Slashdot Mirror
Simply GNUstep Delivers UNIX, Simply
Eugenia writes "A new, Linux-based operating system released recently, called Simply GNUstep and it is based on the GNUstep architecture, originally built by NeXT (OpenSTEP) and is now also used by MacOSX (Cocoa). The alpha version of the x86-based OS is available for download and boots off the 110 MB bootable CD. The cool thing about Simply GNUstep is its partial source compatibility with MacOSX programs (further compatibility is still worked on) and its clean infrastructure, as it only includes GnuSTEP graphical applications like WindowMaker, Mail.app etc. You can read an introduction article of the OS at OSNews."
That's not meant as a flame. I'm just curious why they chose Linux instead of a BSD. I don't exactly care to follow the licensing terms behind all the different open/free software so I'm unaware if that's an issue, but using a BSD would seem to be a wiser choice being that they're "trying" to get an OS X on x86.
I dunno.
No sig for you!!
[sigh], if only OS X and GNUstep were just a little bit more like NeXTstep, and this wouldn't be a problem.
NeXTstep has something called "fat binaries". It means that the same application that runs on my 68040 NeXT machine will also run on my HP and Intel machines running NeXTstep. When I compile my application, I simply tell the compiler to compile for all these platforms.
Sure, the binaries end up being bigger (about 50% larger for each platform, if I recall), but it was completely painless, and you could use "lipo" to reduce the binary size (and make the application only run on a single platform) if you wanted.
Yes sir, this was state-of-the-art back in 1992. 10 years later, it's all but disapeared.
"Tomorrow's forecast: a few sprinkles of genius with a chance of doom!" - Stewie Griffin
NeXTstep is state-of-the-art GUI design, circa. 1985-1990. At the time, it was easily 10 years ahead of anything else available.
:-)
:-) years ago.
But here we are 17 years later, and everyone has finally had a chance to catch up. (Except for Apple, who is now a good 5 years ahead of everyone else by basing their system on NeXTstep
The beauty of NeXTstep was the underlying Objective-C APIs and the dev tools. Amazing, simply amazing. The "build a text editor in under 1 minute without even compiling" example was always a winner.
Using GNUstep w/ WindowMaker is pretty close to the look of NeXTstep, but just seems to be lacking in the "feel" department (that from running GNUstep on my PC, right next to NeXTstep on my 68040 NeXTstation.)
Of course, that was over a year ago, so things might have changed a lot with GNUstep/WIndowMaker since then.
If you can get yourself a NeXT machine cheap, go for it. Heck, you might even be able to find an old Intel of HP machine with NeXTstep on it. But don't expect it to be a dailt use machine. I recently picked up a Mac Classic at a garage sale for the same reason: it's fun to remember what computers were like just a few (well, 12-15
"Tomorrow's forecast: a few sprinkles of genius with a chance of doom!" - Stewie Griffin
Mac OS X still has fat binaries, actually. You should be able to build binaries which will run on Mac OS X or Darwin (PPC) and Darwin (Intel).
Obviously the Darwin/Intel contingent is pretty small, so I haven't seen this in practice.
On my Mac OS X 10.1.2 box:
% which lipo
/usr/bin/lipo
I don't read ACs: If a post isn't worth so much as a nom de plume to its author then I wont bother either.
The problem is that many of the newest video cards do not support the VESA 2.0 standard.
As this was mainly a demo cd, I was more concerned with getting it to run on as many machines as possible with minimal effore (therefore, VESA)
The actual installation disc I'll end up creating will not try to use the VESA framebuffer.
Chad
ACK!!'s feedback is actually important and valuable. The more people who adequately represent their own needs, the easier it is to create templates for profiling users and develop systems that fit those profiles. Yes, technically, with enough customization, 95% of distros can be made to fit 95% of the needs of 95% of the people, but relying too much on customization means that a lot of redundant work goes on (if over half the users are spending 2 hours making the same customization, wouldn't it make sense to make that customization available as a default?) and a lot of people who don't have the time to customize will go elsewhere (might not bother you, but personally I believe in network effects - the more people who use the platform I use, the more development will happen on that platform, and the more goodies I get.) Since open source development doesn't have focus groups and useability labs and market research, forums like this are frankly pretty good ways to present wishlists, complaints, and the like.
It's not just coincidence that having the menu appear below the pointer is a lot faster, or that buttons along the edge are faster to access because you can't overshoot. This phenomena is an example of Fitts' Law (check out usability guru Bruce Tognazinni's article here ). One of the ways that GNUstep truly thrashes KDE usability-wise is that the GNUstep environment has really large buttons often with text right under the icon. By the nature of their size, these buttons can be accessed with a mouse far faster than the really tiny toolbar buttons you often see in other desktop environments. The labels for the buttons also give a clear indication as to what action the button performs; there is no need for the user to try and decipher what a particular icon stands for.
KDE, on the other hand, blindly copies microsoft's system of extremely tiny, unlabelled toolbar buttons that have extremely slow mouse access times and extremely small and cryptic icons whose true nature can only be discovered by either clicking on the toolbar button and possibly performing a destructive task or painstakingly holding the mouse over the toolbar button for several unbearable seconds to get the tooltip. "But Microsoft spends zillions of dollars on usability research" some say. And they spend tens of zillions on security research with results just as good. Microsoft is by far the most frequent inductee into the user interface hall of shame , and such windows UI shennanigans as multi-level tabs, window in window MDI, and Window XP/2000's dynamic menus have been frequently and harshly criticized in the UI design community. "But Windows users coming to Linux will be familiar with lots of really tiny, confusing, toolbar buttons with slow access times" they say. Windows users are certainly familiar with the Blue Screen of Death--maybe we should put stuff in the linux kernel to make it crash so they'll feel right at home. Yes, I know that there are options in KDE to have icons and text appear together. But this is not done by default. And probably 90% of users end up using the default which is installed with their application/OS. If you don't believe me, just ask Netscape. In the cold, hard reality of end-user desktop UI design, not doing something by default is really the same thing as not doing it.
I challenge the KDE Usability project to, by default, give KDE have large, labelled toolbar buttons that are fast to access and easy to understand. They of course don't have to take this challenge; some people would prefer linux not to get on the desktop.
I've got to run to work (I'm late!), but if one of you could please put a post on the slashdot article explaining the problem I would be very
/., your project has
l
thankful!
It seems I broke a rule at sourceforge which limits file sizes to 100MB
If someone could offer a mirror site I would be very thankful for that as well!
Thanks! Chad
Here's what sourceforge had to say
--------------
Greetings,
My name is ********; I am the Quality of Service Manager for
SourceForge.net. This message is directed to you since you are
designated as a project administrator for the Simply GNUstep project on
SourceForge.net.
First, we would like to take a moment to congratulate you on your recent
press exposure on Slashdot.org -- we love to see Open Source projects
succeed, and press exposure of this nature is always of great benefit.
Since the announcement of your project efforts on
received roughly 2241 downloads of the ISO image provided through your
download page at: http://simplygnustep.sourceforge.net/Download.htm
It has come to our attention that you are making use of SourceForge.net
project web services as a mechanism to release file materials in excess
of 100MB in size. Each project hosted on SourceForge.net is provided
with project web services as to ensure that they may adequately provide
an online description and information regarding their project.
... etc
Sourceforge has a policy of not allowing DLs of files over 100MB, I broke that limit with my 110MB ISO image, so they took read access form the file to everybody but me.
If someone would like to offer a mirror I would appreciate it very much!
Chad Hardin