Hi,

Here's an algorithm to prepare for a clean room implementation of a standard 
track RFC:

1. Prepare a set of documents as the RFC to implement and the transitive 
closure of all the normative references in that RFC.
2. for each document in the set, strip all the non-normative parts.
3. print all the modified documents in TEXT format.

The result, together with the reference manual for the compiler and its 
standard library and maybe a couple books on algorithms, is the minimum needed 
to do a clean room implementation.

We can see that if the RFCs are published in RFCXML then step 1 can be easily 
mechanized.  But step 2 is unfortunately not as simple, as deciding which parts 
are normative and which are just informative have to be done currently using 
heuristics, at least for RFCs.  So my suggestion would be to modify the RFCXML 
format to add a flag on <t> elements (and maybe <table> elements) to clearly 
indicate which ones are normative (i.e., required to implement an interoperable 
piece of software) and which ones are not.

Note that solving the problem of having that flag rendered into the derived 
formats is not part of this request.

Thanks.

-- 
Marc Petit-Huguenin
Email: [email protected]
Blog: https://medium.com/@petithug
Profile: https://www.linkedin.com/in/petithug

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature

_______________________________________________
rfc-interest mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to