Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 13 Jan 2014, at 04:04, Niels Grewe wrote: > There’s no tracing, there’s no attempt made at detecting cyclic references, > so you still have to think carefully about memory management. I think it's slightly more subtle than that. With ARC, you have to care about object ownership, not about memory management. You get memory management for free, but only if you get object ownership right. This isn't too different from tracing schemes, where you still have to think about ensuring that references to an object don't outlast their last use if you want to not have to think about memory management. David ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Am 13.01.2014 um 11:29 schrieb Wolfgang Lux : > Niels Grewe wrote: > >>> Am 12.01.2014 um 00:06 schrieb "Riccardo Mottola" : >>> >>> ARC instead is more a "taste". It is a new addition in the GC discussion. I >>> personally prefer ref-counting. >> >> This seems to be a common misconception: ARC *is* reference counting; it has >> very little to do with garbage collection. > > Sorry for nitpicking, Niels. Don’t be :-) > ARC is indeed reference counting, but reference counting *is* one way of > performing garbage collection :-) Very true, though reference counting alone has certain deficiencies as a garbage collection mechanism... So let’s say my point is that ARC doesn’t turn Objective-C into a language where you don’t have to care about memory management (which, I assume, is what people commonly understand when they hear the phrase ‘a garbage-collected language’).There’s no tracing, there’s no attempt made at detecting cyclic references, so you still have to think carefully about memory management. Cheers, Niels ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Niels Grewe wrote: >> Am 12.01.2014 um 00:06 schrieb "Riccardo Mottola" : >> >> ARC instead is more a "taste". It is a new addition in the GC discussion. I >> personally prefer ref-counting. > > This seems to be a common misconception: ARC *is* reference counting; it has > very little to do with garbage collection. Sorry for nitpicking, Niels. ARC is indeed reference counting, but reference counting *is* one way of performing garbage collection :-) Wolfgang ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
> Am 12.01.2014 um 00:06 schrieb "Riccardo Mottola" : > > ARC instead is more a "taste". It is a new addition in the GC discussion. I > personally prefer ref-counting. This seems to be a common misconception: ARC *is* reference counting; it has very little to do with garbage collection. I also think the new ownership model fits the overall language design pretty well: Just as you distinguish between method calls (object world) and function calls (C world) in traditional Objective-C, ARC-support enables you to explicitly distinguish between object references and references to untyped blobs of memory, which (coincidentally) allows the compiler to do most of the reference counting for you. Cheers, Niels ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Riccardo, They may be crap to you, but they are in common usage and this deserve our support. The attitude you display towards them is at once non-productive and not conducive to attracting developers. They are your opinions, not facts. You forget when you say things such as the forgoing that the entire point of a programming languages is to build programs as effectively and efficiently as possible. Calling developers lazy because the want to use ARC or other features of objc 2.0 is like calling C programmers lazy because they don't want to use assembler. O_o To the point Ivan was discussing objc 2.0 is the most recognized name which we can use, but it's important to remember to say "the objective c 2.0 language" since "objective-c" itself is trademarked. :/ Thanks, GC On Saturday, January 11, 2014, Riccardo Mottola wrote: > Hi, > > Ivan Vučica wrote: > >> I definitely wouldn't go with anything like Objective-C+ARC since I, for >> one, don't think ARC is nearly as an important addition to the language as >> @synthesize. And five years from now, any arguments against naming it >> relative to Objective-C 2.0 will stand against naming it Objective-C+ARC >> or >> similar. >> > well, I think it is poinlessin arguing in what is more important and what > not. > To me, they are all crap. The new language additions are dirty, have a > terrible syntax and are appeal to lazy programmers. > ARC instead is more a "taste". It is a new addition in the GC discussion. > I personally prefer ref-counting. > > The point for me is making a clear statement about which runtime you can > use and which features you get, so that somebody porting Apple code knows > how much is supported, which features he can use with which compiler mix, > to estimate, for example, the porting effort. > > This can't be written in stone. You don't know what Apple will invent to > appeal its lazy iOS developers in the future, if and what Objective-C 2.1 > or 3.0 will be. Workstations aren't relevant anyway today... > > > Riccardo > > ___ > Gnustep-dev mailing list > Gnustep-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/gnustep-dev > -- Gregory Casamento Open Logic Corporation, Principal Consultant yahoo/skype: greg_casamento, aol: gjcasa (240)274-9630 (Cell) http://www.gnustep.org http://heronsperch.blogspot.com ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Hi, Ivan Vučica wrote: I definitely wouldn't go with anything like Objective-C+ARC since I, for one, don't think ARC is nearly as an important addition to the language as @synthesize. And five years from now, any arguments against naming it relative to Objective-C 2.0 will stand against naming it Objective-C+ARC or similar. well, I think it is poinlessin arguing in what is more important and what not. To me, they are all crap. The new language additions are dirty, have a terrible syntax and are appeal to lazy programmers. ARC instead is more a "taste". It is a new addition in the GC discussion. I personally prefer ref-counting. The point for me is making a clear statement about which runtime you can use and which features you get, so that somebody porting Apple code knows how much is supported, which features he can use with which compiler mix, to estimate, for example, the porting effort. This can't be written in stone. You don't know what Apple will invent to appeal its lazy iOS developers in the future, if and what Objective-C 2.1 or 3.0 will be. Workstations aren't relevant anyway today... Riccardo ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
