Update device client hint information

[hamishwillee]

This updates DPR HTTP header with directives section, and additional info from the last spec I can find in which it was mentioned (in the next version there is info that says headers were removed to their own specs, but I cannot yet find.

It also adds info about using Vary, and what happens if the value appears multiple times in response. That's from the spec.

I removed info about Accept-CH-Lifetime from this page, as that header is deprecated.

This is part of addressing #1408 and follows on from #5883.

• BCD work to test if all client hints should have the secure header tagging here: HTTP DPR header is secure context browser-compat-data#11144
• BCD work to mark headers deprecated here: Content-DPR, Viewport-Width, Width: set deprecated true browser-compat-data#11186

@Elchi3 Feel free to review/merge this. If you don't, I'll continue to add the other device client hints docs until done. But hoping for sanity check. Note, I'm also after advice on how to find the spec for this.

[github-actions]

Preview URLs

• https://pr6097.content.dev.mdn.mozit.cloud/en-US/docs/Web/HTTP/Headers
• https://pr6097.content.dev.mdn.mozit.cloud/en-US/docs/Web/HTTP/Headers/Width
• https://pr6097.content.dev.mdn.mozit.cloud/en-US/docs/Web/HTTP/Headers/Device-Memory
• https://pr6097.content.dev.mdn.mozit.cloud/en-US/docs/Web/HTTP/Headers/Content-DPR
• https://pr6097.content.dev.mdn.mozit.cloud/en-US/docs/Web/HTTP/Headers/Viewport-Width
• https://pr6097.content.dev.mdn.mozit.cloud/en-US/docs/Web/HTTP/Headers/DPR

Flaws

Note! 5 documents with no flaws that don't need to be listed. 🎉

URL: /en-US/docs/Web/HTTP/Headers
Title: HTTP headers
on GitHub
Flaw count: 4

• macros:
  • /en-US/docs/Glossary/COEP does not exist
  • /en-US/docs/Glossary/COOP does not exist
  • /en-US/docs/Glossary/CORP does not exist
• broken_links:
  • Can't resolve /en-US/docs/Web/API/Save_Data_API

External URLs

URL: /en-US/docs/Web/HTTP/Headers
Title: HTTP headers
on GitHub

No new external URLs

---

URL: /en-US/docs/Web/HTTP/Headers/Width
Title: Width
on GitHub

• https://datatracker.ietf.org/doc/html/draft-ietf-httpbis-client-hints-07 (1 time) (Note! This may be a new URL 👀)
• https://developers.google.com/web/fundamentals/performance/optimizing-content-efficiency/client-hints (1 time) (Note! This may be a new URL 👀)
• https://wicg.github.io/responsive-image-client-hints/ (1 time) (Note! This may be a new URL 👀)

---

URL: /en-US/docs/Web/HTTP/Headers/Device-Memory
Title: Device-Memory
on GitHub

• https://developers.google.com/web/fundamentals/performance/optimizing-content-efficiency/client-hints (1 time) (Note! This may be a new URL 👀)

---

URL: /en-US/docs/Web/HTTP/Headers/Content-DPR
Title: Content-DPR
on GitHub

• https://datatracker.ietf.org/doc/html/draft-ietf-httpbis-client-hints-07 (1 time) (Note! This may be a new URL 👀)
• https://developers.google.com/web/fundamentals/performance/optimizing-content-efficiency/client-hints (1 time) (Note! This may be a new URL 👀)
• https://wicg.github.io/responsive-image-client-hints (1 time) (Note! This may be a new URL 👀)

---

URL: /en-US/docs/Web/HTTP/Headers/Viewport-Width
Title: Viewport-Width
on GitHub

• https://datatracker.ietf.org/doc/html/draft-ietf-httpbis-client-hints-07 (1 time) (Note! This may be a new URL 👀)
• https://developers.google.com/web/fundamentals/performance/optimizing-content-efficiency/client-hints (1 time) (Note! This may be a new URL 👀)
• https://wicg.github.io/responsive-image-client-hints/ (1 time) (Note! This may be a new URL 👀)

---

URL: /en-US/docs/Web/HTTP/Headers/DPR
Title: DPR
on GitHub

• https://datatracker.ietf.org/doc/html/draft-ietf-httpbis-client-hints-07 (1 time) (Note! This may be a new URL 👀)
• https://developers.google.com/web/fundamentals/performance/optimizing-content-efficiency/client-hints (1 time) (Note! This may be a new URL 👀)
• https://wicg.github.io/responsive-image-client-hints/ (1 time) (Note! This may be a new URL 👀)

(this comment was updated 2021-06-25 01:42:58.975625)

[Elchi3]

If there is no spec, then I think we should remove {{Specifications}}.

[hamishwillee]

@Elchi3 OK, this updated and I have added the other screen client hints. It is ready for re-review. There are bits that are "thin" but should be correct (I don't want to go into too much investigation on the deprecated API e.g. Content-DPR).
If you're OK with this it can be merged as the result of any answer below can be applied iteratively.

Following a helpful response on WHATWG matrix group I now know that the screen client hints moved to https://wicg.github.io/responsive-image-client-hints/. What it looks like happened is

• Width, Viewport-Width, DPR changed to Sec-CH-Width, Sec-CH-Viewport-Width, Sec-CH-DPR
• New header added: Sec-CH-Viewport-Height
• Content-DPR removed and no equivalent in new spec but is somewhat covered (according to Yoav Weiss) by Allow an image to indicate its own density and correct its intrinsic size whatwg/html#5574

But Chromium not yet made the change.

My assumption is that rather than attempt to rename/merge the new variants on the header we would just create new pages and new BCD entries for them and deprecate the old ones (@ddbeck). Is this correct?

What would be the trigger for this? Feels a bit early - not in any browser and there are still questions in the spec.

[ddbeck]

My assumption is that rather than attempt to rename/merge the new variants on the header we would just create new pages and new BCD entries for them and deprecate the old ones (@ddbeck). Is this correct?

This is right. The schema does have a notion of true renames (i.e., prefixes and alternative_name), but if it's not actually a rename—just a successor—then create a new feature.

What would be the trigger for this? Feels a bit early - not in any browser and there are still questions in the spec.

For BCD, avoid introducing all false features, unless there's evidence that implementation is in progress (i.e., if you can't link to an open implementation bug, then don't bother). We don't have a strict guideline about this, but mdn/browser-compat-data#10619 is open to answer that question.

[Elchi3]

Thanks Hamish, great work! (again!)
Some nits inline.

If there is no implementation of the new Sec-CH-* headers, then I think documentation work should wait. The spec can change again and so it would be a catch-up game with something that isn't real yet and so just work for us with no user benefit.

[hamishwillee]

Thanks very much @Elchi3 . Accepted all your commits and integrated your suggestions. Ready for another look.

[Elchi3]

Awesome! 👍