Just a few comments:

>> portion of that revenue in advertising. Google ads might do
>> for 50
>> units/year, for the remaining 450 what would you do?
>>
> The market is primarily with field service engineers. These
> guys drag
> around enough stuff, and want a CERTAIN,
> RELIABLE way of upgrading products. Are you saying they would
> like
> dragging around an ICD2, MPLAB, and

---------snip----------

You didn't answer the question!  How do you get the other
450 field service engineers - or more accurately, the
companies they work for to learn about the existence
and benefit of your product?

> The other application is on the production floor. The
> engineering
> department releases the updated firmware for the
> whatsit which is almost ready to be boxed up and shipped. The
> updated
> firmware goes into the bottom of the
> production programmer each morning. The programmer charged at
> its holder
> all night, and its LiIon or NiMH
> battery pack is topped off. The QC engineer notes which product
> needs
> final updates, dials it in, plugs in a tiny ICSP
> plug, and updates are done in a few seconds; NOT just the
> program
> section, but ALL of it; the config bits as well.

Except for really small shops, programming is done with PC
controlled systems.  Engineering releases software, production
verifies it will load, and then puts it onto a production server.
Units in process that have not been programmed go through
a software download area that typically includes a PC, barcode
reader (so production automatically loads the correct code into
the correct units) and a vacuum or clamp fixture that probes the
appropriate signals and can also apply power.  Units already
assembled into housings typically are downloadable (bootload as
used here).


> Here are the pitfalls with bootloaders:
>
> 1. The actual firmware can be intercepted and stolen easily.

True, however this is a tradeoff of complexity and security.
If a customer really wants your code for free, there are ways
to get it. Freely available downloads are easier on the
customer and your business.

>
> 2. The client won't routinely use it. You can lead them to
> water, but
> they may not drink. Maybe because the encryption process
> is too complex?

Clients WILL NOT upgrade unless there is an issue that effects
them.  No business is going to take either the time to upgrade or
the risk inherent with "new and improved" software unless they
need something fixed.

>
> 3. Rarely are changes ONLY firmware changes.A product shipment
> or a
> field-service visit is usually needed.
>
Not true at all.  Most changes in software are for bug fixes,
or new software features.  The typical PC is an extreme
case.  The cost of a field service visit just to download
code will make the customer unhappy.

Ken

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist