At 5:48 PM -0400 6/18/07, Joel M. Halpern wrote: >Clarification below. > >At 06:16 AM 6/18/2007, Dave Cridland wrote: >>On Mon Jun 18 08:30:00 2007, Simon Josefsson wrote: >>> > If you do believe the ABNF needs special licensing in >>>> this case, I am sorry to say that your remedy is not sufficient. >>>This >>>> document imports ABNF from other documents (from RFC 3986, >>>> to take one important example). Those documents do not have >>>> anything like what you suggest. >>I disagree, I think they do have what Simon's suggesting. But then, I think RFC2192bis does, as well. > > >Given existing wording, it is somewhere between an extremely good idea and necessary to indicate that the ABNF and the code (and yes, I agree with Simon that it needs to apply to both) can be extracted and modified as necessary by anyone. And while we are trying to solve this for the future, that unfortunately can not help current documents. I think we're in real danger of going down a very deep rathole. We can start dancing on the "what is code" pin-head and hope to avoid it, but it's a pretty faint hope. To refresh your memory, here is what the IETF trust FAQs that Dave has pointed to say: >6. Am I allowed to reproduce extracts from RFCs? >It is common to use extracts from RFCs that are in >the form of computer code by incorporating them >in software. This is the only usage formally allowed >by the current IETF rules (RFC 3978). > >Generally speaking the IETF Trust will tolerate fair use >of other extracts, but you must indicate the source of the >extract and you must mention the original copyright statement >if present. > >7. Am I allowed to publish modified extracts from RFCs? >It is acceptable under the current IETF rules (RFC 3978) >to modify extracted code if necessary. > >Modification of other extracts requires the permission of the original >authors. The IETF Trust does not in general grant the right to >create derivative works of RFCs; in fact it does not have >the right to do so, under the current IETF rules (RFC 3978). > >The IETF is currently discussing various possible >modifications of its rules to permit the publishing >of modified extracts in certain circumstances. (As noted, 3978 is where the IETF gets these rights from Contributors). If ABNF is code, in other words, we are covered. So, I believe we all also agree a human can read the RFC and code an implementation after having done so without a copyright violation. If the ABNF is specification, in other words, we are covered provided the process to turn it into code is in wetware. If the process is in other code, the question becomes whether the syntax expressed in ABNF in the relevant document is code or input on which the code works. I certainly think that can be treated as code, as do others. If Simon wants to write a one-line RFC that says "ABNF [RFC4234] productions in IETF RFCs are code for the purposes of BCP 78[BCP 78]", I will think it redundant but harmless. But having this conversation on the IETF list for every document that contains ABNF as it goes by is not really how this should go, and having everyone have to make up new license terms to include can run into its own problems, as Simon and I both know all too well. regards, Ted _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf