> John Payson wrote a some comments and I wanted to add my $.02
>
> >The following is a brief description of my personal taste; others' may vary.
>
> >Labels:
> I usually put the acronym of the current subroutine first.
>
> SendChar
> SC_Loop
>
> This allows you to use basic labels easily.

I forgot to mention that I do this (in both .asm and .c) if the code size
warrants.  For small programs, it can look a little ridiculous.

> >Hexadecimal numbers
> >  The one proper format for hex numbers is $ABCD.
>
> Boy, the first sentence of this paragraph is just begging to be flamed!

Hey, I do enough productive posting here, I can be allowed a troll now and
again, can't I? >:*3

> John, I think you have to be more tolerant of other people's preferences.  I
> started out with Motorola Assembly and the '$' format, so I know where
> you're coming from.  Actually, some time ago Andy Warren made this statement
> and I couldn't honestly remember *ever* using '$'.  Most of my programming
> since then has been with "C" and Intel Processors.  I think you'll get at
> least as many (if not more) people who think that "0x0" is the "correct"
> format.  I'm sure that there are people out there that are lost if they
> don't see in a number in the format "h'xx'".

Sorry I was a bit overly strong in my statement, but every other format I've
seen has problems:

[1] Default radix == 16
    Often requires spurious leading zero for constants whose leading digit
    is greater than nine.  Obfuscates the distinction between length of a
    number and its size (e.g. in "$" format, a 16-bit number may be denoted
    as $ABCD or $0015; an 8-bit number can be denoted as $12 or $DE.  In
    default-radix 16 format the first and last of these would become 0ABCD
    and 0DE, making them appear 20 or 12 bits long.

    In addition, having a default radix of 16 without punctuation can easily
    cause confusion between decimal and hex numbers.

[2] Trailing "h" [Intel]
    No longer has the decimal-vs-hex ambiguity of default-16, but still has
    the spurious zero at the beginning.  This format adds two characters to
    the length of the hex number (e.g. 165(dec)==0A5h

[3] Unix-style (h'A5', also used by Microchip)
    Unambiguous, but exceptionally verbose; adds three characters to each
    hex number.

[4] C-style (0xA5)
    Second-best choice.  Other than default-16 and leading-$ formats, all
    other radix notations add at least two characters.  C adds two characters
    without adding ambiguity, so it's the best of the non-Mot formats.

[5] Leading-$
    Only adds one character to a hex number; a comma-delineated list of 16
    8-bit numbers can fit on an 80-character line in, e.g., a .db directive.
    If a number is written out to its full size (e.g. a 16-bit number using
    four digits) then visual observaion will immediately indicate the size
    of the number (byte vs word, etc.)

Perhaps you can identify some way in which leading-$ format is sub-optimal?