• ziomus (9/9) Aug 03 2003 I read phobos library sources.... D is portability.... for Windows or Li...
• Ilya Minkov (19/29) Aug 03 2003 We've got a (currently defunct) project to bind the front-end to GCC or
• Rich C (4/6) Aug 03 2003 Just curious as to what you don't like about it?
• Ilya Minkov (16/17) Aug 04 2003 Oh, there's enough to it.
• Mike Wynn (9/13) Aug 04 2003 I looked through the Sather docs sometime back, and although I do argee ...
• Sean L. Palmer (16/29) Aug 05 2003 You certainly had to type a lot more in Pascal/Delphi to get the same th...
• Ilya Minkov (14/25) Aug 05 2003 Not really. Because most of function calls in C++ would be replaced by
• Sean L. Palmer (26/43) Aug 04 2003 D is currently better than C.
• Mark Evans (3/6) Aug 04 2003 Counterarguments:
• Mark Evans (2/4) Aug 15 2003
• Ilya Minkov (27/36) Aug 04 2003 It simply coudn't be different, else it wouldn't be OCaml. :) And it has...
• Frank Wills (16/26) Aug 04 2003 I've been using C/C++ since around '84. Anything that kind of
• Frank Wills (3/20) Aug 04 2003 correction: Anything that _isn't_ kind of like C/C++ is out of
• Walter (4/10) Aug 05 2003 I'm with you there. I've used it so much it is just second nature to me.
• Frank Wills (5/12) Aug 05 2003 Yeah, I was a real Pascal hater, which we had to use
• Walter (9/18) Aug 05 2003 I wrote several pascal programs long ago. Most of the programming time w...
• Ilya Minkov (4/11) Aug 05 2003 No they don't!
• Mike Wynn (27/30) Aug 07 2003 the
• Walter (5/11) Aug 07 2003 robustly
• Walter (12/22) Aug 07 2003 was

ziomus <ziomus_member pathlink.com> writes:

I read phobos library sources.... D is portability.... for Windows or Linux....
bleah.... portability ? 

And what at BSD, SunOS... other UNIX-Systems... and general others system. ?

portable code for linux is not portable ;ppp

when will be dmd portable ?

I don't understand chapter "The C Preprocessor vs D"

Why you compare MACRO to functions and constant variables ?, this are different
things


..?

Ilya Minkov <midiclub 8ung.at> writes:

ziomus wrote:
 I read phobos library sources.... D is portability.... for Windows or Linux....
 bleah.... portability ? 
 
 And what at BSD, SunOS... other UNIX-Systems... and general others system. ?

We've got a (currently defunct) project to bind the front-end to GCC or 
output compilable C or C++ code. Then we'll be very portable.

The only library part hindering portability at large i'm aware of is a 
GC. Some other parts will probably requiere bugfixing after there's a 
cross-platform compiler. :)

 portable code for linux is not portable ;ppp
 
 when will be dmd portable ?

As soon as some people join in which have enough time and skill to do 
something, and actually come to write it. You won't be alone, you've got 
experienced Jan Knepper on your side, who is quite busy, and probably 
also undermotivated by the lack of people who'd help.

I first considered helping, but now i think i don't like D that much 
after all.

 I don't understand chapter "The C Preprocessor vs D"

What don't you understand specifically? D doesn't include a C-like 
preprocessor. This chapter explains, how you can solve the tasks in D, 
for which you use a preprocessor in C.

 Why you compare MACRO to functions and constant variables ?, this are different
 things

Hopefully yes. :) but when developing in C, you are often forced to 
emulate constants and lightweight functions through macros, which is... 
well, you answer your own question. :)

-i.

"Rich C" <no spam.com> writes:

"Ilya Minkov" <midiclub 8ung.at> wrote in message
news:bgjbkv$2l4m$1 digitaldaemon.com...

 I first considered helping, but now i think i don't like D that much
 after all.

Just curious as to what you don't like about it?

Rich C.

Ilya Minkov <midiclub 8ung.at> writes:

Rich C wrote:
 Just curious as to what you don't like about it?

Oh, there's enough to it.

