Valued fascicle update

[Peter McGoron]

Hello WG2,

I have been working on the text of the Valued Fascicle, and I now
believe that is feature-complete. That is, it should contain a
description for each feature, syntax, procedure, etc. that the fascicle
should cover. This Fascicle is big, probably the biggest in the
Foundations, (500KiB, 70KiB compressed) so I might have missed some
things. It also doesn't include some things we haven't voted into it yet.

You can view it here:

https://florida.moe/r7rs/docbook/valued-fascicle/unofficial-draft/

or compile it yourself by cloning the R7RS repository and running `make
compiled/valued-fascicle.xhtml`. (I recommend Gauche when compiling.
Chibi should work, but is extremely slow.)

There are a lot of errors, if you find any please tell me. In
particular, I have to add a lot more auto-generated section numbers to
make the fascicle more print-friendly.

Some editorial notes:

1. The fixnum and flonum procedures that correspond to generic
arithmetic procedures are included in the fascicle.

2. Fixnum and flonum procedures that I consider to be fundamental
(predicates, accessors, and constructors) are included in separate
sections. Other procedures should be in the Batteries.

3. The core bitwise operations are added: the rest are in the batteries.

4. The formal syntax of each type is included in the section at hand.
DocBook is flexible enough that we should be able to combine them all
and put it in an appendix, but I think putting the formal grammar next
to the informal description of the syntax makes things easier.

5. Where different, R6RS syntax is included, under the assumption that
an implementation supports `#!r6rs` to switch to R6RS lexical syntax mode.

6. `read` and the `write` family are included because they are closely
related to datum representation.

7. The “Changes to the R7RS” section is extremely incomplete. I will
probably fill it in closer to the release of this Fascicle.

-- Peter McGoron

[Zhu Zihao]

Hi, Peter.

Here're some of my comments on the updated draft

1. In Editorial Convention. we should explictly state that "start" and
"end" denotes *exact integer* in the list, following the previous
standards.

2. Exact bitwise arithemtic should operates on "exact integer", not just
"integer" (denotes by "n"). I also suggest to use "ei" to denote "exact
integer", following the R6RS convention [R6RS-Bitwise].

3. bitwise-and have following semantics

> Returns the logical AND of each bit in the argument list. That is, the
> nth bit in the returned value is 1 if the nth bit in each of the
> arguments is 1, and 0 otherwise. This operation is associative.

(similar semantics applied mutatis mutandis to inclusive OR and exclusive OR)

"Returns the logical AND of each bit in the argument list" doesn't make
sense here, because there're actually exact integers in the argument
list.

Maybe say "Returns the bit-wise logical AND of each exact integer in the
argument list"?


[R6RS-Bitwise]: https://r6rs.org/final/html/r6rs-lib/r6rs-lib-Z-H-12.html#node_sec_11.4



--
Retrieve my PGP public key:
执行下列命令以获取我的 PGP 公有密钥：

gpg --recv-keys B3EBC086AB0EBC0F45E0B4D433DB374BCEE4D9DC

Zihao / 閱卜錄

[Peter McGoron]

> 1. In Editorial Convention. we should explictly state that "start"
and "end" denotes *exact integer* in the list, following the previous
standards.

> 2. Exact bitwise arithemtic should operates on "exact integer", not
just "integer" (denotes by "n"). I also suggest to use "ei" to denote
"exact integer", following the R6RS convention [R6RS-Bitwise].

Thanks, I have incorporated these changes.

> Maybe say "Returns the bit-wise logical AND of each exact integer in
the argument list"?

I have fixed the argument names. I have shortened the first sentence to
"Returns the logical AND of the arguments." etc.

-- Peter McGoron