On Feb 9, 2008, at 1:46 PM, Dave Crocker wrote: > My own assessment is that it has improved the documents. The > proofreaders have > their own views of what is correct and that sometimes requires > discussion, but > mostly I consider their intervention to have a positive impact. > > The question to me -- and which I am posing to everyone else -- is > whether the > improvement is worth the cost? Speaking strictly for myself... Over the past year or so, while we have been using this service, the RFC Editor has been able to clean up the backlog. To my knowledge, we have made no other changes to the RFC Editor's office. Speaking absent data - the RFC Editor's office is largely a black box to me - that suggests that improving the quality of what goes in reduces the time it takes to produce what comes out. Now, that brings several possibilities to mind. One is to suggest widening the pre-processing effort. Another is to suggest that the pre-processing effort is successful with the resources applied, and so needs no widening. A third is to suggest that the RFC Editor increase its copy-editing staff, drop the *separate* pre-processing, but encourage the RFC Editor to skim documents that show up in the IESG's workload and opportunistically start its editing processes early. I have occasionally found myself wondering whether a grammar checker that could read our XML files (about 1/3 of our posted drafts have XML source posted with them) and make suggestions would be of value. In converting what is now RFC 1716 to RFC 1812, which was a HUGE editing task, I used a brand new tool that is now a popular grammar checker. It complained about "which" vs "that", which is neither here nor there, but also complained about misspellings, passive voice, and sentences that it couldn't parse. Apart from the technical differences (changing "subnet" to "prefix" and applying then-new CIDR concepts, deprecating RIP, etc), the vast majority of the changes I made could be described as copy-editing in response to those complaints, and folks seemed to think the result was an improvement. With the new posting tool, the idnits checker is run against every submission, which simplifies the life of the document shepherd. If the upload includes an XML file, would it make sense to similarly grammar check it, which the result being advice for the document author? As you say, the heavy lifting has to be done by people that understand the semantics. But there are ways to improve the syntax, and as I say, the available data would suggest that improving what goes in to the RFC Editor helps them produce the output. _______________________________________________ Ietf@xxxxxxxx http://www.ietf.org/mailman/listinfo/ietf