Some reasons for trusting programs which don't come with source:

1) If enough /other/ people are using an unknown executable, then I will be
more likely to trust it.

2) If the program comes from a sufficiently reputable source, then I will be
more likely to trust it.

3) Combining the first two: if the program is likely to have come under
attention from the sort of people who routinely read IA32 machine code (OK,
perhaps they disassemble it first ;-) then I will be more likely to trust it.

4) If the program does something desirable and unique, then I'll be more likely
to trust it.

5) If I simply have no choice (e.g. elements of the OS if I'm on Windows), then
I'll be more likely to trust it.

JAD (which I do use, btw) has at least (1) and (4).  (Decompiling is a good
deal harder than disassembling.)   Your program, at least as far as I know,
does not have any of (1)...(5).  If it becomes popular then it will get (1),
but I doubt if that will happen unless there is something special (4) about it.
Perhaps it is blazingly fast (although it's difficult to imagine speed making
much difference for any real application), or perhaps it has an elegant and
flexible macro system.  Or perhaps it has some other unique advantage.
Whatever, if you want people to use it, then you'll have to tell them what
advantages it has over other assembler/disassembler packages.

(BTW, whatever the advantages you can cite, it's unlikely that I'll use it
myself -- nothing personal, and not a criticism of your program, it's just that
I prefer generate bytecodes directly (from code), rather than via an
assembler.)

   -- chris