"Sebastian Reitenbach" wrote: > I don't know, some unambigous matrix showing what features are available with > a given compiler/runtime combination. Yes please. Perhaps like https://developer.apple.com/library/ios/releasenotes/ObjectiveC/ObjCAvailabilityIndex/index.html Ross. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Wednesday, January 8, 2014 14:47 CET, Stefan Bidi wrote: > On Wed, Jan 8, 2014 at 3:14 AM, Riccardo Mottola > wrote: > > > Hi, > > I would not like to do that, we remain "runtime neutral", the runtime is a > > dependency. You can use GCC with its GNU runtime, if you use Clang you must > > install libobjc2, but you can (or at least, could, I did that a couple of > > months ago) to use GCC+libobjc2. > > > > The point I was trying to make is that some of the features GNUstep > implements are highly dependent on the runtime. Have 2 competing runtimes > with slightly incompatible implements of the Objective-C language might be > confusing. If both implementations are going to be supported, we need to > be very explicit as to which features we support with one and the other. > Stating "GNUstep supports all modern Objective-C features, including ARC > and blocks" is only half true because we only support those features if the > correct runtime and compiler are in use. Not to mention, some of these > features are only supported by specific versions of the compilers. The > current Debian stable, for example, includes clang 3.0 only, and if I > remember correctly, this version doesn't support certain runtime features. > A few years back I also temporarily switch to Slackware and it still > doesn't officially support clang/llvm. With regard to the runtime, we should recommend to use clang/libobjc2, for people that want to use those shiny new features. For people that for whatever reason cannot or want not use clang/libobjc2, or have too old versions, it should be stated that GNUstep still can run/work, but it will not support all the features. I don't know, some unambigous matrix showing what features are available with a given compiler/runtime combination. Sebastian > > I know a significant number of GNUstep developers use and actively support > the BSDs (I develop corebase on a Debian Testing machine, and boot into > FreeBSD to test), however, you still need to admit that GNU/Linux is the > most used FOSS OS on desktops. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Wed, Jan 8, 2014 at 3:14 AM, Riccardo Mottola wrote: > Hi, > I would not like to do that, we remain "runtime neutral", the runtime is a > dependency. You can use GCC with its GNU runtime, if you use Clang you must > install libobjc2, but you can (or at least, could, I did that a couple of > months ago) to use GCC+libobjc2. > The point I was trying to make is that some of the features GNUstep implements are highly dependent on the runtime. Have 2 competing runtimes with slightly incompatible implements of the Objective-C language might be confusing. If both implementations are going to be supported, we need to be very explicit as to which features we support with one and the other. Stating "GNUstep supports all modern Objective-C features, including ARC and blocks" is only half true because we only support those features if the correct runtime and compiler are in use. Not to mention, some of these features are only supported by specific versions of the compilers. The current Debian stable, for example, includes clang 3.0 only, and if I remember correctly, this version doesn't support certain runtime features. A few years back I also temporarily switch to Slackware and it still doesn't officially support clang/llvm. I know a significant number of GNUstep developers use and actively support the BSDs (I develop corebase on a Debian Testing machine, and boot into FreeBSD to test), however, you still need to admit that GNU/Linux is the most used FOSS OS on desktops. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Hi, Stefan Bidi wrote: The first step for anyone new to GNUstep, after they realize the distro supplied packages are sorely out-of-date, will be to compile it using the standard compiler/runtime supplied to build GNUstep. This is partly true. The problem is that Debianpackages suck and in consequence Ubuntu. I don't know how RH/Fedora and OpeNSuste are. Sadly, this is a big market percentage. However e.g. OpenBSD and FreeBSD are quite up-to-date. NetBSD seems acceptable too. Gentoo is also quite reasonable, offering even the choice of all 3 backends! The Application scenario is also quite up-to-date Kudos, Bernard! Riccardo ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Wednesday, January 8, 2014 10:14 CET, Riccardo Mottola wrote: > Hi, > > Stefan Bidi wrote: > > > > Are we now saying libobjc2 is the "preferred" runtime? If so, it > > implies that clang is the preferred compiler. I have no arguments for > > or against it, I just want to clarify. > I would not like to do that, we remain "runtime neutral", the runtime is > a dependency. You can use GCC with its GNU runtime, if you use Clang you > must install libobjc2, but you can (or at least, could, I did that a I don't think this is true. You can use clang with the gcc runtime. IIRC, I