www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - D User Poll - Ready for 1.0?

reply clayasaurus <clayasaurus gmail.com> writes:
The poll is simple.

Do you think the D compiler is ready to be called 1.0?

Say yes if D is working great for you and explain what you have done 
with it.

Say no if D is unusable in its current state, or if there is something 
about the compiler that really bothers you that should be fixed before 1.0.

I posted this poll out of a general curiousity to see how many out there 
think D is ready and if not why.

---------------------------------------------------------------------------

Yes.

    My honest opinion, from the little bit I've used D for, is that the 
D compiler is ready for 1.0. Granted, I'm a little college student who 
is just trying to learn game programming (and I'm not too far into all 
the fancy algorithms like the big shots are), so take my opinion with a 
grain of salt.

    I've never run into any show stopper bugs recently, for a while, and 
I've used some basic templates, modules, bindings, libs, and the build 
utility (a BIG selling point for D IMO, especially if you get a 
graphical frontend for it, will make the newbs very happy, and make 
makefiles look like a waste of time, allow you to concentrate more on 
programming, which is D's purpose, increase programmer productivity).

    I think D is really great and it can definitely holds its own 
against other languages. I'm sure compiler bugs could be fixed for the 
next three years or so (bop the gofer, fix one another pops up).

   The general acceptance of D can only happen when 1.0 emerges, 
otherwise it is just a beta language, and it is percieved as a moving 
target, so writing large amounts of code for it can seem dangerous.

And for those who need more features, be patient and wait for 2.0?

--------------------------------------------------------------------------

Just trying to generate some discussion, and see other peoples views. :-)

--------------------------------------------------------------------------

Also, I know some do not want to discriminate between compiler/lib, but 
IMO it would be hard to rollout a solid lib if the compiler is under 
construction. I think phobos is solid as it is (maybe i'm ignorant), and 
that any great/radical improvement will have to come from a great open 
source lib, like ares or mango, or a behind the scenes redesign by the 
top-notch library writers floating around here.
Mar 30 2005
next sibling parent reply Daniel Siegmann <gandalf optonline.net> writes:
clayasaurus wrote:

 The poll is simple.
 
 Do you think the D compiler is ready to be called 1.0?


No. I think the big "1.0" should only happen when the language
specification, compiler (DMD) and standard library (Phobos) have all
reached that stage.

Being in beta doesn't necessarily keep something from being used, but
reaching 1.0 can be used as a sort of marketing tool to push adoption.
Look at Firefox - I've been using it since Phoenix 0.5 and it's always
been pretty stable, but it only became 1.0 when it was damn ready. Hell,
I've been using Gaim for awhile, and they've only recently moved to 1.0
- seemingly just 'cause they felt it was about time. Look at all the
tools Google has that are technically beta. ;)

No, I feel best to wait until we're *sure* D is ready to be 1.0 before
it becomes so. After all, you only reach 1.0 once. :)

I do wish Gentoo had an ebuild available though. :(
Mar 30 2005
next sibling parent reply Brad Anderson <brad dsource.dot.org> writes:
Daniel Siegmann wrote:

 
 I do wish Gentoo had an ebuild available though. :(


http://forums.gentoo.org/viewtopic-t-163922-highlight-dmd.html

Look for the post by Genone

BA
Mar 30 2005
parent reply =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Brad Anderson wrote:


 I do wish Gentoo had an ebuild available though. :(


 
 http://forums.gentoo.org/viewtopic-t-163922-highlight-dmd.html
 Look for the post by Genone


If you are referring to the missing versioned zipfiles,
and the licensing, that was fixed a month ago (DMD 0.113)

http://bugs.gentoo.org/show_bug.cgi?id=46806#c12

Currently the only problem is the lack of a "make install"
target in the linux.mak, and of course the missing x86_64...

--anders
Mar 30 2005
parent reply Brad Anderson <brad dsource.dot.org> writes:
Anders F Björklund wrote:

 Brad Anderson wrote:
 

 I do wish Gentoo had an ebuild available though. :(



 http://forums.gentoo.org/viewtopic-t-163922-highlight-dmd.html
 Look for the post by Genone


 
 
 If you are referring to the missing versioned zipfiles,
 and the licensing, that was fixed a month ago (DMD 0.113)
 
 http://bugs.gentoo.org/show_bug.cgi?id=46806#c12
 
 Currently the only problem is the lack of a "make install"
 target in the linux.mak, and of course the missing x86_64...
 
 --anders


I know the ebuilds are out there, but not included in Portage, because 
of the licensing.  That was my only point.  If it's fixed, maybe they'll 
include it so 'emerge dmd' works.

BA
Mar 31 2005
parent reply =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Brad Anderson wrote:


 I know the ebuilds are out there, but not included in Portage, because 
 of the licensing.  That was my only point.  If it's fixed, maybe they'll 
 include it so 'emerge dmd' works.


There should not be any problem with the ebuild, as far as I can tell:


PORTDIR_OVERLAY=/usr/local/portage

mkdir -p /usr/local/portage/dev-lang/dmd
cd /usr/local/portage/dev-lang/dmd


wget -N -O dmd-0.119.ebuild \
http://bugs.gentoo.org/attachment.cgi?id=52307&action=view

ebuild dmd-0.119.ebuild fetch
ebuild dmd-0.119.ebuild digest


ACCEPT_KEYWORDS="~x86" emerge dmd

 dev-lang/dmd-0.119 merged.






