Re: A Sugar Free Diet?

[Michael Richter <ttmrichter@...>]

I think you guys will find that "sugar" in this context means "tastes good but is nutritionally dubious".  Drawing parallels to nutritional theory, etc. is a bit ... whimsical.

Syntax sugar is nice.  It makes syntax taste good.  Too much of it, however, is "empty calories" which should be avoided in large quantities.

On 4 May 2014 21:36, Andrew Starks <andrew.starks@trms.com> wrote:

On Sunday, May 4, 2014, John Hind <john.hind@zen.co.uk> wrote:

I have developed a visceral aversion to the use of the term "sugar" by the Lua community.

 

Firstly the metaphor is upside down (I did not understand what it was driving at when I first encountered it in PiL). Sugar is one of the basic food groups essential to life, but pretty unpalatable in its raw state. If anything sugar is analogous to the byte codes while the essential constructs would be the fruit, vegetables and honey which make the sugar palatable or the language usable for human beings. The extra frills that make the language pleasurable to use would be analogous to fancy restaurant ingredients like truffles and garlic, but to ease the transition, let us call them "pavlova" from now on!

 

Secondly, the way "sugar" is deployed is  the rhetorical tactic of "political correctness" - rather than debate the merits of a feature or proposal in terms of its practical consequences, just label it with a disparaging term that no "right thinking" person could possibly  favour eliding right over the question of whether the label is actually meaningful. When I first looked at the Lua sources I expected to find something like the C pre-processor or the early implementations of C++: a text processor that pre-digested the pavlova and emitted honest peasant food for the parser to further digest into bytecodes. This does not exist: the pavlova and the peasant food are digested simultaneously in the same stomach. So the distinction is actually a tendentious conceptual one imposed by the critics in a way intended to pre-determine the argument they are trying to make.

 

An argument can be made that, say, "function foo() ... end" is *conceptual* pavlova while "foo = function() ... end" is nearer to the structure of the bytecodes and hence more conceptually honest. But at the level of *syntax* both are equal and which is preferred is a matter of taste. If we follow the argument to its logical conclusion, we should all be programming directly in bytecodes!

 

I read your point as saying that the two function statements, to use your example, are equal citizens in the language, because they are not equivalent. 

The equivalent of the non "conceptual" form is:

local foo

foo = function() end

And this is the point. The "function foo() end" form does a "thing" (in the language, not byte code)  that is done in another way already. It's there because the two-line version is common enough that a short cut is welcomed. It's the same as with : and for that matter:

x = 3

_ENV.x = 3

The term "sugar" is just a short form for "syntax in the language that encapsulates a longer form of or multiple statements that do the exact same thing." 

Byte code does not enter into it. 

As long as everyone has the same "conceptual" definition in there head, I don't see a problem with using that conversational "sugar."

-Andrew 

--
"Perhaps people don't believe this, but throughout all of the discussions of entering China our focus has really been what's best for the Chinese people. It's not been about our revenue or profit or whatnot."
--Sergey Brin, demonstrating the emptiness of the "don't be evil" mantra.