www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Is it time for D1 to die of natural causes?

reply Justin Johansson <no spam.com> writes:
Now that Andrei's much anticipated publication of TDPL is out, is it 
time that D1 should now perish?

My personal feeling is that by cremating D1, time and effort can then be 
better expended and focused on solidifying D2.

Cheers
Justin Johansson
Jun 23 2010
next sibling parent "Vladimir Panteleev" <vladimir thecybershadow.net> writes:
On Wed, 23 Jun 2010 16:29:19 +0300, Justin Johansson <no spam.com> wrote:

 My personal feeling is that by cremating D1, time and effort can then be  
 better expended and focused on solidifying D2.
As in, stop updating DMD1 with bugfixes? Is porting DMD2 fixes to DMD1 that great an effort to warrant this? -- Best regards, Vladimir mailto:vladimir thecybershadow.net
Jun 23 2010
prev sibling next sibling parent reply Don <nospam nospam.com> writes:
Justin Johansson wrote:
 Now that Andrei's much anticipated publication of TDPL is out, is it 
 time that D1 should now perish?
 
 My personal feeling is that by cremating D1, time and effort can then be 
 better expended and focused on solidifying D2.
 
 Cheers
 Justin Johansson
I estimate that D1 has at least a year before it becomes "the legacy version". In any case, there's not much effort going specificially into D1 at the expense of D2. But it's probably time to change statements about D2 being an unstable beta version, and we should start to encourage newcomers to use D2 rather than D1. In particular, we should make sure that tutorials and documentation should be updated to D2, or else clearly marked as applying only to D1.
Jun 23 2010
parent reply Justin Johansson <no spam.com> writes:
Don wrote:
 Justin Johansson wrote:
 Now that Andrei's much anticipated publication of TDPL is out, is it 
 time that D1 should now perish?

 My personal feeling is that by cremating D1, time and effort can then 
 be better expended and focused on solidifying D2.

 Cheers
 Justin Johansson
I estimate that D1 has at least a year before it becomes "the legacy version". In any case, there's not much effort going specificially into D1 at the expense of D2.
 But it's probably time to change statements about D2 being an unstable 
 beta version, and we should start to encourage newcomers to use D2 
 rather than D1.
Agree that would be a positive step
 In particular, we should make sure that tutorials and documentation 
 should be updated to D2, or else clearly marked as applying only to D1.
Well should the current version be called just D rather than D2 and simply pretend that D1 does not exist?
Jun 23 2010
next sibling parent Trass3r <un known.com> writes:
 In particular, we should make sure that tutorials and documentation  
 should be updated to D2, or else clearly marked as applying only to D1.
Well should the current version be called just D rather than D2 and simply pretend that D1 does not exist?
It should be turned around. D2 becomes "normal" D and D1 becomes the old, "soon"-to-be-deprecated version. This also means the documentation on digitalmars.com should directly point to the D2 versions and D1 would be extra.
Jun 23 2010
prev sibling parent reply Lutger <lutger.blijdestijn gmail.com> writes:
Justin Johansson wrote:

 Don wrote:
 Justin Johansson wrote:
 Now that Andrei's much anticipated publication of TDPL is out, is it
 time that D1 should now perish?

 My personal feeling is that by cremating D1, time and effort can then
 be better expended and focused on solidifying D2.