--anders
Mar 31 2005
parent reply Brad Anderson <brad dsource.dot.org> writes:
Anders F Björklund wrote:

 Brad Anderson wrote:
 

 I know the ebuilds are out there, but not included in Portage, because 
 of the licensing.  That was my only point.  If it's fixed, maybe 
 they'll include it so 'emerge dmd' works.


 
 
 There should not be any problem with the ebuild, as far as I can tell:
 

 PORTDIR_OVERLAY=/usr/local/portage
 
 mkdir -p /usr/local/portage/dev-lang/dmd
 cd /usr/local/portage/dev-lang/dmd
 
 
 wget -N -O dmd-0.119.ebuild \
 http://bugs.gentoo.org/attachment.cgi?id=52307&action=view
 
 ebuild dmd-0.119.ebuild fetch
 ebuild dmd-0.119.ebuild digest
 
 
 ACCEPT_KEYWORDS="~x86" emerge dmd
  >>> dev-lang/dmd-0.119 merged.
 
 --anders


I understand that the ebuild works.  However, it's not in Portage.  See 
these search results:

http://packages.gentoo.org/search/?sstring=dmd

BA
Mar 31 2005
parent reply =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Brad Anderson wrote:


 I understand that the ebuild works.  However, it's not in Portage.


That's up to "politics", and the Gentoo maintainers...

I just showed how to actually install the software. :-)

--anders
Mar 31 2005
next sibling parent J C Calvarese <jcc7 cox.net> writes:
Anders F Björklund wrote:

 Brad Anderson wrote:
 

 I understand that the ebuild works.  However, it's not in Portage.


 
 
 That's up to "politics", and the Gentoo maintainers...


It might be politics, but if we could persuade Walter to let D run for 
office D might "emerge" as the winner in a landslide. (Nyuk, nyuk, nyuk.)

Of course I'm not a Gentoo citizen, so I'm probably not allowed to 
vote--or even make any campaign contributions.

If I've read this thread correctly, Walter could get D on the ballot by 
making only one of these concessions:

1. Change the license to allow redistribution.
2. Versioned tarball (does the versioned zip count?).

Or maybe D is already eligible, but can't get nominated (I kind of got 
lost in this thread)...?


 
 I just showed how to actually install the software. :-)


That's always helpful, too. ;)


 
 --anders



-- 
jcc7
http://jcc_7.tripod.com/d/
Mar 31 2005
prev sibling parent reply Daniel Siegmann <gandalf optonline.net> writes:
Anders F Björklund wrote:

 Brad Anderson wrote:
 

 I understand that the ebuild works.  However, it's not in Portage.


 
 
 That's up to "politics", and the Gentoo maintainers...
 
 I just showed how to actually install the software. :-)
 
 --anders


I appreciate that, and was in fact aware of this already. Thanks to
whoever set up that ebuild.

However, it really needs to be in Portage. I like to know that all the
software on my system is updated when I type "emerge sync && emerge
world -vuDa". I know, I'm being lazy, but I'm hesitant to adopt any
software not easily available through Portage.

If there are still some political issues stopping DMD from being put
into Portage I would like to see them resolved. Is there anything I can
do to help?

--Daniel Siegmann
Apr 01 2005
next sibling parent reply Georg Wrede <georg.wrede nospam.org> writes:
Daniel Siegmann wrote:

 If there are still some political issues stopping DMD from being put
 into Portage I would like to see them resolved. Is there anything I can
 do to help?


Aren't the packages in Portage mostly Production level?

Meaning, you don't usually risk breaking something with "unnecessarily 
updating"?

Currently that is not the case with D, IMHO.

What's the politics on that issue?
Apr 02 2005
parent Daniel Siegmann <gandalf optonline.net> writes:
Georg Wrede wrote:

 Aren't the packages in Portage mostly Production level?
 
 Meaning, you don't usually risk breaking something with "unnecessarily
 updating"?
 
 Currently that is not the case with D, IMHO.
 
 What's the politics on that issue?


Portage can contain more than just stable software because ebuilds (even
newer versions of the same program) can be masked. Masked ebuilds will
not be considered available for the user unless the user specifically
unmasks them (requires one entry in one file).

There are two ways to mask a program: place it under the unstable
keyword (i.e. ~x86, instead of x86), or hard mask it. Some people may
run their system using all ~arch software, but it's not recommended.

Different versions of a program can have different masks. Often packages
will end up in ~arch first for testing, and then be moved to arch once
any bugs are worked out.

DMD is fairly stable - at least, it's not about to destroy anyone's
system (I hope), and it's not like there's a more stable version. So
stick it in ~x86 to start with, but I do think it should be in x86.

For comparison, I believe Phoenix was originally stuck in ~x86, but was
available even as far back as 0.6 (possibly earlier). However, it ended
up in x86 before 1.0.