did that on OpenBSD for some testing. Sebastian > couple of months ago) to use GCC+libobjc2. > > Riccardo > > ___ > Gnustep-dev mailing list > Gnustep-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/gnustep-dev ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Hi, Stefan Bidi wrote: Are we now saying libobjc2 is the "preferred" runtime? If so, it implies that clang is the preferred compiler. I have no arguments for or against it, I just want to clarify. I would not like to do that, we remain "runtime neutral", the runtime is a dependency. You can use GCC with its GNU runtime, if you use Clang you must install libobjc2, but you can (or at least, could, I did that a couple of months ago) to use GCC+libobjc2. Riccardo ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 8 Jan 2014, at 00:28, Lundberg, Johannes wrote: > However, for someone writing apps for iOS / OS X and using features like > automatic @synthesize (no need to write @synthesize) and the, according to > David, brain dead array / dictionary literals :) it would be nice to know > that you can transfer your code to GNUstep without having to rewrite all your > NSArray initializers or add @synthesize everywhere... (maybe most people > would take this for granted...) That's my thinking. If we say we support modern features including ARC and blocks, then it's pretty much implicit that we support everything else that's simpler. David -- Sent from my IBM 1620 ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Hi All I mostly agree with David about what features to highlight. ARC and blocks are definitely the two big features that have affected the way I program on iOS. However, for someone writing apps for iOS / OS X and using features like automatic @synthesize (no need to write @synthesize) and the, according to David, brain dead array / dictionary literals :) it would be nice to know that you can transfer your code to GNUstep without having to rewrite all your NSArray initializers or add @synthesize everywhere... (maybe most people would take this for granted...) -- Johannes Lundberg BRILLIANTSERVICE CO., LTD. On Wed, Jan 8, 2014 at 2:29 AM, Ivan Vučica wrote: > On Tue Jan 07 2014 at 4:31:57 PM, David Chisnall > wrote: > >> On 7 Jan 2014, at 16:13, Ivan Vučica wrote: >> >> > I definitely wouldn't go with anything like Objective-C+ARC since I, >> for one, don't think ARC is nearly as an important addition to the language >> as @synthesize. >> >> I completely disagree. @synthsize is a small improvement. It lets you >> generate methods, but you can already do most of what it does with a macro. >> It's nice, but it's not that special. ARC, in contrast, is a big change >> to the language. It fundamentally changes how you think about writing >> code, letting you think about object ownership rather than reference >> counts. I'd put the non-fragile ABI in the same category as synthesised >> properties - it makes code cleaner and easier to maintain, but doesn't >> really change how you use it. Blocks are in the same category as ARC: >> they're a big change to the language and change how you write code. If we >> were to highlight two features that we support, blocks and ARC would be top >> of the list. >> > > You can do it with macros, but a lot of people handwrote the getters and > setters. > > ARC does change the way you write code. It tries to turn Objective-C into > a "magic" language to satisfy people used to garbage collection, or people > who trust the compiler to do the right thing. > > As we discussed at devmeeting 2013, my personal reasons for not trusting > ARC are admittedly irrational: I was bit by GC many years ago, and one of > the reasons why I approached Objective-C was that refcounting was > consistently present, consistently ubiquitous and publicly exposed. > > ARC hides it. > > I don't like hidden reference counting. > > >> > And five years from now, any arguments against naming it relative to >> Objective-C 2.0 will stand against naming it Objective-C+ARC or similar. >> >> If our web site is still saying the same thing in five years, then we >> have other problems. >> > > Let's compose something that makes it irrelevant if we change the website > or not. Let's compose something that doesn't depend on one specific feature. > > First big changes in the language came with what was called Objective-C > 2.0. There is no need to invent our own point in time where the changes > started to happen and where it is important to point out that we support > post-1990s things. > > At this point I think I gave enough input and will drop the subject. > > ___ > Gnustep-dev mailing list > Gnustep-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/gnustep-dev > > -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- 秘密保持について：この電子メールは、名宛人に送信したものであり、秘匿特権の対象となる情報を含んでいます。 もし、名宛人以外の方が受信された場合、このメールの破棄、およびこのメールに関する一切の開示、 複写、配布、その他の利用、または記載内容に基づくいかなる行動もされないようお願い申し上げます。 --- CONFIDENTIALITY NOTE: The information in this email is confidential and intended solely for the addressee. Disclosure, copying, distribution or any other action of use of this email by person other than intended recipient, is prohibited. If you are not the intended recipient and have received this email in error, please destroy the original message. