Slashdot Mirror
Windows and Linux User Interfaces
Anonymous Coward writes "Greg Raiz, Boston based interface designer and former Microsftie takes a look at Linux and outlines key shortcomings and strengths of an OS that could take on a giant."
← Back to Stories (view on slashdot.org)
You can have your cake and eat it too.
Linux has to made more useable "out of the box", expert users can always strip-down their install or use only certain tools or pick "expert" distros.
No harm is done (to expert users) if a smart company decides to release a user-friendly linux distro.
"The core mantra should be: "Simple and easy in everything we do, but give me a command line and I can move the world.""
I'm guessing he hasnt spent a lot of time in OS X then. Especially since he says in the article that Apple took the simplistic (ie not technical) approach.
For the second point, the photo system would be entirely dependent on the window manager and basic shell suite, and I know that Gnome has thumbnailing. I personally almost never use the default photo management stuff, opting for better software than baseline, but I can understand the author's argument.
The productivity suite one is a difficult one, as it'll require unrelated projects to have some kind of common backbone that may require extensive editing. It also won't be consistent to web-delivered rich-text editors that are common in forums that allow fonts and formatting. Even more annoying would be if it were difficult to remove or supplant with a better productivity suite.
As for contacts, while I'll agree that a baseline system would be nice, I'm inclined to specifically avoid something that's across-the-board for privacy and security purposes. I'd rather not have some malicious software that gets in through some exploit manage to retrieve my entire list of contacts and their types, only to then try to spread to them or to spam them.
The thing that the author doesn't address is that these responsibilities are the job of the distributions moreseo than the application developers. The distributions could very easily hire their own developers to take a project or application and modify it to meet these requirements. It might cost some money, but that's where RedHat or SuSE can 'value add' their part.
Do not look into laser with remaining eye.
___ In the words of Gen. Douglas McArthur: "I'll be right back."
On se Internetz nobody noes your German.
Like on Apple OS X. If you want to install an Application you can normally drag and drop it in your application folder (easy enough). But if you wanted to you can go into the terminal and do a ./configure; make; sudo make install
Just because there is a easy way it doesn't stop you from doing it the hard way. Having an easy option usually save the person time in deployment.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
My fonts look fine. I live in the "real" world, and have spent approximately zero time screwing with fonts. What, exactly, is the problem with your fonts?
Ce n'est pas un vrai mouvement de robot!
Of course, as reasonable people, we all understand that he means LINUX distributions rather than the LINUX kernel.
I think he makes a good point that the clean solution is always better than trying to support older decisions that turned out to be less than ideal. But the problem is, users aren't interested in details. Details don't matter. They only want something to work, and keep working.
Most modern interface elements are implemented by most toolkits. I think a solution would be to take the concept of the X server, which implements low level functionality available via byte stream communications, and implement much higher level concepts using the same idea. Rather than linking in libraries (and tying your graphical concepts to one language - C for GTK, C++ + custom weirdness for QT) have an X server analog that can speak in terms of Menus, Canvas w/ Scrollbar, Button, Text Input, Text Output, etc. Instead of Xlib (or clx in Lisp) you would have a much, much higher level communication protocal. Language bindings for C, C++, what have you would build on the primatives and higher level widigets provided by this X-server plus, and themes and other details would no longer be different because of what language binding you happened to be using. Translating code between languages would also be much easier, since concept names in different languages could all build off of the standard in the server.
Look sometime at the problems people have writing Python bindings for QT. I think the idea of a high level graphical object server with server side theme configuration would take us a long way towards a common desktop look and feel.
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
There ya go! This is the core of the Linux problem. Everyone sees Linux vs. Windows, and in this battle, I have little doubt Linux would win. But that's just not what it's all about. Linux is a kernel, not an OS. The problem is the Linux OSes fragmentation. There is not one, but many Linuxes fighting for the crown, and this is weakening their common kernel: Linux.
;)
If only SuSe, Red Hat, Debian and Mandrake could just agree on some STANDARDS !!! For crying out loud, everyone is bashing microsoft for not adopting an "open" standard (actually plenty of them) but the key distros cannot even agree on a common way to distribute and install an application. How can anyone blame Microsoft when the exact same idiocy is happenning in their supposedly "perfect" open-source world?
</rant>
Write boring code, not shiny code!
