#1202: Question: Could the TAG help define "The Web Platform" vs "The Web" or some other useful distinctions?

Visit on Github

Opened Mar 3, 2026

Recently there was a request to "Reconsider W3C Recommendation status of XSLT 2.0 and XSLT 3.0" (see https://github.com/w3ctag/obsoletion/issues/10).

In the process Dan Brickley made what I thought was an interesting comment:

I take this as evidence in favour of distinguishing 'Web Platform' -oriented RECs from 'World Wide Web' RECs.

The former is concerned heavily with implementability, usability, coherent design, attack surface and technical debt issues. The latter is humanity's planet-wide shared memory. Once deployed, its data formats do not really ever go away, and thoughtful specs need to respect this reality. Many W3C efforts have aspects of both flavour of web standard, but are often closer to one cluster.

XSLT appears for now to be primarily a 'World Wide Web' standard, even if 0.02% of page loads is still a significant number for the parties trying to use those pages. XSLT will also continue chugging away in the background, unseen but enabling many other page views.

I don't want to reopen this github issue, but it is worth trying to come up with a non-polarizing framing for where things like XSLT fit in the web standards landscape, and how W3C handles its ageing XML portfolio.

(see https://github.com/w3ctag/obsoletion/issues/10#issuecomment-3637735401)

I think this is an especially interesting point when we also consider that we have a mix of other things deployed in the world too, which aren't just in the browser: Web Views (and embedded views), IWAs, miniapps and so on... It would be interesting to see if there are non-polarizing (maybe more importantly some potentially helpful) ways to define these things...

<!-- Content below this is maintained by @w3c-tag-bot -->

Track conversations at https://tag-github-bot.w3.org/gh/w3ctag/design-reviews/1202

Discussions

Log in to see TAG-private discussions.

Comment by @hadleybeeman Mar 3, 2026 (See Github)

We discussed this at lunch at our W3C TAG f2f in London. Lots of interest in exploring this topic. No conclusions yet on what we could write — many different scopes are possible.

Other outside-the-browser technologies we mentioned: DIDs, verifiable credentials, RDF, RSS

Comment by @bkardell Mar 27, 2026 (See Github)

@ylafon and I discussed this in the Atlantic breakout today and kind of sketched out (I hope that I am articulating this correctly).

It would be useful to define some terminology. We kind of came up with some "buckets" that we think we can think about things in terms of:

  • The Web Platform for stuff that is for inside browsers.
  • The embedded Web Platform - for stuff that uses web engines, and largely inherits from the Web Platform but might have some different characteristics
  • Structured data - this is all of the formats around XML etc
  • Linked Data

We thought it might be good to discuss wether people think there is some value in discussing whether these distinctions/labels could be helpful to make in terms of guidance around designing features, deprecating things, discussing working groups, etc. It could be.

Discussed Mar 30, 2026 (See Github)

Lola: I'd like to follow this.

Brian: Is it OK if we review this, because I think this is not hard to get your head around. I'd appreciate Mike's input on this one. This started with the deprecation (attempt) of XSLT by Chrome in WHATNOT. All of the engines agreed that they'd like to get rid of it. It is problematic. Someone opened a thing suggesting W3C should obsolete XSLT. Then there's the question of lots and lots of other things that that opens. Then Danbri wrote in and suggested that we have difficulty defining this because there's a long tail of other stuff that's valid, and reliant on things like RDF triples and semantic web ideas, even if not done in RDF, and there are semantic things that are used outside of the browser. We need to come to how we reason about them.

... Relevant to my interest in the UA Finding, and the current CRA standards that are being built that talk about different kinds of user agent, and what has to be supported. The question is how do we do that, and what would we do? Yves and I spoke a bit about this last week. We condensed it into a tiny space that I think fits nicely.

... The MiniApps and other WGs are coalescing with WebViews. The concept of 'Embedded Web' is emerging. Can we paint some advice around this?

Lola: Sarven, Yves, and I also discussed this a little bit in a round-about way last week. It'd be good to hear how others want to respond to Brian's comments. I have a question: Jeffrey: you are working on the Web Platform Architecture doc.

Jeffrey: Aspirational, but yes.

