I kept expecting you to make a jab at the wa-puro ro-mazi…

Nah, that's just orthodox Labovian language change transposed to orthography. I don't like it, but barring a serious economic rebound, the anime-'n'-manga otaku will be the standard bearers for romanized Japanese going forth.

Can we do both?

You mean, retain kanji while abandoning them? Sure, I think Joshu may have said something about that.

I sometimes think of starting a Wahabi-Saudi style alliance with whoever will promise to help me clear the earth of those who use the romanization “jya.”

Of course, it would start that way, with the best of intentions, but then it would end with the sword put to anyone who had ever written “romanji” and I’d be left to plead “I was young!” before the revolution turns on me like Robespierre.

Maybe still worth it.

‹Jya› doesn't make sense – ‹j› already implies palatalization, there's no distinction between ‹jya› and ‹ja› – but I for one am deeply in love with ‹zya› and the rest of Nippon-shiki. Just the right mix of phonemicity (as opposed to the barbarian phoneticness of Hepburn) and diacronicity to delight my tastes! Nippon > Kunrei ≫ Hepburn.

I don't mind "rōmanji" because it would be "rōmānae littera" and the nasalization did spread to the preceding vowel which is precisely what's denoted by the Japanese ‹n›.

Damn! “DiaCHronicity". "Extending through time", not "extending through (the realm of?) father-castrating child-devouring guy".

But Nippon-Shiki doesn't let you write the romanization of things like "DVD" and "film" as pronounced by the younger generation. Are you just going to throw in English spellings like some kind of barbarian?

Hm… tricky. I don't mind writing /fi/ as ‹fi›, if the language got a new phoneme, it gets a new grapheme (though /hu/ = [fu] confuses mattersl are [fu] and [hu] ever distinguished?). But I don't know what to do with /t,d,s/.

Just like in Japanese, in my own native language /da/ is [da] but /di/ is [dʒi]. And we write the latter as ‹di›, naturally, for the same good reasons that English speakers write the [t] of “stop” and the [tʰ] of “top” both as ‹t›. So we have no way, in my native orthography, of distinguishing [di] from [dʒi]. That usually doesn't matter, because we didn't import [di] as a phoneme sequence – when pronouncing English borrowings, like, say, the title of the videogame Diablo, we just say it as [dʒi]. (Brazilians say “tease” like “cheese” :) )

If modern Japanese is incorporating [ti, tu, di, du, si] etc. as sequences phonemically distinct from [tʃi, tsu, dʒi, dzu, ʃi] – then there's no going around it, we'll need graphemes for both sides, and the alphabetic principle dictates that ‹ti, tu, di, du, si› be used for the former group. The problem is, I believe that if you ask a child or illiterate speaker something like "if 'kaka' becomes 'kiki', then 'tata' becomes?", they'll likely answer [tʒitʒi] – /ti/ & friends still feel peripheral… I don't want to lose the intuitive relation between [ta] and [tʃi], nicely represented in Nippon– and Kunrei-siki as ‹ta,ti› – not just for the sake of seldom-used peripheral phonemes; but I can't think of any other good way of representing [ti] except as ‹ti› :(

(voice onset time typo! 'tata' becomes [tʃitʃi])

Represent [ti] as ‹thi›? Like Italian ‹ci› vs. ‹chi›.

That's not bad! We can go with that. Also, Leo, I asked a couple of native speakers and though they were comfortable with differentiating /hi/, /fi/, and /fui/, none felt that /hu/ and /fu/ were distinctive.

So, I guess we have the new and improved standard wrapped up. Let's not forget to create two subtly different variants based on how the moraic nasal is treated before /b/ and /p/, though.

If Wâpuro Rômazi keeps getting traction, perhaps we'll start seeing [ti] = 'texi'? To my horror, there are already a few true-positive Google results for "dexisuko" (mostly URL paths and usernames). I hope it at least evolves to 'txi', 'sxi' etc.

From what I've read, ti, tu, di, du vs. chi, tsu, dji, dzu are standardly differentiated in loanwords, but si and zi vs. shi and ji are not.
So in terms of loanword accommodation, the spellings si = shi and zi = ji would never be problematic.
I suppose this also means you could write chi as tsi and dji as dzi if you want them to share a letter with "ta" and "da".

Back in ye olde ALT days, I remember students occasionally romanizing つ as THU from time to time-- I had the suspicion that it had to be in imitation of some particular example (some athlete?), but either way it shows some awareness that "t" plus something X was perhaps +desirable to "tu".