I have been extolling the virtues of Obj-C and Apple/NeXT's Cocoa's APIs and its legendary user interface design since I first saw it. When I found out about GNUstep about a year ago, I kept asking myself why nobody else had picked it up.
I realized that the people who liked Obj-C and Cocoa just used OS X. GNUstep was a pointless waste of time for a Cocoa developer, but with OS X quickly seeming like the replacement for Windows on the desktop, I'm glad it's finally picking up steam.
But, as we said prior, the people that care just use OS X. Their target platform is OS X. The Linux users are happy with WINE and GTK+. There's nowhere for the project to go, and I fear that this Kickstarter will be a failed effort.
I'm putting down my cash for this. I want to see it succeed. I know it probably won't.
If it does, well, the rest is history. Maybe it will jumpstart Obj-C x-platform like Mono did for C#.
------
also, for those interested in the NeXT-style look and feel, the Window Maker (GNUstep WM) LiveCDs have had consistent snapshots and are always updated; based on Debian testing: http://wmlive.sourceforge.net/
It's not about NeXT or WindowMaker or anything like that. I don't care about the NeXT look and feel and as you can see from the video and the screenshots GNUstep has moved away from that look.
I should point out that there are anumber of companies who have moved their products to Windows using GNUstep. THIS is the point of GNUstep, to allow both development of new projects and porting of existing projects to other platforms without the need for much code change. There are two examples of companies which do this today:
http://www.testplant.com - Testplant ported their EggPlant app (a testing app which is script driven) which is the core of their company to Windows. NOW the revenue from the Windows version makes up more than 2/3rds of the company's revenue.
http://www.apportable.com - Got an Android? Got an android that runs any game which is familiar to you from iOS... well, guess what... chances are these guys probably ported it... and they used GNUstep to do it. GNUstep is currently running on millions of devices out there.
So the sentiment "just get a mac" in this context, makes absolutely no sense at all and, in fact, completely misses the point.
Using GNUstep to develop Android apps is a HUGE win. You really should push this angle more. Look at the success Xamarin have with their 1 language for multiple mobile devices solution. You have the same thing going.
you really should put the essence of this commend on the kick starter page. It makes it absolutely clear why I should care and why it is a cool project.
edit:
Both company's that you mention should just sponsor the project and/or pay you to work on it full time as this is the backbone of there business
Even on OS X the potential of NeXTstep remains largely unrealized as most apps do not take advantage of the APIs in a way that allows leveraging heterogeneous apps into a conceived workflow.
As a result, a fantastic opportunity to create an environment that is composable in a consistent, homogeneous manner is wasted.
What is worse is that such an environment does not exist anywhere in computing.
The closest thing that exists to realizing an entirely composable or scriptable environment is emacs, and that environment has not incorporated progress in graphics into its ecosystem.
Consequently, the capacity to compose workflows that interact with graphics is not possible, leading to a reliance on console style UIs for the most part.
I think that GNUstep losing steam has less to do with people who care choosing OS X and more to do with Gnome and KDE having a head of steam already by the time GNUstep came about, a characteristic that has lead to many cases of used technology being a consequence of circumstance rather than merit.
I just want an entirely scriptable, composable computing environment anywhere, and it is maddening that such does not exist as everyone would benefit from such.
Currently users either have to remain attached to a UNIX CLI userspace that does have composability or only have that composability in a small subset of applications due to the lack of a uniform interface in which to do so.
What computers are best at is what is least accessible in computing presently, leading to people having to do automatable tasks manually.
It's a tragedy that such is the case after decades of examples of how to achieve that being partially implemented.
Computing has even regressed greatly in this regard in the past decade due to giving our generated data to gatekeepers, when the data was local machines at least then we could apply rules to the data to transform it according to the user's whims, such as filtering out noise in a RSS feed, automating archival of useful information etc.
Now users are entirely reliant on the gatekeepers allowing them to transform information, often which is neither possible or comes accompanied with distracting ads or other attention parasites.
