Sorry for the interruption, but I would rather ask: “What is the reason for giving up on many users just to implement one useless and frankly even annoying feature at all costs?”
Because, let’s be honest, who cares about managing alternative colors, given that most users follow forums that use your platform for their content and not for aesthetic appearance?
If you look closely, the “relative color syntax” feature is essentially the only one that is not supported by many of the older browsers, while the other two are supported (or at least, even my old Firefox passes the tests for the other two, as useless as they might be, and only fails on the relative color syntax, which is TOTALLY unnecessary for forum content).
It’s not so hard to decide not to implement something totally useless, right?
Also consider that, according to some sources, Windows 7 is completely insecure (which is false, but let’s leave that aside for now), and that it is no longer used by anyone… According to various online analysis sites (none of which give exact numbers, so it’s an estimate), there are still between 60 and 100 million users for various reasons (backward compatibility, software that no longer works on Win10/11, lack of financial means to upgrade, CNC machines that run on Win7Pro but not on 10/11, and others).
Although this is a small percentage compared to the estimated 1.5 billion users who use a PC (NOT all Windows, including Mac, Linux, etc.), it’s still not insignificant… Considering that, since discourse is " infiltrating" many environments, even just 1 or 2% of these users who use any service based on discourse (probably more, but let’s be conservative), and if they have not yet switched to something better, like newer machines or operating systems, they may be unable or unwilling to do so now. This represents a potential user loss (excluding bad publicity and loss of trust) that could range from a minimum of 60,000 to a maximum of 200,000 (or more). I find it hard to believe that the discourse management cares little about this… Am I wrong?
Wouldn’t it be better to simply remove the pointless implementation of the only “feature” (let’s call it that, because I don’t want to be rude) that causes most of the problems, and only implement the other two, provided they are really useful?