The Postel's suggestion works because it's profitable for the device makers.
Be strict at what you send, as much as you can. This maximizes the probability of your widget working with everything else. Be lenient in what you accept, and your widget will likely work with everyone else's widgets, even if everything else's ideas of the correct protocol slightly diverge.
Following the two approaches makes your widget more compatible than a strictly-adherent, inflexible widget your competitor might produce.
The more compatible your widget is, the fewer unhappy customers you have: it fails to work in fewer cases that a strictly-adherent widget would. It may not work ideally in every case, but it may work well enough for the customer to keep it, or even for you to deny a refund.
With low-price items like cables, it again plays into the cable maker's hand: all the various cables are okay because they demonstrably work (not "broken"), but now you can differentiate between very cheap cables that can do little but suffice in simple cases, and premium cables that offer high speed, high power, etc. You can put a cheap cable into a toy as a charging-only low-power cable and save the all-important few cents. You can charge $29.99 for a cable that can do it all™, for customers annoyed by cheap underperforming cables (which they also bought from you!), who decide to finally splurge on the real thing.
A world where there'd be one and only maximalist format, and where the smallest deviation from it would render a device / cable useless, would be (even) more painful for the customer, more expensive, and with some more preice-sensitive segments underserved. The variability that follows from the "be lenient to what you receive" corresponds to the need to serve many varied market segments.
Be strict at what you send, as much as you can. This maximizes the probability of your widget working with everything else. Be lenient in what you accept, and your widget will likely work with everyone else's widgets, even if everything else's ideas of the correct protocol slightly diverge.
Following the two approaches makes your widget more compatible than a strictly-adherent, inflexible widget your competitor might produce.
The more compatible your widget is, the fewer unhappy customers you have: it fails to work in fewer cases that a strictly-adherent widget would. It may not work ideally in every case, but it may work well enough for the customer to keep it, or even for you to deny a refund.
With low-price items like cables, it again plays into the cable maker's hand: all the various cables are okay because they demonstrably work (not "broken"), but now you can differentiate between very cheap cables that can do little but suffice in simple cases, and premium cables that offer high speed, high power, etc. You can put a cheap cable into a toy as a charging-only low-power cable and save the all-important few cents. You can charge $29.99 for a cable that can do it all™, for customers annoyed by cheap underperforming cables (which they also bought from you!), who decide to finally splurge on the real thing.
A world where there'd be one and only maximalist format, and where the smallest deviation from it would render a device / cable useless, would be (even) more painful for the customer, more expensive, and with some more preice-sensitive segments underserved. The variability that follows from the "be lenient to what you receive" corresponds to the need to serve many varied market segments.
Sorry!