2 ms·

Re: > Fixing known bugs is also a kind of continuous maintaining of a standard (this is what the W3C and other standard bodies do). And: > Fixing bugs and pu

Re: > Fixing known bugs is also a kind of continuous maintaining of a standard (this is what the W3C and other standard bodies do). And: > Fixing bugs and publishing errata is common for standards. The wording implies that this is something the W3C does not do. There are countless W3C specs that don't have errata kept up-to-date, and, hell, the 5.1 errata is just <https://github.com/w3c/html/issues> https://github.com/w3c/html/issues> (if they fix a bug in 5.2 drafts, will they keep the issue open so it continues to appear on that (paginated) list of issues or not? because both options seem terrible!). Similarly, the W3C almost never issues revised recommendations to incorporate errata, which means even for specs where there have been no errata for years you have to cross-reference two documents. Another example is CSS 2.1, which doesn't have errata published any more: they're just fixed in the editor's draft of 2.2, and left to linger in the 2.1 Recommendation and errata. For plenty of W3C specs the only "errata" is an editor's draft of the next revision.

wolfgke 10y ago
Counterexample: XML 1.0, which is currently in Fifth Edition: > https://www.w3.org/TR/xml/ https://www.w3.org/TR/xml/ (click under "Previous versions" to go to older editions).
- gsnedders 10y ago
  Oh, yes, there are absolutely counter-examples. It's essentially down to each WG whether they issue errata or revised recommendations, and when plenty of (admittedly more obscure) WGs have their charter allowed to expire when all of their deliverables have become recommendations there's at times nobody to maintain the document per the W3C process. An aside: XML 1.0 Fifth Edition is actually a weird case: there are documents that are well-formed XML according to the Fifth Edition and not well-formed according to the Fourth Edition, because the characters allowed in places like tag names went from being based on Unicode 2.0 to Unicode 5.0-or-any-later-version (which was originally the difference between XML 1.0 and XML 1.1, except 1.1 was originally 3.0-or-any-later-version); as such, two XML 1.0 implementations are no longer guaranteed to consider the same set of documents well-formed, as it depends on the Unicode version each uses. As a result, plenty of implementations of XML 1.0 have no plan on changing away from their Forth Edition implementations.

wolfgke 10y ago

Counterexample: XML 1.0, which is currently in Fifth Edition: > https://www.w3.org/TR/xml/ https://www.w3.org/TR/xml/ (click under "Previous versions" to go to older editions).

gsnedders 10y ago
Oh, yes, there are absolutely counter-examples. It's essentially down to each WG whether they issue errata or revised recommendations, and when plenty of (admittedly more obscure) WGs have their charter allowed to expire when all of their deliverables have become recommendations there's at times nobody to maintain the document per the W3C process. An aside: XML 1.0 Fifth Edition is actually a weird case: there are documents that are well-formed XML according to the Fifth Edition and not well-formed according to the Fourth Edition, because the characters allowed in places like tag names went from being based on Unicode 2.0 to Unicode 5.0-or-any-later-version (which was originally the difference between XML 1.0 and XML 1.1, except 1.1 was originally 3.0-or-any-later-version); as such, two XML 1.0 implementations are no longer guaranteed to consider the same set of documents well-formed, as it depends on the Unicode version each uses. As a result, plenty of implementations of XML 1.0 have no plan on changing away from their Forth Edition implementations.