First, i *hate* C syntax. Second, as i ran over D, i have cut myself 
hard on C, and was afraid to learn C++. D is much easier.

Now, that i've learned C++, the need in D is not that major.

Besides, i have discovered Sather, which has a syntax i like, as well as 
a sane semantics, which is gorgeous. :) It has not only much more 
theoretical beauty, it feels also very simple, natural and convenient, 
without sharp corners.

Like, take OCaml. A magnificcent language, but overcomplicated and 
*with* sharp semantic corners. And a horrible syntax, which implies 
hundreds of operators and wild name disambiguation. Another example with 
sharp corners is a paranoid Eiffel. Though seemingly similar to Sather, 
it's actually not.

So, don't let me hold you back from using D.

-i.

"Mike Wynn" <mike.wynn l8night.co.uk> writes:

 Besides, i have discovered Sather, which has a syntax i like, as well as
 a sane semantics, which is gorgeous. :) It has not only much more
 theoretical beauty, it feels also very simple, natural and convenient,
 without sharp corners.

I looked through the Sather docs sometime back, and although I do argee with
you view on Sather to me it suffers from the fault as Delphi ... great
features, 90% of the syntax is good but lack the symbolic short cuts and has
different use of ';'
I like the C `if`  (prefer the perl `if <bool><block>[<elseif/else>]`|
<statement> if <bool> )
I always got driven mad by if <bool> then <stat> else <stat> ';'
and resorted to begin and end with every if delphi code.

don't you miss a++ and b += foo; etc ?

"Sean L. Palmer" <palmer.sean verizon.net> writes:

You certainly had to type a lot more in Pascal/Delphi to get the same thing
to happen.

I don't see how anybody can think begin/end is more "readable" than { / }.
It's certainly less writable!!

Yes, a = succ(a) certainly seems less readable than a++.  (slightly
different semantics though).  In Delphi you can't put assignments anywhere
in an expression, they have to be their own statement.

I like syntactical shortcuts, so long as they don't cause ambiguity.  If you
can replace a symbol/keyword with a shorter one and maintain everything else
being equal I'd rather them be shorter.  I can use the saved keystrokes to
make my identifiers longer.  ;)

Sean

"Mike Wynn" <mike.wynn l8night.co.uk> wrote in message
news:bgls6m$1sq6$1 digitaldaemon.com...
 Besides, i have discovered Sather, which has a syntax i like, as well as
 a sane semantics, which is gorgeous. :) It has not only much more
 theoretical beauty, it feels also very simple, natural and convenient,
 without sharp corners.

 I looked through the Sather docs sometime back, and although I do argee

with
 you view on Sather to me it suffers from the fault as Delphi ... great
 features, 90% of the syntax is good but lack the symbolic short cuts and

has
 different use of ';'
 I like the C `if`  (prefer the perl `if <bool><block>[<elseif/else>]`|
 <statement> if <bool> )
 I always got driven mad by if <bool> then <stat> else <stat> ';'
 and resorted to begin and end with every if delphi code.

 don't you miss a++ and b += foo; etc ?

Ilya Minkov <midiclub 8ung.at> writes:

Sean L. Palmer wrote:
 You certainly had to type a lot more in Pascal/Delphi to get the same
 thing to happen.

Not really. Because most of function calls in C++ would be replaced by
property assignments and reads in Delphi. None of these ugly () midst in
expressions, which already have enough of parenthesis.

But well... the library wasn't that advanced, and the lack of templates
caused conceptually more typing, that's true. Not syntax.

 I don't see how anybody can think begin/end is more "readable" than {
 / }. It's certainly less writable!!

/methinks begin...end is cool. Readable and writable. I have a hard time
stretching my fingers to get shift, ctrl, and alt right, to type in
these special symbols! while hatural keywords are very fast and
strain-free to type.

 Yes, a = succ(a) certainly seems less readable than a++.  (slightly 
 different semantics though).  In Delphi you can't put assignments
 anywhere in an expression, they have to be their own statement.

Who uses succ sux! The right way is "Inc (a);"!!!