 Cheers
 Justin Johansson
I estimate that D1 has at least a year before it becomes "the legacy version". In any case, there's not much effort going specificially into D1 at the expense of D2.
 But it's probably time to change statements about D2 being an unstable
 beta version, and we should start to encourage newcomers to use D2
 rather than D1.
Agree that would be a positive step
 In particular, we should make sure that tutorials and documentation
 should be updated to D2, or else clearly marked as applying only to D1.
Well should the current version be called just D rather than D2 and simply pretend that D1 does not exist?
Not too be lame, but what do you think is the benefit of that? Do you think D1 harms D2? At digitalmars.com/d, D2 is currently being called 'the next generation under development' and D1 is designated as 'stable' Personally I would say D must always mean the current stable release of the language, which should soon have version number 2 hopefully :) Last week I was on a RoR course, typing something into irb version 1.9.something. The teacher demonstrated it did not work, but he was on 1.8.something. He also remarked that Ruby backwards compatibility can break even between minor versions, it is a source of annoyance but clearly not of dread. So there is some precedent here. Related to this topic: it would be nice to know how the language will move forward from this point, specifically wrt backwards compatibility. From reading the TDPL I get the feeling the language is somewhat more stable and focus has shifted, but not set in stone at all.
Jun 23 2010
next sibling parent bearophile <bearophileHUGS lycos.com> writes:
Lutger:
 From reading 
 the TDPL I get the feeling the language is somewhat more stable and focus has  
 shifted, but not set in stone at all.
Future evolution of D2 will be more backward compatible (breaking backward compatibility only in small situations), but the only languages set in stone are the dead ones. So some change and evolution will be present, also because no very large D2 systems have being written yet, so it's not battle-tested yet, so some changes can turn up being useful only lot of people start using it for real. In Bugzilla there are some bugs already known for example about array that contain objects that need to be addressed before D2 can be considered stable enough for serious OOP use. Fixing this important bug can require some small changes in D2. Bye, bearophile
Jun 23 2010
prev sibling parent Bernard Helyer <b.helyer gmail.com> writes:
On Wed, 23 Jun 2010 19:53:33 +0200, Lutger wrote:

 Do you
 think D1 harms D2?
I does in so far as people making libraries and wrappers don't support D2. I think getting Tango onto D2 would be a very positive step. Unfortunately, I don't think the maintainers are interested at all in this, the users of Tango wouldn't be, either, I'd imagine. So movement in that area has to come from D2 users, but I don't know from whom. As for myself, I have no interest whatsoever in porting a large library I do not use.
Jun 24 2010
prev sibling next sibling parent SiegeLord <none none.com> writes:
Justin Johansson Wrote:

 Now that Andrei's much anticipated publication of TDPL is out, is it 
 time that D1 should now perish?
 
 My personal feeling is that by cremating D1, time and effort can then be 
 better expended and focused on solidifying D2.
 
 Cheers
 Justin Johansson
To whose benefit? You are advocating nerfing a good, stable, usable language that has an excellent (de facto) standard library, excellent compilers (on Linux, at least), 64 bit support, tons of libraries already written for it etc, and replacing it with a buggy language, buggy standard library, buggy linker, no 64 bit support and only a handful of libraries to support developing in it. How do users benefit from your suggestion? I have scientific computing that must be done *right now* as well as games I want to develop *right now* and I don't want suffer the performance losses that come from DMD's substandard optimizer (at least, relative to LLVM), as well as Phobos 2's abstractions (although a better compiler might solve those issues) and not using 64 bits. I just tried installing dmd2 on my 64 bit Linux box, and while it installed, I couldn't compile "Hello World", since I have no 32 bit versions of 32 bit pthread libraries installed. I am not going to install 32 bit versions of every library I use in my C++ development to use them in D2, that is simply unreasonable: I will use D1 instead. And I'm not worried about investing into D1 and then have my work be obsolete: I am fully planning on porting my things to D2 *when* it becomes better than or even as good as D1 for my purposes. It's easy to do... just don't use too many deprecated features. D2 will only become the dominant version if it is actively better than D1, not by making D1 "disappear." Users are not going to encounter those issues and start spending their "time and effort" improving the standard library and compiler tools, it isn't worth their time: they will go use a different language instead. For D2's future's sake, that different language should be D1. This reminds me of the Vista debacle :P I'm using XP on my other computer too, hehe... so removing support will not kill D1 anyway ;) -SiegeLord
Jun 23 2010
prev sibling next sibling parent Jonathan M Davis <jmdavisProg gmail.com> writes:
Justin Johansson wrote:

 Now that Andrei's much anticipated publication of TDPL is out, is it
 time that D1 should now perish?
 
 My personal feeling is that by cremating D1, time and effort can then be
 better expended and focused on solidifying D2.
 
