This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Discussion: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2014-May/296824.html It sounds like both Google and Mozilla are okay with removing section-* since there is now billing/shipping tokens which were the main driver for sections from Google.
One benefit of the section-* tokens is that you can ask for multiple fields of the same type, as in: <p><label>Japanese name: <input name="j" type="text" autocomplete="section-jp name" inputmode="kana"></label></p> <p><label>Romanised name: <input name="e" type="text" autocomplete="section-en name" inputmode="latin-name"></label></p> If you didn't have section-*, it's not clear how you would indicate to the user agent that these fields shouldn't have the same value.
True, but when you do have section-* in the example shown, it's not clear how you would indicate the different choices the user needs to make if the site is using rAc. I don't think rAc could make this example any easier for the user compared to the website using simple browser autocompletion. See my questions about similar UX issues on the mailing list. Perhaps we should move the discussion back there to avoid parallel discussion? I thought there was consensus on this so I filed the bug.
If rAc() is faced with something it can't work out how to display, it can just return "disabled". That occurs with far more than just "section-*". For example, what would you do if the form had just a bday-day, cc-type, and a tel-local-suffix field? It's hard to imagine that rAc() could do anything useful with that. I responded on the thread as well. If you'd like to continue the conversation there rather than here, I recommend closing this bug in the meantime. Thanks!
http://lists.w3.org/Archives/Public/public-whatwg-archive/2014May/0163.html (the e-mail cited in comment 3) gives reasons to keep these, and hasn't received a reply yet, so for now I'm closing this bug. Feel free to reopen this bug if you'd like to continue the discussion here rather than by e-mail.