"Linux to be 100% RISK FREE. If you don't like it you need to be able to easily uninstall and your computer will be exactly the same as before you started." So linux should use a fat32 fs so it can be "uninstalled"? Try installing XP and "restoring" win98 or win2k. If you want to tryout linux use knoppix, it's stupid to expect _any_ OS to adopt a deprecated fs from another OS in order for a user to be able to "restore" his old OS. Don't put absurd requirements for Linux that the latest MS Windows can't accomplish..
I can do the same under Windows, yet Windows has a central installation architecture. Why is that? Dependencies, for one. If a program has no dependencies or externals, I can just extract and run the binary (in fact, a lot more programs under Windows work this way than one would think).
The OS should ensure that applications are installed before they can be executed.
I don't know of any OS that requires that one. However, that isn't an arguement against the former part of the excerpt, which is the only part I'm addressing.
Linux as a whole needs one. Something that lets joeuser download a file (ONE) from the net, run the file, and it installs everything needed for the program downloaded and runs correctly the first time on any modern linux platform. I still can't do that on SUSE even for most programs.
Compare to the Windows world. You can download a single file (a self-installing executable file) from the net and assume it will work on any modern windows platform (2K/XP/2K3). In fact, the only things I can think of as exceptions off the top of my head are ports from Linux to Windows.
--- Ãther SPOON!
He makes it sound so simple, doesn't he?
Writing a complete VB clone isn;t as simple as writing an IDE. VB 6 worked because of the underlying Windows infrastructure — ADO, Access, COM, and all those other acronyms that could be glued together with VB to make an application. VB provides a great environment for hacking together in-house and vertical market applications. It's good for rapid prototyping, too.
The Unix world has some very strong biases that make cloning VB difficult, not the least of which is a general prejudice that all VB code sucks. I've worked in shops with VB programmers (I'm a C++ guy), and saw some darned ugly code; the anti-VB prejudice has some basis in fact. Be that as it may, VB is a powerful force that locks many developers into Windows. If any of this code is to move to Linux, we would need to replicate the entire foundation of acronyms used in VB programs — a daunting task that most Unix-oriented folk will find unpalatable.
In part, Mono was trying to accomplish Windows-Linux interoperability, albeit using .Net as the foundation. Mono, however, does not address the vast quantity of VB 6 applications. And Mono's viability is still open for debate, given Microsoft's proprietary attitudes.
A while back, I was tried to sell the idea of a FOSS Access and VB to several major Linux "players", without success. Perhaps my pitch just wasn't that good, or maybe, just maybe, Unix people really are letting their prejudices get in the way of a Really Good Idea.
All about me
All of which is easy to do in Linux if you do it the same way that 99.99999% of the world's Windows users install Windows: they just buy a machine with it pre-installed and set up for the hardware in the machine.
Windows is not easy to install in the sense of it automatically sensing everything and never having to download drivers etc. But noone cares because almost noone actually ever installs Windows.
And, THAT is why Microsoft leans so hard on people like Dell and others to prevent them selling computers with Linux or even FreeDOS on them.
TWW
"Encyclopedia" is to "Wikipedia" what "Library" is to "Some people at a bus stop"
Some good ideas, but Greg is really out of touch with Linux and free software development in general.
He seems to miss the idea that (a) we can't throw out diversity of applications. It's confusion, but it's also a fact. (b) There *are* different distribution brands, though they try to lean on the common Linux name (RedHat Linux, Debian GNU/Linux, Gentoo Linux). (c) Most importantly, it's up to independent distributions to make the system into a cohesive user experience, and the success of GNU/Linux systems is precicely *because* of the ability for lots of independent developers to create software packages for it, not some central Linux authority. Linux *is* just the kernel,. It's up to other people to make more complete systems. He makes the very common mistake of confusing "Linux" with "OS consisting of a Linux kernel and GNU libraries and other tools with some user-oriented desktop environment".
There can be no Linux Inc. creating The One And Only Linux Desktop System. It shouldn't happen, and it fundamentally *can't*. There can only be a variety of Desktop Systems that are based on GNU/Linux.
This is yet another article saying that "If only Linux was a single unified force, it would be good." with a few sentences about interfaces at the front.
Let me dispell some of the myths. First, people don't want things to change. It's wrong.
If that were true, no one would have moved to OS X. "iPod, what's an iPod? I listen to music on my walkman."
People aren't going to change to GNU/Linux for no reason, but once they make the switch, we don't need it to look and behave like Windows. We have our own interfaces, and they work. OS X doesn't look like Microsoft Windows and people don't have that much trouble using it.
Some of the interface integration ideas he presents are allright- some of them are already in place, and others will take more work.
But the idea we need to drop KDE or GNOME, and drop distributions is old and tiring.
