GNUstep Project Gets New Chief Maintainer

stivi writes "OSNews is reporting that Gregory Casamento has accepted the position of GNUstep Maintainer. Adam Fedor, former GNUstep leader writes: 'After over 15 years of being the Chief Maintainer for GNUstep, I've found I have too many other responsibilities to devote as much time to GNUstep as is necessary. I still plan on contributing to GNUstep in the future in a lower capacity.' Gregory has been a prolific developer for GNUstep for the past seven years and is currently the maintainer for Gorm (the graphical interface designer) and the GUI library. I think he will make a great choice to lead GNUstep in the future. New plans for change have been set up already. Thank you Adam for the past, congratulations Gregory to the future."

Comments

Direction

[bsharitt]

Hopefully we see GNUStep get some definitive direction to show the world that it's still alive. Most people think it's a legacy development tool kit which at one time was meant to replicate OpenStep, but is now dead, though that is not the case, but they need to let the world know they are alive. Also, they either need place nicer with the rest of the Linux/Unix desktops(Gnome or KDE) or either acknowledge that they are indeed their own little enviroment(the site still tries to pass it off as development libraries and tools)

Re:Direction

[Somnus]

There is at least one effort in the direction of a desktop environment. Theming is provided by the Camaelon bundle.

On the toolkit/tools side of things, the major hurdles are:

• app icon (perhaps could be integrated with/translated to the f.d.o. systray standard)
  
• the work it takes to set up GNUstep just to launch an app (e.g., setting up paths -- thankfully handled transparently in my Gentoo setup)
  
• incompatibility between GNUstep services and dbus, etc. etc.
  
• non-standard build system (still easy enough to making into ebuilds/RPMs/etc.) and monolithic libraries

Sidestep is an experiment in addressing these issues.

Re:gnustep

[bsharitt]

If they get serious about more complete source compatibility with Cocoa, it could go a long way to attracting Mac developers to Linux if they can accomplish ports of many Cocoa apps with simple recompiles.

Alive

[bartvh]

The previous news item in this category was exactly 4 years ago. He could start by sending out more 'press releases'.

Re:Alive

[borgheron]

This would be more the case, yes, thanks. :)

GJC

Re:Alive

[stivi]

well ... if only the GNUstep related articles were not rejected or left pending for weeks... this article is my second sumbission, not counting submissions of other steppers.

Re:Alive

[grahamlee]

I've had more GNUstep stories accepted at Macslash than Slashdot :-/

Ok I read TFA

[Psychotria]

And... I think this dude is a complete moron. He types (on the page and in his blog) all this "business" speak gibberish which, in the end, means nothing. He says that he thinks KDE and Gnome are "amateurish" but doesn't bother to explain his reasoning behind the assertion. To me, this implies that he doesn't actually have a reason behind the assertion and that the whole silly blog is propaganda. I find his "business speak" patronising, transparent and meaningless. It may work for Microsoft, but propaganda does not work for the audience he is (supposed) to be targetting.

Re:Ok I read TFA

[jcr]

He says that he thinks KDE and Gnome are "amateurish" but doesn't bother to explain his reasoning behind the assertion.

If you had ever used NeXTSTEP, you would know what he means. It's one of those things that's hard to explain to someone who doesn't have the experience to tell the difference.

-jcr

Re:Ok I read TFA

[Psychotria]

Actually I have used NeXTSTEP. This doesn't change my opinion of the self-proclaimed new "Chief" though. Seriously, why does he have to keep reminding us (the poor serfs) that he is the "Chief"? And why all the business speak crap? Why not speak to us as equals, rather than speaking down to us like we're idiots? I don't think I'm an idiot. I don't think you're an idiot. But he thinks (judging from his blog) that we're both idiots. That is what is irritating me. I am a technical person. I can see through propaganda and when I do, I don't particularly appreciate it. I'd rather that he was honest and "spoke" (in his writing) to me as a peer.

Re:Ok I read TFA

[mungtor]

"He says that he thinks KDE and Gnome are "amateurish" but doesn't bother to explain his reasoning behind the assertion."

He does to some degree. He clearly states that neither look like they started with a clean vision of what the desktop environment should be and have simply "evolved" to what you see now. He also states that the code base which makes up most of Gnome is a nightmare. I can't speak about KDE too much since I don't use it (strikes me too much as Windows, which I don't particularly like).

GNUStep looks something like the Sun OpenWindows desktop used to... Icons and apps minimize to the desktop, not the the taskbar area.

Either way, I just hope that it will finally be easy to customize the behavior of windows... For example:

I want my Xterm window to maximize to the vertical height of the screen without changing width when I double-click the title bar. How would you tell a non-programmer to accomplish that in Gnome or KDE? Will it be easier in GNUStep?

(and I am sure the Gnome answer is to navigate some XML file to find the variable Window.click.title.bar.some.other.arbitrary.and.me aningless.string.that.you.will.not.know.unless.you .coded.it and change the default value from 1 to 3. Excellent usability there)

Re:Ok I read TFA

[jcr]

speaking down to us

WTF are you talking about?

-jcr

Re:Ok I read TFA

[squiggleslash]

The funny thing is that the same criticisms can be made of GNUstep as of GNOME. Further, while the codebase issue may be true, GNOME is far from "amateurish" in 2006. It's a slick, professional, desktop, that, IMHO, is more usable and friendly than Microsoft's Windows, even if it isn't yet up to Mac OS X standards.

Does GNUstep have direction? Does it hell. There is an official line (that it's supposed to look and feel like OpenStep 4), and an unofficial argument about where it's going. The look and feel issues are being more and more muddied, as the influence of the Mac gains on the project without being officially embraced.

The code base is, despite a "1.0" release many years ago, still in flux with recent changes breaking older code. Compiling it is a PITA.

What's different? Well, the GNOME project really has had something resembling a direction over the last few years, and the system is now extremely usable, consistant, and well polished. I don't like all of the decisions they've made, but compared to GNUstep, it's a professional desktop. It most certainly isn't "amateurish", it might have been in the miguel days, but it's come a long way from then.

GNUstep, by comparison, is a mismatch of "nearly finished" applications. Glitches such as fights between menus as to who's on top, and translation issues ("Informations"?), not to mention confusion as to who manages the dock (GWorkspace.app has one, WindowMaker has another, and they work completely differently. And why is the Dock integrated into either?) and even the desktop, are commonplace and jarring.

I tried, when I switched back to GNU/Linux earlier this year, to start off with GNUstep, because it felt it was a more interesting system, but I ended up, despite prejudices heavily against it, going to GNOME. GNUstep is just too beta-quality at the moment. It's not there yet.

For anyone involved in GNUstep to describe GNOME as "amateurish" takes an astonishing amount of gall. At least release a finished, polished, version first with all the critical components (including the WM and GWorkspace.app) complete, working, and integrated, before criticising a project that's complete, working, and integrated as "amateurish".

Re:Ok I read TFA

[mungtor]

Oh, I definitely agree that Gnome looks very polished and professional. I was only pointing out that the new GNUstep maintainer does elaborate some reasons why he considers them to be sub-optimal.

FWIW tho, I am still annoyed at how hard it is to find the right variables to change in Gnome to accomplish some tasks. Having Folders open in the same window in the file manager shouldn't be nearly as buried as it is. It doesn't bother me since I hardly ever use it, but it is very annoying and difficult to change for somebody trying out Gnome for the first time.

Re:Ok I read TFA

[amorsen]

I want my Xterm window to maximize to the vertical height of the screen without changing width when I double-click the title bar. How would you tell a non-programmer to accomplish that in Gnome or KDE? Will it be easier in GNUStep?

This is actually a window manager thing, not something to do with Gnome or KDE. It's not particularly hard to accomplish with old-school window managers, but modern ones like Metacity deliberately prevent such configurability. Personally I'm just waiting for a modern window manager which lets me turn off click-to-front off.

