[COFF] Joys of ASN.1 [Was: Re: tangential unix question: whatever happened to NeWS?]

[Nemo Nusquam]

Migration to COFF, methinks

On 30/01/2021 18:20, John Cowan wrote:
> Those were just examples.  The hard part is parsing schemas, 
> especially if you're writing in C and don't know about yacc and lex.  
> That code tends to be horribly buggy.
True but tools such as the commercial ASN.1 -> C translators are fairly 
good and even asn1c has come a long way in the past few decades.

N.

>
> But unless you need to support PER (which outright requires the 
> schema) or unless you are trying to map ASN.1 compound objects to C 
> structs or the equivalent, you can just process the whole thing in the 
> same way you would JSON, except that it's binary and there are more 
> types.  Easy-peasy, especially in a dynamically typed language.
>
> Once there was a person on the xml-dev mailing list who kept repeating 
> himself, insisting on the superiority of ASN.1 to XML.  Finally I told 
> him privately that his emails could be encoded in PER by using 0x01 to 
> represent him (as the value of the author field) and allowing the 
> recipients to reconstruct the message from that!  He took it in good part.
>
>
>
> John Cowan http://vrici.lojban.org/~cowan 
> <http://vrici.lojban.org/%7Ecowan> cowan at ccil.org <mailto:cowan at ccil.org>
> Don't be so humble.  You're not that great.
>   --Golda Meir
>
>
> On Fri, Jan 29, 2021 at 10:52 PM Richard Salz <rich.salz at gmail.com 
> <mailto:rich.salz at gmail.com>> wrote:
>
>     PER is not the reason for the hatred of ASN.1, it's more that the
>     specs were created by a pay-to-play organization that fought
>     against TCP/IP, the specs were not freely available for long
>     years, BER was too flexible, and the DER rules were almost too
>     hard to get right.  Just a terse summary because this is probably
>     off-topic for TUHS.
>