I would gladly switch to an ecosystem that realized the vision NeXTstep et al. had so that managing complexity and transforming data did not entail abandoning solidly engineered GUI applications in the process, sadly this is another case where computing has been unsuccessful in achieving ambitious goals.
grimwire.com/local is trying to solve that by turning workers into web servers and letting them power user plugin systems. I'm in the process of redesigning the docs and packaging it with webrtc tools, so it's getting some work atm, but here are some videos of things I've tried while working on it [1][2][3].
> But, as we said prior, the people that care just use OS X. Their target platform is OS X. The Linux users are happy with WINE and GTK+. There's nowhere for the project to go, and I fear that this Kickstarter will be a failed effort.
A modernised GNUStep could allow easy ports of a lot of MacOS software, both open and closed-source. That in itself might be enough to kickstart a non-MacOS ObjC/Cocoa ecosystem.
I don't really see why; GNUStep is already an established FSF project, and I see no need for a company. What it needs is a bit of attention, which is what this kickstarter will hopefully achieve.
I'm sorry to be a downer, but there needs to be a reality check, because future open source projects can learn a thing or two from this.
This project is unlikely to be successfully funded. In fact, "unlikely" is a mild word for just how unlikely it is to succeed.
Problem #1 - No clear value demonstration to the end user. The video is... not very good. The primary reason it's not very good is because it's asking people to read. People react instinctively to faces, to the sound of a person's voice, and to feeling connected with them. It's no coincidence that virtually every successful Kickstarter campaign contains monologues by their campaign creators. Thus, while there may be a value proposition embedded somewhere in the video, it's probably lost on most people who watch it because they just aren't really identifying with what is being presented.
Problem #2 - a $25 minimum price point. The gamedev industry has repeatedly proven that by enabling people to contribute $5 and $10, you reap about 20-40% more income than you otherwise would have. This truth isn't restricted to the domain of gamedev. The evidence for this is that virtually every successful Kickstarter project has low-tier contribution levels, often contributing a significant amount to the bottom-line of projects that aren't going to deliver a tangible product. E.g. this may not be so important for projects like Soylent, where the end-user will receive something tangible, but it's pretty important for most projects that weren't set up with the goal of taking preorders.
Problem #3 - Too high of a threshold for funding. $50k is not reasonable when the value proposition is so unclear. I could see this project reaching $5k or maybe even $10k. But it's not reasonable to calculate how much it would cost you to work on it, and then use that as the funding goal. "What would the crowds be willing to pay?" takes priority over "How much would this cost me in terms of my time, and what is my time worth?"
GNUstep isn't a game, it's a platform. Additionally, since we're on the subject of games there have been games which have asked for $50,000 and gotten $2M. I don't think my project will get that much, but I do believe people understand that bringing GNUstep which is mostly up to about 10.4 compatibility to 10.6 is a non-trivial thing.
Problem #1, the video is all I could do. I am not very good at doing videos, it's not my profession. I have little time to hire a cameraman or do professional video editing and I'm not all that photogenic to boot, so it might be a good thing you don't see me on camera.
Problem #2: Easily fixable. I can add other incentives while the project is running.
Problem #3: If the goal isn't reached, then it becomes a matter of it's not even something that's reachable since I must have enough time freed (by having money) to complete what needs to be done. Without the time, the work can't be done at an accelerated rate since I will need to continue to do it in my spare time. So, while the funding goal is high, it is not unreasonable for a project this size.
You're adding to the problem here instead of taking the constructive criticism to heart and taking action to make things better.
This project is now a business seeking outside funding and should be conducted as such. There are no excuses and perception is everything. Instead of an itemized (note: this projects as petty) response, you need to fix it and say "thank you" to sillysaurus for the valuable input.
You've since added $1 minimum contributions. Awesome! What does that person get? An honorable mention? Their name on the 'Founder's List'? There's nothing in the tier descriptions...
The overall message is still unclear. How does Johnny MacBook benefit from all of your hard work? What is the point?
I want this project to succeed, and that is where the feedback comes from. It's obvious to me and everyone else here that you've put a lot of time and effort into this project. Please don't make the mistake of thinking that it's enough to build it. If they don't know what it is and what it does for them, they won't come.
