[Cryptography] Is ASN.1 still the thing?

Dennis E. Hamilton dennis.hamilton at acm.org
Sat Nov 18 19:25:44 EST 2017

Um, Why not ASN.1 Octet String? 

It would be valuable to supply an identifier for the floating-point representation case, but the ASN.1 type is already there for packaging the value.  Of course, the identifier needs to be very explicit about what the precise floating-point format is. 

It seems to me this is not a matter that requires attention of the ISO/IEC with respect to ASN.1 itself.  Interchange serialization for hashing and signature purposes seems orthogonal.

-----Original Message-----
From: cryptography [mailto:cryptography-bounces+dennis.hamilton=acm.org at metzdowd.com] On Behalf Of Ray Dillinger
Sent: Saturday, November 18, 2017 13:04
To: cryptography at metzdowd.com
Subject: Re: [Cryptography] Is ASN.1 still the thing?

[ ... ]

Why, indeed, should anyone be using a serialized representation of
IEEE754 floats that doesn't roundtrip?  You can express them in hex with
a hex point if you like (some) human readability, but the serialization
is now exact and specifies an unambiguous meaning.

Of course, nobody can propose that to the ASN.1 standardization
committee now, and if one could they'd probably have rejected it.


More information about the cryptography mailing list