Re: Workload constants [was I-D Action: draft-rsalz-termlimits-00.txt]

[Colin Perkins <csp@xxxxxxxxxxxxx>]

On 22 Oct 2021, at 18:08, Salz, Rich <rsalz=40akamai.com@xxxxxxxxxxxxxx> wrote:

  There was a huge difference in another metric, the time from initial

   draft to published RFC. Back in 1998, the median delay was one year. It
   was already more than 3 years in 2008, and still is. Much of the delays
   happen in the working group, from initial draft to last call.

So the common beliefs that things are taking longer is wrong (for the past 13 years!) and that the delays are because WG's are sending less-good documents to the IESG is wrong.  My theory for the possible increased mail traffic is that folks are doing more business between meetings, and the increased use of GitHub which increases volume because of increased participation.

The measurements we’ve done recently show that the median time to publish RFCs is still trending slowly upwards, although there’s a lot of variation. The number of drafts it takes before an RFC is published is also trending up, as is the number of citations from RFCs to other drafts and RFCs (i.e., there are more cross references between documents).

One view might be that there’s a lot more legacy protocols and a bigger installed base that we need to take into account now, so RFCs are more complex and take longer to produce. I’m sure there are other factors involved too. 

Lots more data in https://sodestream.github.io/paper-characterising-the-ietf-through-the-lens-of-rfc-deployment.html if people are interested. 

-- 
Colin Perkins
https://csperkins.org/