 I like syntactical shortcuts, so long as they don't cause ambiguity.
 If you can replace a symbol/keyword with a shorter one and maintain
 everything else being equal I'd rather them be shorter.  I can use
 the saved keystrokes to make my identifiers longer.  ;)

Short keywords are good, as long as they are used vey often. Nothing
against "bool", "int", ...

-i.

"Sean L. Palmer" <palmer.sean verizon.net> writes:

D is currently better than C.

That isn't saying much.  Alot of it is unfinished, either in design or
implementation.  It still lacks polish and a certain amount of elegance.
There are certain points in the language design that I don't think are ever
going to be resolved to my satisfaction either because of irrational
backward compatibility with C or that Walter and I just don't see eye to eye
on many issues.  ;)

I don't think D is ready for prime time yet.  The core language design still
needs work.  The primary implementation is missing several language
features.  There needs to be more implementations or at least a GCC frontend
for D.

D has great potential, but for now, I have to use C++.

There are languages coming out that are better, but none that "have it all",
so to speak, and so I can't really switch to them either.

I agree with you Ilya about OCaml.  It could be a great language, but I
think it was designed by crazed Frenchmen;  it's a really alien feeling
language to people used to C++.  It seems complicated, and relies much on
convention for naming of conversions and print methods.  I really dislike
the syntax.

Sather is better in many ways but still don't like it.  It's just a gut
reaction;  maybe I can overcome it.

I wish one of the array languages like Nial would get popular or merge
somehow with one of the good functional languages.  ;)

Sean

"Ilya Minkov" <midiclub 8ung.at> wrote in message
news:bglnqq$1o8p$1 digitaldaemon.com...
 Rich C wrote:
 Just curious as to what you don't like about it?

 Oh, there's enough to it.

 First, i *hate* C syntax. Second, as i ran over D, i have cut myself
 hard on C, and was afraid to learn C++. D is much easier.

 Now, that i've learned C++, the need in D is not that major.

 Besides, i have discovered Sather, which has a syntax i like, as well as
 a sane semantics, which is gorgeous. :) It has not only much more
 theoretical beauty, it feels also very simple, natural and convenient,
 without sharp corners.

 Like, take OCaml. A magnificcent language, but overcomplicated and
 *with* sharp semantic corners. And a horrible syntax, which implies
 hundreds of operators and wild name disambiguation. Another example with
 sharp corners is a paranoid Eiffel. Though seemingly similar to Sather,
 it's actually not.

 So, don't let me hold you back from using D.

 -i.

Mark Evans <Mark_member pathlink.com> writes:

I agree with you Ilya about OCaml.  It could be a great language, but I
think it was designed by crazed Frenchmen;  it's a really alien feeling
language to people used to C++

Counterarguments:
http://caml.inria.fr/archives/200104/msg00013.html
http://caml.inria.fr/caml-list/1542.html

Mark Evans <Mark_member pathlink.com> writes:

Try 
http://merjis.com/richj/computers/ocaml/tutorial/


http://caml.inria.fr/archives/200104/msg00013.html
http://caml.inria.fr/caml-list/1542.html

Ilya Minkov <midiclub 8ung.at> writes:

Sean L. Palmer wrote:

 I agree with you Ilya about OCaml.  It could be a great language, but I
 think it was designed by crazed Frenchmen;  it's a really alien feeling
 language to people used to C++.  It seems complicated, and relies much on
 convention for naming of conversions and print methods.  I really dislike
 the syntax.

It simply coudn't be different, else it wouldn't be OCaml. :) And it has 
nothing to do with crazy frenchmen, since i don't appear to know any 
frenchman to use OCaml. :) Though some to use C++.

And the problem is not with the syntax itself - the tokens and their 
order can be changed. It's in the semantics. For global type infererence 
to work, all names have to be disambiguated -- and that refers not only 
to identifiers, but also to operators. Every datatype defines its own 
set of operators. A reference defines its own special kind of 
assignment... and so on. And i'm just not ready to take time to study 
deeply another language, which is not even a bit less complex than C++.

