meta Meta language released

[Kaj de Vos]

It's caused by

flags: at-head flags

That doesn't do anything here, anyway, but it's also incorrect because you can't reassign a constant.

 

It should be caught by the Meta compiler instead of the C compiler.

 

[ilmenit]

@Kaj de Vos, ping here when you have a fix ready so I can continue to play with the language.

 

Do not take it as a rant or criticism but observation, please. From what I see so far, the language has been officially released too soon.

On the page you write that the language is already useful ("Is it useful for anything yet? Yes! This website and the downloads are served to you by the language itself. ".

It also states that "Meta currently has no bugs, it only has unimplemented features, so it is not useful yet to report them" but clearly it's has not only unimplemented features but also a bugs in the ones implemented.

The language has zero documentation available and pointing to REBOL documentation is misleading, because almost everything is different in Meta (even basic keywords/functions like loop/repeat/forall/print/write) and trying to write even a simple console program is a guessing game (or blackbox testing).

A while ago there was a Zig language released and the documentation, while incomplete, was enough to use the language. Other example of language is Crystal - take a look what was in the first official release and how many examples were there.

For Meta, considering conceptual similarities to Rebol at least list of implemented control functions/keywords and their parameters should be available, something like this would be enough. Also a few more complex examples that use e.g. arrays/series would be useful, because the current examples are too basic and give impression that there is nothing implemented in the language yet (while there is already a lot). You have added a lot of information on the page (about the language), but this is not the information for the programmers that want to try the released language.

[Kaj de Vos]

Just remove the line that doesn't work. Your code runs for me.

 

We have gone over this at length. All you say is true, but Meta was released three weeks ago. Languages take decades to mature. If you focus on what's there instead of what's not there, you will see that I am ticking off all your wishes as fast as I can.

 

I didn't release Meta to the world at large, but to the smallest group I could: this sub-forum. The feedback here is changing the order in which I proceed before enlarging this circle.

[TGB1718]

3 minutes ago, Kaj de Vos said:

If you focus on what's there instead of what's not there

But that's exactly the problem, you released the language, but without the most basic instructions

so it's almost impossible to get started

[ilmenit]

1 minute ago, TGB1718 said:

But that's exactly the problem, you released the language, but without the most basic instructions

so it's almost impossible to get started

I'd rephrase it: we simply have no idea what is there and what is not.

[Kaj de Vos]

Yet, Zbyti wrote programs in it, and Ilmenit a Sieve of Eratosthenes, which I advised him was too advanced to try yet.

[Kaj de Vos]

Just now, ilmenit said:

I'd rephrase it: we simply have no idea what is there and what is not.

Yes, you have, I keep stipulating to focus on the published examples, because that is what is there and has been tested.

[Kaj de Vos]

https://www.joelonsoftware.com/2001/07/21/good-software-takes-ten-years-get-used-to-it/

[ilmenit]

1 hour ago, Kaj de Vos said:

We have gone over this at length. All you say is true, but Meta was released three weeks ago. Languages take decades to mature. 

 

7 minutes ago, Kaj de Vos said:

https://www.joelonsoftware.com/2001/07/21/good-software-takes-ten-years-get-used-to-it/

All correct, majority of the languages needed at least 6-7 years from the first release until they gained traction. However these languages on the first release had the basic stuff working, examples and at least some documentation that allowed to use the language for something else than running the compiler and receiving compilation errors. The target audience of the language are developers and maybe we just do not understand why you did release at this state of the language and what kind of feedback you expected.

The main point we are trying to express is that you released the language too early. Do not take it as an attack on you or your work, please - I'm really keeping my fingers crossed for success of Meta. You would avoid all the comments we are posting here just by postponing the release or highlighting better that the language is unusable now. Claim that "This website and the downloads are served to you by the language itself." gives impression that the language is almost "production ready". You even posted the language in Windows Tools for Atari Programming thread giving impression that it can be used for software development on Atari - nowadays it can't. Pointing to Rebol documentation and writing "subtle differences" is also misleading, because almost no function from Rebol can be used as in the documentation.

[Kaj de Vos]

Nowhere do I claim "production ready". I say "Is it useful for anything yet?". Meta runs its own project, it's very useful.

 

I released it to let interested early birds try it, after more than a year of lively discussion. If you would rather have nothing than something, just pretend it isn't released yet. However, you're just one semicolon away from getting your Sieve to run.

