On 2018-09-13 08:51, Stephen Farrell wrote: > > Hiya, > > On 12/09/18 19:10, Adam Roach wrote: >> Thanks for weighing in, Bob. I have a request for clarification. >> >> On 9/11/18 1:31 PM, Bob Hinden wrote: >>> I worry about the >>> >>> "abstract should contain enough detail to help readers decide if >>> they need to read the rest of the RFC†>>> >>> This may result in a long Abstract depending on the nature of the >>> “update†and defeat the purpose of the abstract if it gets too long. >> >> >> Based on other people's comments, you're not alone in such a concern, >> and I'd like help understanding why. My conception of document abstracts >> in general is that they serve the singular purpose of letting people >> know whether they should read the rest of the document. This makes the >> text you quote almost tautological. >> >> Clearly, you (and others) have a different notion of the purpose of the >> abstract section, and that understanding is driving these concerns. I >> don't think we can reach a reasonable consensus unless we surface what >> those differences in understanding are. >> >> Can you concisely describe what purpose you believe an abstract serves? > > There are more than just functional properties. There's the issue of > folks having to jump through bureaucratic hoops. For some documents, > it'll be just fine if this information is somewhere other than the > abstract. Exactly. Common sense should apply in all cases. If the abstract is too long or looks like a catalogue of boring stuff, it will itself become TL;DR and therefore defeat its own purpose. So all - that's *all* - of these guidelines should be *guidelines* that are varied when that seems appropriate. If a document updates 8 others, then the abstract could say: This document updates all the RFCs underlying Foobar. Or for a current topic, perhaps: This document updates numerous RFCs that mention the IAOC. As Spencer says from time to time, do the right thing. When the rules or guidelines are wrong, do the right thing. (Possible update to RFC2119: if MUST is clearly nonsense in a particular case, ignore it, especially in a BCP.) Brian
