When I first got in touch with the Linux world – it was back in 1998 -, I started with an old Slackware distribution. Of course I cannot say it was a wise choice for a beginner, and I soon moved to Red Hat, and then tried various SuSE, Mandrake and Debian flavours; now I am using this good Slackware 8.1. Everyone can see the Linux world has been making giant steps towards usability during the last five years. Applications have dramatically improved, too, both in quality and quantity: OpenOffice, AbiWord, Mozilla, Evolution, Gimp are examples of packages allowing Linux to be used for everyday work. So, is Linus Torvalds’ creature now ready to conquer the desktop?In my opinion, the answer is “not fully, not for everyone”, and the reason is the lack of solid standards, which results in the following problems.
- Beginners get confused by the amount of “duplicated” software bundled into distributions.
- Programmers waste their efforts (and ability, time, money) in creating similar projects, and related configuration tools, documentation, translations.
- Software houses cannot develop something and be sure it will work everywhere.
- Hardware manufacturers may face similar problems when thinking about porting drivers.
Is the situation recoverable? It certainly is, but the Linux Standard Base organization must impose a bunch of severe standards to be strictly respected.
I am going to to summarize the most common problems encountered by newbies and desktop users on Linux. Please note that most of the following advices fit perfectly to other *nix platforms, including BSD and the upcoming Hurd.
Desktop/window managers and GUI libraries
I do not want to start another “KDE vs Gnome vs WindowMaker vs …” war, but having multiple window managers means that beginners will never master one environment only.
Not that I like people to learn things mechanically, but it has to be recognized – this is just a silly example – that most vendors provide detailed instructions (“click here”, “click there”, etc.) on CD covers for installing their programs: if you have 3 or 4 widely used window managers, this becomes pretty impossible for the vendor, and thus Linux will always be seen as “more difficult than Windows”.
So, what to use? The final choice is obviously between KDE and Gnome, the most mature projects up-to-now. I would choose the latter.
First of all, let me thank the KDE group for the big job they have been doing. KDE is probably more complete than Gnome at this moment, but this one has a huge advantage in its turn: it is based on the LGPL-licensed Gtk+ libraries, while KDE is based on the Qt ones, which are commercial under Win & Mac and commercial or GPL’ed under X11 (Linux, BSD, Solaris, etc.). This means you cannot write closed-source Qt software (I personally hate it, but some Companies think different) under Linux without paying big money to Trolltech, and you cannot write cross-platform (Win+Mac+X11) Qt software, neither opensource, without spending a lot of money in licenses.
Also, Gnome seems to me cleaner and newer, less bloated and ready to help people affected by disabilities (see Atk for this).
Moreover, the most important opensource projects are released for (or will be ported to) the Gtk+/Gnome world.
Packages and Linux File System
Installing a new application under Linux is a big pain. Even if one is not interested in recompiling from sources, there are a lot of other problems to face: various package formats, dependencies, files location on disk.
I really would like LSB to adopt a single package format (I suppose the choice is .rpm) and a strict set of specifications for decompressing the package content. At the moment, LSB only recommends them, and this is not enough.
Also, one between GNUpdate and Autopackage should be completed (it should be in every distribution’s interest to do that) and adopted as default package management system.
No multiple user applications in the same field
Is there any point – for the majority of us – in providing multiple Text Editors? (KDE – for instance – installs KEdit, KWrite and Kate!)
Not to mention the image viewers army, 3 or 4 email clients, a certain amount of browsers, and so on.
People (especially beginners, who represent – like it or not – the biggest part of desktop users) usually need a single application for every field: one text editor, one image viewer, one email client (Evolution?), and an affordable and fast browser (Galeon?) to rely on.
No multiple system applications in the same field
CUPS is widely spread and supports a lot of printers. Why do some distributions prefer other choices?
And what about the various system configuration tools? Why not just finishing the promising Gnome System Tools and integrating them in Gnome, so that we would enjoy a clean and complete desktop with a single and powerful set of system tweaking utilities?
Similarly, Sane and gPhoto should be adopted as default scanning and photo systems (which is quite already true at the moment).
More libraries installed by default
Currently (see developerWorks) LSB supports libc, libdl, libm, libutil, libcrypt, libz, libpthread, libncurses, libX11, libXext, LibXt, libICE, libSM, and libGL.
Anyway, standards should also include basic libraries for executing Gtk+, Gtk–, Inti, Gnome, Qt, KDE, wxWindows applications, as well as libSDL for games and advanced multimedia applications. This would avoid many dependencies requirements, and that problem would be partially solved.
Conclusion
I do not agree with Linus Torvalds when he says that Linux will be ready for the desktop in 2006. In my opinion this could happen in a couple of years.