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Tue Jan 07 2014 at 4:31:57 PM, David Chisnall wrote: > On 7 Jan 2014, at 16:13, Ivan Vučica wrote: > > > I definitely wouldn't go with anything like Objective-C+ARC since I, for > one, don't think ARC is nearly as an important addition to the language as > @synthesize. > > I completely disagree. @synthsize is a small improvement. It lets you > generate methods, but you can already do most of what it does with a macro. > It's nice, but it's not that special. ARC, in contrast, is a big change > to the language. It fundamentally changes how you think about writing > code, letting you think about object ownership rather than reference > counts. I'd put the non-fragile ABI in the same category as synthesised > properties - it makes code cleaner and easier to maintain, but doesn't > really change how you use it. Blocks are in the same category as ARC: > they're a big change to the language and change how you write code. If we > were to highlight two features that we support, blocks and ARC would be top > of the list. > You can do it with macros, but a lot of people handwrote the getters and setters. ARC does change the way you write code. It tries to turn Objective-C into a "magic" language to satisfy people used to garbage collection, or people who trust the compiler to do the right thing. As we discussed at devmeeting 2013, my personal reasons for not trusting ARC are admittedly irrational: I was bit by GC many years ago, and one of the reasons why I approached Objective-C was that refcounting was consistently present, consistently ubiquitous and publicly exposed. ARC hides it. I don't like hidden reference counting. > > And five years from now, any arguments against naming it relative to > Objective-C 2.0 will stand against naming it Objective-C+ARC or similar. > > If our web site is still saying the same thing in five years, then we have > other problems. > Let's compose something that makes it irrelevant if we change the website or not. Let's compose something that doesn't depend on one specific feature. First big changes in the language came with what was called Objective-C 2.0. There is no need to invent our own point in time where the changes started to happen and where it is important to point out that we support post-1990s things. At this point I think I gave enough input and will drop the subject. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 14:13, Stefan Bidi wrote: > The first step for anyone new to GNUstep, after they realize the distro > supplied packages are sorely out-of-date, will be to compile it using the > standard compiler/runtime supplied to build GNUstep. Both OpenBSD and FreeBSD ship with GNUstep packages that are up to date and compiled with sensible options. I've not tried the OpenBSD ones, but on FreeBSD things like ARC and libdispatch work out of the box. We should perhaps have a prominent list of systems with up-to-date packages so that people know what is sensible to install in a VM if they want to play with GNUstep. David -- Sent from my Cray X1 ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Am 07.01.2014 15:13, schrieb Stefan Bidi: > The first step for anyone new to GNUstep, after they realize the distro > supplied packages are sorely out-of-date, will be to compile it using the > standard compiler/runtime supplied to build GNUstep. I'm working on this to make this "first step" obsolete. Because I think time is invested much better in building new packages than in debating on how to work around the old ones. First results are here: https://launchpad.net/~mah-jump-ing/+archive/ppa Not flawlessly, yet, because the old packaging interferes a lot with recent sources, but I'm continueing. With some luck, a recent gnustep-make is available tomorrow, which should act the same way as if you had compiled/installed it by hand. Currently it uses gcc/libobjc, switching to clang/libobjc2 should be straightforward. Markus -- - - - - - - - - - - - - - - - - - - - Dipl. Ing. (FH) Markus Hitter http://www.reprap-diy.com/ http://www.jump-ing.de/ ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 16:13, Ivan Vučica wrote: > I definitely wouldn't go with anything like Objective-C+ARC since I, for one, > don't think ARC is nearly as an important addition to the language as > @synthesize. I completely disagree. @synthsize is a small improvement. It lets you generate methods, but you can already do most of what it does with a macro. It's nice, but it's not that special. ARC, in contrast, is a big change to the language. It fundamentally changes how you think about writing code, letting you think about object ownership rather than reference counts. I'd put the non-fragile ABI in the same category as synthesised properties - it makes code cleaner and easier to maintain, but doesn't really change how you use it. Blocks are in the same category as ARC: they're a big change to the language and change how you write code. If we were to highlight two features that we support, blocks and ARC would be top of the list. > And five years from now, any arguments against naming it relative to > Objective-C 2.0 will stand against naming it Objective-C+ARC or similar. If our web site is still saying the same thing in five years, then we have other problems. David -- Sent from my brain ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
I get your viewpoint, but I disagree and have more faith in humanity than that :-) Post-2007 is post-2007. It's not "in 2007". It's not "in one year after 2007". Does the concept of post-PC include wearable computing? By the aforementioned standards, it would seem to me personally that no -- it would have to be strictly relating to the devices being introduced along with the term 'post-PC'. Does a requirement 'Windows XP or later' mean the program will run on Vista, 7 and 8.x? How about '512mb RAM or more'? Apart from extreme cases (Windows 3.1-or-later programs running on 64-bit Windows 8), I think we can agree that more is better. But let's say that people will truly think 'oh, it's just things added in 2008'. How about "numerous features introduced to language and language runtime after 2007, supporting even additions in 2013"? Not really pleasant. I still think Objective-C 2.0 is more concise than the above contraption. If anything, even "features from Objective-C 2.0 and later" sounds way better than the above. I definitely wouldn't go with anything like Objective-C+ARC since I, for one, don't think ARC is nearly as an important addition