I prefer that all the software on my system is managed by it.
Apr 02 2005
prev sibling parent =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Daniel Siegmann wrote:


 If there are still some political issues stopping DMD from being put
 into Portage I would like to see them resolved. Is there anything I can
 do to help?


There were a couple of earlier issues stopping DMD:

1) The DMD zipfiles were not being versioned
    (they are now, example being dmd.119.zip)

2) The license was not known to the Portage system
    (added to the gentoo bug, as LICENSE="DMD")

3) The license does not permit any re-distribution
    (the ebuild is now flagged RESTRICT="nomirror")

4) Man pages were not included in the DMD archive
    (they are now, after some lobbying)

5) The Makefile (linux.mak) lacks a "make install"
    (currently using a ugly workaround)

None of these should stop it from entering ~x86 now ?

--anders

PS. http://bugs.gentoo.org/show_bug.cgi?id=46806
Apr 02 2005
prev sibling parent Vladimir <Vladimir_member pathlink.com> writes:
I do wish Gentoo had an ebuild available though. :(  


I'm successfuly using ebuilds from http://bugs.gentoo.org/show_bug.cgi?id=46806

for dmd and from http://bugs.gentoo.org/show_bug.cgi?id=48136 for gdc (but with 
small bug: I had to build one missing file (config.d) by hands).
Mar 30 2005
prev sibling next sibling parent Dejan Lekic <leka entropy.tmok.com> writes:
No


-- 
...........
Dejan Lekic
  http://dejan.lekic.org
Mar 31 2005
prev sibling next sibling parent reply "David Barrett" <dbarrett quinthar.com> writes:
No, I don't believe it's currently 1.0 in that I wouldn't seriously 
recommend to my boss that we start using it.

I do believe it's ready for personal "fun" projects, but I think too much 
risk surrounds it before I would recommend building a business atop it:

- I don't know of any significant commercial application successfully 
written and released in D, and I don't want to be the first.

- Phobos seems disorganized and undirected, thus leading me to believe it 
might change significantly.  At this point, I'd probably insulate myself 
against future Phobos changes by creating my own "standard library" that 
wraps Phobos and "libc" (or I might use Ares).  But this means I and my team 
would be building up expertise in how to program with my library (or Ares), 
not Phobos, and I'd really prefer to avoid that.

- I perceive a lack of end-to-end development and debugging tools, and the 
tools that do exist don't feel "1.0".  Furthermore, I see little progress in 
integrating with existing development tools and environments under Win32 
(some great starts, but they seem to have stagnated).

- I perceive a lack of interest among the D developer community.  I see lots 
of started projects, but few completed ones, and little progress in the 
meantime.  I can only assume *something* turned off these people who gave D 
a shot, and this increases my perceived risk around D.

Note that no amount of compiler improvement will alleviate my fears. 
Further note that most of my fears could probably be alleviated through 
improved documetnation and better organization of what's already out there 
in a scattered form.

In fact, the least risky component in my mind is the D compiler itself.  I 
bet it works pretty good.  It's just all the "low priority" stuff that's 
hanging me up.

-david
Mar 31 2005
next sibling parent reply "Matthew" <admin.hat stlsoft.dot.org> writes:
 - I perceive a lack of interest among the D developer community.  I see lots
of started projects, but few completed 
 ones, and little progress in the meantime.  I can only assume *something*
turned off these people who gave D a shot, 
 and this increases my perceived risk around D.


An excellent, and important, point. Working out the reason(s) for this would be
a good use of our collective time.

But be careful, David, lest you be marked a flamer, and become a
Cassanandra-like presence. ;/
Mar 31 2005
parent reply "Regan Heath" <regan netwin.co.nz> writes:
On Thu, 31 Mar 2005 20:21:53 +1000, Matthew <admin.hat stlsoft.dot.org>  
wrote:


 - I perceive a lack of interest among the D developer community.  I see  
 lots of started projects, but few completed
 ones, and little progress in the meantime.  I can only assume  
 *something* turned off these people who gave D a shot,
 and this increases my perceived risk around D.


 An excellent, and important, point. Working out the reason(s) for this  
 would be a good use of our collective time.


Perhaps the compiler stopped these projects. In my experience finding out  
that the compiler does not allow something (yet) stops my ideas dead in  
their tracks.


 But be careful, David, lest you be marked a flamer, and become a  
 Cassanandra-like presence. ;/


I found his post both direct and honest. Less flame like, more knife like.

Regan
Mar 31 2005
parent "David Barrett" <dbarrett quinthar.com> writes:
"Regan Heath" <regan netwin.co.nz> wrote in message 
news:opsoh0jlec23k2f5 nrage.netwin.co.nz...

 But be careful, David, lest you be marked a flamer, and become a 
 Cassanandra-like presence. ;/


 I found his post both direct and honest. Less flame like, more knife like.


Yikes.  I hope I'm neither knife nor flame.  Perhaps instead a cool, 
refreshing breeze... :)
Mar 31 2005
prev sibling next sibling parent reply Stewart Gordon <smjg_1998 yahoo.com> writes:
David Barrett wrote:


 No, I don't believe it's currently 1.0 in that I wouldn't seriously 
 recommend to my boss that we start using it.
 
