Bug in rules: link for RFC2119

[svgeesus]

Link for "Key words for use in RFCs to Indicate Requirement Levels" is pointing to unreliable domain or legacy service "https://tools.ietf.org/html/rfc2119".

Okay but it works, and the warning should say what to link to instead.

For example I could use https://www.ietf.org/rfc/rfc2119.html but compared to the "legacy" url it does not have the "updated by" or "errata exist" annotations.

Maybe https://www.rfc-editor.org/rfc/rfc2119.html is a good one?

Found while checking https://www.w3.org/TR/2021/PR-webaudio-20210506/.

[jennyliang220]

Hi @svgeesus , please refer to #1178:

we should generate warnings (or errors?) if one uses https://tools.ietf.org/.

[svgeesus]

Hi @jennyliang220 I think you missed my point.

I understand why the warning is given. My point is that, instead of saying "don't use this" the warning should say "don't use this, use that instead".

[deniak]

If the legacy link is managed by W3C, then it would be safe to list alternatives but for outside links, it's tricky to maintain an up-to-date list. We don't necessarily follow how these resources are handled and I think it's the role of the editors to make sure he's using pointing to the right resources. We can only warn about unreliable services.

[svgeesus]

It doesn't seem hard to issue W3C-wide guidance on how to link to RFCs.

Forcing individual editors to each independently solve this is severely suboptimal and will just result in ignored warnings.

[xfq]

I sent speced/bikeshed#2076 , which might be related to this issue.

[svgeesus]

Thanks @xfq that fixes it for CSSWG and other Bkeshed users.

for outside links, it's tricky to maintain an up-to-date list

Right, we have no idea who IETF are and its not like we have any ongoing coordination meetings with them or anything. Come on.

Your point is valid in general, but invalid in this specific case.

[jennyliang220]

@plehegar, are we going to suggest alternative services if there are any? If so, who will maintain the list？