to the language as @synthesize. And five years from now, any arguments against naming it relative to Objective-C 2.0 will stand against naming it Objective-C+ARC or similar. On Tue Jan 07 2014 at 4:02:48 PM, David Chisnall wrote: > On 7 Jan 2014, at 14:19, Ivan Vučica wrote: > > >> > "Post-2007 features in language and language runtime"? > >> > >> I read this as 'we support a 6-year old version of the language! Yay!' > > > > Last time I checked "post-2007" does include features introduced in > 2011, 2012, 2013... :-) > > It does, but that's not how marketing works. If something is greater than > x, you say it's greater than x, for the largest possible x. If you say > 'Under $10' then people assume you mean $9.99. If you say 'post-2007', > people assume you mean 2008. > > David > > -- Sent from my Apple II > > ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 14:19, Ivan Vučica wrote: >> > "Post-2007 features in language and language runtime"? >> >> I read this as 'we support a 6-year old version of the language! Yay!' > > Last time I checked "post-2007" does include features introduced in 2011, > 2012, 2013... :-) It does, but that's not how marketing works. If something is greater than x, you say it's greater than x, for the largest possible x. If you say 'Under $10' then people assume you mean $9.99. If you say 'post-2007', people assume you mean 2008. David -- Sent from my Apple II ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Tue Jan 07 2014 at 10:37:43 AM, David Chisnall wrote: > On 7 Jan 2014, at 10:26, Ivan Vučica wrote: > > > "Post-2007 features in language and language runtime"? > > I read this as 'we support a 6-year old version of the language! Yay!' > Last time I checked "post-2007" does include features introduced in 2011, 2012, 2013... :-) ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Tue, Jan 7, 2014 at 6:55 AM, Sebastian Reitenbach < sebas...@l00-bugdead-prods.de> wrote: > How about just calling it as simple as: "GNUstep Runtime" > That GNUstep runs on Objective-C, should be clear from the general > description > about what GNUstep is anyways. > I would just like to add that it is very likely that there is still some confusion around the "GNU runtime" and the "GNUstep runtime". I'd imagine this is a problem for people new to the GNUstep eco-system. Most, if not all, GNU/Linux distros still include GCC as the standard compiler, and do not support the new features. The first step for anyone new to GNUstep, after they realize the distro supplied packages are sorely out-of-date, will be to compile it using the standard compiler/runtime supplied to build GNUstep. Then a page dedicated to the runtime, could go into the gory details of all > the nice features it provides. > Are we now saying libobjc2 is the "preferred" runtime? If so, it implies that clang is the preferred compiler. I have no arguments for or against it, I just want to clarify. cheers, > Sebastian PS: It would be nice if we could get the GNU runtime deprecated within GCC. I think I remember David tried integrating the GNUstep runtime with GCC and got frustrated, but GCC is still the "GNU compiler" and the standard compiler on GNU/Linux, having 2 slightly incompatible runtimes is confusing, even for me who has followed the project for over 10 years. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
> On Jan 7, 2014, at 1:23, Richard Frith-Macdonald > wrote: > > However, I'm not sure that we can use the term ARC as a big selling point, > simply because I'm not sure people will understand how good a feature it is. True -- but when I have described ARC to non-ObjC developers/colleagues, most are amazed by the idea and have trouble believing that it's actually even possible. The fact that you can have such a feature so well integrated in a C-based language (and that GNUStep brings this to other operating systems) is a very attractive. Patryk ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Tuesday, January 7, 2014 13:23 CET, Richard Frith-Macdonald wrote: > > On 7 Jan 2014, at 12:16, Niels Grewe wrote: > > > (Sorry, forgot to CC the list) > > > Anfang der weitergeleiteten Nachricht: > > >> Von: Niels Grewe > >> Betreff: Re: Intermediate Summary: Re: GNUstep.org website redesign > >> proposal > >> Datum: 7. Januar 2014 13:00:12 MEZ > >> An: David Chisnall > >> >> >> Am 07.01.2014 um 12:06 schrieb David Chisnall : > >> >>> On 7 Jan 2014, at 10:49, Richard Frith-Macdonald > >> >>> wrote: > >>> >>>>> I read 'Objective-C 2' as 'we support the features Apple introduced > >>> >>>>> in 2005! Yay!' > >>>> >>>> You may be unusual in that. > >>> >>> Possibly, but the 2006 WWDC was the first and last time Apple > >>> >>> referred to a set of new Objective-C features as Objective-C 2. The > >>> >>> term Objective-C 2 does not appear in current Apple docs, > >> >> > >> That’s not quite true: Some of the documents still talk about something > >> called "Objective-C 2.0“, most notably the runtime programming guide: > >> >> https://developer.apple.com/library/mac/documentation/Cocoa/Conceptual/ObjCRuntimeGuide/Articles/ocrtVersionsPlatforms.html#//apple_ref/doc/uid/TP40008048-CH106-SW1 > >> > >> >> But that doesn’t make the term any more intelligible. I think we should > >> >> prefer to use a concise description of the feature set over any > >> >> marketing mumbo jumbo. > I sympathise/agree with that ... but we have, on the website , a consice > description of what gnustep is, and yet just had Doc O'Leary complaning > because we didn't have a marketing mission statement. > > Probably we need both ... a good description and a marketing phrase to refer > to it. How about just calling it as simple as: "GNUstep Runtime" That GNUstep runs on Objective-C, should be clear from the general description about what GNUstep is anyways. Then a page dedicated to the runtime, could go into the gory details of all the nice features it