 I do believe it's ready for personal "fun" projects, but I think too much 
 risk surrounds it before I would recommend building a business atop it:
 
 - I don't know of any significant commercial application successfully 
 written and released in D, and I don't want to be the first.


I guess people haven't really had time to write a commercial-quality 
application in D.  Or maybe they're waiting for the language to 
stabilise a bit before committing themselves to it.

<snip>

 - I perceive a lack of interest among the D developer community.  I see lots 
 of started projects, but few completed ones, and little progress in the 
 meantime.


What do you mean by "completed" exactly?  Every project should be 
continually under development.


 I can only assume *something* turned off these people who gave D 
 a shot, and this increases my perceived risk around D.


Or are you talking about projects that seem to have been abandoned 
without so much as something suitable to be a 1.0 release?


 Note that no amount of compiler improvement will alleviate my fears. 
 Further note that most of my fears could probably be alleviated through 
 improved documetnation and better organization of what's already out there 
 in a scattered form.


<snip>

Same here.

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on 
on the 'group where everyone may benefit.
Mar 31 2005
parent reply "David Barrett" <dbarrett quinthar.com> writes:
"Stewart Gordon" <smjg_1998 yahoo.com> wrote in message 
news:d2gnc1$otv$1 digitaldaemon.com...

 David Barrett wrote:
 <snip>

 - I perceive a lack of interest among the D developer community.  I see 
 lots of started projects, but few completed ones, and little progress in 
 the meantime.



 I can only assume *something* turned off these people who gave D a shot, 
 and this increases my perceived risk around D.


 Or are you talking about projects that seem to have been abandoned without 
 so much as something suitable to be a 1.0 release?


Yes.

Naturally, everything is a work in progress.  But I'm a Win32 programmer, 
and the last time I tried D (8 months ago) I spent more time hacking C 
header files than writing D code .  Anyway, I'd hoped the Core32 library 
would solve my needs, but it's still marked "Alpha", and the last forum post 
was in October of '04.

(Oye, I had totally forgotten about that experience.  Add 
"importing/converting C header files and linking with C libraries" to my 
"must have" list.  I mean, it's great that D can link with them, but so what 
if I can't use the header files?  h2d looks like a great, albeit abandoned 
start, but even the author asserts that the right place to do it is in the D 
compiler itself.)

And a further discouragement was in this thread of the Core32 project:

http://www.dsource.org/forums/viewtopic.php?t=360

"I'm not in the mood to make large changes to parts of the library that 
already work. I signed up to be a caretaker, not an overhauler."

It sounds like Core32 -- an essential component of any D Win32 
application -- has already been put on life support.  Hopefully I'm reading 
too much into an off comment.  But I note that the user whose patch was 
rejected never posted again.

And all languages have a high rate of abandonware.  But it's disheartening 
to look at Dsource and see:

Server Resources:
Mango The Mango Tree - Stable/Production
DSP - Pre-Alpha
CORBA - Alpha
D Server Platform - Pre-Alpha

Development Tools (15)
eclipseD - Alpha
Dig Dug - Alpha
DCoder Visual Studio.NET extensions - Beta
DCC D Compiler Collection - Planning
h2d .h to .d conversion tool  - Alpha
Bindings - Pre-Alpha
ddoc - Planning
Orion - Planning
leds - Beta
DGrammar - Planning
Hydrus - Pre-Alpha
molt - Alpha
Build - Beta
DManager - Alpha
GDB - Alpha

Libraries-GUI (8)
DWT - Pre-Alpha
Sinbad - Pre-Alpha
L8night - Alpha
Derelict - Alpha
unDig - Beta
dcouple - Pre-Alpha
CursesTUI - Pre-Alpha
DUI - Beta

Libraries-System (11)
Docoa - Planning
Concurrent - Alpha
Core32 - Alpha
Deimos - Alpha
swig4d - Alpha
lr-lalr - Pre-Alpha
dragon-stl - Pre-Alpha
sdbo - Alpha
dcurses - Pre-Alpha
dool - Beta
Walnut - Pre-Alpha

Games (5)
Deliria - Planning
Claytek 3D - Alpha
Rylix Rylix - Pre-Alpha
WarBots - Planning
FragBots - Planning

Misc. Apps (1)
dBerBible - Alpha
Mar 31 2005
next sibling parent Derek Parnell <derek psych.ward> writes:
On Thu, 31 Mar 2005 14:20:48 -0800, David Barrett wrote:


[snip]

 Build - Beta


Build is at release level 1.17 and is only 'beta' because there is one
known outstanding issue that I'm awaiting user feedback from. Otherwise it
would be in 'production' status. When that issue is resolved (actually I
have a fix already done, just waiting confirmation) I'll move Build to 2.0
and call it 'production'.

(Note to 'pragma': Did you confirm if Build was inserting two "LIBRARY"
entries into the .DEF file for DLL targets?)

-- 
Derek
Melbourne, Australia
1/04/2005 9:13:24 AM
Mar 31 2005
prev sibling next sibling parent reply J C Calvarese <jcc7 cox.net> writes:
David Barrett wrote:

 "Stewart Gordon" <smjg_1998 yahoo.com> wrote in message 
 news:d2gnc1$otv$1 digitaldaemon.com...
 

