ja, tokenizing works like this:

C:\> kix32 /t myscript.kix

this produces a file called "myscript.kx" (note the new extention .KX) ... this file contains byte code, looks like this:


՝֕ ƲŤֳ˨ڵƩϻٷӼ˸ ҷ٭8Д_ˀ
́ݎՖ2¡ Ϋڵ۽Գƴաȧ;h;
T2F1P"GV?\.A2];OD-C'H?LS&T&C-Yjkm_b~ct({vvVw`A$W<H'WWqhM

your standard unreadable blob, then you can run this tokenized file back through kix32 like this:

C:\> kix32 myscript.kx

Depending on the script, and what its doing, you can get anywhere from 0 to 50% "compression" in size ... plus its already mostly "pre-parsed", so it should be faster ...

-Shawn