Transcription, Transliteration, Latinization

One of the problems with a script as far removed from Latin as Khmer is how to transcribe/transliterate it into something intelligible (and whether it’s even worth the trouble of transcribing it when the native spelling is so unphonetic).
The most popular of these seems to be the IPA-based one used by Huffman. It is, more or less, the same one used on seasite. It also by far is the most accurate in representing the sound of the language.
Colloquial Cambodian uses its own transcription based on English spelling, and is very innacurate, writing very different sounds the same way. “ Cambodian for Beginners” also has its own system, which while a step above that used in Colloquial Cambodian (it is also IPA-based) also doesn’t go far enough in distinguishing between different sounds.
There are other transliteration/transcription for Khmer, like Akea Khmer or the French-based Latinization scheme imposed on Khmer during the French rule of Indochina (and which AFAIK is still used by immigrants living in France) but none seem to have caught on. Just as with other languages, most native speakers do not follow any system and transcribe according to their whims when writing webpages or chatting, which can be a big problem for someone unfamiliar with the language (and can cause difficulties even for natives).
Keeping this in mind I have chosen to use the Huffman transcription, with a few minor changes to make it more web-friendly:

ɛ (open e: Unicode-025B SAMPA-E) = ê
ɔ (open o: Unicode-0254 SAMPA-O) = ô
ɨ (barred i: Unicode-0268 SAMPA-1) = ı
ɑ (script a: Unicode-0251 SAMPA-A) = â

All said, no matter how accurate, a transcription system can only serve as a crutch – it can never fully represent the sounds of a language. And no matter how difficult it may be, the alphabet will at some point have to be learned if one ever hopes to become literate and learn the language beyond the tourist-phrasebook level.