It's a peak (one of the possible peaks) of what a functional language 
can reach, and it's impressiing. But it's so complex... so that when i'd 
have to decide between C++ and OCaml, i'd still and alwas choose C++, 
which is at least widely used. And with all the advantage which follows 
it. Including somewhat crazy but experienced frenchmen giving me tips on 
my personal projects. :) Well, it's probably not very realistic, but the 
libraries on my hands are.

 Sather is better in many ways but still don't like it.  It's just a gut
 reaction;  maybe I can overcome it.

It took me some time to realise that OCaml is an interesting and 
powerful language. :) However, i still wouldn't like to get acquainted 
with it any further.

 I wish one of the array languages like Nial would get popular or merge
 somehow with one of the good functional languages.  ;)

Hm... There are still so many languages i haven't heard of... dispite 
Mark always dropping a link, and myself always following them.

-i.

P.S.

http://www.csse.monash.edu.au/~damian/papers/
Human factors in programming languages.

Frank Wills <fdwills sandarh.com> writes:

Ilya Minkov wrote:
 Rich C wrote:
 
 Just curious as to what you don't like about it?

 
 
 Oh, there's enough to it.
 
 First, i *hate* C syntax. Second, as i ran over D, i have cut myself 
 hard on C, and was afraid to learn C++. D is much easier.
 

I've been using C/C++ since around '84. Anything that kind of
like C/C++ is out of my comfort zone, except assembly, which
I like, but don't need so much.

D works for me because it is essentially a modernization of C/C++,
with a new OO foundation. Also, since it is more of an open source

that no one will be as able to turn it into some kind of monster


I can look forward for writing my own libraries, or using others
that are developed as open source software, and can refine
everything as much as I want for my own use. It's a language that
I can "own", rather than getting "owned" by some corporation
because I bought into their tools. I imagine that this is the
general case for many others using D. It's also the reason I
am doing all I can to use *BSD and Linux.

Frank Wills <fdwills sandarh.com> writes:

Frank Wills wrote:
 Ilya Minkov wrote:
 
 Rich C wrote:

 Just curious as to what you don't like about it?



 Oh, there's enough to it.

 First, i *hate* C syntax. Second, as i ran over D, i have cut myself 
 hard on C, and was afraid to learn C++. D is much easier.

 
 I've been using C/C++ since around '84. Anything that kind of
 like C/C++ is out of my comfort zone, except assembly, which
 I like, but don't need so much.

correction: Anything that _isn't_ kind of like C/C++ is out of
my comfort zone.

"Walter" <walter digitalmars.com> writes:

"Frank Wills" <fdwills sandarh.com> wrote in message
news:bgmbtr$2c97$2 digitaldaemon.com...
 Frank Wills wrote:
 I've been using C/C++ since around '84. Anything that kind of
 like C/C++ is out of my comfort zone, except assembly, which
 I like, but don't need so much.

 correction: Anything that _isn't_ kind of like C/C++ is out of
 my comfort zone.

I'm with you there. I've used it so much it is just second nature to me.
Looking at pascal syntax just sets my teeth on edge <g>.

Frank Wills <fdwills sandarh.com> writes:

Walter wrote:
 "Frank Wills" <fdwills sandarh.com> wrote in message
 news:bgmbtr$2c97$2 digitaldaemon.com...
 
 I'm with you there. I've used it so much it is just second nature to me.
 Looking at pascal syntax just sets my teeth on edge <g>.
 
 

Yeah, I was a real Pascal hater, which we had to use
for all of our college lab assignments. I was also a
self-appointed C evangelist, which they weren't yet
offering as a course.

"Walter" <walter digitalmars.com> writes:

"Frank Wills" <fdwills sandarh.com> wrote in message
news:bgnsb4$pgl$1 digitaldaemon.com...
 Walter wrote:
 "Frank Wills" <fdwills sandarh.com> wrote in message
 news:bgmbtr$2c97$2 digitaldaemon.com...
 I'm with you there. I've used it so much it is just second nature to me.
 Looking at pascal syntax just sets my teeth on edge <g>.

 Yeah, I was a real Pascal hater, which we had to use
 for all of our college lab assignments. I was also a
 self-appointed C evangelist, which they weren't yet
 offering as a course.

