Well, of course I like the abacus number system (biquinary) but not the mix with thee additive digits. I would rather have two more digits. (‘si' and ‘po’, say). The binary code for digits is a simplification in some ways but makes for even longer numbers. And dropping the initial ‘ala’ sets the whole scheme off right at the start. So I prefer my version of the same general idea. I also like order-of-magnitude words to deal with long complex cases. All that aside, this is one of the better suggestions.
The second system is in some ways even better, except that it can’t be done. The second syllable of a word cannot begin with the null consonant. (basic rule). To be sure, the null consonant is occasionally a glottal stop between words, but not consistently and never within a word. On the other hand, except when reading out strings of digits (phone numbers, say) zero is almost never required, because the place is specified for each digit, without needing to rn through all the zeros. But the rest of the words here are generally impossible by the other rules of phonology: there cannot be any of ‘ji’, ‘wo’, ‘wu’ or ’ti’ and many of the remaining syllable are already assigned standard tp words, which we should try to avoid mucking with if possible. At best, the system could word with ‘a’ and ‘e’, but that is not much help.
Why are the consonants not in alphabetic order (or some other reasonably obvious order)? Adding memory issues to. the rest of the system seems just self-defeating.
This system takes up to 100,000 very tidily (previous problems aside) but does not suggest any way to extend from there on. (The obvious way to do it is just to start over again, with an ‘a’ digit come just before a ‘u’ one telling us that ths addition has taken place. The similar treatment of decimals is brilliant.)
What is the ‘e’ before the digits for? Just to warn that these are numbers?
I really like the compactness and clarity of this system and wish I could think of a dodge around the phonological issue, but in the dozen years or so since something like this first came up, no one has managed a workable suggestion (even noting that the forbidden syllables only need be absent in certain positions doesn’t help, since as digits, they have to occur in those positions, too.)
Here is a germ, however. Treat each digit as a. CVCV pattern with the same C (the digit) and the Vs being ‘a’ and ‘e’, in a fixed code to cover the cases through E4. This unfortunately doubles the length of a number. But we actually only need to do this for four numbers, ‘ji, wo, wu, ti’. The trick now is to get a CVCV pattern, especially a VV pattern, that does not occur. For a single iteration of the system, ‘ae’ satisfies the condition. But, of course, no pattern will work for an iterated system (above E5) unless we mark the beginning of iterations and that turns out to be hard, but probably not insurmountable.