| 05:38 | <Domenic> | One thing I just thought of that's worth keeping an eye on is the degree of conversion. E.g. if you leave all our raw markup as-is, it might be faster to build than one that converts the markup to markdown and thus incurs Bikeshed's markdown processing. |
| 10:04 | <keithamus> | annevk: I spent the morning splitting out some of the command/commandfor spec PR; https://github.com/whatwg/html/pull/10961 and https://github.com/whatwg/html/pull/10960 are the result. To me these seem like sensible discrete changes, but I'd love to get your feedback - if you think they're unwarranted I can close them out and go back to the one PR. |
| 10:40 | <Noam Rosenthal> | Added some (non-authoritative) notes, if this helps always feel free to ask (I think we're in the same time zone) |
| 11:17 | <keithamus> | thank you! I've addressed your comments, very helpful |
| 11:58 | <jmdyck> | Are you open to converting the HTML spec to markdown (provided the build is still fast)? |
| 12:33 | <Noam Rosenthal> | You might want to start with something less ambitious... converting the entire HTML spec to a format like markdown is a project that holds so many details and pitfalls that would likely end up making it infeasible |
| 13:02 | <Ms2ger> | Note that the question is "is the end point where html is written in markdown desirable at all", not "should this be the first step" |
| 13:04 | <Noam Rosenthal> | I get that, and to me it feels like too many steps ahead to answer. Markdown is great but the devil is in the details |
| 13:06 | <Domenic> | It seems ~pointless to use Bikeshed if we're just writing raw markup instead. You get some small benefits around cross-referencing, but the big one of matching other Bikeshed spec styles and being easier to author is missing. If we just wanted better cross-referencing we could add that to html-build's Rust component in a few working days I think. |
| 13:55 | <Luke Warlow> | Just as an FYI it would be nice to hear the opinions of others regarding (https://github.com/whatwg/html/issues/10953) (cc annevk and smaug) |
| 14:03 | <jmdyck> | Noam Rosenthal: Note that I wasn't offering to do the conversion! I'm not even sure it's a good idea. (I think it'd make some things easier, some harder.) I was just asking about editorial disposition towards the idea. (It's probably come up before, but I haven't been around here long.) |
| 14:11 | <Noam Rosenthal> | jmdyck: I think I totally misunderstood what you originally meant by converting to MarkDown, sorry. Got it now. |
| 14:38 | <foolip> | This is a good point. With more work the conversion tool could convert a lot of raw HTML to Bikeshed-isms, but it's worth measuring if it makes a difference for build times. |
| 14:50 | <annevk> | FWIW, I used to prefer HTML because you lost a certain amount of control about the output with Markdown, but we probably reached the point where imperfect output should be addressed by the tooling instead. Still, a mix of Markdown and HTML makes it unclear how it's going to be parsed. |
| 16:36 | <Dominic Farolino> | I'd love to ask something of folks that touch the HTML Standard. Could you please start using the specfmt tool (https://github.com/domfarolino/specfmt) on your PRs? I'd like to get its output to be part of the HTML Standard CI soon, and that requires more people using it on their PRs so we can find bugs in that tool and harden it accordingly. |
| 16:58 | <Luke Warlow> | Oooh is this the great rewrapper as a CLI tool? I'll definitely start using it. |