 Cheers
 Justin Johansson
I don't think that Walter has any intention of ceasing support for D1 any time soon. Even if D2 were perfectly stable and better than D1 in every respect, I don't think that Walter would be stopping support for D1 just yet. And as far as I know, that's restricted to bug fixes of dmd for D1 - at least some of which come up when working on D2 or are at least bugs in both D1 and D2. If you want efforts towards replacing D1 with D2, then we need to be doing work to replace references to D which are really D1 (be it tutorials or the online docs or whatever) with references to D1 and references to D2 with references to D - thereby treating D2 as D and D1 as D1. Even though I have no interest in D1, I see no reason to get rid of D1. It works and people use it. What we need to do is work on clearing away the confusion of what D is and make it so that stuff referring to D is referring to D2 or is clear that it's referring to D1. Leaving users of D1 in the cold isn't going to help anyone. - Jonathan M Davis
Jun 23 2010
prev sibling next sibling parent Stewart Gordon <smjg_1998 yahoo.com> writes:
Justin Johansson wrote:
 Now that Andrei's much anticipated publication of TDPL is out, is it 
 time that D1 should now perish?
 
 My personal feeling is that by cremating D1, time and effort can then be 
 better expended and focused on solidifying D2.
It won't, because the outstanding D1 spec problems affect D2 as well. There are many people, myself included, who are still using D1 and still waiting for the language to be finished. Stewart.
Jun 23 2010
prev sibling next sibling parent reply "Nick Sabalausky" <a a.a> writes:
"Justin Johansson" <no spam.com> wrote in message 
news:hvt27i$hd4$1 digitalmars.com...
 Now that Andrei's much anticipated publication of TDPL is out, is it time 
 that D1 should now perish?

 My personal feeling is that by cremating D1, time and effort can then be 
 better expended and focused on solidifying D2.
Other people brought up other issues with doing that, but I'll add this: D1 users are primarily Tango users. And until Tango goes D2, those user's migration paths to D2 would be fairly large (certainly not impossible, but notably harder than it would really need to be).
Jun 23 2010
parent reply dsimcha <dsimcha yahoo.com> writes:
== Quote from Nick Sabalausky (a a.a)'s article
 "Justin Johansson" <no spam.com> wrote in message
 news:hvt27i$hd4$1 digitalmars.com...
 Now that Andrei's much anticipated publication of TDPL is out, is it time
 that D1 should now perish?

 My personal feeling is that by cremating D1, time and effort can then be
 better expended and focused on solidifying D2.
Other people brought up other issues with doing that, but I'll add this: D1 users are primarily Tango users. And until Tango goes D2, those user's migration paths to D2 would be fairly large (certainly not impossible, but notably harder than it would really need to be).
Sometimes I feel like D1/Tango and D2/Phobos really should evolve as completely separate languages. D1/Tango feels very much like Java++, while D2/Phobos feels a lot more like (C++)++. If there's enough manpower for it, it'd be very interesting to see how this would play out.
Jun 23 2010
next sibling parent "Mike James" <foo bar.com> writes:
"dsimcha" <dsimcha yahoo.com> wrote in message 
news:hvthsj$253t$1 digitalmars.com...
 == Quote from Nick Sabalausky (a a.a)'s article
 "Justin Johansson" <no spam.com> wrote in message
 news:hvt27i$hd4$1 digitalmars.com...
 Now that Andrei's much anticipated publication of TDPL is out, is it 
 time
 that D1 should now perish?