Re:Ok I read TFA

[C0vardeAn0nim0]

"GNUStep looks something like the Sun OpenWindows desktop used to... Icons and apps minimize to the desktop, not the the taskbar area."

kinda off-topic, but... what i'd like to see is a desktop environment that allows me to minimize apps to a _folder_ and a file manager that can tell me when a file is open by hashing it's icon, just like OS/2 warp used to do.

"I want my Xterm window to maximize to the vertical height of the screen without changing width when I double-click the title bar. How would you tell a non-programmer to accomplish that in Gnome or KDE? Will it be easier in GNUStep?"

I'm at the office right now (WinXP only. aaaahrgh!!!), so you'll have to wait until i get home to check, but IIRC, KDE has an option easily accessible through the control panel where you can choose the behavior when you double-click the title bar. options include full-screen maximization (like in windows), roll up the window (like in MacOS classic), minimize (like in MacOS X) or maximize verticaly.

Re:Ok I read TFA

[mungtor]

Ahhh... and that's another thing. Even using Linux for a few years I still don't have a good idea of where Gnome stops and Metacity starts...

Why would Metacity want to prevent the configurability? Unless it's stupendously hard to code, isn't this the same lack of choices people generally deride Microsoft for?

Re:Ok I read TFA

[welshmnt]

In KDE you can middle click on the maximise button to max vertically.

KDE REALLY is very good.

The standard themes look very windows ish, (thats because if it's not then everybody bitches about how strange it looks and how hard it must be to adapt to....) but it can look like any system you've seen ('cept plan 9) or like nothing on else on earth.

Lots of power hiding underneath an (optionally) pretty face :)

Re:Ok I read TFA

[mrsbrisby]

He says that he thinks KDE and Gnome are "amateurish" but doesn't bother to explain his reasoning behind the assertion

They are amateurish! Almost every rough edge in KDE and GNOME can be attributed to the fact that nonprofessionals have made significant UI decisions.

GNOME is at the very least attempting to change that, but if you've ever used a professional user interface (such as NeXT/OpenSTEP, CDE, or OS/2's WPS) you'd see just how lacking it can be to use something else.

Note that this isn't the same thing as approachable- you'll note the examples I gave are often considered relatively unapproachable interfaces- but they are extremely user friendly.

I find his "business speak" patronising, transparent and meaningless.

The smartest and dumbest thing the GNUstep people did was start GNUstep: It made an excellent developer framework, and using it means they get access to one of the best workflow systems ever (the kind that Microsoft has been playing catchup with their Cairo product- announced in 1989 still hasn't made good on), but it also means that the uninformed don't really have all that much to do.

You see, because GNOME and KDE aren't designed yet, anyone can offer a suggestion on what needs to be done. With GNUstep, the number of people who can offer productive suggestions is much smaller- and up until recently was comprised almost exclusively of programmers. After all, if anyone needed to wonder what GNUstep needed next- it was just a matter of picking one of the OpenStep headers that had implement-me markers all over it.

o me, this implies that he doesn't actually have a reason behind the assertion and that the whole silly blog is propaganda

This is obvious. You don't use GNUstep, or OpenStep, and you don't know anything about either, so this is a horrible way for you to find out about them. The story editor and submitter chose very poorly when putting this up here on slashdot, and as a result, you have very strong negative feelings about GNUstep.

Re:Ok I read TFA

[amorsen]

Why would Metacity want to prevent the configurability?

You'll have to see if you can find their reasons on the web somewhere. Searching for gnome configurability on Google and Groups got me some results, but not something definite to link to. Perhaps you'll have more luck...

Anyway, to solve your specific problem, you can actually assign a keyboard shortcut to "maximize vertically" to Metacity with gconf-editor. Not as good as what you really want perhaps, but it might tide you over. In the process I discovered that click-to-front is known as raise-on-click in Metacity, and that it can be turned off! Yay!

Re:Ok I read TFA

[pimpimpim]

I STILL use windowmaker, as I don't want junk on my Desktop and windowmaker got installed by default on all machines at work. I have made a keyboard shortcut to maximize vertically.

To do this, put in:

~/GNUstep/Defaults/WindowMaker

the line:

VMaximizeKey = "Mod1+Mod4+F12";

Or whatever you want. You can probably also fix this in the graphical configuration interface of windowmaker.

Re:Ok I read TFA

[claes]

This is not what you asked for but something handy to know: middle-clicking the "Maximize" button maximizes vertically. Left-clicking it maximizes horizontally. In KDE, at least.

Re:Ok I read TFA

[rg3]

Someone has already replied this can be done under KDE (kwin to be precise). The specific procedure follows:

* Click on the application icon in the title bar and select "Configure window behaviour", OR
* Go to the control center or KDE settings and the same options can be found under Desktop, Window behaviour.
* There, go to the "Titlebar actions" tab.
* Change the action of "Titlebar double-click" to "Maximize (vertical only)".

Re:Ok I read TFA

[MobyTurbo]

...I'm not certain whether that's a Gnome thing or an Ubuntu thing. Thank God for the online community...
it's certainly less hassle to put up with Gnome than to install an alternative. Of course, it's more of a hassle to have GNUstep on Ubuntu than GNOME. That's because Ubuntu is basically a snap-shot of Debian unstable with the latest GNOME desktop installed by default. The equivalent is the live GNUstep CD, http://livecd.gnustep.org/ to have GNUstep by default. Or use Debian, on which both are equally hard: "aptitude install gnustep" or "aptitude install gnome-desktop".

Re:Ok I read TFA

[Raenex]

He types (on the page and in his blog) all this "business" speak gibberish which, in the end, means nothing.

Examples? I read the blog post that was linked to in the article summary and he was very concrete in what needed to change.

Re:you

[timmarhy]

in the same manner that i would spit at a crazy person waving a sign as i drive past them, yes.

The *big* problem with GNUStep...

[muecksteiner]

was (and is), that few people realised how great the original NeXTStep environment - which GNUStep attempts to clone - was.

I was already around as a CS major at the time NeXTStep basically failed in the marketplace due to a) asinine marketing/pricing on the part of NeXT Inc. and b) the fact that everytime we showed the NeXTStep environment to fellow CS students and CS faculty, you would mostly get blank stares, and a few polite remarks. But no more.

Few "got it" how easy this was to use - concepts like the seperation of the user interface specification from the core logic of a program simply did not register with people weaned on TurboVision ("one line per code for each UI element"), and Apple has (probably rightfully so) more or less given up on educating people on how great the current successor to NeXTStep (Cocoa) is.

Nowadays, people code for OS X because OS X is seen as a hip system with a small but viable installed base, and the fact that the dev tools are extremely nice is just an added bonus.

So if GNUStep is just an Open Source version of something that is obsolete, why care at all?

Well, because the likes of KDE could have had it so much easier if they had used something like GNUStep (the structure of which is pretty revolutionary), instead of toolkits like QT, which were developed to be just a "better Win32" API.

Make no mistake, QT/KDE et al. turned out to *be* a better Win32/Foundation class environment, but I guess that most folks who were ever proficient in developing for the NeXT environment will agree, that a widely used and enhanced GNUStep would have been even more productive than that.

And still could be someday - after all, Linux desktops are such a melting pot of different toolkits and environments, that perhaps some "killer GNUStep apps" (graphics apps, like an Illustrator clone would be a good start) could get people to notice GNUStep again.

One can always dream... :-)

Just my $0.2E-32

A.

Re:The *big* problem with GNUStep...

[jcr]

So if GNUStep is just an Open Source version of something that is obsolete, why care at all?

Well, if you care whether Linux is going to make a dent in Microsoft's market share, you should care very much about GNUStep. For my part, I'll just keep using the Mac, so GNUStep is mostly a matter of nostalgia.

-jcr

Re:The *big* problem with GNUStep...

[listen]

Did you ever actually use Qt? It is not to be sniffed at.
Its main disadvantage is the same one as all C++ toolkits : the vast majority of people programming are absolute beyond belief raving idiots and will destroy their lives very quickly when they are given as many avenues of complexity as C++ gives them. Of course, this will probably make them feel that they are experts. I've recently realised after working on a few Python projects in groups that the same sadly applies to dynamic languages : most people programming can not be trusted with anything more expressive than Java. They will happily screw that beyond hell as well, but Java tools are good enough to help you out of the mess, and any other languages tools are not. Lets not bring up the pitiful smalltalk refactoring tools that people who have never used them rave about. They were crap because they relied on type inference - and the code you really want to refactor is shit code, ie has no consistent typing. So, worse is better... sigh.