Excuse me, but I am not one to simply say "thank you" when some of the suggestions therein don't make sense.
The video IS all I could manage to do. I don't have the equipment or the skills to make a professional looking video.
The second point does make sense, I'm acting on this.
As for the third... he acts like I'm supposed to ask for $10,000 for such a goal. GNUstep is an very large project. Getting it to 10.6 compatibility is going to take time and I can't take the time off of work unless I haver the money I need to make it happen. Asking me to do this for anything less than what I projected is ludicrous beyond belief.
So, yes, I realize where the input comes from and I respect that. I will, however, not simply take those comments and act on them when some of them are unsupportable and untenable.
All of that being said I am planning on adding a video in the next couple of days (most likely tomorrow) of me talking about the project so that people have a face to put to it.
I seriously do thank all of you for your input and I apologize if my responses seem combative. It was my understanding or impression, however, that this was not simply a case of me taking input and mindlessly acting on it, but, instead, a discussion of what's best to help the project succeed and that does include my responses to your feedback.
> Excuse me, but I am not one to simply say "thank you" when some of the suggestions therein don't make sense.
You should. There's no downside, but there's plenty of downside to being brusque. You don't need to act on the suggestions.
> As for the third... he acts like I'm supposed to ask for $10,000 for such a goal. GNUstep is an very large project. Getting it to 10.6 compatibility is going to take time and I can't take the time off of work unless I haver the money I need to make it happen. Asking me to do this for anything less than what I projected is ludicrous beyond belief.
He's saying that unless you recontextualize the project, you're asking for too much. Potential donors will have their expectations anchored by (a) their ideas about Kickstarter projects in general, (b) their ideas about similar Kickstarter projects, and (c) their understanding of the value you'll be providing relative to their needs.
If your goal is high relative to similar Kickstarter projects you have to make it appear worth that premium. If it's unclear what value your project will add to your donors' lives that will make your job that much harder.
That's what he's saying.
By the way, if you're asking for donations to pay for the time you're working on it and not the hard costs of the project then it's absolutely critical you have a video with your face on it and not just text. Folks are really donating for your labor, not to for this abstract project.
You're correct. Much of this does need to be clarified. It wasn't my intent to be brusque, just disagreeing where I saw cause to. Thank you for your feedback.
I know this is HN, but not everyone is trying to "make a business" from his hobby project. Couldn't you consider that the campaign author is just an engineer that wish he could afford to spend more time on something he likes to do and is passionate about? That's why he is able to offer a year of work on undercompensated rate. The same moment when he tries to "do it like a business", the fun is over, the joy is over, and it's just a work. And I doubt any remotely competent software engineer needs to bother with kickstarter to do the boring worky thing.
Those were always there. They are always there. The top-level commenter was complaining about reward tiers. Apparently he thinks developers aren't incentivized enough by the work getting done on the platform they want to use. I think that's pretty bizarre.
> How does Johnny MacBook benefit from all of your hard work?
He doesn't in any way you could successfully communicate to "Johnny MacBook". The drive is squarely aimed at exactly the people who would most directly benefit from the work being done, and be the most likely to contribute: Developers.
> This project is now a business seeking outside funding
Huh? No it isn't. It's sponsored improvement of an existing open-source project; it might very well be useful to businesses, but is in no way a business itself.
I'd like to contribute to this, but perhaps you could outline how you plan to go about it and if you have buyin from many other GNUStep developers? Apologies if this is in the video; can't watch that at the moment.
To be honest, I wouldn't worry too much about the rewards; most people contributing to this will likely do so because it's something they'd either like to see or would find useful themselves.
> Problem #3 - Too high of a threshold for funding.
I looked at this and had the opposite reaction. $50K is between 50 and 75% of a full-time developer's salary in NYC for a single year (I believe it's similar in SF). Are we expected to believe that that's all it would take to get Cocoa libraries running seamlessly on GNU?
Granted, I've never dealt with Cocoa at all, so maybe I'm overestimating this, but it seems like a lot of work.
