Actually the discussion about periods has led to more than how to just represent units of time; but also concepts like miles per hour, meters per second-squared, etc.
I have been told about the existence of two experimental cmavo, {pi'ai} (multiplication prefix, in selma'o KE) and {te'ai} (exponentiation subscript, in selma'o XI) that can be used for specialized tanru. So 30 miles per hour = lo pi'ai minli cacra te'ai ni'upa be li cino , and 9.8 meters per second-squared = lo pi'ai mitre snidu te'ai ni'ure be li sopibi . Unfortunately these constructions can be a little long.
One crazy idea I thought of was... what if we had a PA cmavo for every unit of measurement? After all, they are treated much like algebraic constants in calculations. So assuming [mai'i] = minli and [cai'ai] = cacra, you can have have 30 miles per hour = li cino pi'i mai'i fe'i cai'ai , or assuming [nai'a] for year and [jei'i] for day, three years and four days = li ci pi'i nai'a su'i vo pi'i jei'i (fau'u rodo I forgot if Lojban mathematical precedence works this way, or if I need mathematical brackets which I forgot how to use). Do we nearly have the cmavo space for this though, and would the semantic hooks be enough? Also they might still be a tad long - do we really need all these pi'i-s?
You see, the main barrier to Lojbanizing mathematical and scientific conventions is that mathematical variables are so often overloaded, as there are only so many letters to go around. Even the
familiar symbol of π is given many other meanings besides the ratio of the circumference to the diameter.
But Lojban, on the other hand, has a theoretically-infinite number of words to assign to these variables to. Unfortunately though, syllabicity is still a practical issue. The constraints of cmavo space is another; consonants provide more bits than vowels and diphthongs do, but a cmavo is only allowed one consonant at the beginning.