Moment will soon come when Courts will forbid vendors to preload an operating system when selling a pc. Linux distributors must be ready to take advantage of it and offer a stable, secure, usable and uniform product, while differentiating in documentation, technical assistance, presence on the territory, courses, partnerships. Maybe there could even be points of sale for Linux-related hardware and software (like Apple’s ones).
I hope we will not miss that opportunity, and fragmentation will not spoil all the good work that has been done.
Let me add a note for those who think that my advices are a limitation for the “freedom of choice”: you are certainly right, but remember that too much freedom means no rules at all, and no beginners playing with Linux, and no important software being ported, and therefore less people adopting it.
On the other hand, I am not asking for the moon: I would just like all the distributions to default-install some packages instead of others, and let expert users the ability to modify this default choice and tweak their environments.
I strongly believe this would be a fair price to pay for seeing Linux rule in the desktop, don’t you?
About the author
I am Marco Zanon, 23, living in Vicenza (Italy) and currently studying Electronic Engineering. I am interested in Linux, from both user and programmer perspective, but I am also looking forward for OpenBeOS to be ready. I love to study usability issues in modern operating systems.
But isn’t .deb more flexible?
Out of the great chaos will come something truly amazing.
Something beyond the crap that people use today. And by crap, I include Windows XP, OS X, and Linux. They are all crap and make a human being do far too much work to get anything done. None of these operating systems handles a lot of information, many disparate types of information, or does any chores for the user automatically. I am typing in a little box right now? About 40 characters wide? On a 1280×1024 ClearType LCD screen? With tons of light gray green space all around? With nothing else of import on the screen. With Excel running on my second display which I’m not using at the moment. Hmmmm. If one really looks at the way a computer works, any of them, it is a very poorly designed and engineered system.
It will be fun to see how the new Windows looks and works in 2004/2005. If people like it, the Linux crowd will be cloning it for the next 3 years. If Microsoft’s patent attorneys don’t sue them to death.
So I agree with Linus. It will take until 2006. But along the way we’ll get a good show, some glimpses of interesting breakthroughs.
The next few years are going to be an amazing fun adventure in the land of operating systems.
–ms
<ramble>
The various UIs are not a major problem. KDE GNOME and IceWM (these are the ones I use, so are probably the only ones I have the right to comment on) are all pretty straight forward for anyone (yes, including Newbies) to navigate. You just need them to all have a common menu layout, like Mandrake does. Also, KDE’s use of QT does not mean that every program you run while using KDE has to be coded in QT (i.e., Mozilla, OpenOffice, Xine, etc.). As far as which package tool to use I am surprised he chose .rpm over .deb, but that can be debatable, so I won’t attack that.
The author does have some very good point about the LSB needing to be much more strict. right now distro A’s file system is different from distro B which is also different from C….. After all, look at the coding nightmares that The Kompany and Mosfet have had when it comes to writing QT apps for Redhat. I also agree with the problem of having to many similar apps install by default, major distros should look at user friendly distros like Lycoris, and only install the one or two most popular apps in each category by default (leave the others available for the picky techies like myself).
</ramble>
Please not another “what Linux needs” article… it does not change anything anyway. Why? Because the only way to make a change in the Free Software world is to contribute.
Contribute by writing code. Contribute by writing documentation. Contribute by helping others to set up their systems. Contribute by making designs that are so convincing that coders will pick it up. Contribute by paying people to write code, or by donating.
An important thing to understand is that you can not make any difference by making destructive criticism. Like saying “Linux must drop all other desktop systems”. This is simply not going to happen, as people will not stop using or contributing to the system of their choice. What do you think will happen when the LSB says that Gnome is the official desktop environment? Will all KDE users and developers abandon KDE or LSB? Certainly the latter…
What you can do is contribute to the system of your choice, help to improve it and make it so good that using it becomes compelling.
Please not another “what Linux needs” article…
I agree. For me, the Linux world is about choice, for the devs just as much as for the users. The goal is not world domination, but providing alternatives and sharing wisdom.
I really hate those “we have to drop 90% and unify everything!” articles. Dumbing things down isn’t the way to go. Educating the users is. Usabitlity should be about helping a user to understand the concept behind a system so he can chose the best way of action for himself, not limiting him to a single one. For everything else, use an appliance, a tool dedicated to a single task, not a desktop PC.
Do you think Joe User is so stupid, that he cannot understand the fact that there is more than one email client? Do you think this is different in MacOS/Windows?
Go to download.com oder versiontracker.com and see how many windows mail clients there are. Even windows xp delivers 2 text editors (Notepad and Write) in default installtaion but have you ever heard someone talking “I cannot use Windows, because it has 2 text editors.”?
What distributions need is a good text describing the mail application. A screenshot would not hurt either.
A average user can decide between 2 applications, you just have to give him information to base his decision on.
/bla
Awa ! Good ideas !
One email client, one browser, one desktop …
one government, one plannet ??!!
The point of free software is about choices.
Do not pretend to be God, and decide what is good
for everybody.