I mean, how can you take a 15-year, multi-person project, and then assume it's going to be completed with a single person-year[0]?
[0] I don't mean literally a single person - I'm referring to a person-year being a unit of "work" (ie, the theoretical work equivalent of one person for a year).
I'm also skeptical, but I think the comparison with regular dev salaries overestimates the money that would be needed, at least for some projects. There are many volunteer contributors to open-source projects who would be willing to move to working on them full-time, or at least more-time, for much less than market rate. For example one of the SBCL devs did a lot of work on the compiler for only $16k (http://www.indiegogo.com/projects/sbcl-threading-2011), and the Git-Annex Assistant guy worked on it for a whole year for $25k (http://www.kickstarter.com/projects/joeyh/git-annex-assistan...).
Granted, if you have to hire devs not already part of your project at market rate, things will go way up from that.
> Are we expected to believe that that's all it would take to get Cocoa libraries running seamlessly on GNU?
That's not what the project plans to do. Apple's implementation of Cocoa isn't open-source. GNUStep already has large parts of the Cocoa APIs implemented, but they're very outdated. I think fixing this on $50k is extremely ambitious, but note that the project says they'll be brought up to at least 10.6 (which would be good enough to allow easy ports of a lot of Mac software), not all the way to 10.9.
I think GNUStep already implements the public Cocoa SDK for 10.4. Though I still think the jump to 10.6 will take longer than a year, if that means supporting Core Data, GCD, bindings, etc.
That's not the point. They're still human and humans still react better to faces than to words. I had the same complaints. I think GNUStep is a great project, but I watched the video expecting to learn what made the developers excited about it and instead I got a powerpoint.
I don't. I have no interest whatsoever in what your face looks like, whereas I might be interested in your ideas. In this case I actually skipped the video and just read the text summary, expecting it to have better information density. My suggestion for improvement would be a little more discussion of what concrete benefits would be provided by the availability of the proposed software, and evidence that the work can be done for such a small sum of money.
Just because someone's an engineer doesn't magically make people throw money at their project. It's _understandable_, perhaps, that an engineer might overlook the need for good marketing, or not know how to go about marketing something well - after all, for probably the vast majority of software engineers (or other types of engineers!), marketing is just not something they do often. But that doesn't make it 'lame criticism' to point out why such a campaign may do poorly; hopefully the author of the campaign uses this criticism to improve on the campaign so that it might succeed.
>> Just because someone's an engineer doesn't magically make people throw money at their project.
No one said that. It is a lame criticism - focus on whether there is a need for the project, is it possible in the time frame, etc. Substance over style, please.
It's also worth mentioning that the project is already %2 funded approximately 5 hours into the campaign and word has not yet hit slashdot or other websites yet.
Most Kickstarter donations come in at the very beginning and very end of a campaign in an "inverted U" shape. If you're expecting the contributions to come in linearly, I'd recalibrate and plan accordingly.
I think, here, the rewards are more tokens than anything else; this isn't really a product-oriented Kickstarter, except arguably for people who are interested in Darling. Min contribution is actually $1; there's just no reward.
> Too high of a threshold for funding. $50k is not reasonable when the value proposition is so unclear.
I'm not too sure about this. A lot of people, myself included, would like to see GNUStep brought to parity with modern Cocoa. It's never going to be a mass appeal project, but I can see it making its goal.
Étoilé svn is actually rather active. I suspect the reason the site has languished so badly is that the focus largely shifted away from the desktop environment and towards producing a development platform with tighter involvement in GNUstep's own development.
I am the chief maintainer of GNUstep. I wrote most of AppKit and all of the InterfaceBuilder equivalent Gorm. I run the GNUstep free software project, I was appointed as it's lead by Richard Stallman.
Gorm keeps giving me segfaults when I try to drag files from it to ProjectCenter (or vice versa). Even when I open any interface builder file in Gorm, it segfaults. I run Debian Wheezy.
Ouch! I'll test this and see if I can recreate/fix it. One thing to make certain of is that you're using GNUstep from subversion. The debian packages are almost always about a year out of date, unfortunately.