I wrote several pascal programs long ago. Most of the programming time was
spent *fighting* the compiler, as it thwarted me at every turn. It was so
bad I was turning to using assembly language. Then I chanced on a copy of
K+R, and it was like a breath of fresh air. No longer was I fighting the
compiler, as the language allowed me to do what I wanted.

I also simply liked the terseness of C. When you write a lot of code, the
long wordy keywords of Pascal just get annoying.

Ilya Minkov <midiclub 8ung.at> writes:

Walter wrote:

 I wrote several pascal programs long ago. Most of the programming time was
 spent *fighting* the compiler, as it thwarted me at every turn. It was so
 bad I was turning to using assembly language. Then I chanced on a copy of
 K+R, and it was like a breath of fresh air. No longer was I fighting the
 compiler, as the language allowed me to do what I wanted.

You had a wrong compiler. :) Nowadays they don't have these problems.

 I also simply liked the terseness of C. When you write a lot of code, the
 long wordy keywords of Pascal just get annoying.

No they don't!

-i.

"Mike Wynn" <mike.wynn l8night.co.uk> writes:

"Ilya Minkov" <midiclub 8ung.at> wrote in message
news:bgp2e4$1vin$1 digitaldaemon.com...
 I also simply liked the terseness of C. When you write a lot of code,


the
 long wordy keywords of Pascal just get annoying.

 No they don't!

Yes they do!! 6 months of Delphi was enough for me,

I think it is very much personal taste, having had to write in C, C++, Java,
Delphi, Perl and PHP all sections of the same project often all in the same
day, and with some code having to be ported from one lang to another, you
start to develope strong personal preferences over how to do things.
personally I lean towards the Java way for most stuff, it have the shortcuts
and it defines the order of eval of params to functions.
delphi as I remember will not allow you to write
if  i := func( foo ) then .... ;
in Java you can write i = ar[idx++];
or even func( ar[idx++], ar[idx++], ar[idx] ); => func( ar[idx], ar[idx+1],
ar[idx+2] ); idx+=2;

pointers are another stumbling point to me with pascal/deplhi/modular2 I
always found them a little confusing and akward, once I leant C I found
pointers much more useable.
and although passing 'var' params is nice I prefer the C *, &foo or c++ &
I do miss sets, they where useful but that is realy the only good thing I
can say about pascal based langs

pascal based langs also try to give the illusion of robustness by (imho)
hiding much of what they are doing away from the programmer, C makes no
illusion about letting you walk of either end of your array!
on the whole I would say C makes you aware that you have to program robustly
and unlike c++ of delphi there is little going on outside your code that you
have to be aware of.

"Walter" <walter digitalmars.com> writes:

"Mike Wynn" <mike.wynn l8night.co.uk> wrote in message
news:bgu082$k41$1 digitaldaemon.com...
 pascal based langs also try to give the illusion of robustness by (imho)
 hiding much of what they are doing away from the programmer, C makes no
 illusion about letting you walk of either end of your array!
 on the whole I would say C makes you aware that you have to program

robustly
 and unlike c++ of delphi there is little going on outside your code that

you
 have to be aware of.

You're right that what-you-write-is-what-you-get in C can be very appealing.

"Walter" <walter digitalmars.com> writes:

"Ilya Minkov" <midiclub 8ung.at> wrote in message
news:bgp2e4$1vin$1 digitaldaemon.com...
 Walter wrote:
 I wrote several pascal programs long ago. Most of the programming time


was
 spent *fighting* the compiler, as it thwarted me at every turn. It was


so
 bad I was turning to using assembly language. Then I chanced on a copy


of
 K+R, and it was like a breath of fresh air. No longer was I fighting the
 compiler, as the language allowed me to do what I wanted.

 You had a wrong compiler. :) Nowadays they don't have these problems.

What those old pascal compilers did was implement the language per Wirth's
spec. It wasn't until years later that compiler companies figured out that
pascal, as specified, is an unusable language. Each then put in a boatload
of mutually incompatible extensions to fix that, making pascal a very
unportable language.

 I also simply liked the terseness of C. When you write a lot of code,


the
 long wordy keywords of Pascal just get annoying.

 No they don't!

<g>