 My personal feeling is that by cremating D1, time and effort can then 
 be
 better expended and focused on solidifying D2.
Other people brought up other issues with doing that, but I'll add this: D1 users are primarily Tango users. And until Tango goes D2, those user's migration paths to D2 would be fairly large (certainly not impossible, but notably harder than it would really need to be).
Sometimes I feel like D1/Tango and D2/Phobos really should evolve as completely separate languages. D1/Tango feels very much like Java++, while D2/Phobos feels a lot more like (C++)++. If there's enough manpower for it, it'd be very interesting to see how this would play out.
That could be useful to have D1 associated with Tango and D2 associated with Phobos. It would stop the arguements that there are 2 competing "standard" libraries for D. -=mike=-
Jun 23 2010
prev sibling parent reply Eric Poggel <dnewsgroup yage3d.net> writes:
On 6/23/2010 1:56 PM, dsimcha wrote:
 == Quote from Nick Sabalausky (a a.a)'s article
 "Justin Johansson"<no spam.com>  wrote in message
 news:hvt27i$hd4$1 digitalmars.com...
 Now that Andrei's much anticipated publication of TDPL is out, is it time
 that D1 should now perish?

 My personal feeling is that by cremating D1, time and effort can then be
 better expended and focused on solidifying D2.
Other people brought up other issues with doing that, but I'll add this: D1 users are primarily Tango users. And until Tango goes D2, those user's migration paths to D2 would be fairly large (certainly not impossible, but notably harder than it would really need to be).
Sometimes I feel like D1/Tango and D2/Phobos really should evolve as completely separate languages. D1/Tango feels very much like Java++, while D2/Phobos feels a lot more like (C++)++. If there's enough manpower for it, it'd be very interesting to see how this would play out.
I think this type of fragmentation could only hurt things--and has already, except that phobos has become more open. With fragmentation you have less effort for each. One complete product is far better than two half-implemented ones.
Jun 23 2010
parent Jesse Phillips <jessekphillips+D gmail.com> writes:
Eric Poggel Wrote:

 I think this type of fragmentation could only hurt things--and has 
 already, except that phobos has become more open.
 
 With fragmentation you have less effort for each.  One complete product 
 is far better than two half-implemented ones.
The thing is that people will work on what they want no matter what the "community" says. If they are interest in developing and improving D1+Tango they may have no interest in D2+Phobos (even if D1+Tango didn't exist) I think it would be interesting to see D1 continue to progress (release of D1.1?) But since it hasn't caught on I can't imagine very many new users would choose it now that D2 is entering maintenance. Personally I really like the new Phobos and wouldn't want to go back to Tango.
Jun 23 2010
prev sibling parent Bane <branimir.milosavljevic gmail.com> writes:
SiegeLord Wrote:

 Justin Johansson Wrote:
 
 Now that Andrei's much anticipated publication of TDPL is out, is it 
 time that D1 should now perish?
 
 My personal feeling is that by cremating D1, time and effort can then be 
 better expended and focused on solidifying D2.
 
 Cheers
 Justin Johansson
To whose benefit? You are advocating nerfing a good, stable, usable language that has an excellent (de facto) standard library, excellent compilers (on Linux, at least), 64 bit support, tons of libraries already written for it etc, and replacing it with a buggy language, buggy standard library, buggy linker, no 64 bit support and only a handful of libraries to support developing in it. How do users benefit from your suggestion? I have scientific computing that must be done *right now* as well as games I want to develop *right now* and I don't want suffer the performance losses that come from DMD's substandard optimizer (at least, relative to LLVM), as well as Phobos 2's abstractions (although a better compiler might solve those issues) and not using 64 bits. I just tried installing dmd2 on my 64 bit Linux box, and while it installed, I couldn't compile "Hello World", since I have no 32 bit versions of 32 bit pthread libraries installed. I am not going to install 32 bit versions of every library I use in my C++ development to use them in D2, that is simply unreasonable: I will use D1 instead. And I'm not worried about investing into D1 and then have my work be obsolete: I am fully planning on porting my things to D2 *when* it becomes better than or even as good as D1 for my purposes. It's easy to do... just don't use too many deprecated features. D2 will only become the dominant version if it is actively better than D1, not by making D1 "disappear." Users are not going to encounter those issues and start spending their "time and effort" improving the standard library and compiler tools, it isn't worth their time: they will go use a different language instead. For D2's future's sake, that different language should be D1.
Excellent points. I do believe, at the moment, D1 is used for real work, not D2. So idea of shooting D1 so half finished D2 could 'gain momentum' is idiotic one.
Jun 24 2010