A range of possibilities

[janKipo]

tp offers a range of possibilities for other than alphabetic writing systems and other than oral face to face communications. To begin with the obvious. the restricted set of syllables, so nicely laid out in a grid, suggests a syllabary -- starting with Japanese (complete with the -n marker) and then going on to others satisfying various aesthetics. Equally obvious, the fact that there are only 123 words (count on 09/10/17) makes it natural for a picto-/ideo-/logo- graphic system, ala Chinese -- perhaps with a syllabic or alphabetic side system to cover the unofficial proper adjectives. And again, aesthetics sets various criteria for these: how much ideo and how much phono for starters, then what ideos are clearest and so on. This same feature encourages a sign language (or several) since there is an easily learned set of signs (plus finger spelling, of course) and, again, there is the issue of what best represents the words involved (again ideo v phono in part): borrowing from any of the recognized SLs (including Iron Eyes Cody's or some one else's version of Plains Indian) or creatio de novo for this new language. In addition (is this on Sonja's list?), there is the question of secret tp although, with only 123 word (today) things like Pig Latin and Alfalfa aren't going to be very concealing (not that those work very well), but hashbash has some promise, as does something like Emrys and Fink's hand-holding pressure code (a bit spelling centered but probably adaptable).
Now, I am perfectly happy with the Latin alphabet, but tp is simple enough to learn that the additional effort of a new orthography or dactylography doesn't seem out of the question. Go to it!

[janKipo]

Oh, yeah, I forgot to think about script v. print, which, as calligraphers in every language know, opens up another whole range of interesting appearances only loosely related to one another (even if you know the stroke order for Chinese characters, figuring out what is going on is not a trivial pursuit -- and the harder it is the more elegant the hand).

[jan-ante]

Equally obvious, the fact that there are only 123 words (count on 09/10/17) makes it natural for a picto-/ideo-/logo- graphic system, ala Chinese -- perhaps with a syllabic or alphabetic side system to cover the unofficial proper adjectives.

that is japanese kanji and kana. i remember you were critical about this system, but here its advantage is obvious.
there is a proposal of alternative hieroglipics, but in my opinion kanji is a better options.
first of all, learning the 123 kanji will make one familiar with japanese and chinese writing system and indirectly with their culture.
the second point is that for some words there are no kanji. these are la li o e pi. and this is really a good chance, because these symbols should be graphically distinct from the rest of vocabulary. so, one can use the symbols | > !< { respectively. this will make the text structure evident. btw this is a reason why i do not advocate the usage of Bliss system or LoCoS for toki pona. as a planned logographics (or even ideorgaphics as they claim), they have clear advantages over traditional millenia-aged system. but it is really a problem to introduce graphicaly distinct la li o e pi in these systems.
the use of logographic has another important consequence. it could inourage the development of dialects. e.g., using the logographic one can read the toki pona words in monosyllabic manner. this will make all the expressions much shorter and probably clearer. this could also serve as a model of language evolution.

[janKipo]

Yeah, the Japanese system is the first thing everyone thinks of and it fits pretty well. I seem to recall that the kanji for some tp words are complex but there are probably alternate and simplified versions. For the words without kanji, I wonder if that is really true, either in Chinese or especially in Japanese, where there seem to be a lot of words with purely syntactic (as opposed to semantic) functions (their occurrences is a quick way to spot the text as Japanese rather than Chinese, for example), But maybe not,. In any case, the suggestions given look good, except I would use ! for o and maybe just | for li. Bliss and the like are just not very language like and so don't fit into,languages too well, though adding the needed bits is as easy as you just did. Of course, I am quite content with the Latin alphabet. (I won't even go into the horror that the suggestion of monosyllabic versions of tp fill me with -- there aren't syllables enough without changing the phonology and when you do that, the possibilities of minimally distinct -- so interchangeable in noisy, i.e., real, environments -- pairs increases logarithmically and tp has more than enough of that already. Plus, of course, it wouldn't be tp then at all.)