I had my first exposure to Objective-C using GNUstep on a Linux box in about 2002. I am grateful that it existed and it was there for me to be exposed to it; I was pretty young and probably could not have afforded a Mac in those days.
As objc and Apple itself has gained so much steam over the years, it's kind of funny to have learned it long ago via a clone of NeXT stuff. Most folks working in the language have no idea that it could possibly exist outside of Apple and tend to think of all the framework underneath them as a black box. I think maybe some of these folks would learn something by looking at an open source re-implementation of these NS* classes, or attempting such an endeavor themselves, even if on a small scale.
Also: objc is just too interesting for Apple to hog it all for themselves, or for it to be a one-implementation language.
I don't know if I'll go quite that far; I'm not going to put on objc blinders to the point where I'm unwilling to see what other platforms do well.
But man, seeing objc and comparing it with its peers across multiple decades... In a world where most people were doing C++ and doing it badly[1], or doing Java and doing it badly, it was very eye-opening to see this tiny little runtime that adds objects to C and doesn't do a bad job of it.
[1] and I'm talking the pre-STL, pre-C++11 at that.
My biggest problem with GNUstep is that there is never any mention of an OS X build.
There was one, but it is pretty broken and answers to inquiries were basically answered with "why would you even want to do that"?
Quite simply: for testing purposes! Most Cocoa applications are developed on Macs, so being able to build them with the "other" framework on the same machine would help alot.
While I'm glad to see the effort to enable more cross-platform application development, I wonder why the open source community doesn't produce its own cross-platform toolkit to rival those made by Apple and Microsoft. It seems that far too often the vision behind open source projects is limited to copying successful closed-source projects. I understand it's much harder to launch an open source project in an altogether new category, but it's that much more exciting when it does happen. How can innovation be encouraged in open source? It seems somehow there needs to be a greater potential for financial reward to encourage such developments.
Loads of radical innovation takes place as Open Source. Git (the software and the philosophy it embodies) being a really good example.
As to innovation in software frameworks, it is not going to take place on the Desktop, it’s taking place in the browser—where every major burgeoning framework is Open Source.
The video says Elegant, then a picture appears and all I see is "old".
Then the video says Simple, and other qualities, and all I see is pictures of something appearing "old".
I mean, I was there 15 years ago, and Windowmaker on Debian 2.[0|1|2] was my environment of choice.
I loved it.
But as I write from my new MBPr, I see pictures of something appearing "old".
That's why I hope the campaign is funded, but I know that it can and will be funded mainly by developers. And developers not caring much for that part of the UX which we call look & feel.
Unless of course somebody is doing a better video with more modern screenshots while I'm writing.
I've backed this because I'd like it to succeed. However, I think Cocotron had a far better chance at succeeding with their philosophy of using Xcode and cross compiling to other platforms. This drastically reduces the barrier to entry for the majority of Cocoa developers and lets them use Apple's (excellent) developer tools. It's a shame there isn't more collaboration between the two projects but I understand this is down to licensing differences.
Allow me to get this straight... you would like to see a complete analysis of everything that's missing between the current state of GNUstep and 10.6? I can produce that, no problem but it is very boring and not worthy of either my blog or a kickstarter campaign.
No, it certainly wouldn't be worth posting in either place, but a link would be cool, even if it was to a mailing list archive where it was previously placed.
A copy of the GNUstep Live CD with GNUstep and all applications pre-installed on an instance of the latest version of Debian Linux.
Estimated delivery: Feb 2014
GNUStep has been stagnant for years, it is obvious that nobody finds it interesting. I understand that developers are willing to put up with some anachronisms to make money off iOS/OSX , but they clearly opted away from them for open source. I don't see how money is going to change its appeal.
"GNUstep features a cross-platform, object-oriented development environment. Like Apple Cocoa, GNUstep also has a Java interface, as well as Ruby, Guile and Scheme bindings."
If you are able to get WebObjects 4.5.1, then you can take a glimpse of how GNUStep should've been done on the Windows.
Sorry, but detachable menus, while great (and I love them in Autodesk Maya, but they start out as regular ones) would not make lots of people happy. Even Apple more than 15 years ago realized that, and their menus on Windows looked like Windows.