provides. cheers, Sebastian > > > ___ > Gnustep-dev mailing list > Gnustep-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/gnustep-dev ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 12:16, Niels Grewe wrote: > (Sorry, forgot to CC the list) > > Anfang der weitergeleiteten Nachricht: > >> Von: Niels Grewe >> Betreff: Re: Intermediate Summary: Re: GNUstep.org website redesign proposal >> Datum: 7. Januar 2014 13:00:12 MEZ >> An: David Chisnall >> >> >> Am 07.01.2014 um 12:06 schrieb David Chisnall : >> >>> On 7 Jan 2014, at 10:49, Richard Frith-Macdonald >>> wrote: >>> >>>>> I read 'Objective-C 2' as 'we support the features Apple introduced in >>>>> 2005! Yay!' >>>> >>>> You may be unusual in that. >>> >>> Possibly, but the 2006 WWDC was the first and last time Apple referred to a >>> set of new Objective-C features as Objective-C 2. The term Objective-C 2 >>> does not appear in current Apple docs, >> >> >> That’s not quite true: Some of the documents still talk about something >> called "Objective-C 2.0“, most notably the runtime programming guide: >> >> https://developer.apple.com/library/mac/documentation/Cocoa/Conceptual/ObjCRuntimeGuide/Articles/ocrtVersionsPlatforms.html#//apple_ref/doc/uid/TP40008048-CH106-SW1 >> >> >> But that doesn’t make the term any more intelligible. I think we should >> prefer to use a concise description of the feature set over any marketing >> mumbo jumbo. I sympathise/agree with that ... but we have, on the website , a consice description of what gnustep is, and yet just had Doc O'Leary complaning because we didn't have a marketing mission statement. Probably we need both ... a good description and a marketing phrase to refer to it. ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 10:49, Richard Frith-Macdonald wrote: >> I read 'Objective-C 2' as 'we support the features Apple introduced in 2005! >> Yay!' > > You may be unusual in that. Possibly, but the 2006 WWDC was the first and last time Apple referred to a set of new Objective-C features as Objective-C 2. The term Objective-C 2 does not appear in current Apple docs, nor is it a language dialect that you can select in the compiler. If we make a big deal about supporting Objective-C 2, then there's an immediate question of why we're not saying anything about supporting newer features. Simply saying that we support all of the modern Objective-C language features, including ARC, tells people that we support the thing that they're most likely to care about, and we also support the other new stuff, irrespective of how you choose to brand it. David -- Sent from my Cray X1 ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Tuesday, January 7, 2014 12:02 CET, Markus Hitter wrote: > Am 07.01.2014 09:37, schrieb David Chisnall: > > - ARC > > - Blocks > > - Properties > > Good points. Please write the full wording, not everybody knows what > "ARC" stands for. And how "Properties" is different from what C++ gave > us in 1990. With the Objc getting an own page, there likely plenty of room for detailed explanations ;) Sebastian > > > - Braindead array and dictionary syntax with poorly thought out semantics > > added to appease Python programmers > > Uhm, yes. Good wording for alienating Python hackers. :-) > > > Markus > > -- > - - - - - - - - - - - - - - - - - - - > Dipl. Ing. (FH) Markus Hitter > http://www.reprap-diy.com/ > http://www.jump-ing.de/ > > ___ > Gnustep-dev mailing list > Gnustep-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/gnustep-dev ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Am 07.01.2014 09:37, schrieb David Chisnall: > - ARC > - Blocks > - Properties Good points. Please write the full wording, not everybody knows what "ARC" stands for. And how "Properties" is different from what C++ gave us in 1990. > - Braindead array and dictionary syntax with poorly thought out semantics > added to appease Python programmers Uhm, yes. Good wording for alienating Python hackers. :-) Markus -- - - - - - - - - - - - - - - - - - - - Dipl. Ing. (FH) Markus Hitter http://www.reprap-diy.com/ http://www.jump-ing.de/ ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 10:37, David Chisnall wrote: > On 7 Jan 2014, at 10:26, Ivan Vučica wrote: > >> "Post-2007 features in language and language runtime"? > > I read this as 'we support a 6-year old version of the language! Yay!' > >> I'd however go with Objective-C 2.0. Apple may have stopped using the name, >> but it is still a good historical descriptor without a suitable replacement. >> It is also a good name to point out a difference from a much simpler >> language (and runtime) that existed prior to that point. > > I read 'Objective-C 2' as 'we support the features Apple introduced in 2005! > Yay!' You may be unusual in that. > Neither of these are the impression that we want to give. How about 'Objective-C: The Next Generation' :-) ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 10:26, Ivan Vučica wrote: > "Post-2007 features in language and language runtime"? I read this as 'we support a 6-year old version of the language! Yay!' > I'd however go with Objective-C 2.0. Apple may have stopped using the name, > but it is still a good historical descriptor without a suitable replacement. > It is also a good name to point out a difference from a much simpler language > (and runtime) that existed prior to that point. I read 'Objective-C 2' as 'we support the features Apple introduced in 2005! Yay!' Neither of these are the impression that we want to give. David -- Sent from my PDP-11 ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
