World: r3wp

btw... if I do print float, I get result like: 100.00000000000000 
. I guess you know this issue.

By default, PRINT for floats maps to `printf ["%.14f" f]`, I guess 
the formatting string passed to PRINTF could be improved.

Since Red is related to REBOL, prin and print seem natural choices.

LF by default: sure, but the issue is coming up with a good name 
for PRINT without LF then. We already had a debate here about that 
a few months ago, and having just PRINT ended up being the less problematic 
solution (but it's not perfect I know).

I suppose "%14f" would avoid the extra zeros? (haven't checked the 
PRINTF docs).

What is wrong with PRIN? It's not more confusing than LF

I wonder if Carl's choice for PRIN was planned or added later when 
the need for a PRINT-WITHOUT-LF arised...

And I think in 99% you want the LF :) At least I forget to add it 
all the time.

Anyway.. not important now.

The only issue with PRIN is that it looks odd to all newcomers, not 
knowing about REBOL. I guess also we all found that odd in REBOL 
while learning it. I am used to it now, but it could be seen as "weird" 
choice by people first looking at Red/System.

Also PRIN is breaking the naming conventions used in REBOL, like 
having full words.

Since Red is related to REBOL, prin and print seem natural choices.
 - I agree.

like having full words
 - do words like REFORM not also violate that rule?

reform
 is a valid english word, no?

;-)

but in REBOL, it is a short for REDUCE-FORM.

Yeah, I know. We will use the same words in Red for maintaining compatibility 
with REBOL. I wish we had stricter rules for naming, but I guess 
that it won't hold in practice.

I think I would just be annoyed that Red tries to "fix" the wording 
of a REBOL function, as it means that you need some kind of alias 
table or must fix scripts or remember two sets of functions in your 
head.

I guess it's too soon for int64, isn't it?

Agreed, and I plan to keep Red as compatible as possible with REBOL 
wrt the syntax, but Red/System, as a dialect, can be a bit different.

int64: too soon yes, adding support for it would break all the current 
integer math code in backends, so not a trivial task. I have not 
planned to support it in this first Red/System version but add it 
only in the v2 (the rewrite in Red). OTOH, if we hit some wall in 
supporting an important feature or binding, I might reconsider that.

It should be possible though, to support it for bindings, but without 
any math operations.

it is used in iMagick. It would be fine without math.

I think it can be simulated with struct but I don§t know how difficult 
it is to implement it just without math.

It would take 2-3 days probably.

But I don't want to delay the work on Red compiler anymore, I should 
have started around new year, and I'm still on Red/System...

I must invest which routines are affected, but I think it's not a 
big priority like the float.

What about enums? What is the best way how to work with them in Red/System?

No plan for enums for now. You can use defines instead.

that would be quite a lot defines: https://gist.github.com/1683212

What about adding #enum into the loader?

I would like enumeration constants to be first class. Possibly integrated 
with a symbol type

no refinements support in Red/System, that will come in Red only.

 - just curious - isn't such a feature still a valid entry for future 
 Red/System enhancements? It is still in 19.7. section - should be 
 updated in doc as a dismissed feature then ...

Regarding PRINT, I would suggest simply defining a PRINT-LINE as 
an alternative to PRINT [LF}. There already is one in the C library 
binding, but it just prints a simple string

I don't agree... it's loger than print [lf] and as I said.. at least 
im my case 99% I need print with LF

Yes, [lf] is really not that long, so it's hard to find a good shorter 
alternative. In many cases you'll need a block anyway, so it's just 
" lf"

How about ?? as an alias for print-line, if it really must be ultra 
short for debugging?

Pekr: the section name in the documenation is "Possible Evolutions", 
not "Planned Evolutions", so features listed there might or might 
not make it in Red/System.

??
 sounds like a good addition.

Regarding integer64!, wouldn't it currently be possible to fake it 
with float! to pass them around to bindings?

I think it should work.

+1 for "??"

I guess someone can contribute that to github, no need for me to 
write it? ;-)

FEAT: added `print-line` function to runtime, with `??` as alias. 
Works like `print`, but adds a line-feed character at end.
 

https://github.com/dockimbel/Red/commit/82099d117e790859606697c33f90c35ef87cf5b6

Just imo, but:


1) hmm, I am with oldes - if there is 99% case, where you want your 
"print" to print including the newline, it follows REBOL rule = most 
used case goes first

2) There is nothing wrong with following REBOL compatibility, especially 
if we claim, Red is inspired by REBOL. So 'prin and 'print worked, 
new user would learn their stuff

3) If Red itself will have some REBOL compatibility, I would expect 
both 'prin and 'print there. Now imagine, how messy the translation 
is going to be: print -> print-line, prin -> print ...

it's not a big deal for me, so regard it being just my opinion ...

As Doc is a guru here, the final decision is of course upon to him 
:-)

There are going to be quite a few differences between Red and REBOL, 
anyway. It's a fundamentally different language. You can't expect 
to run REBOL programs unchanged, except in incidental cases

I used to be of the opinion that REBOL clones should be compatible, 
but Red is not a clone, REBOL itself is incompatible between R2 and 
R3, and REBOL in general is fading into irrelevance

Pekr: you should really make the distinction between Red and Red/System, 
they don't have the same goals. People currently coming to Red/System 
are usually C coders, not REBOL-only coders. Things like "PRIN" might 
give a negative feeling on the language design, and it would be a 
pity, because PRIN is not my idea/choice, but Carl's one. At Red 
level (when it will be available), we'll have a wider compatibility 
with REBOL, so to ease the transition for rebolers, PRIN will be 
there (but maybe just as an alias).