WHATWG is considering to deprecate XSLT

[jobbautista9]

https://github.com/whatwg/html/issues/11523

I only noticed this because someone in the fediverse mentioned the Pale Moon account for an opinion, I've tried my best to give what I think would be our position in this kind of situation: https://sociale.network/@oblomov/115031649278311653 (full conversation), https://outerheaven.club/notice/AxFlFCfzzgRRpvubVw (response itself, but doesn't show full conversation for logged out users)

[vannilla]

Everything must be Javascript.

[Moonchild]

Everything must be Javascript.

Goes right along with "everything must be https" (DoH, removal of FTP, etc.)

EDIT: Come to think of it, they probably want to use the Streams JS API to replace XSLT. Because why not overload the JS engine with more pointless tasks that are done better and faster in native implementations...

[Moonchild]

Also, I'd chime in myself on the WhatWG issue but it was locked as "too heated" 2 days ago, so the only thing I could do was add to the overwhelming majority of downvotes on the initial post. I guess contextless voting is all we're left with at the moment.

I see the point was at least raised before it was shut down that similar efforts of killing mature, stable web technologies was attempted with SMIL, MathML, etc.
Just because something can be done with js or a polyfill (in js or wasm) doesn't mean the native implementation is somehow bad. JS has a much, much much larger and less maintainable attack surface than individual implementations in native code. The "security" angle is BS.

[jobbautista9]

Ironically they don't even trust their JavaScript and WebAssembly implementations enough to even consider shipping a wasm polyfill implementing XSLT 1.0 inside their browsers. So yeah the security argument is definitely bs. They'd really rather break a lot of websites (whether it's old ones their authors have forgotten about back in the good 'ol XML days, or new ones which style syndicated feeds)... At that point they might as well drop HTML 3 and 4 support like the old frame tags as well as all deprecated JavaScript methods like document.write().

[Moonchild]

they don't even trust their JavaScript and WebAssembly implementations enough to even consider shipping a wasm polyfill implementing XSLT 1.0 inside their browsers.

From an end-user point of view I wouldn't want a wasm implementation anyway. Transpiled WASM is like a black box binary component (the only difference is it's delivered as bytecode instead of natively executable) unlike a native implementation that can easily be audited, especially if it's mature and doesn't see a lot of code churn. I don't see this working for any high-sec environment as a result, either. So instead of a fast, safe native implementation they are asking us to trust some arbitrary blob that "gets shipped with" but isn't part of, the browser. What was this whole fight against NPAPI plugins about again, if that's the direction they are going in?

[vannilla]

In light of this announcement I've seen a growing number of articles talking about Google's attempts at undermining the open web. Earlier today I've even read one making a historical record of the major, and some minor, events.
Beyond the contents of the articles, one detail that really caught my eye is nobody knows about Pale Moon or Basilisk or UXP. One of these articles (well, it was a thread of posts on the Fediverse but same difference) even claimed the only web browser with built in support of RSS was Vivaldi.
It's honestly insane that these people fighting the good fight don't bother looking further than Firefox or a Chrome clone.

[Moonchild]

Beyond the contents of the articles, one detail that really caught my eye is nobody knows about Pale Moon or Basilisk or UXP.

It depends on where these articles are published. If they are published on any of the IGN/ZD tech news sites it's a good possibility that it's actively being avoided even if they know about UXP and apps building on it.

[vannilla]

It depends on where these articles are published. If they are published on any of the IGN/ZD tech news sites it's a good possibility that it's actively being avoided even if they know about UXP and apps building on it.

Most of what I read was on personal websites, not news outlets. They were not quite blog posts, hence "articles".

[Moonchild]

I see. But, it's a broad effect anyway. Writers won't know to check if they don't know. Then they won't write about it, and others won't know, etc.
Of course the bouts of animosity from various groups in the past doesn't help either, making the threshold to get anything published about it higher.
Either way, I'm just going by what little information I have.

[jobbautista9]

And the deprecation begins: https://groups.google.com/a/chromium.or ... s4EsD5AQAJ

It's funny they noted that the feedback from web developers is overall negative, and still went with the deprecation just because Gecko and WebKit are fine with it. Really goes to show how WHATWG is just an implementers' cartel who don't care about what actual users think, unlike the W3C.

Also XSLT's specs are still owned by W3C right? How is WHATWG "deprecating" something they don't even handle? Like XSLT 1.0 will still be a web standard, whether the implementers like it or not. W3C only gave WHATWG authorship over HTML and DOM...

[Moonchild]

Maybe someone needs to clue in the W3C that their authority is being undermined...