Comments on: arguing over semantics vs. syntax

By: apotheon

apotheon — Sat, 17 Feb 2007 07:13:16 +0000

Funny — I'm pretty sure those books don't contradict my statements that syntax is structure and semantics is meaning. Maybe I should go double-check.

By: tjalfi

tjalfi — Sat, 17 Feb 2007 00:34:39 +0000

The definitions you are using are different from those in the following books: Semantics of Programming Languages, the Formal Semantics of Programming Languages, and Theories of Programming Languages. The only formalism that defines what you're referring to as synax with what most people call syntax is W-grammars. There is a Russian saying that comes to mind, "when three men tell you you're drunk, go home."

By: apotheon

apotheon — Fri, 16 Feb 2007 22:11:06 +0000

Thanks for the input, tjalfi.

What you're failing to include in your analysis is that the man or boy test actually tested for differences in behavior. The syntax used in the test is the same for differing results because the syntax and semantics implemented in the language are different.

By: tjalfi

tjalfi — Fri, 16 Feb 2007 21:36:59 +0000

You are mistaken. The canonical example is Knuth's man or boy test. The results of this code are different depending on whether call-by-value or call-by-name are used. The syntax is identical.

By: Chipping the web - Free as a bird -- Chip’s Quips

Chipping the web - Free as a bird -- Chip’s Quips — Thu, 15 Feb 2007 21:13:29 +0000

[...] When it comes to the expressive power of a programming language, syntax (the form of statements) makes a huge difference. Semantics (the meaning of terms) are relatively inconsequential, so long as they are precisely defined. [...]

By: sosiouxme

sosiouxme — Thu, 15 Feb 2007 13:39:11 +0000

though perhaps now i'm confusing which is syntax and which is semantics :-) or just relating a different dichotomy.

By: sosiouxme

sosiouxme — Thu, 15 Feb 2007 13:35:49 +0000

very insightful. i think there is still this prominent misconception that form and function can be separated, and that any given function can be swapped for an "equivalent" one without impacting form.

By: apotheon

apotheon — Wed, 14 Feb 2007 23:30:58 +0000

Yeah . . .

As someone pointed out to me in IMs (I think it was Allison, but I don't really recall off the top of my head), the issue with pjdelport's (mis)understanding may arise from the simple fact that we (programmers) often trivialize syntax by how we make reference to it when discussing programming languages. For instance

Err, wait, found the direct quote. It was indeed Allison:

(15:28:52) Allison: I suspect the other language issue that one hits, when discussing syntax v. semantics, is that people often use syntax in the "I don't know Java, but I know C++, and after that it's just a matter of learning the syntax." That is, there's a third level: that of programming *paradigm*, and it trumps any of the lower-level language details.

(15:29:46) @: True . . . but that isn't really a "design" issue — it's a bit more fundamental than that, and is largely manifest in syntax.

(15:31:13) Allison: Not design, but architecture, yes. It's just that the way we talk about these things tends to trivialize syntax, except where language holy wars are taking place.

(15:31:26) @: True, that.

(15:31:40) @: I thought the Derrida bit was a more amusing possible explanation, though.

By: Sterling Camden

Sterling Camden — Wed, 14 Feb 2007 23:10:26 +0000

pjdelport gets stuck on the "form vs meaning" distinction. s/he doesn't realize that the form is what creates the expressive power of a programming language, and the meaning you express with it is secondary.