David Barrett wrote:
<snip>


- I perceive a lack of interest among the D developer community.  I see 
lots of started projects, but few completed ones, and little progress in 
the meantime.



I can only assume *something* turned off these people who gave D a shot, 
and this increases my perceived risk around D.


Or are you talking about projects that seem to have been abandoned without 
so much as something suitable to be a 1.0 release?


 
 
 Yes.
 
 Naturally, everything is a work in progress.  But I'm a Win32 programmer, 
 and the last time I tried D (8 months ago) I spent more time hacking C 
 header files than writing D code .  Anyway, I'd hoped the Core32 library 
 would solve my needs, but it's still marked "Alpha", and the last forum post 
 was in October of '04.


If you don't like Core32, you might prefer Y Tomino's efforts:
http://hp.vector.co.jp/authors/VA028375/d/windows.h.html

It's a solid port and I think it's more complete than Core32.

 
 (Oye, I had totally forgotten about that experience.  Add 
 "importing/converting C header files and linking with C libraries" to my 
 "must have" list.  I mean, it's great that D can link with them, but so what 
 if I can't use the header files?  h2d looks like a great, albeit abandoned 
 start, but even the author asserts that the right place to do it is in the D 
 compiler itself.)
 
 And a further discouragement was in this thread of the Core32 project:
 
 http://www.dsource.org/forums/viewtopic.php?t=360
 
 "I'm not in the mood to make large changes to parts of the library that 
 already work. I signed up to be a caretaker, not an overhauler."
 
 It sounds like Core32 -- an essential component of any D Win32 
 application -- has already been put on life support.  Hopefully I'm reading 
 too much into an off comment.  But I note that the user whose patch was 
 rejected never posted again.


Was there a reason (or reasons) why the patch was rejected? I believe 
so. I don't think I offended the person who offered the patch, and if I 
did I'm sorry.

The original project leader of Core32 is MIA so I offered to fix it if 
it breaks. The "patch" that you seem to be referring to was a wholesale 
rewrite. I didn't see a problem that was being fixed. I didn't agree 
with the merit of the changes, so I declined making the changes to 
Core32. I was blunt, but I didn't mean to be rude.

I haven't checked to see it if still compiles with the latest DMD, but 
if someone alerts me to a problem I'll try to fix it.


 And all languages have a high rate of abandonware.  But it's disheartening 
 to look at Dsource and see:


I suspect that part of the reason for this is that the newness of D 
attracts people that like to try new things. (Note that I said "try".) 
As D has been maturing there have been some growing pains. It can be 
annoying when you write code for DMD 0.65 and have to patch it for DMD 
0.69 (and then it needs patching again for DMD 0.74). If you're like and 
don't produce code that fast, it's easy to fall off the upgrade 
treadmill. It's discouraging to spend more time

With D's current relative maturity, that's less of a problem, but some 
people from the early days may still be burned out. Personally, I 
haven't been writing much code recently.

And, besides, wxD isn't vaporware and DWT just produced a "Hello, 
World", so the sky's the limit. :)

-- 
jcc7
http://jcc_7.tripod.com/d/
Mar 31 2005
parent Stewart Gordon <smjg_1998 yahoo.com> writes:
J C Calvarese wrote:
<snip>

 If you don't like Core32, you might prefer Y Tomino's efforts:
 http://hp.vector.co.jp/authors/VA028375/d/windows.h.html


None of them cut it.  But I have Tomino's version and I'm kind of 
working on cleaning it up.

<snip>

 I suspect that part of the reason for this is that the newness of D 
 attracts people that like to try new things. (Note that I said "try".) 
 As D has been maturing there have been some growing pains. It can be 
 annoying when you write code for DMD 0.65 and have to patch it for DMD 
 0.69 (and then it needs patching again for DMD 0.74). If you're like and 
 don't produce code that fast, it's easy to fall off the upgrade 
 treadmill. It's discouraging to spend more time


Yes, there is still this issue.  But once 1.0 comes out it'll've 
stabilised a bit.


 With D's current relative maturity, that's less of a problem, but some 
 people from the early days may still be burned out. Personally, I 
 haven't been writing much code recently.
 
 And, besides, wxD isn't vaporware and DWT just produced a "Hello, 
 World", so the sky's the limit. :)


At least SDWF hasn't died down development-wise since it started.

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on 
on the 'group where everyone may benefit.
Apr 01 2005
prev sibling next sibling parent Mike Parker <aldacron71 yahoo.com> writes:
David Barrett wrote:


 And all languages have a high rate of abandonware.  But it's disheartening 
 to look at Dsource and see:
 
 Server Resources:
 Mango The Mango Tree - Stable/Production
 DSP - Pre-Alpha
 CORBA - Alpha
 D Server Platform - Pre-Alpha
 
 Development Tools (15)
 eclipseD - Alpha
 Dig Dug - Alpha
 DCoder Visual Studio.NET extensions - Beta
 DCC D Compiler Collection - Planning
 h2d .h to .d conversion tool  - Alpha
 Bindings - Pre-Alpha
 ddoc - Planning
 Orion - Planning
 leds - Beta
 DGrammar - Planning
 Hydrus - Pre-Alpha
 molt - Alpha
 Build - Beta
 DManager - Alpha
 GDB - Alpha
 
