You could argue that there was nothing to gain from the various hurricane reanalyses they've done over the years (some of which for the data is very sparse). The actual changes to the hurricane climatology have been more or less slight AFAIK. Yet they've done them over a long period and will probably keep doing them in the future.Things don't happen without there being some motivation for them to be done. From the NWS perspective there's nothing to be gained by going back over the things they've already gone back over, which leaves it to those more interested in the subject to do that. The NWS could consider the deeper research done by others; after all one person would be enough to verify what has been said and if need be alter the database to a more accurate status.
Apparently they see little point in accurate records when they fail to do even this. That naturally makes me wonder about their accuracy of what happens today.
Of course there are some differences. Tornadoes were a lot more underreported before the WSR88D network was complete, which massively increases uncertainty. Tropical cyclones don't suddenly dissipate with a new one forming shortly afterwards on the same track. They're much bigger and don't do weird things like split into multiple vortices which might fool people into thinking a new one's formed. The physical relationships are simpler and less dependent on subjective data.
But there are people right now writing and publishing papers making claims based on the offical tornado data, which is not that accurate especially if you want to do any kind of analysis that uses paths rather than points. Our pre-April 27 knowledge of what might be possible in Dixie Alley was significantly influenced by our knowledge of the outbreaks in 1884, 1908, 1920, 1932 and 1936. Whereas the early part of the official period was evidently rather quiet down there.