The simple fact is that when you consolodate for the sake of a unified force, you remove what makes the Free Software world great- competition.
If we'd all consolodated with Slackware in 92, we wouldn't have had packages. If we'd consolodated behind, say GNUStep, we wouldn't have had KDE, or GNOME, and so on.
The idea of lots of distributions and lots of interfaces and lots of every app is to let them all go, find which work best, borrow ideas, and, in the end, everyone benefits.
If we'd decided to "consolodate" and make an incompatible change, then that change would have to be left out. Once that happens, progress stops, and then someone else comes along and steals the rug from right under us.
Even "Consolodated" OSes like FreeBSD are, like GNU/Linux, collections of programs from other places.
I can't believe these articles still make it to Slashdot. They're old and tired.
Time for the FOSS community to start working on OpenOSX.
"The MAJORITY of all new servers today are slated to run Linux.
This is not going to change, I repeat, it will NOT change."
Right, because nothing better than Linux at running servers will ever come along. By the way, what OS always has the best uptimes? Linux right...
"How can you call Linux a "niche" OS?"
Because all of the estimates I have found place Linux desktop usage somwhere between 3 and 6%. If you need to ask now, it's because you don't understand what "niche" means.
"First off, Linux, or rather GNU/Linux, is an operating system KERNEL."
Man, you lost that battle years ago, give it up already.
"But more importantly, it is hugely successful and I am personally offended"
And NOW, FINALLY we get to the crux of the issue. YOU are personally invested, so rather than accept honest criticism, you get "offended" because someone spoke unfavorably about your hobby, and you feel that criticizing Linux is equivalent to criticizing you.
I'm feeling snarky, so I'm tempted to say that's patheitc. But I won't.
What I will say is that of that's the best argument you have, then Linux had better have someone else arguing for it. Luckily it does.
They only made one.
I do admit that VB may not be the best thing around, but it is fairly easy to learn, and it can get some simple jobs done. And I would venture that VB meets the minimum set of requirements of a programming language as defined by Turing himself.
So, VB is not for you. Thank you for deciding what is best for the whole world. For your next trick, please tell us what the solution is for world hunger or poverty.
Next, I suppose that you will be saying that VI is the only real editor, and anybody who uses anything else is dumb.
I apologive for appearing trollish, but this sort of attitude is quite annoying.
"-1 Troll" is the apparently the same as "-1 I disagree with you."
Some of his comments are good, some are abysmal:
There is an opportunity for the open source community to create a VB compatible IDE that could compile applications for both to Windows and Linux.
Please, by all the gods you believe in, NO!. The very last thing we need is all this crappy VB stuff on Linux. VB is - by rights - famous for the shoddy software created with it. And don't tell me you can write good software with VB, the fact is that the vast majority of VB software are abominations that should've never left the author's imagination.
Entice users with well thought out end to end solutions
That entire chapter would've been much shorter if he had simply written: "Look to OSX for ideas on how to do it right, and to Windos for ideas on what to avoid at all costs".
Users are forced to untar, un-gzip, copy, configure and sometimes compile in order to properly install software.
Has the dude used any Linux distribution during the past 5 years or so? Now I do compile stuff occasionally, but then I want to be on the bleeding edge and some of that stuff was written by me. Almost all actual applications I use rely much more on apt-get and dselect than on tar and gzip.
Linux should stop copying Microsoft feature for feature and embrace the differences and features that advanced users love.
YES. Besides some of the stupid comments, he's got the basics right. Hey, wait. Some of us have been saying this for years. The problem is that too many decision makers in both KDE and Gnome believe copying windos is the road to heaven.
Assorted stuff I do sometimes: Lemuria.org
First of all, there is no "Linux community" in that they have a single answer or position on things. I wish people would stop generalizing like that.
/usr or /usr/local.
It's true that it's a generalization. However, my point is that enough of the community stands in the way of some forms of progress to prevent it. Generally Linux is seem as "Elite". Anything that bolsters that leetness is well received while anything that threatens it is heavily challenged by a portion of the community. It can be challenged so heavily, that it can give the impression that the challengers speak for the entire community. Their position also doesn't help the rest of the community in moving forward on many user-friendly ideas.
I would say Linux's answer has been: "If you move to Linux, your computer will run for years without getting slower, there is no risk of viruses or spyware, you can customize the look of your computer or just stick with the elegant default, the operating system comes with hundreds of applications that are of high quality and absolutely free."
