5 ms·

Good question. If they don't release the source, then this, as awesome as it is, is just a product and not a language.

by deletes 13y ago

colanderman 13y ago
Languages don't have source code, they have documentation. I haven't looked through it all, but this appears to be documentation specifying (to some degree) the Wolfram Language. Therefore the language exists, regardless of whether any implementations of interpreters or compilers thereof exist and are open-source.
- muyuu 13y ago
  A formal spec would be good enough too. But since he's specifying concrete output and graphics as part of the language, it's rather hard to specify binary-compatibility without a reference implementation. This is why there are no interpreted languages without a reference implementation that haven't failed miserably.
  - colanderman 13y ago
    SQL, Prolog, Scheme, and VHDL seem to be doing just fine. And if you leave off the "interpreted" qualifier, which is (a) a property of an implementation, not a language, and (b) is rather meaningless anyway (can you actually define "interpreted"?), well, then C ought to count too.
    - muyuu 13y ago
      None of these are established as a single compatible language but rather as a family of languages. Wolfram obviously isn't aiming for that. Go to Wolfram's page. That is no spec, it's a rough description of a particular implementation of a (closed source) system.
      - abjorn 13y ago
        It's not as if they started as a "family" of languages, each one of those was just one discrete language at one point.
        muyuu 13y ago
        I'm a big fan of Prolog and Scheme, but seriously the lack of an established solid spec has seriously hurted them. There's no way Wolfram can succeed with its much more complex feature set without a spec and a reference implementation that's not closed source. Flash pulled that because they caught that opportunity window in the browsed, but neither CDF as a format nor Wolfram (Mathematica language) stand a chance as a generic language. Even Matlab stays niche.
        colanderman 13y ago
        You think ISO Prolog and R6RS aren't "established" and "solid"?
        muyuu 13y ago
        Definitely. I use both and nobody uses ISO Prolog and much less pure R6RS scheme. Here's hoping R7RS will be better although for starters it will be divided in two (small and big). I've been following the process quite closely. R6RS completely failed to establish itself, to this day R5RS is more popular. You'd be hard pressed to find a more divided language than Scheme. As for ISO Prolog, it's not even the most popular standard. It might be in America, in Europe it's definitely Edinburgh Prolog and there's where it's used the most. You guys are giving me great examples of how interpreted languages without a clear spec + reference implementation completely fail to establish a single standard. It can be argued that in the case of Scheme this is a feature, but for something like Wolfram with an existing closed runtime you definitely don't want this.
      - colanderman 13y ago
        Uh, all ISO Prolog programs run on all ISO Prolog implementations, regardless of what extensions a particular implementation has. Same for RxRS Scheme. And VHDL and C99 definitely aren't "families" of languages. GCC, Clang, ICC, and maybe even MSVC all interpret the same code semantically equivalently. Unless you can either show C99 to be unsuccessful, find a reference implementation of C99, or define "interpreted" in a meaningful way, I think your assertion doesn't hold.
    - RBerenguel 13y ago
      APL, too
    - khyryk 13y ago
      I'm trying to figure out how VHDL fits in that list. It's essentially a spec in itself, only for hardware rather than software, is it not?
      - colanderman 13y ago
        It's a language, it's formally specified (by the IEEE), and it has no preeminent reference implementation. Its domain is irrelevant to the assertion I countered.
        muyuu 13y ago
        Whereas Wolfram isn't. I'm trying to understand the analogy or how does this contradict my assertion, if that was the point at all.
  - bunderbunder 13y ago
    VBScript and AppleScript come to mind. Oh, how I wish they had, some days. But they didn't.
    - muyuu 13y ago
      Both qualify as failures to me, but they did have at least a spec.
  - jordigh 13y ago
    > This is why there are no interpreted languages without a reference implementation that haven't failed miserably. Matlab? I'm not exactly sure what you mean by "reference implementation", though.
    - muyuu 13y ago
      Yeah Matlab is Matlab. You are locked in to them or try to use an attempt of compatible runtime. Far from desirable and this is probably Wolfram's wet dream. I doubt it will happen though. This can be considered a commercial success, although it's pretty much a failure for the user, the student, academia, ... everybody else.
akater 13y ago
The language exists because it has distinctive features, like its evaluation model and symbolic nature. The rest is a matter of implementation. http://mathics.net http://mathics.net is an attempt to create a free one.

