ATS3: syntax

[gmhwxi]

I will be working on documenting the syntax of ATS3 from time to time:

https://github.com/githwxi/ATS-Xanadu/tree/master/docgen/SYNTAX

This is also time for me to clarify my own thoughts. Whenever I implemented

ATS in the past, I felt rushed to get things done. This time I am trying to take

time to do a bit of thinking before coding :)

Obviously, some changes to the syntax will be made over the time if needed.

Please bear in mind that ATS3 is meant to be both a source language and a target language.

Its current syntax is on the verbosity side. More syntactic sugar could be added later.

As always, you are welcome to give your comments, suggestions, feedback, etc.

Feel free to ask questions, too.

Cheers!

--Hongwei

[Kiwamu Okabe]

Why not introduce Proof stuff at `overview.txt `?

I believe ATS3 stands on:

* Dynamics
* Proofs
* Statics

http://cs.likai.org/ats/ml-programmers-guide-to-ats

And also the Proofs are important parts for me.

[Kiwamu Okabe]

I think we need `ATS-Xanadu/docgen/SYNTAX/README.md` for table of contents.

[Kiwamu Okabe]

If `ATS-Xanadu/docgen/SYNTAX` is a specification document, not a
tutorial, we do not need to choose any format such like markdown, TeX,
SGML, XML, HTML, etc...
Let's choose your style as plain text such like RFC:

https://www.rfc-editor.org/in-notes/rfc-index.txt

But we need TOC.

[gmhwxi]

 

Proofs are part of dynamics. As far syntax design is concerned, writing a proof

is pretty much like writing a program.

[Kiwamu Okabe]

On Mon, Oct 22, 2018 at 9:04 AM gmhwxi <gmh...@gmail.com> wrote:
> Proofs are part of dynamics. As far syntax design is concerned, writing a proof
> is pretty much like writing a program.

OK. I should re-draw my big picture:

http://jats-ug.metasepi.org/draw/flow.png
--
Kiwamu Okabe at METASEPI DESIGN

[gmhwxi]

I don't want to take the RFC route. At least not for now. ATS is too rich

and complex to be handle in this way.

I am writing some notes now. In these notes, I plan to mostly talk about

various syntax entities in ATS3 and also use concrete examples to illustrate

them.

Once I have written enough notes, I (or someone else) can quick prepare

a TOC based on these notes.

[Kiwamu Okabe]

As a translator English to Japanese, I would like to choose `po4a` tool.

https://po4a.org/index.php.en

It can maintain translations on gettext, and be easy to track original update.
Could you choose your document format from followings which are supported
by po4a:

* manpages
* POD
* XML (generic, DocBook, XHTML, Dia, Guide, or WML)
* SGML
* TeX (generic, LaTeX, or Texinfo)
* text (simple text files with some formatting, markdown, or AsciiDoc)
* INI
* KernelHelp

Regards,

[Hongwei Xi]

>> text (simple text files with some formatting, markdown, or AsciiDoc)

I choose this one: Text + Markdown is what I like most these days :)

--
You received this message because you are subscribed to the Google Groups "ats-lang-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ats-lang-user...@googlegroups.com.
To post to this group, send email to ats-lan...@googlegroups.com.
Visit this group at https://groups.google.com/group/ats-lang-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/ats-lang-users/CAEvX6d%3DO2Vu%2BSggZ0UA38DK_gCBQzDtbHjzd9qiVb7wv_a_YvA%40mail.gmail.com.

[Vanessa McHale]

You can use pandoc to get manpages, TeX, etc. if you'd like :)

To view this discussion on the web visit https://groups.google.com/d/msgid/ats-lang-users/CAPPSPLrgq9MWRAQFtmMObkW8oRuZDUPC7Sm9tbQiYouXWoaZPw%40mail.gmail.com.

[Kiwamu Okabe]

On Mon, Oct 22, 2018 at 9:21 AM Hongwei Xi <gmh...@gmail.com> wrote:
> I choose this one: Text + Markdown is what I like most these days :)

:+1:

[Kiwamu Okabe]

I have question: what is different on following?:

* #infix0
* #infixl
* #infixr

How about introduce your infix syntax such like Haskell?