"Post-2007 features in language and language runtime"? I'd however go with Objective-C 2.0. Apple may have stopped using the name, but it is still a good historical descriptor without a suitable replacement. It is also a good name to point out a difference from a much simpler language (and runtime) that existed prior to that point. If it comes up in conversations (and it does, because it's much easier to write and pronounce), then we can and should use it. No need to follow Apple's lead here. Besides, Apple's replacement in the context of the runtime is '64-bit runtime'. Not as applicable to us, I'm sure you'll agree, as we can easily build the libobjc2 runtime with varying 'amounts of bits'. On Tue Jan 07 2014 at 9:23:31 AM, Richard Frith-Macdonald < richardfrithmacdon...@gmail.com> wrote: > > On 7 Jan 2014, at 08:37, David Chisnall wrote: > > > On 7 Jan 2014, at 08:23, Sebastian Reitenbach < > sebas...@l00-bugdead-prods.de> wrote: > > > >> * better OBJC2 support, some more proper gs-make support > > > > A minor point, but Apple hasn't used the term 'Objective-C 2' for over > five years. Possibly because they were mocked for describing the version > of Objective-C that came after Objective-C 4 as Objective-C 2... > > > > The main point that we want to be making today is that we support ARC. > We might want some bullet-point features, such as: > > > > - ARC > > - Blocks > > - Properties > > - Braindead array and dictionary syntax with poorly thought out > semantics added to appease Python programmers > > > > Or, more simply, all of the language features that are supported on OS X > with the latest Apple tools. > > Good point ... I hadn't really considered the branding of the > language/runtime. > > I agree that ARC is the killer feature. The others are, IMO relatively > minor refinements not suitable to be the headline feature, or > braindead/bloat in some way (even though they have possible good > applications). > > However, I'm not sure that we can use the term ARC as a big selling point, > simply because I'm not sure people will understand how good a feature it is. > How can we brand the latest objc language/runtime so that it both sounds > impressive without being either too technical (arc) or too vague (modern)? > > On the other hand, maybe calling it ObjectiveC-ARC is OK if OSX developers > all understand it? > ___ > Gnustep-dev mailing list > Gnustep-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/gnustep-dev > ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 09:23, Richard Frith-Macdonald wrote: > Good point ... I hadn't really considered the branding of the > language/runtime. > > I agree that ARC is the killer feature. The others are, IMO relatively minor > refinements not suitable to be the headline feature, or braindead/bloat in > some way (even though they have possible good applications). > > However, I'm not sure that we can use the term ARC as a big selling point, > simply because I'm not sure people will understand how good a feature it is. I think OS X / iOS developers will, given how much effort Apple has spent on selling it. And people who remember life before ARC will know how painful it is to go back to the pre-ARC world[1]. > How can we brand the latest objc language/runtime so that it both sounds > impressive without being either too technical (arc) or too vague (modern)? > > On the other hand, maybe calling it ObjectiveC-ARC is OK if OSX developers > all understand it? I would say 'All of the latest Objective-C features, including ARC'. People who know about ObjC will most likely be explicitly looking for ARC on the feature list, people who don't will just see the general point. And it's brief enough to fit in a bullet point. David [1] On which note, I wonder if GNUstep is losing potential contributors by requiring code in the core libraries to use manual RR... -- This email complies with ISO 3103 ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 08:37, David Chisnall wrote: > On 7 Jan 2014, at 08:23, Sebastian Reitenbach > wrote: > >> * better OBJC2 support, some more proper gs-make support > > A minor point, but Apple hasn't used the term 'Objective-C 2' for over five > years. Possibly because they were mocked for describing the version of > Objective-C that came after Objective-C 4 as Objective-C 2... > > The main point that we want to be making today is that we support ARC. We > might want some bullet-point features, such as: > > - ARC > - Blocks > - Properties > - Braindead array and dictionary syntax with poorly thought out semantics > added to appease Python programmers > > Or, more simply, all of the language features that are supported on OS X with > the latest Apple tools. Good point ... I hadn't really considered the branding of the language/runtime. I agree that ARC is the killer feature. The others are, IMO relatively minor refinements not suitable to be the headline feature, or braindead/bloat in some way (even though they have possible good applications). However, I'm not sure that we can use the term ARC as a big selling point, simply because I'm not sure people will understand how good a feature it is. How can we brand the latest objc language/runtime so that it both sounds impressive without being either too technical (arc) or too vague (modern)? On the other hand, maybe calling it ObjectiveC-ARC is OK if OSX developers all understand it? ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On Tuesday, January 7, 2014 09:37 CET, David Chisnall wrote: > On 7 Jan 2014, at 08:23, Sebastian Reitenbach > wrote: > > > * better OBJC2 support, some more proper gs-make support > > A minor point, but Apple hasn't used the term 'Objective-C 2' for over five > years. Possibly because they were mocked for describing the version of > Objective-C that came after Objective-C 4 as Objective-C 2... > > The main point that we want to be making today is that we support ARC. We > might want some bullet-point features, such as: > > - ARC > - Blocks > - Properties > - Braindead array and dictionary syntax with poorly thought out semantics > added to appease Python programmers But what would then be a "better" name for it? Just "modern objc runtime"? Or something that would describe it in one word, suitable as a menu entry? cheers, Sebastian > > Or, more simply, all of the language features that are supported on OS X with > the latest Apple tools. > > David > > -- Sent from my Difference Engine > > > ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