Anyway, away from that semi-relevant rant, Qt itself is pretty well done and only things that I could see as real OpenStep advantages are scope related - it deals with a lot of interapplication stuff that Qt doesn't. I would really not dismiss Qt as an MFC++ or a Win32 wrapper - MFC is so foul it shouldn't be mentioned in the same breath as any vaguely sane toolkit.

Re:The *big* problem with GNUStep...

[Blakey Rat]

I think a "big" problem with GNUStep and Apple's Cocoa is its reliance on Objective-C, a language most people have never heard of. Apple put some work into building Java compatibility into its Cocoa environment, but they seem to have given up on that. (To the chagrin of projects like Adium, which used it extensively.) I would also argue that Java is a terrible choice for Mac programmers, who seem to favor more elegant and "artsy" results.

If you could code in Cocoa using C++, C# or even a language like Python, PHP or Ruby, I think it would be a lot easier to convince programmers. I know that personally I was very taken-aback by Objective-C and it's frankly mystifying and strange syntax compared to the C++ I was used to.

It also doesn't help that designing a Cocoa app is conceptually hard and has a relatively high learning curve. I've tinkered on and off with Cocoa with three years now, and designed several finished apps with it, and I still have to get help remembering which way to draw my lines in Interface Builder. (Maybe it's a case of "old dogs", but designing a UI with something like RealBasic or Visual Studio is much easier to my brain.)

Re:The *big* problem with GNUStep...

[lieven_dekeyser]

Leopard (Mac OS X 10.5) will have official support for writing Cocoa apps in Python and Ruby

Re:The *big* problem with GNUStep...

[hey!]

the vast majority of people programming are absolute beyond belief raving idiots and will destroy their lives very quickly when they are given as many avenues of complexity as C++ gives them.

Which is just a stuck up way of saying that C++ doesn't meet the neds of most projects or programmers.

And even smart people make mistakes.

Re:The *big* problem with GNUStep...

[shrewdsheep]

In terms of capabilites Qt and GNUstep/Cocoa are equivalent (roughly). The beauty of ObjC (the language underlying GNUstep/Cocoa), however, is its simplicity yet power which dwarfs C++ quite substantially in terms of ergonomics. If you are not familiar with ObjC you can best compare it with Java. As a matter of fact the Java environment is quite inspired by OPENSTEP which was developed jointly by NeXT and Sun, after which Sun moved to Java and eventually dropped OPENSTEP (for no good reason ;).

Re:The *big* problem with GNUStep...

[listen]

No, I really mean what I said. Most people programming do not understand the basics of what they are doing. They have a huge problem grasping concepts if they can not envisage in their minds directly how the code they are writing will be evaluated, and they have a very limited understanding of possible computational evaluation models. So they work at a very very low level of abstraction.

The code they are writing "works", but generally has huge performance, extensibilty and maintainability problems. The reason why I have become a Java advocate ( for corporate projects and for "sexy" open source projects - in some ways I abhor the strictures of Java when I am the only person on a project ) is that the Java market place (maybe not through conscious design) is the only one where the central concept driving changes is that almost all people programming are utter dunces. If you have to cope with tragic code, whilst not breaking it, it is very hard to do so without the tool support Java has, and it is an open research problem to make tool support as thorough as Java's or even vaguely workable with either a more dynamic (Smalltalk, Python, Ruby, JS) or more expressive static type system (C++, ML, Haskell).

Of course the best solution here is to stop employing muppets as programmers. This is certainly never going to happen in our lifetimes.

We have to start thinking of code more as data to be manipulated and reshaped over time, i.e. it is ok to have a lot of shit code as long as it can eventually be reshaped by someone with a clue when the inevitable problems come about.

Re:The *big* problem with GNUStep...

[borgheron]

GNUstep implements many parts of Cocoa. It's not *simply* an implementation of OPENSTEP. It's also an implementation most of the Cocoa additions to OPENSTEP as well. GNUstep is a cross platform API first and foremost.

GJC

Re:The *big* problem with GNUStep...

[listen]

Er, yeah. My point was, when someone contrasts Qt and OpenStep, all the "big" advantages are mainly just not having to deal with people getting into a mess with C++ ( in which it is easy to express some very neat abstractions, and ridiculously hard to express others - but this is not the way to judge a language you are choosing for other people).

If you really look at ObjC objectively, it is kind of disgusting. Don't get me wrong, it is usable, and will "normally" get muppets in to less trouble than C++, but it is C with bolt on bits of smalltalk syntax, and a very odd type system. And the tool support is quite, quite tragic.

Re:The *big* problem with GNUStep...

[hey!]

is that the Java market place (maybe not through conscious design) is the only one where the central concept driving changes is that almost all people programming are utter dunces.

Which is a stuck up way of saying that Java meets the needs of more projects than C++.

Of course the best solution here is to stop employing muppets as programmers. This is certainly never going to happen in our lifetimes.

One implementations of your best solution would be to hire, exclusively, graduates of the Vulcan Science Academy. Since your solution is predicated on assumptions contrary to fact, it is not in actuality a solution.

While having more competent programmers would, in any world real or imagined, be a good thing, I think you also underestimate the problem of realism in project goals. Your goals and who you hire should be closely interrelated. If you aren't willing to hire people with the skills needed to complete your project with reasonable quality and the time alloted, your project's goals are not realistic.

Hiring expensive talent and using a less "idiot-proof" language is not necessarily more virtuous than hiring median talent and using a more "idiot-proof" language. Nor are these the only two alternative strategies. The point is to have a strategy for a project that is sustainable, effective, and economical.

Re:The *big* problem with GNUStep...

[Blakey Rat]

Of course, now after dropping Java like a hot potato, Apple's going to have to convince developers that if they write a Python app, it'll run longer than a few versions.

Re:The *big* problem with GNUStep...

[dosius]

What I think would be nice for GNUstep is source-level compatibility for OSX apps, though that's probably a pipe dream. GNUstep on Darwin might be closer, but I don't know.

-uso.

Re:The *big* problem with GNUStep...

[listen]

Erm... you have a truly unique talent for missing the point. My "best solution" was intended as a "this is NEVER going to happen" alternative. It is quite sad that you've wasted so much time attacking an opinion that no one professes to hold. To make it simple for you - I think it is very nearly impossible to hire a sufficient number of competent programmers for *any* project, mainly because the testing of competence in programming is incredibly expensive, and competent programmers are INCREDIBLY rare. So you have to accept that up front. You can couch this in management terms if it makes you feel better paid.

Restating every argument as a variation of "which is a stuck up way of saying " also makes you look like a bit of a loon. Having technical knowledge and opinions is not, in fact, "stuck up". Argue on the level of the discussion, rather than just spouting management text book garbage.

Re:The *big* problem with GNUStep...

[jcr]

It's nostalgia as far as I'm concerned, because I've moved on from OpenStep to Mac OS X. For Linux though, GNUStep is still the best chance it has for a decent GUI and OO development framework, so it's import if Linux is going to move beyond the server room and embedded apps to any significant degree.

-jcr

Re:The *big* problem with GNUStep...

[nuzak]

> Few "got it" how easy this was to use

If you have to "get" how easy something is ... it isn't.

I like my menus horizontal, and my scrolling list of icons, well, not to scroll, but vertical if possible. The newfangled "mousewheel" thingie makes vertical things even more desireable. I also think grey 3-d relief checkmarks on grey checkboxes is not really a good UI decision. And speaking of 3d effects, tastes change, so let's ditch the grey "bumpy" look already, mmkay?

Re:The *big* problem with GNUStep...

