Re: [ANN] Lua 5.2.0 (alpha-rc2) now available

[HyperHacker <hyperhacker@...>]

On Sat, Nov 20, 2010 at 18:55, KHMan <keinhong@gmail.com> wrote:
> On 11/21/2010 4:59 AM, Luiz Henrique de Figueiredo wrote:
>>>
>>> It is not a good reason to CAPTIALIZE a FUNCTION name.
>>
>> The all-caps names in bit32 are clearly not ideal but I don't see that
>> they warrant such a fuss. Yes, they look like assembly directives and
>> that's on purpose. Of all the alternatives we considered (like the ones
>> suggested here), this one sucked less and had the clear advantage of
>> being different yet consistent with common asm use. If you think the
>> all-caps names are shouting, please do use your own local names (or
>> even rename them directly in bit32). No one wants to write code with
>> complicated expressions involving bit32.AND and bit32.XOR, whatever the
>> actual names are.
>
> Oh good, the high priesthood is talking :-) Good, some feedback is better
> than one-way traffic.
>
> To me it's just a library. No different from the rest. Why does it have to
> look different? We did fine before with lower case.
>
> Leaving aside the aliasing, many people are attuned to having all caps as
> constants.
>
> I think only ancient assembly code users insist on having all caps as
> mnemonics. All caps as mnemonics was never necessary, it was merely due to
> keyboards and terminals having defaults of upper case. Remember COBOL,
> FORTRAN, BASIC? *Everyone* has moved on from all caps. All caps is not
> modern practice. Caps Lock and Shift are being used less and less.
>
> When I write assembly language these days, it's in all lower case except for
> constant equates. This is consistent with common use of constants in C for
> instance, and one becomes attuned to quickly scanning source code patterns
> in that way. For embedded programming people, bit manipulation is routine
> and I don't see why bit32 functions has to look special. High-speed kernels
> like graphics kernels or crypto are often in assembly -- maybe they use
> upper case -- but they are pretty static write-once rarely-touched blocks of
> code and is *very* different from code worked on daily.
>
> I don't think "common asm usage" is a valid rationale at all.
>
> Complicated expressions can be written with bit32.band etc also, if the user
> chooses to do so. When performance is not an issue, then there is no
> problem. No kittens will die. I thought one time the top duo even said using
> globals are very fast and not too bad compared to locals on the list. But
> now you are saying that aliasing using locals is standard practice? Whatever
> happened to simplicity that aliasing using locals is now such a virtue? Also
> when I alias, I try to stick to the same names. Having custom names for
> aliases means different usage in different programs from different places
> and unnecessary comprehension effort.
>
> The arguments put forward to justify the change is not convincing in the
> least.
>
> It's just a library. They are just functions and methods. It's not assembly
> code. Lua looking like assembly code (maybe because it was easy to port
> stuff like that) is not a virtue.
>
> I have said my spiel and I'll slither away for now.
>
> --
> Cheers,
> Kein-Hong Man (esq.)
> Kuala Lumpur, Malaysia
>
>

I'm wondering why they can't just be named "and", "or" etc. Can the
compiler not be made to recognize that the "or" in "bit32.or" is an
identifier rather than a keyword (which wouldn't be syntactically
valid)?

-- 
Sent from my toaster.