 

The feedback is very valuable and essential. I'm constantly updating the website with information from testers.

[ilmenit]

2 minutes ago, Kaj de Vos said:

I released it to let interested early birds try it, after more than a year of lively discussion. If you would rather have nothing than something, just pretend it isn't released yet. However, you're just one semicolon away from getting your Sieve to run.

And I'll try to finish it, and later rewrite it more from Rebol-ish way of operating on series ? I didn't have opportunity in the past to play with Rebol and it taught me interesting piece of computer history. Existing inline assembly should also allow to create a simple benchmark over Sieve code, to track the performance progress in the language development. Such benchmarks were useful e.g. for @tebe for MadPascal improvements and serve also as a series of regression tests (for both language and performance changes).

[zbyti]

On 11/2/2021 at 5:26 PM, Kaj de Vos said:

Surely not everyone will be able to use it right away, that's why I put in all the caveats, and drew up a road map. It's not an option to wait until it's usable for everyone, because that day would never come. It would take decades, and during that time I would have no feedback, so it would go in wrong directions. For one thing, it's quite a surprise to me that, apparently, most people here need much more hand-holding than I expected.

Don't be surprise, most peoples needs (including me) proper documentation, especially if they don't have the source code to see what and how it works or don't.

[Kaj de Vos]

Indeed. And for Sieve, this has already been done. I optimised it to its highest speed. The only thing that might still improve it a bit is the native 6502 back-end.

[ilmenit]

1 minute ago, Kaj de Vos said:

Indeed. And for Sieve, this has already been done. I optimised it to its highest speed. The only thing that might still improve it a bit is the native 6502 back-end.

Do you think also about generating the output C code? 

[Kaj de Vos]

2 minutes ago, zbyti said:

Don't be surprise, most peoples needs (including me) proper documentation, especially if they don't have the source code to see what and how it works or don't.

Sure:

https://language.metaproject.frl/#now

 

Not sure if you would really use 30,000 lines as a reference documentation.

[TGB1718]

2 minutes ago, zbyti said:

Don't be surprise, most peoples needs (including me) proper documentation, especially if they don't have the source code to see what and how it works or don't.

Exactly, even if it was only a list of the primitives/reserved words, anything please

[Kaj de Vos]

2 minutes ago, ilmenit said:

Do you think also about generating the output C code? 

Not sure what you mean? The current release generates C.

[Kaj de Vos]

2 minutes ago, TGB1718 said:

Exactly, even if it was only a list of the primitives/reserved words, anything please

As said, coming up next, but the storm here is really slowing down my work.

[zbyti]

7 minutes ago, Kaj de Vos said:

Sure:

https://language.metaproject.frl/#now

 

Not sure if you would really use 30,000 lines as a reference documentation.

If I was desperate and had to use Meta, I would read  

 

But I would prefer brief documentation.  

 

Look... I even revamp two documentations because I wanted to learn those languages:

 

https://zbyti.github.io/k65-mkdocs/

 

http://mads.atari8.info/doc/en/

[ilmenit]

5 minutes ago, Kaj de Vos said:

Not sure what you mean? The current release generates C.

ok, how to receive it? Is there some command line parameter to compile.ape? It generates only program.xex 

[Kaj de Vos]

No, it's internal.

[Kaj de Vos]

2 minutes ago, zbyti said:

If I was desperate and had to use Meta, I would read  

 

But I would prefer brief documentation.  

 

Look... I even revamp two documentations because I wanted to learn those languages:

 

https://zbyti.github.io/k65-mkdocs/

 

http://mads.atari8.info/doc/en/

Great, then the solution is that you write the Meta documentation. 

[ilmenit]

4 minutes ago, Kaj de Vos said:

No, it's internal.

so to ask differently, do you plan to generate for the programmer the output C code? 

[Kaj de Vos]

On 11/11/2021 at 2:06 PM, ilmenit said:

Also in why section you wrote "It is known that the number of lines of program text a software developer can write per unit of time is largely independent of the programming language." - any source for this, because I'd say it's quite opposite (e.g. ASM vs C vs Python vs C#)?

Found the source of what's in my head:

http://www.paulgraham.com/power.html

[Kaj de Vos]

2 minutes ago, ilmenit said:

so to ask differently, do you plan to generate for the programmer the output C code? 

I plan to reserve that for paying customers. It was hard to implement, so I consider it a trade secret until the project is firmly established.