[Cryptech Tech] Alpha USB interface(s)

[Peter Stuge]

Fredrik Thulin wrote:
> The FTDI chips we will use for USB<->UART

Has that been decided without transparent discussion since my offer
to contribute an alternative solution with strong, obvious technical
benefits and - let's be honest - very few drawbacks?


> EEPROM

That's a programmable chip too, so now it seems that my suggested
approach really has no drawbacks at all.


Guys - if this project is going to be able to stay credible then
decisions can not be made without either a) learning actual details
or b) trusting those who already have - as happens with the other
aspects of the project. This can not be any different.


I am more than happy to educate about USB in general and the small
simple Cortex-M3+firmware I propose to use in particular. I have
done so several times before in two-three hour long workshops at
events in different countries, and it was exceptionally well received
each time.

I really do think that restricting UART communication to onboard
between the little Cortex-M3 and the STM32 brings real significant
technical advantages; that way the serial protocol stays in a far
more controlled environment, and speaking a real USB protocol with
the host allows the host software to be much simpler, without that
USB protocol requiring any design effort to speak of.

If you are able to describe real issues with my suggestion then I am
very eager to learn about them. Maybe they can even be resolved.


Finally, supporting FTDI in any way is both poor politics and
technically risky.

In their latest driver update which is pushed through Windows Update
they started injecting ASCII text into the serial bytestream if the
driver considers the connected chip to be a clone.

FTDI sells a poor product at too high a price and are being
increasingly shitty about other companies reusing their sucky
protocol, in a way which will affect *end users at runtime*.

They don't deserve any more business, not mine, not yours.


//Peter