the government be more willing to be gouged
than anyone else? What's the fight over the mandate about? There
are many _inside_ the government who are voting (or trying to vote)
with their feet.

>Well, these are only my opinions and observations, but sometimes people 
>on this board sometimes "bash" Ada vendors who are only trying to make
>a profit and compete in a relatively small market.  

I realize it's all water under the bridge, but after ten years the
market should not still be "relatively small". Call it "bashing", if
you will, but I've had enough discussions with vendor principals
over the last ten years to be pretty confident in my statement
that the vendors created a self-fulfilling prophecy that is now 
biting them good and proper, now that DoD is downsizing (and seemingly
unable to _really_ enforce a mandate).

It's not like there weren't folks out here trying to help them see 
that they were really stifling Ada's growth. That they're seeing it 
now tells me that, even if we are fuzzy-headed academics, we were not 
so far off the mark. I've certainly been told privately that we
were pretty close to right.

Mike Feldman
------------------------------------------------------------------------
Michael B. Feldman -  chair, SIGAda Education Working Group
Professor, Dept. of Electrical Engineering and Computer Science
The George Washington University -  Washington, DC 20052 USA
202-994-5253 (voice) - 202-994-0227 (fax) - mfeldman@seas.gwu.edu (Internet)
"Pork is all that stuff the government gives the other guys."
------------------------------------------------------------------------