Doug Conn wrote:
> I think the C6C is nifty, but would you and Joe have gone through all the
> effort if you had only planned to use it yourselves and not make it a
> product ? In that case it probably would have made more sense to shell out
> the extra $100 for a commercial product.
Negative. The C6C is clearly a hobby product (I'm definitely not going
to retire - or even eat well - on the profits from it), but it is still
worth while to develop on an inexpensive, modular platform.
After over 25 years as a professional programmer, I can assure you that
the amount of time spent coding a cheap micro-controller is basically
the same as coding an expensive micro-computer ... neither one saves a
significant amount of development time. Compilers, available libraries,
development environments, and the testing cycle is roughly the same
these days for all platforms.
But, when you start designing things with modular hardware, you get a
*big* payoff in terms of flexibility and architectural evolution over
the long term. That is, modular hardware lets you move up and down the
complexity spectrum very easily, while larger, more powerful monolithic
hardware prevents you from easily solving simple problems.
For example, Joe built his ESC first to solve an immediate need. An ESC
is a medium-easy component and all the lessons learned from it apply up
and down the spectrum. When someone asked about a simple relay trigger
one day, Joe grabbed a smaller PIC, grabbed some routines from the ESC
software and quickly developed a production-ready servo relay switch.
Then, when we chatted about gamepads as input devices, he jumped up on
the complexity spectrum, grabbed a more powerful PIC, copied some
routines from the previous projects and developed the first C6C. Three
different problems, three different complexity levels and three
different inexpensive solutions ... all using the same development
techniques, baseline code routines and all interoperable.
Therein lies the power of the micro-controller approach.
The only thing a micro-computer platform (or these "packaged" PIC
platforms) does for you is eliminate the need to create your own PCBs,
but that also locks you into solving problems just with the hardware the
package gives you. So, they cram more and more hardware options on the
board, thereby driving the price up, even though it is highly unlikely
that *any* application actually uses all of the hardware on the board.
I have been monitoring the new generation of modular high-level
components (e.g., tinker-toy electronics) which provide a family of
well-defined components that can be integrated with each other easily.
There are lots of good benefits to such an approach and the only wiring
required is point-to-point or bus based, thereby eliminating the need to
create PCBs. But, unfortunately, they all seem to be aimed at the
educational robotics market currently, with a hefty price tag per
component, making the final solution even more expensive than a
micro-computer approach.
Frank P.
--~--~---------~--~----~------------~-------~--~----~
You are currently subscribed to the "R/C Tank Combat" group.
To post a message, send email to [email protected]
To unsubscribe, send email to [EMAIL PROTECTED]
Visit the group at http://groups.google.com/group/rctankcombat
-~----------~----~----~----~------~----~------~--~---