 Libraries-GUI (8)
 DWT - Pre-Alpha
 Sinbad - Pre-Alpha
 L8night - Alpha
 Derelict - Alpha
 unDig - Beta
 dcouple - Pre-Alpha
 CursesTUI - Pre-Alpha
 DUI - Beta
 
 Libraries-System (11)
 Docoa - Planning
 Concurrent - Alpha
 Core32 - Alpha
 Deimos - Alpha
 swig4d - Alpha
 lr-lalr - Pre-Alpha
 dragon-stl - Pre-Alpha
 sdbo - Alpha
 dcurses - Pre-Alpha
 dool - Beta
 Walnut - Pre-Alpha
 
 Games (5)
 Deliria - Planning
 Claytek 3D - Alpha
 Rylix Rylix - Pre-Alpha
 WarBots - Planning
 FragBots - Planning
 
 Misc. Apps (1)
 dBerBible - Alpha
 
 


What are you getting at? A number of those projects are active. Some are 
updated periodically. Many of them never got off the ground. Just check 
the repositories to see how many made an initial commit (which is 
usually revision 3 or 4) then never did anything again. That in itself 
says nothing tome about the quality of D, but says volumes about the 
motivation of the project maintainer. And motivation be affected by many 
factors (lack of time, laziness, lack of commitment, whatever), which 
*might* or *might not* include dissatisfaction with the language. This 
is par for the course on any free project hosting site. Care to count 
the number of abandoned projects on Sourceforge?

Or are you getting at how many projects are marked as 
planning/pre-alpha/alpha? For the large majority of them it's fairly 
accurate. Where's the problem there?

I just don't see how the number of abandoned/incomplete projects could 
be directly indicative of the failure of D to grab people and hold on to 
them. If a significant number of those projects were abandoned after a 
significant amount of effort had been put in to them (i.e. the 
maintainer demonstrated commitment, dedication, and strong potential to 
complete the project), then questioning if the language had something to 
do with it would make sense.
Mar 31 2005
prev sibling parent "Walter" <newshound digitalmars.com> writes:
"David Barrett" <dbarrett quinthar.com> wrote in message
news:d2ht41$2265$1 digitaldaemon.com...

 And all languages have a high rate of abandonware.  But it's disheartening
 to look at Dsource and see:


I don't think it's disheartening at all. It takes a lot of commitment, time
and work to take a project from initial prototype to a professional product,
far more than most realize. D offers big productivity gains over C++, but it
isn't a miracle language <g>, and the need for putting in the time to do a
polished product isn't going to go away.

What I see is a lot of activity going on in D that is very heartening. This
also shows in the big increase in bug reports in the D.bugs newsgroup. That
means that lots of people are using it, and are pushing it hard.

I'm very familiar with how much work it takes to do a professional product,
which is why I only very rarely start new ones. I keep upgrading and
extending existing ones, heck, parts of DMD (the back end) were written back
in 1982 <g>. I believe that DMC++ is the oldest compiler in existence still
actively maintained. The Empire game started out in 1976 in BASIC, went to
FORTRAN, then PDP-11 assembler, then C, then C++, and now there's a D
version out. The amusing thing is there's still some of the old BASIC DNA in
the code.
Mar 31 2005
prev sibling next sibling parent "Lynn Allan" <l_d_allan adelphia.net> writes:
 - I perceive a lack of interest among the D developer community.  I


see lots

 of started projects, but few completed ones, and little progress in


the

 meantime.  I can only assume *something* turned off these people who


gave D

 a shot, and this increases my perceived risk around D.


Good point.  Although not at all due to lack of interest.

I've the instigator/developer for one dSource project, a completely
separate SourceForge D project, and another completely separate
project to which I'm trying to recruit people to help work on "real
soon now"

... all more or less on the backburner waiting D maturity

... and I consider myself a D zealot who eagerly looks forward to
better debugger support and gui lib(s) .. which are getting better all
the time.

... and could probably benefit from some focus <g>

I've been aware of D for some time ... the way it seems to work out is
that several times a year, I pick up where I left off 3-4 months
previously, get frustrated, and put a "to-do" reminder in my calendar
to check back in 3-4 months.

Getting about due, I suppose. Probably overdue. I'm noticing some very
positive info about gui libs
Apr 01 2005
prev sibling parent "Lynn Allan" <l_d_allan adelphia.net> writes:
 No, I don't believe it's currently 1.0 in that I
 wouldn't seriously
 recommend to my boss that we start using it.


At first reading, this seems reasonable ...

but when I try to think of ANY version 1.0 development tool that I
could have "seriously recommended to my boss", I can't come up with
any. Nada.

Version numbering seems more conservative in the Linux world, but the
"rule of thumb" was to NEVER have an important project be dependent on
version 1.0 software. That had the potential to be a "career
shortening move".

So perhaps that's not a valid criteria?

To me, version 1.0 declares, "Early Adapters: Proceed With Caution"

FWIW: I vote to release as 1.0 RC-1 (release candidate)
Apr 01 2005
prev sibling next sibling parent Stewart Gordon <smjg_1998 yahoo.com> writes:
clayasaurus wrote:

 The poll is simple.
 