[be-fan]

It's a bullshit excuse C++ programmers use to divert attention from the fact that C++ is a crappy language. First, C++ is not particularly powerful. Any language in the Lisp family is an order of magnitude more powerful than C++. Second, its not complex because its powerful, its complex because its poorly designed. Template "metaprogramming" makes C++ people feel elite because of its sheer obscurity, but you can do more a lot easier with a Lisp-like macro system.

Re:The *big* problem with GNUStep...

[644bd346996]

Right now, the biggest barrier to GNUStep having source-level compatibility with OS X programs is the build system - there is not a good replacment for xcodebuild. GNUStep recently added support for OS X nibs, so GUIs do not have to be rebuilt. As far as the actual source, GNUStep has all most of the important stuff, but it lacks things like quartz and coreaudio.

Source level compatibility is not a pipe dream by any means. Already, porting from OS X to GNUStep is fairly straightforward, with a few caveats. Porting from GNUStep to OS X is trivial.

Re:The *big* problem with GNUStep...

[644bd346996]

First, the comment about "getting it" was about GNUStep/Cocoa as a development platform, not just as a desktop. InterfaceBuilder was revolutionary when it was released, and most clones are not as good because the underlying apis are ill-suited to the task. Even today, Xcode/IB allows some really neat stuff that cannot be done easily on any other platform. Case in point: creating a simple web browser without writing a single line of code. That is 100% code reuse. Once you stop trying to write windows programs in cocoa, it really is easy.

Second, the look of the GUI is just superficial. Most of what you don't like can already be changed, and plenty of people are working to make themeing simpler. Mac-style menus are possible, but the window manager doesn't quite behave with them yet. I'm not sure where specifically you think scrolling is horizontal but should be vertical. Please elaborate on that one.

Re:The *big* problem with GNUStep...

[borgheron]

There is a tool in the GNUstep toolchain known as pbxbuild which works well for this purpose. It's currently experimental, but it is making progress.

GJC

Re:The *big* problem with GNUStep...

[lieven_dekeyser]

PyObjC and RubyCocoa are both open source projects, and they will remain open. Cocoa-Java apps still run on 10.4, and will continue to do so, but new features will not be ported to Java.

Re:The *big* problem with GNUStep...

[KonoWatakushi]

If you really look at ObjC objectively, it is kind of disgusting. Don't get me wrong, it is usable, and will "normally" get muppets in to less trouble than C++, but it is C with bolt on bits of smalltalk syntax, and a very odd type system. And the tool support is quite, quite tragic.

I do not think you are looking at ObjC objectively; in no way is it disgusting. It is a very simple, elegant, and powerful extension of the C language. The dynamic typing system may seem odd to you, but it is extremely flexible, and largely responsible for the beauty of Cocoa and GNUStep. The Smalltalk infix syntax may feel foreign to a C programmer, but it in the end, it is compact, and expressive. Once you use it, you will never want to go back. It is easier to code, and much more readable.

As far as tool support, it is standard in gcc; what more do you need?

By the way, if ever anything had a bolted-on feel to it, that would be C++.

Re:The *big* problem with GNUStep...

[fm6]

asinine marketing/pricing on the part of NeXT Inc.

Marketing, I don't know about. But as for pricing, they probably did the best they could. When you create a proprietary hardware platform, you're forgoing the economies of scale you get with a commodity hardware platform, and that makes everything more expensive. Which is why there are fewer proprietary hardware platforms every year.

The Wikipedia article on the company says that NeXTStep was originally meant to be a Windows toolkit. If that's true, then their expensive detour into the hardware world was their fatal mistake. By the time they finally created the toolkit that they should have started with, Microsoft and Borland owned that market. When you create a new technology, you have to get into place before the market window closes. After that, it doesn't matter how superior your technology is, you can't fight the lock-in that your competitors have created. Call it the QWERTY effect.

And even when you hit the market at the right time, you have to persuade people to retool to use your technology. Which means that you do not make it any harder than necessary for people to code against your API. Which means you don't totally reinvent your programming language and paradigm the way the NextStep people did.

Re:The *big* problem with GNUStep...

[644bd346996]

I've used it. It is good, but it is not ready to be advertised as the way to migrate an xcode project to GNUStep.

Re:The *big* problem with GNUStep...

[Art Tatum]

I think a "big" problem with GNUStep and Apple's Cocoa is its reliance on Objective-C, a language most people have never heard of.

If you could code in Cocoa using C++, C# or even a language like Python, PHP or Ruby, I think it would be a lot easier to convince programmers. I know that personally I was very taken-aback by Objective-C and it's frankly mystifying and strange syntax compared to the C++ I was used to.

I hear this all the time, and it still doesn't make any sense to me. I have yet to understand why anyone in this profession would run screaming into the night because they must learn something new. And, when it comes to syntax, it's not even that different. Object.DoThis() becomes [Object doThis]. What's so difficult about that? I also find the ObjC syntax much easier to scan through quickly, but that might just be me.

Maybe it's a case of "old dogs", but designing a UI with something like RealBasic or Visual Studio is much easier to my brain.

I've worked with Visual Studio 2005 and Interface Builder (both on OPENSTEP/Mach and Mac OS X). I hated Visual Studio's designer component. Could. Not. Stand. It. But IB always seemed so intuitive to me. It fit my mental model like a glove, and was, incidentally, the inspiration for Delphi, Visual Basic, and all the other RAD tools that followed. I always wondered why Borland, Microsoft, and everybody else who copied the idea couldn't achieve the same simplicity and elegance. I'm quickly coming to the conclusion that it's just two different development communities with completely different ideas of what software is all about. Full stop.

Re:The *big* problem with GNUStep...

[Blakey Rat]

I have yet to understand why anyone in this profession would run screaming into the night because they must learn something new.

You don't have to understand it for it to be true. Besides, I'm not "in the profession." I'm just a hobbyist.

It's not that it involves learning something new, it's that it involves something that conceptually I'm just not that good at grasping. I can't learn calculus, either, I've learned through many, many disappointing grades in college math courses. That's just the way I am. C++ I get. The Visual Basic and RealBasic window designers I get.

Interface Builder I don't get. I need to ask my more knowledgeable buddies about 50 questions about how it works every time I touch it. Then the next time I touch it 6 weeks later, I need to ask all the same questions, because my brain just simply doesn't absorb it. I apologize that I don't meet your standards, but that's just the way it is.

And, frankly, since .net does it all and runs on more computers, I think Apple's in trouble when it comes to development tools. If they really wanted to make an impact, they'd allow XCode cross-compile/link for Windows.

I hated Visual Studio's designer component. Could. Not. Stand. It.

That's fine, but you *do* realize that you aren't the alpha and omega of the entire computer industry, right?

I guess I really don't get the point of your post. I thought it was basically a "well, everyone who doesn't think and act like me is wrong!" Slashdot Special, but after re-reading it I'm not so sure.

Re:The *big* problem with GNUStep...

[Art Tatum]

I guess I really don't get the point of your post. I thought it was basically a "well, everyone who doesn't think and act like me is wrong!" Slashdot Special, but after re-reading it I'm not so sure.

Well, chalk one up for your intuition and mark me down with a loss for communication. :-)

I suppose I feel a bit confused as to why there seem to be so many people who, when coming into contact with NeXT technologies, have such a visceral dislike of them; while at the same time there is a smaller number for whom it not only seems comfortable, but demonstrably much more productive. There used to be a similar phenomenon with Blender, I think. I'm not really a 3D modeller/animator, so the discussions were always peripheral for me. But I seem to recall people who had worked with 3DS Max and some other professional 3D tools being blown away by how productive Blender made them, once they got past the unusual GUI layout. At the same time, there were a lot of other people who tried it once, said something along the lines of "I don't like this because it's too different," and walked away. Is it just that they didn't take the time to learn how to be effective with the tool? Or is there really some difference in cognitive psychology that makes certain tools more natural to certain people? (You could say, "why not just change the GUI to be more traditional, then?" The problem, I would presume, was that the "unusual" GUI was one of the keys to the increased productivity.)

