3 ms·
Over time I've concluded that a single standard for anything important simply isn't enough. It's best to accept that 2 standards and optionally a 3rd and 4th s
by AdamN 3y ago
Over time I've concluded that a single standard for anything important simply isn't enough. It's best to accept that 2 standards and optionally a 3rd and 4th slot for upcoming standards are the best that can be done. Fixating on getting to one is actually counter-productive.
This isn't just true of standards, it's also true of implementations of a standard.