[jan Ote]

these are la li o e pi. and this is really a good chance, because these symbols should be graphically distinct from the rest of vocabulary. so, one can use the symbols | > !< { respectively

Good choice, although >< could be inconvenient in XML languages.

[jan-ante]

(I won't even go into the horror that the suggestion of monosyllabic versions of tp fill me with -- ... Plus, of course, it wouldn't be tp then at all.)

this is a strong argument in favour of logographics. you can speak traditional toki pona, i can speak monosyllabic dialect and we can communicate with a logographics

either in Chinese or especially in Japanese, where there seem to be a lot of words with purely syntactic (as opposed to semantic) functions (their occurrences is a quick way to spot the text as Japanese rather than Chinese, for example), But maybe not,.

for pi the japanese use hiragana symbol "no" (genitive case). for e they use hiragana "wo" (accusative case). for la there is a two kanji semi-equivalent "ba ai". for polite imperative they use a word "kudasai" after the predicate (i.e. at the end of the sentence). "kuda" is a kanji "down", "sai" are 2 hiragana symbols. for impolite imperative they use a special form of verb ending with "e".

the possibilities of minimally distinct -- so interchangeable in noisy, i.e., real, environments -- pairs increases logarithmically and tp has more than enough of that already.

note, this is a statistical rule. it is valid for a big number of randomly choosen onservation. it can be invalide for a limited number of specially designed monosyllabic words

[jan-ante]

>< could be inconvenient in XML languages.

probably. i am not sure, but may be a duplication of > or < could help in XML. One can also consider another option. kanji/hanzi contain no round shapes. so, one can use
la - }
li - ~
o - o or *
e - {
pi - @

[janKipo]

Well, no. There is no monosyllabic dialect of tp. Toki pona is not just the concept list or the grammar, it is the whole thing, complete with the built in jokes and the incongruities. We can work to smooth it out a bit, but it remains what it is. It is meant, after all, to be a spoken language (a language, in short), even though we have to learn it -- for now, any how -- in writing. As in Japanese, even if we write a word with a single symbol, it is pronounced with all its syllables.

The problem with getting 123 (or so) monosyllables are 1) the need to expand the phonology somewhere, either adding new sounds or allowing other syllable forms. 2. In either case, you need to make sure that you do not introduce new possibilities for mishearing into the mix -- and this takes a good deal more care than might at first appear, since you have to consider not only problems within words but also across word boundaries. The present array of sounds is about as distinct as possible and the syllable structure heightens that. The only interference problems come with the words themselves (see previous lists of noisy room horrors), A monosyllabic language derived (somehow) from tp, will have all of those problems multiplied -- although it does contain possibilities for elegant solutions (involving inelegant words, alas). I don't think a monosyllabic "dialect" is a good reason to used kanji etc.

[rdmiller3]

Actually, there is... but it was just something I was playing with. I called it "hó ni" (toki pona). I haven't messed with it in months though.

Syllables were consonant-vowel, using eight consonants and five vowels with three tones. I'd have to add something else to handle the new words though, since 8*5*3=120.

I suppose I could allow for final -m/-n like in real Toki Pona, or for consonant clusters or for a fourth tone.

Anyway, I came up with it just to see what playing with an artificial tonal language might be like.

[janKipo]

My point being that that ain't tp, or even a dialect of it, but another language altogether, which may share some history with tp but otherwise not. Tones are a nice device, though tend to cut back on the popularity of a conlang that uses them, and they need to be carefully chosen , not just for bad situations (where they need to be easily exaggerated) but even for perfect transmission. The fully packed 8*5 syllable space is asking for misunderstanding at the best of times (and is hard to learn or to read). I wouldn't take it up again if I were your (which I am not, of course, so ...). tp is actually pretty good from most points of view (the noisy channel problem not so much -- including proof reading).