I have been injured physically by common, but ergonomically poor, equipment designs. Yet when I was shown equipment that is more appropriate for the human body, I disliked it because it felt different from what I was familiar with. Fortunately, I eventually came around and now feel much better because I stuck with it long enough. But "cognitive ergonomics," if there is such a thing, may be different from physical ergonomics.

Re:The *big* problem with GNUStep...

[be-fan]

How do you think people become bitter Lispers? Years of using C++!

Levity aside, I've observed that nearly all Lispers have years (or in many cases decades) of experience in C or C++ (or more recently Java). I personally liked C++ just fine, until using it on a research project over several years. C++ ate shit in that environment. It was terrible for rapidly developing complicated algorithms, and even worse for making modifications in response to changing requirements, which are unavoidable in a research setting where you often don't know what you need until you build it. My appreciation of Lisp developed from seeing all the ways it avoided the weaknesses of C++ we encountered working on that project.

Re:The *big* problem with GNUStep...

[aperezbios]

> And still could be someday - after all, Linux desktops are such a melting pot of different toolkits and environments, that perhaps some "killer GNUStep apps" (graphics apps,
> like an Illustrator clone would be a good start) could get people to notice GNUStep again.

Well, fancy this, there *IS* a GNUstep killer app, called Cenon, available at http://www.cenon.info/ . It also happens to run under OS X.

Re:The *big* problem with GNUStep...

[aperezbios]

There is an EXCELLENT Illustrator alternative for GNUstep, called Cenon. There is also an OS X version. You may find more about it at http://www.cenon.info/frame_gb.html

Please explain if you know...

[pembo13]

what is useful about GNUstep? I read throw the Wikipedia article on it, and I got the impression that it is yet another GUI toolkit. So I am curious as to what makes it more useful in which situations: ie. where does it shine?

Re:Please explain if you know...

[m0llusk]

GNUstep is a GUI based on a toolkit and basic set of system services implemented using Objective-C. Because of the elegance of Objective-C, the design of the toolkit, and the architecture of the services, the experience of use is enhanced for ordinary usage and high level development and points in between. GNUstep emerged from the OPENSTEP standard.

Much of Mac OS X Cocoa was derived from NeXTSTEP, so there is the possibility of some level of compatibility with Mac OS X. In some ways GNUstep might be considered superior because of simplicity resulting in the lack of integration with Carbon which was done on Mac OS X for partial compatibility with previous Mac OS versions.

Re:Please explain if you know...

[Anonymous Coward]

Amusing you assume everyone aspires to to be associated with 'artistic' types. In reality they're mostly vapid poseurs, interested in being seen to be seen and desperate to culture an outward appearance that detracts from their inner-dullness.

Re:gnustep

[WillAdams]

People who value a nice, open development environment and the integrated and synergistic environment which such creates. Consider a typical work-flow in NeXTstep:

  - write an article in TeXview.app
  - select a word, hit = and get a definition / thesaurus entry while writing it
  - create a drawing in Altsys Virtuoso which needs an equation in a label
  - copy the proper equation out of your .tex file from the TeXview.app window
  - paste in the equation into Altsys Virtuoso
  - invoke the Service TeX eq -> eps in Altsys Virtuoso and get a .eps of the typeset equation (you can send the source to a background layer for reference (what I usually do) or delete it.
  - select the address of the journal receiving the article
  - invoke Poste.app to bring up a window from you you can print an envelope to mail it for submission

The environment affords similar integration w/ Mail.app as well if desired.

The commercial developer Nova Mind, http://www.nova-mind.com/ uses it to get a Windows version of their Mac OS X software.

And for those who say just use Mac OS X (I do at work):

(from: http://macslash.org/comments.pl?sid=4190&cid=63590 )

    - monolithic main menu bar w/ wasted blank space between the menus and the (optional) information / settings menus for Airport &c.

    - verbose Mac-style shortcut descriptions w/ arcane symbols instead of concise NeXT-style shortcuts (in NeXTstep, Save is indicated by ``s'' and Save as by ``S'', no Command symbol (it's assumed---Control only as a modifier is reserved for personal shortcuts / Unix-use), Shift by case)

    - Print, Hide, Services and Quit are no longer top-level menus where they made more sense and were quicker to get at.

    - scroll bars on wrong side (this can't be fixed by theming 'cause Carbon apps are responsible for deciding where scroll bars are placed :( having them on the left means a window is more useful when partially dragged off-screen and results in less-frequent need to resize a window

    - no Webster.app (this has since been addressed w/ 10.4), Digital Librarian / Shakespeare or Oxford's Book of Quotations --- in NeXTstep this meant one was guaranteed to have Command = _not_ used in an app so it'd be available for looking things up in Websters

    - Pantone colour library --- used to be this was licensed w/ the system, now each graphic app which needs it has to pay a license, and one _doesn't_ get them in one's office apps (major negative for adhering to corporate identity programs where such is specced)

    - vertical menu

    - pop-up main menu --- this is wonderfully fast / efficient / elegant. For me, ``Punch'' in Altsys Virtuoso is pretty much a gesture, right-click, down a bit, then straight over and release

    - repositionable sub-menus --- no need for inscrutable button bars, and one can make a given command easy to get to as needed (when doing lots of envelopes I tear off the poste.app Services menu, put it in the bottom left corner, then an envelope is merely a selection, mouse move to bottom left, click, shift right to the print menu (also aligned on the bottom edge for this) click away. (takes longer to say / type than to do)

William
(who really should save all that and put it on a web page or something instead of typing it up each time --- check my rants at http://groups.google.com/ in comp.sys.next.advocacy to see if I forgot anything...)

Maybe they can fix ....

[phoxix]

Their annoying usage of a top level conf dir ~/GNUStep (or whatever it is). No other app I've seen does such garbage, dot-dirs all ftw.

Re:Maybe they can fix ....

[m0llusk]

Sure, or you could. The code is open and the software makes it easy to put all that stuff where ever is best. The only problem is agreeing on where is a location, what format is best and so on. There are a number of variations available if you poke around.

Re:Maybe they can fix ....

[squiggleslash]

In theory (I haven't tried it), this is configurable: The GNUSTEP_USER_DIR, the default of which is set in {GNUSTEP_HOME}/Library/Makefiles/GNUstep.sh, has the name of the GNUstep directory relative to your home. ie, if you're prefer ".gs", place "export GNUSTEP_USER_DIR=.gs" in your .profile, or if that doesn't work, in Library/Makefiles/GNUstep.sh.

Re:Maybe they can fix ....

[RAMMS EIN]

Oh well. I never actually look at the files that are put in my home dir; there's too much crap there anyway. And dotfiles are doubly evil, because they both clutter the home dir and get in the way when you don't want them to, and are hidden and often ignored, so that you forget to process them when you need to (e.g. when making backups).

I put everything under three subdirectories of my home directory: local, private, and public (with appropriate permissions), with some frequently accessed things having symlinks there, as do files that are expected to be under $HOME, such as dotfiles. public and private are shared among multiple computers, whereas local (which contains lots of temp files, as well as projects which actually are shared, but through subversion) is not.

Re:Maybe they can fix ....

[dircha]

"Their annoying usage of a top level conf dir ~/GNUStep (or whatever it is). No other app I've seen does such garbage, dot-dirs all ftw."

Insightful?! It's a fricking directory. You have thousands of them. Can you give one single reason why any sane, non-obsessive compulsive person would care that the GNUStep directory doesn't have a dot in front of it?

And this is "garbage"? This is your biggest complaint with the framework?

Seriously. Are you obsessive compulsive?

Objective-C

[metamatic]

Well, having programmed in both, I'd take Objective-C over C++ any day. If you know C and an OO programming language, you can pick up Objective-C in a weekend. The same cannot be said of C++!

That said, I'm looking forward to automatic memory management being added to Objective-C.

Where to begin?

[Tony]

In 1985, Steve Jobs left Apple to found a new computer company. His company attracted many very talented individuals. They created the NeXT computer, a very advanced, very beautiful computer running a Unix-like operating system eventually called NeXTStep, which eventually became OSX.