Lola: How does that fit in to this? And the UAs Finding. Is there space to define this in that Web Architecture document?

Jeffrey: Yes. This touches on a question that Dan Appelquist posted recently - 'what is the web?' There are 3 categories:

  • The web (linked resources)
  • Web browsers (UAs for visiting a subset of those resources)
  • Web technologies (HTML, CSS, JS, XML, ...) which are used to build those resources but they can also be used in other non-linked areas

We should probably write this down.

Brian: I think that document is going to be hard to write already.

Jeffrey: Maybe start with a Finding. Or individual findings on the sub-pieces.

Brian: +1

Lola: Do we want to start breaking this down into what Findings we'd need.

Jeffrey: This issue may be a reaosnable base for a Finding. We don't need to figure out the architecture of the architecture before we start this.

Brian: There are interesting related questions around security. XSLT has a lot of nice characteristics for offline. Putting it in the browser brings characteristics that may be negative.

Jeffrey: The XSLT security issues were totally solvable. But it didn't have the adoption to justify that.

Brian: But it wasn't done.

Jeffrey: It was put in when a C implementation was reasonable. But not enough adoption to justify moving on. Nor to justify implementing versions 2 and 3 in the browser.

Brian: There is some debate over this; a Finding could help.

Jeffrey: We should split this Finding from detailed questions of XSLT. E.g. there was no reason to deprecate versions 2 and 3.

Brian: It helps to have a thing that helps you to ask questions.

Jeffrey: There's a lot of XML formats that are not browsable in web browsers, but are perhaps part of the Web. XML is based on something that came out of the Web, so they're the Web technologies category.

Yves: There's a difference between the web platform, which is everything browser-related, and the web, which is everything linked. A browser can't process everything that is linked, or has linking capabilities. That's the crux of the definition. I think there's something like that in The Architecutre of the Web.

Mike: Would Atom be one of those?

Jeffrey: Yes in my opinion.

Lola: Next steps on this? We're going to write a preliminary Finding based on this issue here?

Brian: I'll continue to work on it.

Lola: If you want to discuss things in these meetings, keep the agenda label.

Comment by @bkardell Apr 1, 2026 (See Github)

Timely post by @torgo on some of this ...

https://www.torgo.com/blog/2026/03/yes-but-is-it-the-web.html

It would be nice to have more thoughts from him. (h/t @jyasskin)

Comment by @bkardell Apr 10, 2026 (See Github)

dan replied via www-tag email list, and I totally missed it

Thanks. I should write this up properly. We are im strange and changing times, and those can put pressure on established technologies to unpack and be explicit about things that are often kind of lumped together. For years everyone knew what “watching television” meant, even as cable, satellite, dvds and so on arrived. Now that staring at a painted wall (projection…) or a wristwatch, or a canvas in a webxr headset are also kind of “tv”, conversations about TV can’t take old definitions for granted. I think we’re headed there rapidly with “web” too.

Dan

ps. Was ftp://ftp.example.net/pub/papers/1988/doe.j.foo.ps a web page before the web was created? There certainly were huge collections of documents and software on ftp sites, and mentions of them posted to mailing lists, usenet news etc., just without compact urls. This is kind of a goofy example but it also goes to the heart of the matter…

We talk about “the web” as if it were an inspectable entity, when it is something more like a set of shared practices you can place on a spectrum. At one end of the spectrum is a thing so daunting and complex that is costs 100s of millions of dollars to implement. At the other is something so simple and pluralistic that it sort of existed before it was named, and which could outlive HTML and HTTP just as it survived and encompasses Flash, Silverlight, Java applets, SGML, Postscript and PDF.

Discussed Apr 27, 2026 (See Github)

Brian: I've continued to have some discussions on this and had a look at the W3C's current catalog of WGs and things, to see if I can figure that out. No updates yet.

Yves: There was a comment from Heather talking about layering things. I'm not sure the originally proposed layers match the definition that we had, but worth exploring.

Brian: Beyond architecture, I think this is relevant to W3C itself (thinking as an AC rep) - it's a little bit like 'Why are we here? Why are we doing this work in this venue, and in this way? Who is "we"?' - but I don't have anything meaninful at this time.