APCA trademark

[svgeesus]

I notice that the APCA W3 README was changed on 25 April 2022 to add a trademark symbol to the APCA name. The commit message was Link to discussions area which seems unrelated to this change

A search on the WIPO Global Brand Database to APCA revealed 20 trademarks, 15 active, and none related to this usage. Similarly a search on the UK Intellectual Property office turned up one hit for APCA registered to a shipping company in Japan.

Raising this issue therefore to see whether all mention of APCA in the WCAG materials should be to APCA™ and also to enquire where and when the trademark was registered.

[Myndex]

Hi Chris @svgeesus

There are two forms of trademark, ™ and ® and neither one is required to be affixed, doing so only indicates the intention to protect (i.e. intention to take legal actions). Similar to how a copyright notice is not actually needed, (however, in the case of copyright it's automatic at creation + affixation to physical medium.)

The ™ has been on a lot of the APCA™ related logos going back to 2019. The reason I was compelled to add in more places and with more restrictive language, is due to misuse and abuse, meaning some bad implementations claiming to be APCA but not even close to correct results or methods.

I was not expecting the sudden popularity and rapid adoption of APCA in tools all over the place. I then learned that not everyone is "appropriately diligent" in implementing code. These "bad" unauthorized implementations where then used by some trolls to harass me personally as well as deride APCA™.

And just today I came across an example of reverse engineering (attempts) of APCA™.

This is all frustrating, as at the moment I am trying to focus on finishing some papers for publication, but that is now amidst having to protect my intellectual property, most of which is eventually intended for a "free use" license... Apparently "public domain" is a terrible idea as it gives rise to counterfeit and inferior implementations.

I suppose this is the consequence of being open about the work during the development phase.

Raising this issue therefore to see whether all mention of APCA in the WCAG materials should be to APCA™

WELL, all things considered, if the W3C would like to help by indicating limitations of use, adding at least the ™ mark, etc., I would be pretty happy—there needs to be a way to maintain code uniformity here as this is intended for standards!!!

Thank you for reading

[svgeesus]

The reason I was compelled to add in more places and with more restrictive language, is due to misuse and abuse, meaning some bad implementations claiming to be APCA but not even close to correct results or methods.

I was not expecting the sudden popularity and rapid adoption of APCA in tools all over the place. I then learned that not everyone is "appropriately diligent" in implementing code

I think the usual way to deal with that situation is to have test cases. In this case, a bunch of (foreground, background) color pairs and the expected correct L^c for each pair (in the order specified).

The issue is that APCA is not an in-browser technology so this wouldn't be suitable for Web Platform Tests (WPT). But I would certainly like to see such tests, especially as I recently implemented APCA myself for color.js. and my next step once the pull request s merged is to add some tests.

[Myndex]

Hi @svgeesus

Just as an FYI, the those tests are in the apca-w3 repo, and the npm package apca-w3

[svgeesus]

Cheers, I put them into the color.js tests. My implementation passes all except the last one:

    #888 vs #fff  •  63.056469930209424
    #fff vs #888  • -68.54146436644962  

    #000 vs #aaa  •  58.146262578561334
    #aaa vs #000  • -56.24113336839742
    
    #123 vs #def  •  91.66830811481631
    #def vs #123  • -93.06770049484275

    #123 vs #444  •   8.32326136957393
    #234 vs #444  •  -7.526878460278154

For text=#234 and bg=#444 I get 0, not -7.526878460278154

So does your APCA contrast calculator would you mind checking this one again?

[svgeesus]

Also that last test result cannot be right because the L^c is negative so the text must be lighter. Swapping them round so text=#444 and bg=#123 I still get 0 on your APCA tool but -7.527 on SAPC implying these are using different constants or a different algorithm in some way.

So what I think happended here is:

• the text and background colors got swapped in the last test
• the expected result is from the SAPC calculator, but this is a test for APCA so the result should be the APCA one

By the way, in the default configuration, what is the difference between the APCA and SAPC constants and formulae?

[Myndex]

DOH!

Hi @svgeesus thank you for catching this... yea, that's not correct...

#123 and #444 were intended to be just above the clamp IIR... so lemme see what up there...

[Myndex]

Okay @svgeesus I see hat happened... the #234 was from an earlier version, and that last test should have been

#444 (text) and #123 (BG) this reports as -7.5 on the APCA tool at https://www.myndex.com/APCA/

Sorry for any confusion, I'm looking through the docs now to see if it is listed this way anyplace else...

[Myndex]

Also, this is a documentation error, it was correct in the automated test scripts.