 Do you think the D compiler is ready to be called 1.0?


<snip>

Definitely not.

http://www.wikiservice.at/wiki4d/wiki.cgi?PendingPeeves

pretty much speaks for itself here.

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on 
on the 'group where everyone may benefit.
Mar 31 2005
prev sibling next sibling parent Georg Wrede <georg.wrede nospam.org> writes:
clayasaurus wrote:

 The poll is simple.
 
 Do you think the D compiler is ready to be called 1.0?


No.

Before we do, some of the DMD bugs should be fixed. Then the library 
should be fixed and polished.

AND THEN the documentation totally rewritten, or better, another set of 
documentation should be written to complement the current one.

When that's done, we may announce D 1.0.

----------------------------------------

I wish:     (and read my lips here!)

The compiler bugs would be sorted in two bins, right now: those we try 
(ehh, Walter tries) to fix before 1.0, and those that WILL NOT GET 
FIXED!!!!!

Every piece of software has bugs known and unknown. That's just life.

 From my own experience, I tend to get this obsession when something is 
"almost perfect": I can't quit polishing it.

Fix the first bin, and freeze DMD. From that day on, Walter could be 
100% on the lib. 5 weeks of that, and it would be good enough. Freeze 
that too, and get the docs polished.

That done, we have a killer 1.0!

(My guess would be at end of Q2.)

Man, even the Linux kernel has bugs. Accepting some for D and knowing 
when to freeze the compiler, the lib, and the docs -- we just have to. 
And in that order.

The bugs left in 1.0 should be in the documentation, with proper 
explanations on when they occur, suggested workarounds -- and, _most_ 
importantly, some explanation for _each_ on why we chose to leave that 
one unfixed for 1.0.  !!

This kind of candor will kick the critics [you know where]!
And it will really make D look professional!

---

Oh Bob, sometimes I really wish I were Bob!!!  ;-(
Mar 31 2005
prev sibling next sibling parent reply "Ben Hinkle" <bhinkle mathworks.com> writes:
"clayasaurus" <clayasaurus gmail.com> wrote in message 
news:d2ftk0$2ppq$1 digitaldaemon.com...

 The poll is simple.

 Do you think the D compiler is ready to be called 1.0?


compiler - yes
phobos - no*
web site - no**

(*) the only complaints I have about phobos is the AA behavior and some 
needed Exception/Error refactoring
(**) the community stuff needs updating. A vesion 1.0 will get new people 
and the community links will be vital for these folks to get oriented. Also 
a more complete spec would help. For example "final" is an attribute but the 
only mention of what it does is in the functions.html for decorating a 
function member. No other use is mentioned (is it legal or undefined...?)
Mar 31 2005
next sibling parent reply "Walter" <newshound digitalmars.com> writes:
"Ben Hinkle" <bhinkle mathworks.com> wrote in message
news:d2h6a0$193m$1 digitaldaemon.com...

 (**) the community stuff needs updating. A vesion 1.0 will get new people
 and the community links will be vital for these folks to get oriented.


Also

 a more complete spec would help. For example "final" is an attribute but


the

 only mention of what it does is in the functions.html for decorating a
 function member. No other use is mentioned (is it legal or undefined...?)


It's a big help to me when people use the wiki links at the bottom of each
page to provide specific feedback. When I update a page, it's easy to click
on the wiki link and incorporate the suggestions.
Mar 31 2005
parent reply "Ben Hinkle" <bhinkle mathworks.com> writes:
"Walter" <newshound digitalmars.com> wrote in message 
news:d2hbs5$1fb6$1 digitaldaemon.com...

 "Ben Hinkle" <bhinkle mathworks.com> wrote in message
 news:d2h6a0$193m$1 digitaldaemon.com...

 (**) the community stuff needs updating. A vesion 1.0 will get new people
 and the community links will be vital for these folks to get oriented.


 Also

 a more complete spec would help. For example "final" is an attribute but


 the

 only mention of what it does is in the functions.html for decorating a
 function member. No other use is mentioned (is it legal or undefined...?)


 It's a big help to me when people use the wiki links at the bottom of each
 page to provide specific feedback. When I update a page, it's easy to 
 click
 on the wiki link and incorporate the suggestions.


I went to modify the Attributes page and I'm confused about what each 
section of the wiki page is for. There's "More 
Information","Examples","Message","Amendments" and "Links". Examples and 
Links I can understand (sortof) but the other three seem the same to me. I 
put my feedback under "Message" but that's because it says "Put your 
comments here". The others are blank and I have no idea who would fill those 
in.

Other questions come to mind:
Is there a summary page?
How often do these wiki's get looked at and what happens to the suggestions?
Does anyone go through the wiki and make sure it makes sense and move stuff 
around if not?

-Ben
Mar 31 2005
parent "Walter" <newshound digitalmars.com> writes:
"Ben Hinkle" <bhinkle mathworks.com> wrote in message
news:d2hnhl$1si1$1 digitaldaemon.com...