"Infix Functions In Haskell"
https://wuciawe.github.io/functional%20programming/haskell/2016/07/03/infix-functions-in-haskell.html

The infix syntax in Haskell has:

A. order such like `infixl 7 *`
B. way on left or right

I believe that you already have B on your infix syntax, but not A.

[Kiwamu Okabe]

On Mon, Oct 22, 2018 at 11:43 AM Kiwamu Okabe <kiw...@debian.or.jp> wrote:
> How about introduce your infix syntax such like Haskell?
>
> "Infix Functions In Haskell"
> https://wuciawe.github.io/functional%20programming/haskell/2016/07/03/infix-functions-in-haskell.html
>
> The infix syntax in Haskell has:
>
> A. order such like `infixl 7 *`
> B. way on left or right
>
> I believe that you already have B on your infix syntax, but not A.

Ah, I already has A:

```
#infixl * / % mod of 60
```

When we use `#infix0`? Why use that?

[Hongwei Xi]

It is all explained here:

https://github.com/githwxi/ATS-Xanadu/blob/master/docgen/SYNTAX/fixity.txt

#infix0 : infix with no associativity

#infixl : infix with left associativity

#infixr : infix with right associativity

The way I handle fixity is very similar to that of Haskell. But a bit more general.

For instance, I allow an operator to be both infix and prefix (or infix and postfix).

Also, you can do

#infixl ++ of +(-1) //  ++ is infixl and its precedence value is 1 less than that of +

--
You received this message because you are subscribed to the Google Groups "ats-lang-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ats-lang-user...@googlegroups.com.
To post to this group, send email to ats-lan...@googlegroups.com.
Visit this group at https://groups.google.com/group/ats-lang-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/ats-lang-users/CAEvX6dm-6B-QuHqDz6u-1RYp%2BjT9y-YCb47VwPPgN5cfdsNq%3DQ%40mail.gmail.com.

[Kiwamu Okabe]

Ah...

```
$ git grep infix0 | cat
docgen/SYNTAX/fixity.txt:#infix0 foo2 bar2
docgen/SYNTAX/fixity.txt:#infix0 baz of 71
prelude/fixity.sats:#infix0 < <= > >= of 40
prelude/fixity.sats:#infix0 = != == !== of 30
prelude/fixity.sats:#infix0 := of 0 // HX: assign
prelude/fixity.sats:#infix0 :=: of 0 // HX: exchange
prelude/fixity.sats:#infix0 :+= :-= :*= :/= // x:=x+a, x:=x-a, ...
prelude/fixity.sats:#infix0 :=+ :=- :=* :=/ // x:=a+x, x:=a-x, ...
srcgen/xats/DATS/lexing_kword.dats:val () = myins("#infix0", T_INFIX_)
srcgen/xats/TEST/DATA/syntax_dyn.dats:#infix0 40 < <= > >=
srcgen/xats/TEST/DATA/syntax_dyn.dats:#infix0 (0) := // HX: assign
srcgen/xats/TEST/DATA/syntax_dyn.dats:#infix0 (0) :=: // HX: exchange
srcgen/xats/TEST/DATA/syntax_dyn.dats:#infix0 (0) :+= :-= :*= :/= //
x:=x+a, x:=x-a, ...
srcgen/xats/TEST/DATA/syntax_dyn.dats:#infix0 (0) :=+ :=- :=* :=/ //
x:=a+x, x:=a-x, ...
srcgen/xats/TEST/DATA/syntax_sta.sats:#infix0 < <= of 40
srcgen/xats/TEST/DATA/syntax_sta.sats:#infix0 > >= of 40
srcgen/xats/TEST/DATA/syntax_sta.sats:#infix0 = == != of 30
```

`#infix0` is only for equation? If so, please note that on `fixity.txt`.

[Kiwamu Okabe]

On Mon, Oct 22, 2018 at 11:59 AM Hongwei Xi <gmh...@gmail.com> wrote:
> It is all explained here:
>
> https://github.com/githwxi/ATS-Xanadu/blob/master/docgen/SYNTAX/fixity.txt
>
> #infix0 : infix with no associativity
> #infixl : infix with left associativity
> #infixr : infix with right associativity

Thanks. It's clear for me, now.