That's pretty clear-cut if one just thinks of the fictional Star Trek universe in terms of what we actually see. Intent counts for nothing if it's offscreen; on the other hand, onscreen often starts out as accidental and becomes intentional only in retrospect, when it's quoted in later onscreen material.
There is no Encyclopedia in Star Trek, even if there is Star Trek in the Encyclopedia.
While I agree with you in a strictly canon sense, I personally don't like to discount the hard work Mike Okuda and Doug Drexler did on the Encyclopedia, Chronology, et. al, and just sweep it under the rug. That's why I have two separate ship lists, one for both canon and "official" ships, and one that's just strictly canon, based ONLY on what we see on screen (including these obscure wall displays, as they
were on screen, albeit in an almost impossible to read fashion).
Also, while it's true that the Encyclopedia does contain several mistakes, I'm sure Mike would be the first person to admit that, and try to fix them accordingly. Unfortunately any future editions of the book seem very unlikely.
We know of no other Hokule'a class vessels; the Tripoli, with an Okudagram-provided low registry from "Unification", has no canonical class identity AFAIK. As far as we can tell, there's no pressing reason for the Hokule'a to be a class ship, or a Hokule'a class to exist - but if the class does exist and has this class vessel, there are no contradictions in that case, either.
Again, I'm going under the assumption that the Tripoli is in fact a Hokule'a class ship. Of course, when this MoaM chart was made, there was no concept yet of conjectural class ships, so Okuda probably just took some names from this chart that he liked, and later used them as class names in the Encyclopedia, while forgetting the actual registries of said ships on the original chart.
No, I mean that if we are to say that the Nebula class Lexington was built as a replacement to the low-registried Lexington from the list, in the mid-2360s, then in light of the Endeavour precedent the new Lexington really should have a 71000 range registry.
Not necessarily. There is proof that at least some vessels with registries of 6XXXX and even 5XXXX were being produced along with ships of 7XXXX registries around the 2360's and '70's.
If so, it was never seen. All the shots that might even in theory have showed the registry were stock footage from "Yesterday's Enterprise", and all the new shots were from the rear. So, the Okudagram registry can freely take precedence in that respect. The only question remaining is, which Okudagram? Tachyon grid or "Measure of a Man"?
Presumably the tachyon grid, as the registry on that diagram matches the registry on the model in the same episode.
...if the explicit dialogue reference to NCC-1305-E is the real deal, then any ships preceding this vessel should have registries in the NCC-1305-Letter format. And probably also any ships succeeding this vessel.
Which is probably one of the reasons why it was changed.
If we dismiss that particular one, then we can pretend that there were multiple different ships. But we still meet a Galaxy class vessel before we get to read a chart with a 20000-range registry, after which we meet a Galaxy class vessel again. It's a bit unlikely that the modern ship Riker claimed was NCC-1305-E was really NCC-24383...
The only explanation that rationally works is that another, older vessel called the Yamato was in service at the same time Varley's ship was. Perhaps the Galaxy Yamato was so newly commissioned that the older ship hadn't been decommissioned yet, and was on its last mission.
I don't see a problem with a NCC-31911 being commissioned in the 2360s.
If registries are completely random, then no, there's not a problem. The problem comes with the design. I just don't think that Starfleet would still be building brand-new Miranda class ships in the 2360's.