I use NeXTStep as proof that Microsoft has set the computer industry back 12 years. NeXTStep used display postscript on both the video display, and for printing. It was fully-preemptive, with a clean, POSIX-compliant system interface. The application framework was extremely advanced, and extremely easy to code for. Using Objective-C as the programming language of choice, NeXTStep had some very advanced programs for the time, such as Lotus Improv, the spreadsheet MS-Excel wishes to become when it grows up.

As it is, MS-Windows still lags behind NeXTStep by a good amount, especially in terms of ease-of-development, ease-of-use, and aesthetics.

Tim Berners-Lee created the World Wide Web on a NeXT.

Anyway, when it became apparent that NeXT was not going to survive, they released a bunch of specifications that together made up the technical documentation for OpenStep, based on NeXTStep. The idea was that OS vendors could implement OpenStep APIs, and application vendors could target a single API for multiple OSs.

GnuStep is an implementation of the OpenStep API, and other programs to recreate the NeXT environment on any Unix-like operating system. Applications written for GnuStep can be recompiled to target OS X with little-to-no work.

Basically, when people say Linux needs an easy-to-use, easy-to-develop-for application environment and desktop, they are talking about GnuStep, whether they know it or not. It's not as flashy as GNOME or KDE, but it's much cleaner, easier to develop for, easier to use, and much more consistent. Where both GNOME and KDE try to be similar to MS-Windows, GnuStep tries to be like NeXTStep, the best application development and user desktop ever created.

Re:Where to begin?

[WillAdams]

Tony said:
>Tim Berners-Lee created the World Wide Web on a NeXT.

Other interesting programs which began on NeXTstep:

  - FreeHand v4 (essentially a port to Windows and the Mac of Altsys Virtuoso v2)
  - Doom
  - Lotus Improv
  - Stone Design's Create
  - sBook

An interesting opensource app w/ NeXTstep roots:

  - http://www.cenon.info/

William

Re:Where to begin?

[gdek]

"GnuStep is an implementation of the OpenStep API, and other programs to recreate the NeXT environment on any Unix-like operating system. Applications written for GnuStep can be recompiled to target OS X with little-to-no work."

Does this mean that the converse is true? i.e. applications written for OS X can be recompiled to target GnuStep with little-to-no work?

Re:Where to begin?

[Frumious Wombat]

Microsoft is certainly one offender, and certainly the largest, but let's not forget our dear friend, X11. Pre-Irix-4.0 SGI ran NeWS, which was also display postscript based. Nexts ran NextStep, VAXStations VWS, all relatively light, efficient, and functional, and everywhere I went the same whine arose, "we can't use this, it doesn't run X11!" So, Nexts weren't purchased, even though given the software and performance, they weren't out of line versus Sun 3/60, Apollo, etc, and SGI had to port everything to X in order to survive. We took a performance hit on every machine that had to run X versus the previous window-manager, had to add megs of expensive (early 1990s) memory just to not hear the disks whine, and generally gained very little in return for adopting this, ahem, standard. Then, if you wanted to see real death by toolkit, running Motif on a Vaxstation 3100/38 that had run smoothly under previous versions of X alone was a good example. We had a program that *somebody* insisted had to be Motif only, and the performance was so apalling we spent ~$15K on an Indy, just to be able to work. That VAX was perfectly fine (and would have still been useful if the programs the lab used had an X11-only, VWS, or even Tek-4107 interface), but it had to be retired due to a bloated toolkit.

The Knight with the Chicken is going to be very busy in the computer industry one of these days.

Failure?

[jcr]

NeXT was bought out for $400 million. How can I make my business a "failure" like that?

-jcr

Re:Failure?

[jcr]

I was never a NeXT employee, I was a NeXT customer. As for your claim that I'm "rewriting history", please point out any factual error I have made.

-jcr

Re:Failure?

[jcr]

it was taken over largely so Apple could get hold of its charasmatic leader.

Actually, your statement above is rewriting history. Gil Amelio, following the recommendation of Ellen Hancock, made the call to take NeXT instead of Be, and he sure didn't do it to replace himself as CEO. Ed Woolard begged SJ to take the CEO position, after the board had decided to buy out Gil's contract and let him go.

-jcr

Re:Failure?

[jcr]

No, Canon invested $100M initially, and another $30M later. Jobs put in about $10M of his own, and Perot was in for $20M. Apple bought them for $400M. That's not exactly a shabby ROI.

-jcr

Re:gnustep

[_|()|\|]

Those interested in GNUstep as a poor man's Yellow Box may be interested in a younger, more focused project: Cocotron. It seeks to clone Foundation and Appkit, and to provide tools to cross compile for other platforms with Xcode. It's a little Windows centric, but support for Linux, Solaris, and others seems to be in the works.

Re:gnustep

[tenco]

TeXview.app, Altsys Virtuoso, Poste.app, Webster.app, Digital Librarian / Shakespeare, Oxford's Book of Quotations

I tried GnuStep a while ago and i never heard of any of these great apps you mentioned. Are they new?

Only app i knew from your description is Mail.app. And i think it's awful. It crashed alot, support for gnupg was buggy, it looked awful and had no real support for using it with keyboard only.

and...

[oohshiny]

it could go a long way to attracting Mac developers to Linux if they can accomplish ports of many Cocoa apps with simple recompiles.

And this is useful... why? What Cocoa apps would actually be of interest to Linux users and wouldn't be so tied into the Macintosh desktop that it would still be a lot of work to port?

A lot of the "big apps" are developed using various compatibility or wrapper libraries anyway (e.g., Skype, NeoOffice, Microsoft Office, AOL IM, Acrobat, Firefox, Thunderbird, Java, etc.)

Re:and...

[Watson Ladd]

Think commercial apps. The Objective-C and Foundation platform would look a lot better then .NET if it works on all platforms. The speed of native code, the FFI abilities of C, and the flexibility of Smalltalk vs. the speed of Java, the FFI abilities of Java, and the flexibility of Java with even less cross platform support. Only additional thing Objective-C needs is Lisp for it to be perfect.

Re:and...

[WillAdams]

TeXshop.

Granted, a lot of its niceness is lost w/ the loss of the synergy one gets w/ things like LaTeXiT, spell-checking, shared completion lists &c., but still, it'd be nice to have a tex previewer / editor w/ synchronization between a .pdf preview and the code.

William

revolutionary? nice?

[oohshiny]

was (and is), that few people realised how great the original NeXTStep environment - which GNUStep attempts to clone - was. [...] GNUStep (the structure of which is pretty revolutionary)

Neither GNUStep nor NeXTStep were "revolutionary"; almost all of the fundamental concepts and designs in those systems came from Smalltalk. They may have been better than C and Motif at the time, but they were still a poor imitation of Smalltalk.

Linux desktops are such a melting pot of different toolkits and environments,

I'm running Gnome right now, and there is not a single non-Gnome app running on my desktop or on my launchbar. I think KDE users have the same experience with KDE. The notion that Linux desktops are an inconsistent mix of different toolkits and interface styles is a myth.

And still could be someday - after all, that perhaps some "killer GNUStep apps" (graphics apps, like an Illustrator clone would be a good start) could get people to notice GNUStep again.

There are several Illustrator clones for Gnome and KDE; what makes you think that GNUStep can deliver something that is better in any way?

Nowadays, people code for OS X because OS X is seen as a hip system with a small but viable installed base, and the fact that the dev tools are extremely nice is just an added bonus.

It sounds to me like you haven't tried modern development tools. I have done some apps in XCode and Objective-C, and I very much prefer Glade+Python, Eclipse, and Monodevelop. As far as I'm concerned, Apple is way behind on development tools, and even their upcoming update isn't going to fix that.

Re:revolutionary? nice?

[99BottlesOfBeerInMyF]

The notion that Linux desktops are an inconsistent mix of different toolkits and interface styles is a myth.

Hm, My Linux install (KDE) has several Gnome apps I require.

There are several Illustrator clones for Gnome and KDE; what makes you think that GNUStep can deliver something that is better in any way?

