Re: ReSpec CANDIDATE CORRECTION wizardry

On 11/7/21 6:48 PM, Marcos Caceres wrote:
> Very few of us have had the opportunity to update a REC, but it's
> something that will become increasingly common (i.e., few of us have
> experience with this so a lot of us a fumbling our way through it).

Since sending that email, I believe I've successfully been able to put
something together that's going to be published on Nov 9th 2021 as an REC with
Proposed Corrections.

It was hella-painful to put together... 12 hours of manual copy-pasting
between an htmldiff to the old rec, creating custom anchors, batching
changesets together, and other manual markup. It took three entire attempts
just to understand what the Team wanted (through no fault of theirs) because
the process was so foreign from what existed before.

Some more thoughts below...

> On the ReSpec side, we only have a bug filed: 
> https://github.com/w3c/respec/issues/3809

Ok, I'll take the conversation there, thanks.

My biggest concern is that I don't know how to get Editors out of the critical
path for marking up candidate corrections/candidate additions/proposed
corrections/proposed additions.

During Verifiable Credentials v1.1, the WG members provided multiple proposed
corrections. None of us knew at the time that we needed to mark these up in
any special way, and so two years of these built up. Luckily, that only
resulted in like 3 proposed corrections... but one of them was a change from
RFC3339 datetime to XMLSCHEMA11-2 datetime, which rippled throughout the
entire spec. Again, that wasn't /that/ bad... but woe to the WG that needs to
search/replace a normative dependency throughout the specification.

We did end up marking all of those PRs with "substantive" github labels, so
perhaps Editors just need to be put on notice that any block-level element
that contains changes needs to be expanded to contain the old text and the new
text? If that's done, then ReSpec can generate the diff between the old text
and the new text... and once the corrections are out, the Editors have to go
back through the document and delete all the old text?

Feels error prone... to the point where it might just be better to do a full
diff of the file. I get that people want to get fancy, but the burden on the
Editors is high and the possibility of error (when doing the cleanup) is
pretty high as well.

I'll put these thoughts in the issue.

-- manu

-- 
Manu Sporny - https://www.linkedin.com/in/manusporny/
Founder/CEO - Digital Bazaar, Inc.
News: Digital Bazaar Announces New Case Studies (2021)
https://www.digitalbazaar.com/

Received on Monday, 8 November 2021 00:40:49 UTC