Paul Boddie wrote:
> I saw Pastilda featured a while back on Crowd Supply, and it looks like the
> campaign has now started:
>
> https://www.crowdsupply.com/third-pin/pastilda
Good ! The more such critters hit the market, the easier it will
be to demonstrate the viability of this class of products.
Let's see how it does compared to Anelok:
Pastilda Anelok
--------------- ----------
Integration-less use (with PC) Yes Yes
Wireless No Yes (1)
Works offline No Yes
Works disconnected (e.g., ATM) No Yes
Secure storage Yes No (2)
Keyboard pass-through USB-USB USB-BT (1)
Price USD 50 USD ~100
(1) Planned, but not implemented in any prototype yet.
(2) Could do it, but it'll be slow, due to having only Full-Speed USB.
But I'd want that in a future version.
There's a bunch of such devices that all need USB. Seems that weaning
them off it is hard. Good :)
> You seem to need
> to use a text field (or similar writable interface) to interact with the
> device.
That seems a little worrisome. If they actually echo secrets (such
as the device's master password) through a text field, they could
easily leak, considering all the server-based auto-completion and
partial search functions that many input fields have nowadays.
By the way, Anelok doesn't have to be a spectator sport. A lot of
things that happen in software can be tested with the simulator.
I've written a number of task cards:
https://gitlab.com/anelok/doc/wikis/Tasks
Especially https://gitlab.com/anelok/doc/wikis/Task_altchill
and most of the 2FA group could be implemented with only the
simulator.
But there are many more things that still need doing, e.g.,
- use uSD instead of internal Flash for the password database,
- the database it currently read-only. After it's been moved to
uSD, implement
- new entry creation,
- change of entries,
- renaming/moving of entries,
- deletion
Or maybe something more researchy ?
- to what extend could Anelok interoperate with GPG and similar
systems ? Anelok could simply store the user's secret key, but
maybe we can do better.
- e.g., for signing, the host could compute the hash and Anelok
could sign that hash, with the GPG secret never leaving Anelok.
Would the algorithms currently in use for that allow such a
split ? And would the operations needed on the Anelok side be
compact enough for the MCU ?
- likewise, for encrypting, decrypting.
- Werner
_______________________________________________
Qi Hardware Discussion List
Mail to list (members only): [email protected]
Subscribe or Unsubscribe:
http://lists.en.qi-hardware.com/mailman/listinfo/discussion