That's a nice thing to say, but that's not the value proposition that's seen by much of the market. Plus it's blatently incorrect. There have been several Linux viruses thanks to security flaws, and your computer is just as much at risk of slowing down due to more background programs and desktop widgets being installed. Older Linux systems can also have a great number of compatibility problems stemming from their insisitence on installing everything in
You think Mac's don't have any viruses?
I think Mac OS X doesn't have any, primarily because not a single one has EVER been identified. Mac OS 9 and less were highly vulnerable, but the two systems are not the same.
friend of mine had OS 10.something destroyed by one
You're going to have to back this up with facts. For example, from this link: "As of now, there are no viruses for OS X Macintoshes."
Regardless of whether or not there will be any in the future, this is an exceptional marketing tool that draws in new users.
Javascript + Nintendo DSi = DSiCade
"You may want to try the latest build of 6.0-RC1 on the Sun Blade 100, or disable ATA DMA if you are using IDE hard drives in that machine. I also have a Sun Blade 100 that throw fits with ATA DMA enabled, but the problem was fixed prior to 6.0-RC1 was released."
"Thanks for the heads up. I was actually putting it on there to test the upgrade procedure from 5.4 to 6.0 before I applied it to the other sparc machines but not Blade 100s."
And you wonder why the average person uses Windows.
Greg's Head is part of my daily reading and he always has interesting things to say about the field of user interfaces. His latest post is Linux Thoughts. While the post has some very shrewd observations, there was something that nagged at me while I read it.
Going back through the post, I think his basic premise is wrong. The first paragraph of the post opens with,
And I think that's where he's wrong. I think there are definitely players in the Linux arena who want to use Linux to compete with Microsoft, but I don't think that's true for Linux as a whole. The drive behind Linux isn't to compete with Microsoft, to replace Windows, or even to provide a mom-and-pop OS to the mainstream. As with almost all open source software, the drive behind Linux is the scratching of an itch. No more, no less.
Looking from that point of view greatly changes some of the things Greg had to say. For instance,
If Linux was trying to be a brand, this would hold very true. However, from the standpoint that Linux progress comes from people scratching an itch, it doesn't hold any water. Part of why Linux is where it is today is because there are "dozens and perhaps even hundreds of different Linux distributions."
If you are trying to understand how Microsoft could compete against Linux, it becomes easier if you take the viewpoint that Linux is trying to be a brand that is competing back against Microsoft. But I believe you would be fooling yourself to take that viewpoint. The simple fact that it's not trying to be a brand is why it's so hard to compete against Linux.
As I said before, there are players in the Linux arena who would like to use Linux to compete against Microsoft, and the most obvious player that comes to mind is Novell. With Linux as a whole, though, there's nothing really for Microsoft to compete against. Microsoft could crush SuSe into the dust and it wouldn't really affect Linux.
Keeping all that in mind, I want to stress that this does not invalidate anything Greg has said about Linux in his post. He brings up some excellent points, lays some very good directions for Linux developers to take, and really hits the nail on the head with a lot of things with regard to how Linux could expand to the masses. But...
As long as there are a handful of programmers who are happy with their Linux distribution and are continuing to tinker with it, Linux will be wildly successful. Why? Because that handful of programmers are scratching an itch. It's as simple as that.
How to type ./configure && make && sudo make install in case of unpackaged apps; 3. Whether your distro uses sudo or su so you can modify Step 2 appropriately.
yea, because that is straight forward to someone who just wants to pick up linux...
I swear to god, you linux types will not get anywhere until you realize that having more than 1 step for getting the installation of programs going is too much.
I don't even know what the fuck those two steps are, but I do know that I installed linux on a machine I had at home and gave up on trying to get it to work because I couldn't get the drivers to work for my wireless network card and installing programs makes fuckall sense.
Why can't I just double click on the executable and have the installation process start? What is so hard about that? Linux works great for servers, but it is so far away from being ready for desktop use by most users it is amazing.
With windows xp, things just work. With 95 and 98 things were a pain, and you needed to know some tricks to get around problems you had and I learned those tricks. Now I'm spoiled. When I want to install a new program I like the fact that I can just pop the CD in my drive and it tells me to start the installation and walks me through it.
In short, there is a lot that is difficult about those 3 steps, mainly because THEY ARE NOT INTUATIVE. Explain to me how knowing how my distro deals with packages, then having to type some bullshit into a command line is easier than putting a CD in the drive and following the steps presented.
Linux may be a really nice OS, but some people just want their computer to be a tool where they are able to get things done quickly and effienctly with a low learning curve. Until linux can offer all those things it will remain a niche OS relegated to servers.