When I want to reference an official specification, it is gennerally not a need. I can normally satisfy my own usecase by hacking together enough knowledge from reverse engineering and public sources (often with less effort then actually reading the spec requires), but I view it as my responsabilty to be spec compliant wherever possible.
A spec is only as good as it is followed. By not making the spec free, it is less valuable to me even if I pay for it, because there are now even more noncompliant systems out there I might need to interlop with.
Indeed! A "standard" that costs $100 to read is not actually a standard at all. The standard is, by definition, what the users think it is.
Basically there are annotations that are considered part of the law but they are behind copyright and can't be distributed in the way you expect public laws to be.
https://www.techdirt.com/articles/20181019/12232640876/appea...
It doesn't seem to me to be a terrible trade-off to say that ad-hoc use via public drafts is fine, but if you need to use the actual spec for whatever reason then you're probably being paid to abide by the spec and it makes sense to charge for it.
Where I strenuously object to non-public specifications is where they're referenced in (and required by) legislation: if ignorance of the law is no excuse, the law had better be freely available.