 I went to modify the Attributes page and I'm confused about what each
 section of the wiki page is for. There's "More
 Information","Examples","Message","Amendments" and "Links". Examples and
 Links I can understand (sortof) but the other three seem the same to me. I
 put my feedback under "Message" but that's because it says "Put your
 comments here". The others are blank and I have no idea who would fill


those

 in.


You put it in the right place.


 Other questions come to mind:
 Is there a summary page?


No.


 How often do these wiki's get looked at and what happens to the


suggestions?

I look at the wiki suggestions when I update the corresponding web page.


 Does anyone go through the wiki and make sure it makes sense and move


stuff

 around if not?


Anyone who wants to can. Me, I'll remove suggestions from the wiki once they
are incorporated. They'll still remain in the wiki history, though.
Mar 31 2005
prev sibling parent reply "Charlie" <charles jwavro.com> writes:
Agreed 100% here.  I've pushed for Ares in the past, ( though admittedly i
still haven't _tried_ it , but I will! ) this time I want to sit back and
see what happens w/respect to phobos.

Another thing, does anyone know who owns :
http://www.minddrome.com/produtos/d/ : , when I search for D Programming
Language on google this is the second result , and all the information is
grossly out of date.

Charlie


"Ben Hinkle" <bhinkle mathworks.com> wrote in message
news:d2h6a0$193m$1 digitaldaemon.com...

 "clayasaurus" <clayasaurus gmail.com> wrote in message
 news:d2ftk0$2ppq$1 digitaldaemon.com...

 The poll is simple.

 Do you think the D compiler is ready to be called 1.0?


 compiler - yes
 phobos - no*
 web site - no**

 (*) the only complaints I have about phobos is the AA behavior and some
 needed Exception/Error refactoring
 (**) the community stuff needs updating. A vesion 1.0 will get new people
 and the community links will be vital for these folks to get oriented.


Also

 a more complete spec would help. For example "final" is an attribute but


the

 only mention of what it does is in the functions.html for decorating a
 function member. No other use is mentioned (is it legal or undefined...?)
Mar 31 2005
parent Thomas Kuehne <thomas-dloop kuehne.thisisspam.cn> writes:
Charlie schrieb am Thu, 31 Mar 2005 12:10:36 -0600:

 Another thing, does anyone know who owns :
 http://www.minddrome.com/produtos/d/ : , when I search for D Programming
 Language on google this is the second result , and all the information is
 grossly out of date.


Daniel Yukio Yokomiso
Mar 31 2005
prev sibling parent John Demme <me teqdruid.com> writes:
No.  Bug fixes, bug fixes, bug fixes.

-We need a feature and syntax free now.

-We need DMD's Linux support to not be sub-par compared to Windows
support.
	-Proper DWARF2 output
	-Shared Object support
	-Probably other stuff

-Phobos sucks... enough said.
	-A minimal DTL should be included with Phobos.  Despite D having cool
array stuff, some minimal wrapper classes  (or templates) would be nice.
	-Completely unorganized, an incohesive

I haven't run in to any show stopping bugs such that I can't get my code
to work, but without _basic_ debugging support, some of my bugs are so
hard to find, that I don't bother.  This is a *necessary* 1.0 feature.

John Demme


On Wed, 2005-03-30 at 23:17 -0500, clayasaurus wrote:

 The poll is simple.
 
 Do you think the D compiler is ready to be called 1.0?
 
 Say yes if D is working great for you and explain what you have done 
 with it.
 
 Say no if D is unusable in its current state, or if there is something 
 about the compiler that really bothers you that should be fixed before 1.0.
 
 I posted this poll out of a general curiousity to see how many out there 
 think D is ready and if not why.
 
 ---------------------------------------------------------------------------
 
 Yes.
 
     My honest opinion, from the little bit I've used D for, is that the 
 D compiler is ready for 1.0. Granted, I'm a little college student who 
 is just trying to learn game programming (and I'm not too far into all 
 the fancy algorithms like the big shots are), so take my opinion with a 
 grain of salt.
 
     I've never run into any show stopper bugs recently, for a while, and 
 I've used some basic templates, modules, bindings, libs, and the build 
 utility (a BIG selling point for D IMO, especially if you get a 
 graphical frontend for it, will make the newbs very happy, and make 
 makefiles look like a waste of time, allow you to concentrate more on 
 programming, which is D's purpose, increase programmer productivity).
 
     I think D is really great and it can definitely holds its own 
 against other languages. I'm sure compiler bugs could be fixed for the 
 next three years or so (bop the gofer, fix one another pops up).
 
    The general acceptance of D can only happen when 1.0 emerges, 
 otherwise it is just a beta language, and it is percieved as a moving 
 target, so writing large amounts of code for it can seem dangerous.
 
 And for those who need more features, be patient and wait for 2.0?
 
 --------------------------------------------------------------------------
 
 Just trying to generate some discussion, and see other peoples views. :-)
 
 --------------------------------------------------------------------------
 
 Also, I know some do not want to discriminate between compiler/lib, but 
 IMO it would be hard to rollout a solid lib if the compiler is under 
 construction. I think phobos is solid as it is (maybe i'm ignorant), and 
 that any great/radical improvement will have to come from a great open 
 source lib, like ares or mango, or a behind the scenes redesign by the 
 top-notch library writers floating around here.
Mar 31 2005