Embedded devices vs "real" computers (was Re: Fwd: Featured Offer of the Month: Free FD100 Terminal!)
ethan.dicks at gmail.com
Wed Mar 14 11:32:41 CDT 2007
On 3/14/07, Chuck Guzis <cclist at sydex.com> wrote:
> Personally, I find little widgets with embedded processors
> Consider a credit-card terminal--it has a modem,
> display, keyboard, can interface to a printer and most have some non-
> volatile memory. The difference between many word processors and
> general-purpose computers is one of nomenclature more than anything.
> Yet no one's really interested in them. An old Artec, CPT or AES WP
> box has the same basic components and vintage as most early
> microcomputer systems, but no one's interested in collecting them.
For me, at least, the difference is that a word processor or credit
card terminal or a gas pump, or whatever, does what it does and is
difficult to have it do something it was not designed to do. If it
were possible to get developer-level documentation on the internals of
said embedded devices, that might be something different entirely.
The one class of devices like this that _is_ thoroughly explored is
Internet/e-mail appliances like the Audrey, the WebPal, the iOpener,
etc. Many of these, of course, are just Intel PCs with a UNIXy OS in
FLASH ROM, but at least they are inexpensive enough and powerful
enough to attract attention from the contemporary hardware hacker
community. Older devices, while interesting in their own right,
aren't as easy to come by, or, prior to the advertiser/subscription
business model, were paid for by the end user, not heavily subsidized.
ISTR IOpeners were costing the company $400 each, but were marketed
at $99 (later $99 plus a mandatory 4 month sub.) For a 200MHz PC
w/USB, 16MB Flash, 800x600 LCD, single-DC-voltage power (19VDC), $99
was a *scorching* price, so lots of hackers picked them up and figured
out how to dump the OS and replace it with Linux or Windows (yes...
people _did_ put Windows98 on them). For your money, you get a nice
little toy that's easy to reverse engineer, and when you do, the
machine becomes quite useful for a variety of applications. A used
credit card terminal for $20 (or whatever the going rate is) can still
be turned into something else, but it's not going to capture the
attention of an army of curious hackers.
I guess the other thing is... for as much work as it would be to delve
the secrets of a credit card terminal, I can whip together an LCD
display and a microcontroller for a few bucks and have 100% control
over its innards. If I don't care to start from scratch, I can begin
with any number of Atmel/LCD projects already out there (and I've done
just that recently). At that point, the advantage of an existing CC
terminal diminishes rapidly. It still has the benefit of being mass
produced and potentially abundant, but anyone can order an ATtiny2313
for a couple of bucks and hang a 2x20 text LCD or a 128x64 graphic LCD
off it and make it display some sensor input or whatever. The CC
terminal has to have some pretty cool stuff in there to pull into the
lead for a hardware project - I'd most likely gut the LCD and possibly
keep the plastic case for some other project than hack the
microcontroller that's already in there, and that's presuming it's not
a mask-programmed part and that I can reflash the chip.
So while I might not turn down a box of embedded devices like a credit
card terminal, hacking a lone terminal isn't as appealing as building
a similar device from scratch and knowing exactly how that new device
works. As I said, though, having developers' docs might tip the
balance, but the chances of that are rather low.
More information about the cctech