LargestContentfulPaint and PaintTiming are separate specs

[andydavies]

LCP and FCP are both paint metrics yet they live in separate specs, and have different Performance Entries interfaces.

This separation leads to confusion

• it doesn't make sense that observing just paint entries won't capture LCP
• LCP has attributes for element, it's on screen rect, natural height, width etc. whereas FCP doesn't but yet aren't they both observing an area of the viewport?

Know it's early days for LCP but if it's a paint metric then surely it belongs with the other paint metrics?

Also commented in LCP spec see w3c/largest-contentful-paint#36 (comment)

[yoavweiss]

/cc @npm1

[npm1]

Thanks for the feedback! To consolidate my thoughts:

• entryType cannot be made the same due to compat reasons. Right now developers get two entries of type paint, FP and FCP. If there were no users of Paint Timing at the moment, using the same entryType for LCP would make sense. But that's not the case so we shouldn't break the existing users.

• Agreed that we could expose more information on PaintTiming. Note that it will not be equivalent to LCP, for instance we can trigger FP with an image without url(), or with an image which is not fully loaded yet (whereas those cannot occur in LCP entries). Filed Add attribution information about what triggers paint #44 to track adding more attribution on Paint Timing.

• They live in separate specs because one is adopted by W3C and the other is in incubation in WICG. I'd be happy to attempt merging the two when LCP is adopted by W3C.

[npm1]

Renamed this, let's keep it to keep track the problem of PaintTiming and LCP being separate specs.