On 7 Jan 2014, at 08:23, Sebastian Reitenbach wrote: > * better OBJC2 support, some more proper gs-make support A minor point, but Apple hasn't used the term 'Objective-C 2' for over five years. Possibly because they were mocked for describing the version of Objective-C that came after Objective-C 4 as Objective-C 2... The main point that we want to be making today is that we support ARC. We might want some bullet-point features, such as: - ARC - Blocks - Properties - Braindead array and dictionary syntax with poorly thought out semantics added to appease Python programmers Or, more simply, all of the language features that are supported on OS X with the latest Apple tools. David -- Sent from my Difference Engine ___ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: Intermediate Summary: Re: GNUstep.org website redesign proposal
Hi, thank's everyone for the input so far. A bit later than anticipated, but here is the promised summary of the tread so far: As the discussion turned out, there are things that need clarification before it makes sense to setup a new web presentation. Which also gives me, and Riccardo time to improve our web skills in the meantime ;) As far as I can see now, decisions need to be made on the following: * A statement is needed, and needs to be agreed on, what "proper" GNUstep actually is. Richard made some good suggestions: * Core GNUstep (make/base/gui/back) * Development tools (Gorm) * Other things that should be added as soon as they are matured: * THEMES, at least one for Linux/Unix-like, and one for Windows bundled with the core system * better OBJC2 support, some more proper gs-make support * MACUSER: a simple mac user porting mechanism is needed, either provide a VM, or some xcode integration, whatever would be better suited * MOBILE: implementing new/improverd graphics stuff and porting to android etc. * Breaking up GNUstep into three things: * GS Core (the core, supported stuff) * Hot future projects (worked on to become core, supported stuff) * i.e. Thematic * opal backend, core animation, etc. * GS Extras: additional libs and tools built on top of GNUstep, currently available in SVN or that only have (lousy) wiki pages ;) * this would include: Steptalk, GDL2, GSWeb, ProjectCenter, EasyDiff, webserver, webservices, simplewebkit, sqlclient, gsldap, corebase, GWorkspace, etc. * Maybe it would make sense to move some of them to GAP, at least the applicationsi.e. GWorkspace, PC, EasyDiff? Citing Doug Simons description of GNUstep, that I really think fits well: GNUstep is an open-source framework modeled on Apple's Cocoa frameworks to provide a cross-platform API to make it easy to create sophisticated modern software. Ports of OS X software to other platforms and new software development in Objective-C are both supported, with or without a graphical user interface. Which means, basically remove all mentioning of NeXT/OPENSTEP from the main page. Likely, most of the current Cocoa developers nowadays don't have an idea of what that is anyways. I suggest a history page should clarify about those roots, and how GNU and GNUstep fits in. * Make GAP a more central hub for GNUstep applications: * i.e. merge the gnustep-nonfsf project into GAP * and maybe the gsimageapps project too * point interested end users that come along on gnustep.org to GAP * GAP likely should also switch from CVS to SVN as RCS Based on above, just citing Richard here again: We want people to know what's there now and well supported We want people to know what's coming soon and where they can help most We want people to know about all the cool stuff that uses GNUstep and makes it look good We also want people to join in and help rather than complaining about problems which aren't part of core ... keeping a clear separation ought to help with that. We ought to make it explicit who is supporting what, and to what extent support/development is being done (and whether we are looking for a volunteer to support any particular thing). When it's more clear, what GNUstep really is, and what's supported, what's worked on, what's extra etc., then designing, structuring the contents of the webpage is likely to become much more easier. * As many have pointed out, the website and the wiki are there to meet different purposes/requirements. However, the wiki contains a lot of old and outdated information, and really needs a cleanup. But maybe before a cleanup is attempted, a decision must be made about what type of contents goes into the wiki, and what goes to the website. * A decision about the softwareindex. Does the softwareindex makes sense, or is it just more work and containing redundant information than it makes sense? I.e. currently, when you release a software, you have to update links/version numbers on the softwares homepage, wiki, softwareindex. Or maybe keep the softwareindex, and maybe remove redundant information about applications that have a nice homepage from the wiki? The look and feel of the softwareindex can likely be enhanced with CSS to make it fit the design of the rest of the website. Feedback the design proposal of the GNUstep website I had: * modelling similar to gtk.org makes sense for at least a few people, the similar type of software is offered, and a similar type of audience is intended * have news on the GNUstep.org web site front page to show off activities within the project. * The dock at the bottom I had on my design, seems to be nice, but seems to be too difficult to recognize as a navigation item. Proposals were to move it to the top, or on the side. I basically dropped it from my current incarnation of design, as