Good question. If they don't release the source, then this, as awesome as it is, is just a product and not a language.

by deletes 13y ago

Good question. If they don't release the source, then this, as awesome as it is, is just a product and not a language.

Languages don't have source code, they have documentation. I haven't looked through it all, but this appears to be documentation specifying (to some degree) the Wolfram Language. Therefore the language exists, regardless of whether any implementations of interpreters or compilers thereof exist and are open-source.

muyuu 13y ago
A formal spec would be good enough too. But since he's specifying concrete output and graphics as part of the language, it's rather hard to specify binary-compatibility without a reference implementation. This is why there are no interpreted languages without a reference implementation that haven't failed miserably.
- colanderman 13y ago
  SQL, Prolog, Scheme, and VHDL seem to be doing just fine. And if you leave off the "interpreted" qualifier, which is (a) a property of an implementation, not a language, and (b) is rather meaningless anyway (can you actually define "interpreted"?), well, then C ought to count too.
  - muyuu 13y ago
    None of these are established as a single compatible language but rather as a family of languages. Wolfram obviously isn't aiming for that. Go to Wolfram's page. That is no spec, it's a rough description of a particular implementation of a (closed source) system.
    - abjorn 13y ago
      It's not as if they started as a "family" of languages, each one of those was just one discrete language at one point.
      - muyuu 13y ago
        I'm a big fan of Prolog and Scheme, but seriously the lack of an established solid spec has seriously hurted them. There's no way Wolfram can succeed with its much more complex feature set without a spec and a reference implementation that's not closed source. Flash pulled that because they caught that opportunity window in the browsed, but neither CDF as a format nor Wolfram (Mathematica language) stand a chance as a generic language. Even Matlab stays niche.
        colanderman 13y ago
        You think ISO Prolog and R6RS aren't "established" and "solid"?
        muyuu 13y ago
        Definitely. I use both and nobody uses ISO Prolog and much less pure R6RS scheme. Here's hoping R7RS will be better although for starters it will be divided in two (small and big). I've been following the process quite closely. R6RS completely failed to establish itself, to this day R5RS is more popular. You'd be hard pressed to find a more divided language than Scheme. As for ISO Prolog, it's not even the most popular standard. It might be in America, in Europe it's definitely Edinburgh Prolog and there's where it's used the most. You guys are giving me great examples of how interpreted languages without a clear spec + reference implementation completely fail to establish a single standard. It can be argued that in the case of Scheme this is a feature, but for something like Wolfram with an existing closed runtime you definitely don't want this.
    - colanderman 13y ago
      Uh, all ISO Prolog programs run on all ISO Prolog implementations, regardless of what extensions a particular implementation has. Same for RxRS Scheme. And VHDL and C99 definitely aren't "families" of languages. GCC, Clang, ICC, and maybe even MSVC all interpret the same code semantically equivalently. Unless you can either show C99 to be unsuccessful, find a reference implementation of C99, or define "interpreted" in a meaningful way, I think your assertion doesn't hold.
  - RBerenguel 13y ago
    APL, too
  - khyryk 13y ago
    I'm trying to figure out how VHDL fits in that list. It's essentially a spec in itself, only for hardware rather than software, is it not?
    - colanderman 13y ago
      It's a language, it's formally specified (by the IEEE), and it has no preeminent reference implementation. Its domain is irrelevant to the assertion I countered.
      - muyuu 13y ago
        Whereas Wolfram isn't. I'm trying to understand the analogy or how does this contradict my assertion, if that was the point at all.
- bunderbunder 13y ago
  VBScript and AppleScript come to mind. Oh, how I wish they had, some days. But they didn't.
  - muyuu 13y ago
    Both qualify as failures to me, but they did have at least a spec.
- jordigh 13y ago
  > This is why there are no interpreted languages without a reference implementation that haven't failed miserably. Matlab? I'm not exactly sure what you mean by "reference implementation", though.
  - muyuu 13y ago
    Yeah Matlab is Matlab. You are locked in to them or try to use an attempt of compatible runtime. Far from desirable and this is probably Wolfram's wet dream. I doubt it will happen though. This can be considered a commercial success, although it's pretty much a failure for the user, the student, academia, ... everybody else.

akater 13y ago

The language exists because it has distinctive features, like its evaluation model and symbolic nature. The rest is a matter of implementation. http://mathics.net http://mathics.net is an attempt to create a free one.