The only argument against GNUStep I've ever heard that makes sense is that it adds bloat for servers. This is a very weak argument in my opinion. Here's my argument for GNUStep. A while back one of my remote coworkers needed an application I had been using. He didn't have a copy and the publisher had stopped distributing it. We were both using OS X. So I dragged the application into our chat session and he ran it. Problem solved. There was no searching for an installer that may or may not have been backed up somewhere.

Here's another argument for it. I upgraded my laptop from a PPC to an Intel and the installer automatically migrated my applications. Some of the applications it migrated already contained native binaries for Intel and they worked seamlessly without me having to download a new binary.

GNUStep/OpenStep is not perfect. It is better in several ways than anything else I've used. I'd like to see it extended even further so that packages may contain source, build instructions, licenses, and repository info for updates and registration info. I don't care about how pretty the windows are, just about real, objective benefits. I just wish people weren't so stubborn about rejecting real improvements because those improvements aren't ones they care about.

Re:revolutionary? nice?

[99BottlesOfBeerInMyF]

First, the two examples (drag-and-drop application copying, processor independence) you give have nothing to do with the *Step APIs, they have to do with other system issues in OS X.

Both of these functions are enabled by use of OpenStep-like APIs. With either Gnome or KDE it is very difficult to recreate this functionality.

Second, drag-and-drop installations and Apple's approach to processor independence are arguably bad designs;

Arguably, round rather than square wheels are a bad design.

the fact that you find them occasionally useful is far outweighed by the headaches they cause for many users.

What headaches? You can't just label something a bad design and then not mention a single, specific problem that that design has ever caused, especially when you're arguing against several specific benefits.

Third, none of those properties you like were pioneered by NeXTStep or OS X, and none are unique to them.

Who cares? I don't care if Adolph Hitler came up with the benefits. I just care which systems provide those benefits to me.

In fact, if you want drag-and-drop installation and processor independence, your best bet at this point is not an OS X app, but a Java .jar file or a .NET assembly.

OS X is an operating system that (usually) gives me these benefits. This is partly due to their use of OpenStep. So you think I should get those benefits instead by convincing all developers of all the applications I create to start developing in Java or .net and switch to whatever platform they use? And this makes sense in your little world?

It's your choice to make your KDE desktop inconsistent;

True, if I want to use the best application for each task, I choose to mix KDE and Gnome apps. That doesn't mean the choice I have to make is a good one.

...and it offers a complete desktop solution that way.

...for some subset of tasks I want to do. Complete means I can do everything I want. That just isn't the case. I'm not arguing this is a major problem, just mentioning that my experience differs from that of the previous poster who claims most people only use one.

For developers, GNUStep is less mature than the alternatives, has limited language and tool support, is considerably more effort to program in than common alternatives, and a tiny developer community.

So GNUStep isn't commonly used and hence has less widespread support and maturity, so developers shouldn't use it? Doesn't that kind of thinking lead to everyone always using the same outdated APIs with the same inherent flaws, forever?

For users, GNUStep is wildly inconsistent with Linux desktop conventions.

Yeah, it is different than what is already being used. That's the point, because anything that is fundamentally better will be different, sort of by definition. I reject your, arguments because they boil down to the fact that OpenStep isn't what is being used now, so it is nonstandard. Almost all improvements require moving away from what is currently dominant. If people use the same thinking process it won't matter because no one will switch away from Windows to a nonstandard OS like Linux anyway.

Re:revolutionary? nice?

[99BottlesOfBeerInMyF]

Drag-and-drop installations don't have dependency management.

They bundle the needed libraries and use dynamic linking. What's harder, always downloading from repositories and using a tool to manage dependencies, or just downloading and not worrying about it? This is a win for drag and drop, not a loss.

And Apple-style fat binaries waste space compared to more modern alternatives.

Yup, this is the only argument that makes sense, but even it is very weak. For a workstation, the disk space used is negligible. If it is a concern on a given platform, the OS can handle stripping out the unneeded binaries so you only lose that functionality if you have that constraint.

Well, here you fail to give a reason.

Part of OpenStep (within PDO I think) is the dynamic linking and object handling that makes both of these features default behavior. Doing the same thing with another set of APIs would require adding that functionality or writing it from scratch for each app. And don't get me wrong. I'm arguing in favor of this as a default, not for any given implementation that provides that functionality.

...because it's better.

Yeah, better by having fewer features and by being harder to use.

GNUStep, Cocoa, and OpenStep are outdated APIs with inherent flaws. That is why people are not using them widely. Java and .NET are better in every respect.

Well, in every respect except they result in inferior programs with lesser functionality, right? Show me the OS + API combo I can use today that gives me both of the above features, for most applications. Gee, that would none of them. That is called "inferior."

The supposed "benefits" you list exist in all modern platforms, but developers don't usually use them because they are considered bad.

The benefits I list aren't included because they're hard, not bad. The reason they exist for most OS X apps is because the OS and dev tools are designed to provide that functionality easily. You can claim it is inferior all you want, but if the end result of one gives me more functionality than the other, then your claims lack any weight.

Re:gnustep

[WillAdams]

I'm afraid those are all NeXTstep apps --- for GNUstep you'd need to see if you could port TeXshop, use Cenon instead, use Affiche instead of Poste (might need to add the envelope-printing service), there was a project called ``mylibrary'' to replace Digital Librarian/Shakespeare and you could use WordNet instead of Webster's. Not aware of an unencumbered book of quotations though --- Fortune w/ an index maybe?

William

Re:gnustep

[jcr]

Don't forget in-house custom app development. The easier it is to develop an app for a given platform, the better chance that platform has of being chosen. My company is a Mac shop, but if there was a 80+% Cocoa library available for Linux, we might very well choose Linux for certain vertical-market deployments. As it is, the GNUStep foundation makes Linux a possibility for our non-GUI apps (data acquisition, etc.)

-jcr

Apple for one...

[Junta]

Well, to be more accurate, they cared about NeXT enough to base OSX off of it. GNUstep is the most API-compatible option that can run on alternative platforms. Of course, its not because of the API that it is so interesting as a desktop platform, and developers and users have come to appreciate it in OSX.

The two main open desktop projects (GNOME and KDE) heavily mimick the user interface paradigm established by MS. GNUstep is a good complement with the NeXT (also OSX) user interface paradigm (separate menu, management of windows individually and by application, applications registering services for more complex/powerful gui actions than what is done by drag/drop, copy/paste, etc).

GNUstep/NeXT/OSX services are the only appropriate equivalent of command line pipes in GUI land, which makes it a highly logical fit for those who understand the beauty and power of pipes in *nix. For example, in Gnome/KDE if an application wants spell check, they need to implement it themselves or at least take in a library and hook things around it. In GNUstep, any text application I can highlight something, click services/spell check if I have a spell check app installed, and it will happen. People complained for a long time about browsers not having spellcheck, but with services implemented and used browsers would have had it for free. It's kinda like piping the output from some command into aspell. All kinds of interesting things have been done with services, and someone implementing something new and different ends up enhancing all the desktop software that is appropriate for it without extra effort.

I have used GNUstep many a time to see how they are going, and if the environment were more complete (i.e. a GNUstep web browser, and IM client, office software) I would use it as my desktop full time. I remember before gcc had objc++ and before gnustep & gorm had nib support, that those two barriers going away was expected to allow all kinds of wonderful porting from OSX (i.e. the OSX Firefox code, one of their IM clients, whatever else). I haven't seen any word on efforts since those developments. I would love to contribute, but my plate is too full.

The downside is that in GNUstep more so than KDE/Gnome, non-native applications are really jarring, without separate menu and not interfacing with services. WindowMaker does a good job grouping windows by application for application hiding, but it isn't enough. Also GNUstep is capable of doing a lot, but fonts, for example, are a pain in the ass (at last check with the decent backend with anti-aliasing you had to package fonts in .nfonts). Also GNUstep could probably accomodate more of the freedesktop specification than they currently do, however I do recognize that freedesktop specifications pretty much have the MS way of doing things in mind and therefore some things aren't appropriate for GNUstep. Also, GNUstep doesn't have a perfect window manager to use with it. WindowMaker is very good, but doesn't render menus/dock in a way that is guaranteed to be visually consistant with GNUstep. WindowMaker is probably the best effort to focus on moving forward, but there is work to be done.

