| 06:56 | <annevk> | keithamus: PRs are now up across DOM, HTML, and WPT. I requested review from you. |
| 13:20 | <annevk> | HTML standard is over 160k lines now. 😅 |
| 13:24 | <Stephen> | Curious what the growth formula is. Exponential with a low exponent? i.e. 5% per year or something? |
| 13:25 | <bkardell> | I think adding code examples with bikeshed grew it a lot, right? |
| 13:26 | <bkardell> | oh no that's not lines necessarily anyway, it was the file size of the output, i think that grew a lot then |
| 13:26 | <bkardell> | I just remember asking about it when it happened because it surprised me :) |
| 13:31 | <annevk> | I wonder if this works: |
| 13:33 | <annevk> | Seems we last managed to remove something in the first half of 2020. And otherwise we typically end up adding several thousand lines each half year. |
| 13:35 | <annevk> | (Draft here is referring to a Review Draft, which seemed like a convenient way to measure as they are pretty much identical to source at a given point in time.) |
| 13:39 | <bkardell> | two extra big leaps in there |
| 13:39 | <bkardell> | 2023-07 and 2026-01 are exceptional |
| 13:43 | <Luke Warlow (back on the 18th)> | I'm guessing the html spec is the single biggest tc39, W3C or whatwg spec? Probably by quite some margin? |
| 13:45 | <Ms2ger> | Time to split into more than one source file? |
| 13:56 | <Tom Schuster> | That would be nice. |
| 14:01 | <annevk> | Would it be? Having to search multi-file does not seem like an improvement per se. As for the earlier discussion, I'm curious what the line count is if you concatenate all of CSS, though this may be hard to do without duplication. |
| 14:29 | <Luke Warlow (back on the 18th)> | I think rather than multiple files it might make sense to move stuff into its own spec document if there's a part of the spec that made sense to do that to. Like workers and worklets as a random idea. But only if the current way isn't working anymore. |
| 14:33 | <annevk> | Possibly, it's been rather useful to have most globals owned by HTML and service workers are in a somewhat worse state because they are not. But yeah, that's been the approach (XMLHttpRequest, parts of Fetch, parts of URL) and works okay. |
| 14:35 | <annevk> | And we moved a bit to Infra. Might be worth evaluating possible targets, but I vaguely recall Domenic at least not finding anything he deemed worth splitting. |
| 14:57 | <Noam Rosenthal> | Some tools choke on that huge file. e.g. can't accept suggestions on github automatically (or is it just me?) |
| 14:58 | <Tom Schuster> | Yeah, I thought I remember something about the github UI complaining. |
| 14:58 | <bkardell> | not just you |
| 14:58 | <Noam Rosenthal> | We talked ages ago about moving general media handling to its own spec, especially images since CSS images rely on them as well |
| 14:59 | <Tom Schuster> | I don't think searching through multiple files is a huge barrier |
| 14:59 | <bkardell> | I think the people who do the editing are easily equipped to do multifile search |
| 14:59 | <Tom Schuster> | This way it's actually easier to find something in a specific section only |
| 15:00 | <Tom Schuster> | (Assuming you split the file by section/chapter) |
| 15:00 | <bkardell> | especially since all of our lives are spent the other way round, looking at small files that are compiled or bundled together :) |
| 15:01 | <bkardell> | it could be pretty easy to generate a concated and annotated file if that is what you wanted to do or even a plugin that does that for you if it's what you preferred as a view |
| 15:13 | <annevk> | Ultimately this is an editor call so dismissing their pov might not yield the desired result |
| 15:17 | <bkardell> | I don't understand the comment annevk ? is it in regard to the splitting or not discussion? |