John's comment re WALS led me to go back through the thread... I had entered the discussion by commenting on Toaq Dzu, which I had not actually considered it to be an IAL as such, and had scanned right past And's reference to my proposal as a 'Speedtalky IAL' without focusing on the IAL bit.
So while my side project is an engelang (albeit a very loosely defined one at this point), it is most definitely NOT intended as an IAL, and while I'm looking for an optimal phoneme inventory, I am not interested in the degree of robustness that I presume would make for a good IAL.
My expectation is that an IAL would optimally have a 5 vowel system and about 20 (±5) consonants - roughly /i, e, a, o, u/ & /p, t, tʃ, k, b, d, dʒ, g, f, s, ʃ, x, v, z, ʒ, ɣ, m, n, ŋ, l, r, j, w/ (note that /tʃ, dʒ/ might not occur as unit phones, and /j, w/ might merge with /i, u/). I realize that some of those contrasts may disfavor some speaker populations, and my own perverse nature would enjoy throwing some lateral fricatives, clicks, and some other uncommon sounds into the mix to even out the playing field.
My goal for my version of Speedtalk (henceforth S64 or S64^2, to give it a label) is a phoneme inventory that balances a relatively large inventory of onsets and rhymes (64 of each) with reasonably robust contrasts, tied more to distinctive feature contrasts than to phonetic contrasts. To give an example of my notion of robustness, lateral fricatives occur in about 9.7% of the languages in WALS, so not common on a global scale (but more common than front rounded vowels at 6.5%), but if the lateral fricatives are dorso-palatal, contrasting with alveolar and post-alveolar/retroflex sibilants on the one hand, and dorsal (velar-uvular-pharyngeal) fricatives on the other hand, I think they'd be sufficiently distinct. Note that I also would not rule out a velar-uvular vs. pharyngeal opposition. (And yes, S64^2 will have some clicks in its inventory.)
(Hopefully my tone hasn't been too snarky - it's been a rough couple of weeks)
bfowol