--On Thursday, 08 January, 2004 23:32 +0100 jfcm <info@xxxxxxxx> wrote:
At 21:45 08/01/04, John C Klensin wrote:A better answer would have been "the term 'request for comment' is historical, dating from a time when the preferred way to make a formal comment on a document involved writing another document, which then was numbered into the series". That mechanism is still available, although usually very slow. But documents that become RFCs are now first posted as Internet Drafts (see http://www.ietf.org/ID); comments on those are both solicited and, usually, handled very quickly.
Today, the RFC Series, despite retention of the original name and numbering series, acts as a permanent, archival, repository of information, decisions taken, and standards published. As such, documents in the series are subjected to review and editing processes (which differ somewhat depending on the type of document and are appropriate for conventional references from conventional documents. Running conversations, logs of comments, etc., are not well suited for that archival and reference role, regardless of their other advantages and disadvantages.
Could it not be useful to have a "List of Comments" (LOC) for each RFC? Where experience about the RFC reading, testing and implementation could be listed by the authors (or a successor) from experience and questions received. It would avoid the same questions to be debated again and again and it would help further thinking. These comments could start with a summary of the WG debated issues, explaining the whys of some options. I suppose the implementation would be easy enough since it would follow the same numbering scheme and titles. Such a LOC being an updated appendix could be reviewed and help preparing replacements.
Jefsey,
There are many ways by which almost any system could be, or could have been, designed or evolved. Typically each has its advantages and disadvantages, even if, overall, the difference in effect is slight. Certainly the Internet standards and definitions process could have evolved with RFCs and LOCs or some other pieces. Certainly the idea of "standards" could have been put in place early in the process, and we could have called what are now "Internet-Drafts" "RFCs" instead, could have had standards in a separate series, and had an additional separate series for experimental and informational documents (we might even have followed the lead of some other bodies and called the latter series "technical reports", which they are definitely not). And certainly we could have adopted a distinct numbering system and document series for standards along with a regular review model, standing WGs to which each standard is applied, and revisions that share the number but have different dates or revision numbers.
In some cases, the fact that we didn't go down those paths is an historical accident. In others, it was a very deliberate decision. And, in some cases, the path that was chosen was very important in shaping how the IETF operates or reflects the nature of those operations (e.g., the fact that we generally close WGs that have finished their work, rather than encouraging them to find new things to keep them busy has, I believe, been very important in keeping us from developing many standards that no one wants and that are out of touch with reality).
But we do things in a particular way that mostly works (although it should probably be more effectively and accessibly documented), and, while we might tune things differently or use different terminology if we were starting over today, what purpose would it serve to develop redundant mechanisms that would require extra volunteer or staff effort to make work and keep working and that would leave people uncertain as to where to go for information? And what purpose is served by proposing such an action?
Regards, john