If you work it, GNUstep is a lot further along than most people realize, but the fact you have to work hard to get a complete environment discourages new users. And even when all is said and done, things are a bit rough around the edges in spots...

I couldn't see much about it..

[Junta]

But is it merely trying to get Mac apps to compile and run, or actually duplicate the whole user interface paradigm? Honestly speaking, the ability to simply run Mac apps isn't exciting, for the most part we already have good equivalents. It is appealing to use those projects under GNUstep because they were designed with that sort of user interface in mind unlike GTK/QT apps, but the applications in and of themselves have no features not implemented in the counterparts.

If you limit yourself to Cocoa stuff this is true

[sqar]

So if you want to port to GNUstep don't use Carbon / QuickTime / CoreBlah stuff in your application. Also keep in mind that GNUstep has a certain lag when it comes to the latest features of Cocoa since the GNUsteppers always have to keep up with Apple (and since Apple NDAs new features GNUstep learns about those at earliest when Apple releases a new Version of OS X). Thats the reason they announce only: "Applications written for GnuStep can be recompiled to target OS X with little-to-no work."

What GNUStep needs is...

[Colin Smith]

Sexiness. It isn't at the moment.

It needs users to go "wow I want that" and for developers to go "wow I want to do that". Take a lesson from Apple and Microsoft here, make it look and sound good.

Re:What GNUStep needs is...

[644bd346996]

It would be easy to make GNUStep look like aqua. It could be made to look more aquaish than Vista. But that would probably piss off the apple legal team, which so far has played nice with GNUStep.

Re:What GNUStep needs is...

[Colin Smith]

1: You can pretty much rip off a GUI as much as you like, nobody's won a court case to prevent competitors copying them.

2: Take a lesson doesn't mean copy. If I wanted Aqua I'd just buy an Apple.

However a great looking GUI along with a genuinely intuitive user interface, which is API compatible with an Apple Mac is a compelling proposition. In fact I reckon it has more potential than Gnome or KDE. The ability to write a GNUStep app and then just re-build on a Mac (and viceversa) vastly increases the market for both systems.

Re:What GNUStep needs is...

[smash]

Agreed, and the more i read about objective-C, the more it seems like the "right way" to do things :) I'm no major application developer, but I do have a bit of programming experience... the flash demo for GORM that's online somewhere is very impressive.

Personally I'd love to see OpenStep take off in a big way - just yesterday in fact I was going through a bunch of old stuff in my email and was thinking "I need to try out openstep again this weekend" :)

Etoile-buntu?

[pschmied]

I've been casually flollowing the etoile development, and I've even gone so far as to (mostly successfully) build GNUStep and Etoile on my OS X-running Powerbook. It's clear that GNUStep has made some strides in recent times. Etoile seems to be proving that there are some with a vision of what a GNUStep *platform* could be. All in all, pretty exciting stuff.

The catch is, that integrating this stuff is a bit more work than your average ./configure, make, make install... I'd like to see someone pick a reference platform and target it for continuous integration that closely tracks Etoile and GNUStep development. My personal favorite would be some kind of BSD, however, I'm a pragmatist and realize that Ubuntu is probably the logical choice for such a task given its ubiquity and its history of eventually making 1st class citizens of derivatives (kubuntu, edubuntu, xubuntu, etc).

Fact is, I think that we'd start seeing more apps show up for GNUStep if we had a supported reference platform. You know, give developers some place where they could port their apps over from Mac OS X in peace without having to worry about spending "hacking day" compiling software rather than writing it.

Re:gnustep

[Alioth]

The game Oolite (see my sig) was ONLY portable to Linux because of GNUstep. If there was no GNUstep, Oolite would be Macintosh only.

It's a shame that GNOME and KDE have become the mainstream desktops - if GNUstep had become dominant instead, we'd have had easy compatibility with Cocoa in Linux.

GNUstep Base is also a very handy Objective-C class library, and is pretty much completely compatible with Mac OS X Cocoa.

Re:gnustep

[Anonymous Coward]

Wow, you're very close to being one of those "older is better because it's older" people.

- monolithic main menu bar w/ wasted blank space between the menus and the (optional) information / settings menus for Airport &c.

By putting it there, it's faster to use. The blank space isn't so bad, since you couldn't put anything in a space that shape/position, anyway. Look at anybody using NeXT and see how much time they waste moving their menu around (not to mention acquisition time for the menu items in the first place, since they just float there).

- verbose Mac-style shortcut descriptions w/ arcane symbols instead of concise NeXT-style shortcuts (in NeXTstep, Save is indicated by ``s'' and Save as by ``S'', no Command symbol (it's assumed---Control only as a modifier is reserved for personal shortcuts / Unix-use), Shift by case)

Yeah, ^S is really verbose. (And sadly, Apple keyboards outside of the US have those symbols on the keys, but that's a hardware issue, not a software one.)

- Print, Hide, Services and Quit are no longer top-level menus where they made more sense and were quicker to get at.

I'm not sure what you mean by this. (Are you upset because you think it takes too long to quit a Mac app? How often do you do that?)

- scroll bars on wrong side (this can't be fixed by theming 'cause Carbon apps are responsible for deciding where scroll bars are placed :( having them on the left means a window is more useful when partially dragged off-screen and results in less-frequent need to resize a window

When you partially drag the window off the right, then left-side scrollbars are more useful. When you partially drag it off the left, then right-side scrollbars are more useful. Result: this is not a deciding factor.

- no Webster.app (this has since been addressed w/ 10.4), Digital Librarian / Shakespeare or Oxford's Book of Quotations --- in NeXTstep this meant one was guaranteed to have Command = _not_ used in an app so it'd be available for looking things up in Websters

As you point out, Mac OS X does have a dictionary service now. If I used it more than once a week, I might wish it had a keyboard shortcut, but I don't. (There's probably a trick to add one.)

- Pantone colour library --- used to be this was licensed w/ the system, now each graphic app which needs it has to pay a license, and one _doesn't_ get them in one's office apps (major negative for adhering to corporate identity programs where such is specced)

I guess this is a downside, but most people really don't care. Mac OS still handles color far better than Windows or Linux.

- vertical menu

This is an advantage ... ?

- pop-up main menu --- this is wonderfully fast / efficient / elegant. For me, ``Punch'' in Altsys Virtuoso is pretty much a gesture, right-click, down a bit, then straight over and release

The Mac has a "mile high" main menu. This has been shown to be faster than a context menu, *even when it's on a different screen* (see Tog on Interface).

Then again, Mac developers always have the choice of using either one. If "Punch" is a common operation, developers are welcome to make it a popup menu, too. On NeXT, you can't make a mile-high menubar at all.

- repositionable sub-menus --- no need for inscrutable button bars, and one can make a given command easy to get to as needed (when doing lots of envelopes I tear off the poste.app Services menu, put it in the bottom left corner, then an envelope is merely a selection, mouse move to bottom left, click, shift right to the print menu (also aligned on the bottom edge for this) click away. (takes longer to say / type than to do)

I guess the Mac way is to use Applescript or Automator so you don't have to do the same mindless thing 1000 times by hand.

Or right-click and choose "Icon & Text" -- then your toolbars are no

I went to C++ in a weekend.

[Marbleless]

When I moved to C++ about 20 years ago, I had a very good book called something like "From C to C++" which outlined the basic bottlenecks/problems etc in C one at a time and then showed how C++ could be used to overcome them.

I remember starting work on Monday fired up with my shiney new C++ knowledge and was really able to push the project I was working on to a new level. It took about a week to convert the project to compile with C++ (sans classes) and then another week to make a big step forward by encapsulating some of the code into classes.

Re:I went to C++ in a weekend.

[metamatic]

C++ has grown enormously since then, and it's widely understood that it takes years of practice to be proficient in it.