Re: [csswg-drafts] Let’s Define CSS 4 (#4770)

This is a great discussion! Allow me to summarise and reply to a few points - all from my perspective as a supporter of the CSS4 plan.

First, some people are afraid that CSS4 will confuse CSS devs because it diverges from the current modular approach, and from earlier statements that there will not be a CSS4. I do not think this is a problem, because the huge majority of CSS devs has no idea how the W3C process works and doesn't want (or need) to know. If they want to learn we should accomodate them, but I don't feel that learning about the W3C process is mandatory for the average CSS developer.

If someone delves deeply enough into the specs that they see that there's a mismatch between module levels and the idea of CSS4, I think they're advanced enough to be let in on the secret that CSS4 is just marketing.

An alternative to CSS4 now and CSS5 in three to five years would be an annual CSS2020, CSS20201 and so on, similar to what ECMA is doing. I do not think that this would be advisable. A three-to-five-year release cycle is more in line with educational needs. With a one-year release cycle it doesn't make sense to produce books and workshops about the new release, because by the time they're finished the next release is just around the corner and your book or workshop will seem outdated. A longer cycle sidesteps this issue.

It is unclear to me why an annual release cycle would be better. I'd like to ask proponents of the annual cycle to clarify the advantages they see over a three-to-five year cycle. (And for me, "being more like ECMA" is not good enough.)

It seems CSS developers also want to use CSS4 as a feature list, giving them guidance as to what browsers support now or will support in the near future. I feel this can be achieved fairly easily.

I think we should take the CSS 2020 snapshot, which I believe is being created right now, remove the features that were added to browsers before 2019, and put all remaining features on the "CSS4 feature list". This should be a fairly easy process for the WG, since the snapshot will be created anyway.

The WG should continue to operate as always: work on modules and create a snapshot once per year. It should not be forced to take on even more work. As far as I can see, adopting the CSS4 proposal would increase the WG workload only marginally, with two extra tasks:

1) Create a short blurb about CSS4, as well as the initial feature list as described above. This would be a one-time extra workload.
2) Update the CSS4 feature list by reviewing recommendations from CSS developers. I think this can be pretty much a rubber-stamp process, where a feature is rejected only if it's too old (say, implemented in 2018 or before), or there are serious other reservations from multiple WG members. This should be only a small amount of extra work (he said hopefully).

That's where I stand right now.

-- 
GitHub Notification of comment by pp-koch
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/4770#issuecomment-590365871 using your GitHub account

Received on Monday, 24 February 2020 15:04:54 UTC