On 24/07/2010 01:21, Jonathan M Davis wrote:
> On Friday 23 July 2010 11:46:47 Nick Sabalausky wrote:
>> It's always bugged me when people use the term "invent" in relaton to a
>> programming language. It's like saying that a musician "invented" a song,
>> or that Mark Twain "invented" a book. Wrong word.
>
> Actully, I believe that invent _is_ the right word here. You write a book or a
> song. With a book or a song, you're actually physically writing something (well,
> in the past anyway - now it may be typing or involve a mouse, but people used
> pen and paper before). With a computer program, you are again writing it (again
> likely typing it, but for pretty much the same reasons, the word write applies).
> However, a programming _language_ is a tool, not something that you write with
> pen and paper. Tools aren't written. They're invented. So, a programming
> language is invented, not written. The compiler itself - being a program - is
> written, but the language itself is invented.
>
> - Jonathan M Davis

As both a programmer and an artist, I completely disagree.

When you craft/create a song/painting/statue, you use techniques that you learned over time and the experience that comes with using them again and again. These techniques have been discovered/invented by someone at some point, however the the piece of art hasn't.

The same holds true for a programming language.

To go with your analogy: the compiler is the tool that allows you to apply the art, not the other way around. The language then is the piece of art and the skill of creating a language is the art itself.

/Max

bearophile wrote:
>> Many companies use it, but they do not advertise their use, Bright said.<
> Really? ^_^ I didn't know this.

Yes, I receive a lot of information via the "back channel" that I can't discuss more than that. It's not that I was asked not to talk about it, I just think it's best to let companies do any such disclosures themselves if and when they choose to.

"Jonathan M Davis" <jmdavisprog@gmail.com> wrote in message news:mailman.0.1279934183.13841.digitalmars-d-announce@puremagic.com...
> On Friday 23 July 2010 11:46:47 Nick Sabalausky wrote:
>> It's always bugged me when people use the term "invent" in relaton to a programming language. It's like saying that a musician "invented" a song, or that Mark Twain "invented" a book. Wrong word.
>
> Actully, I believe that invent _is_ the right word here. You write a book
> or a
> song. With a book or a song, you're actually physically writing something
> (well,
> in the past anyway - now it may be typing or involve a mouse, but people
> used
> pen and paper before). With a computer program, you are again writing it
> (again
> likely typing it, but for pretty much the same reasons, the word write
> applies).
> However, a programming _language_ is a tool, not something that you write
> with
> pen and paper. Tools aren't written. They're invented. So, a programming
> language is invented, not written. The compiler itself - being a program -
> is
> written, but the language itself is invented.
>

I'll grant that "'write' a language" isn't quite correct (unless you're talking about writing the spec, or writing the grammar, etc.), although I was going more for "create" (or "develop", or "design", etc.). "Invent" tends to imply the creation of a new technology or concept. But a new programming language is more an arrangement and implementation of such "inventions" (variables, static typing, LL/LR parsing, etc., although, of course, one could argue that even these are more "discoveries" than "inventions", but that gets into the "Is math/logic a creation or a discovery?" debate, on which I hold no particular opinion.) True, most inventions do include other inventions as their components, but I'd argue that merely using existing inventions doesn't necessarily constitute an invention.

As far as "Tools aren't written. They're invented.", I agree, but only to a point: The hammer is an invention. But if you go and create your own line and brand of hammers, you didn't "invent" Brand X Hammers. "Programming language" itself was an invention. But it's already been invented, and now we're just making specific instances of the technology that is "programming language". A new programming language may intruduce a new invention (like LISP's invention of macros), but even then it's that particular component that's the invention, not the whole language.

I guess one way to sum it up would be:

Common Noun: Invention
Proper Noun: Not an Invention

At least that's the way I see it.

"Walter Bright" <newshound2@digitalmars.com> wrote in message news:i2coqa$jss$1@digitalmars.com...
>
> Since the reporter was recording my conversation with him, I was pretty worried it would come out all wrong. I prefer to do interviews via email based on a couple bad experiences I had. It turned out better than I expected.

I find email to be far more reliable (and practical) than voice for most forms of "business" communication. There's a number of problems I find that voice has for such types of conversations, one of the more notable being the need to blurt out an answer (nearly) immediately without always being able to give proper thought.

"bearophile" <bearophileHUGS@lycos.com> wrote in message news:i2ddai$1rrq$1@digitalmars.com...
> Walter Bright:
>> Inforworld writeup on it mentions D on page 2: http://www.infoworld.com/d/developer-world/mirah-brings-ruby-niceties-java-430
>
> From the text:
>
>>The idea with D is to push complexity into the compiler so user code is simple and elegant, he said.<
>
> It's the first time I read such comment about D :-) I am not sure it's a right comment. Lately I have seen the opposite, moving complex numbers and associative arrays out of the compiler...
>

I think the main point there is that it's "behind-the-scenes" from the viewpoint of the user of the language instead of actually having to be right there in the user's own code. Whether some of that hidden complexity is in the runtime, in the language proper or in the std lib (albiet to a slightly lesser extent) is an implementation detail from the perspective of anyone who isn't actually working directly on one of those three (or on an alternate implementation of one).

Nick Sabalausky wrote:
> I find email to be far more reliable (and practical) than voice for most forms of "business" communication. There's a number of problems I find that voice has for such types of conversations, one of the more notable being the need to blurt out an answer (nearly) immediately without always being able to give proper thought. 

Here's what a journalist printed after a phone interview with me many years ago after I said something stupid, he printed what I said verbatim followed by:

[Walter] ... laughs ...
[Walter] don't print that
[journalist] Ok

Hello Walter,

> Yes, I receive a lot of information via the "back channel" that I
> can't discuss more than that.
> 

Can you give a ball park on how many?

-- 
... <IXOYE><

BCS wrote:
> Hello Walter,
> 
>> Yes, I receive a lot of information via the "back channel" that I
>> can't discuss more than that.
>>
> 
> Can you give a ball park on how many?
> 

Sorry, I never kept track.