Thom Hogan demonstrates for us, in spades, that he's a lazy complainer, who wants to throw around technical-sounding ideas like confetti and get respect as a major commentator. Here's his whiny remarks about an update to EXIF, at which let us look a little bit. To be honest, you can pretty much burn most of his commentary to the ground, but this one is particularly easy, and I am particularly lazy (why yes, I too am a lazy complainer!), so here we are.
This will be a bit technical, sorry. You can skip to the last two paragraphs without missing much.
First he details the updates, which I agree are minor, incremental updates. Ok, so what?
Then there's a complaint about how not enough tags are mandatory. Say what? There's literally no rationale given for this complaint, and I cannot think of one that has any weight whatsoever. I wrote software for a while, and I assure you that I can almost certainly destroy any rationale you come up with for this, but let's keep it short. Will your software accept earlier versions of EXIF data: (of course it will) well then, you've already solved the problem of what to do when there is no Humidity information in the file. Therefore, there is no compelling reason to make the Humidity tag mandatory.
Then there's his usual very very thinly veiled racism, the Japanese control EXIF (no, Thom, a set multinational corporations, each headquartered in Japan, do), as if that was remotely relevant to anything. Honestly, the dude's racism is probably the most tiresome thing about him. Then we get a complaint about how the basic problem is that EXIF is trying to be backward compatible "to a group of CE products mostly defined by these same companies" because, one is forced to deduce, the only things that read EXIF are Japanese built Consumer Electronics (whaaaaat? it's puff puff pass, Thom, not puff puff puff) This is blended with some complaining about 8.3 filenames and the use of ASCII.
These two are actually fair complaints. It is, however, clear that the filenames in EXIF are used, in the general case, not really intended as actual names of actual files in the long term, as the images move around. The fact that path information (what directory/folder the thing lives in) is elided is a clue to this. "File name" in this context should be taken as "a key we can hopefully use to find the thing referred to" (e.g. the Audio file connected to the Image file). The complaint about ASCII is more serious, but honestly I am not seeing any way around this. There's tons of legacy software that is going to choke and die if it runs in to Unicode. As a guy who has actually managed real transitions from one generation of data format to the next with both forward and backward compatibility, I can tell you this ain't no cakewalk. It's doable, but it's a tad complicated and it's not clear that anyone would bother to support it.
In short, while the filename and ASCII complaints are legitimate gripes, it is not at all clear that CIPA could do a much better job, given the very real constraints they operate under here.
Then he wraps up by saying, as nearly as I can tell, that EXIF is fine, you just have to use the optional extended tags for any new kinds of data you might want to attach and then get some industry consensus on what tags mean what. Ok, Thom. But then.. what exactly is the problem?
Hilariously, in all the bitching about supporting new kinds of data, failure to, he comes up with exactly one (1) example. Likes. I cannot, to be honest, make out if he thinks Likes should or should not be in EXIF (they should not: Likes go with a specific copy of a picture is a specific context, they are not intrinsic to The Picture). God knows what all the mysterious metadata "Smart devices of all forms" are generating that "should be in EXIF", but it seems clear that while God does, Thom does not. In reality, nobody really gives a shit about anything Thom is yammering about.
Thom is not a serious commentator. Thom is just a guy who has some sort of problem with the Japanese (or something?) who wants to rail at them, and can't be bothered to think up anything sensible to say.
Sure, Thom knows a tremendous amount about Nikon system technical detail. But it's clear that his grasps of software development, of data architecture, of business management, of new product development, all these things he loves to pontificate about, are very weak.
Plus, let's face it, he's a goddamned racist dipshit.
1. Yes, I know that the J in JEITA (a group that co-authors the relevant standard) stands for Japan. Not relevant. The point is that Thom persistently uses "the Japanese" and "Japanese Culure" as broad brush explanations for all the flaws he perceives in Nikon, CIPA, et al, and that is simply not how grownups talk in this country. Not in public, anyways.