This is what for example ActiveDeveloper, an IDE based on WebObjects looked like:
TL;DR: It would be great to see GNUstep catch up with Cocoa, the guy who run the campaign IS the project lead, but the only thing that may prevent the campaign to succeed is the very poor marketing level.
From what I can understand (which I wanted to also note -- half the time I'm really not sure what GNUstep is -- correct me if my understanding of what he's trying to achieve is wrong), he's trying to make extremely portable code that will run identically in any OS right?
TideSDK has made pretty good progress on that front, by leveraging web tech. He could maybe save a heck of a lot of time by porting TideSDK to cocoa and then pushing that
I think you're misreading the emphasis there. GNUstep is driven by an interest in objective-c/cocoa, not in making yet another cross platform framework.
Because there's something inherently useless about it, or because you personally don't care about the NeXTStep/Cocoa APIs? If it's the second, then it's totally irrelevant to the discussion.
The point of this is:
1) to have a working Cocoa port on Unix, Linux and Windows platforms.
2) to provide a way for programmers creating ObjC/Cocoa apps (and with iOS and the Mac App Store those are hundrends of thousands) to port their apps to Linux/Windows.
3) To move the GnuStep project forward.
4) To provide an additional set of nice, modern APIs on Linux. E.g. GTK3 is nowhere as good or feature complete as Cocoa (even 10.6 Cocoa).
5) To eventually give the possibility for cross-platform development of Mac/iOS apps (e.g games).
The kickstarter page doesn't explain why those things are all worthwhile, or why one system will solve all of them.
If they want to raise a moderately large amount of money they probably need to speak to more than just people who already totally understand what they're doing and why.
> 1) to have a working Cocoa port on Unix, Linux and Windows platforms.
OK but why? That's not a reason in itself.
> 2) to provide a way for programmers creating ObjC/Cocoa apps (and with iOS and the Mac App Store those are hundrends of thousands) to port their apps to Linux/Windows.
This may make sense, but there are important subsidiary questions:
- Will you cover enough of the APIs that a substantial number of apps can be ported?
- Will the coverage be sufficiently high-quality? For example, can you do animation through GNUstep that's as good as on platform native APIs?
- Will those app authors care about porting to Linux/Windows?
- Is this easier for them than just writing a frontend to the native API, as they currently do?
> 3) To move the GnuStep project forward.
This is a tautology. Why does that matter?
> 4) To provide an additional set of nice, modern APIs on Linux. E.g. GTK3 is nowhere as good or feature complete as Cocoa (even 10.6 Cocoa).
OK. Will one guy part time for a year overtake GTK3 to get to parity with Cocoa? What does "nice, modern" mean and why is GnuStep better? Providing the best API on Linux seems in tension with providing fidelity to Apple APIs.
> 5) To eventually give the possibility for cross-platform development of Mac/iOS apps (e.g games).
This seems basically the same as #2, except if you explicitly mention games then covering 3d and input APIs at high quality seems even more important.
I have been extolling the virtues of Obj-C and Apple/NeXT's Cocoa's APIs and its legendary user interface design since I first saw it. When I found out about GNUstep about a year ago, I kept asking myself why nobody else had picked it up.
I realized that the people who liked Obj-C and Cocoa just used OS X. GNUstep was a pointless waste of time for a Cocoa developer, but with OS X quickly seeming like the replacement for Windows on the desktop, I'm glad it's finally picking up steam.
But, as we said prior, the people that care just use OS X. Their target platform is OS X. The Linux users are happy with WINE and GTK+. There's nowhere for the project to go, and I fear that this Kickstarter will be a failed effort.
I'm putting down my cash for this. I want to see it succeed. I know it probably won't.
If it does, well, the rest is history. Maybe it will jumpstart Obj-C x-platform like Mono did for C#.
------
also, for those interested in the NeXT-style look and feel, the Window Maker (GNUstep WM) LiveCDs have had consistent snapshots and are always updated; based on Debian testing: http://wmlive.sourceforge.net/