1) GNU/linux, a kernel won’t do anything as desktop…
2) Pretty simple way to make people install things::
open a terminal, type su press enter and then type your root password,
emerge//apt-get[your favourite package manager] the_thing
OR
tar -xjvf my_tarbz2
cd mydir
./configure –prefix=/usr
make all install
ldconfig
exit
Quite straightforward, isn’t it?
3) having ONE program as solution isn’t good and isn’t useful, different people has different needs! I won’t use evolution and I’m ok with mozilla mail, I know people that are perfectly happy with evolution or mutt or kmail.
A good operating system would present the user with the choices of what is available and have a blurb about each one includes strengths and weaknesses, perhaps even a comparison matrix.
As for the people who think Linux should have “one” of everything… damn, the world is sick of not having choice. Why is Linux extremely popular in the first place? It’s because there is choice for the individual. It’s because no asshole company like Microsoft or Apple dictates to the user what is available on the machine.
Make the software easy to install/uninstall and this fear of choice would be dealt with. Many Windows users are scared of trying out things because they know they only get X installs/uninstalls before their registry/machine/whatever gets corrupted and they have to re-pave.
Let Linux show the world that they have a tough-as-nails install/uninstall system that works 1,000,000 times without a glitch.
–ms
1) GNU/linux, a kernel won’t do anything as desktop…
And yet you advocate choice a few sentences later? Why should the FSF tell me what to call it, when people recognise the name ‘Linux’? GNU won’t do anything as a desktop either, for that matter. Hmm… XFree86 isn’t GNU. KDE isn’t GNU. Mozilla isn’t GNU. ls and date are GNU, but they aren’t desktop tools… GNU/Linux and Linux are as much a desktop as the other.
This brings me to another point. People keep talking of Linux on the desktop, but a lot, if not most, of what they talk about applies equally to (say) FreeBSD. Really, it could be ‘X on the desktop’. But I guess one letter isn’t as user-friendly or as topical as ‘Linux’.
But as to the article: I generally agree. I think that if the LSB says that Gnome (or Galeon or Evolution or TomartoPaste) should be used, it won’t be the end for KDE (or Konqueror or Kmail or TomaytoPuree). It just means that these won’t be as obvious to the end user. But is Linux now? People can’t walk into a department shop and get Linux sitting on their computer, at least not here. But I’m still typing this up on a Linux box…
To get recognised by John A. Citizen or MegaSoftCorp as an alternative, Linux needs to mean something. Currently, it doesn’t. Currently, it just means a bunch of random software thrown together that may or may not work the way I expect it. And that’s bad.
“Programmers waste their efforts (and ability, time, money) in creating similar projects, and related configuration tools, documentation, translations. ”
This is the only part I don’t agree on, If Mozilla, Evolution, and KDE were the only browser, mail client, and Wmanager I’d most likely not be running Linux, or currently working on my own apps. I get your point about it seemingly being a waste of man hours. But what would these people agree on anyway? Thats why there are different apps.
Don’t most distro’s come with a setting for install that goes something like ‘Server, workstation, Desktop” and certain settings throw out most redundant applications? except KDE, it Kseems Kit Kwants a Kapp for Keverything.
Having so many choices is one of the greatest things about Linux. More importantly though, having multiple instances of roughly the same things generates a little competition which is good for everyone involved. For example, I don’t think kde would be where it is today with out having to compete with gnome, and vice versa. The same goes for a lot of other open source projects. At least that’s my opinion.
However I do agree that a common package management system and more standards would be very beneficial.
Open Source programmers are not a willing mass of work force which can be thrown at the project of _your_ liking. They do what _they_ like, unless they work for a Linux distributor. That’s why Linux _needs_ commercial distributors and big company support from IBM and Co.
You can lament all you want, but that’s how Open Source works. Unification and simplification is a value-add service to be done on the distribution level, that’s one item you pay for (or don’t …). Some newer distributors seem to do a pretty good job in this repect.
It astounds me every time that so many people overlook this simple fact. Even I know it and see its inner logic, and I don’t even use Linux …
Linux geeks are so funny
open a terminal, type su press enter and then type your root password,
emerge//apt-get[your favourite package manager] the_thing
OR
tar -xjvf my_tarbz2
cd mydir
./configure –prefix=/usr
make all install
ldconfig
exit
Yes, for a newbie to open a command window and type stuff is just so straight forward. Why not something better like
Find Setup.exe
Double Click
Click Next one or more times
done
Is that just so hard for the Linux crowd to make a program to do that? Make it so that when they double click something it fires off the commands you have written above? Is it just not geek enough to double click, you have to open a command window and type stuff to get something done? You can do it both ways in Windows, from a double click to a command window; why not the same in *nux?
I couldn’t agree more with Marco, Linux big failure was not being able to deliver a consistent OS to end users. Users like my dad who doesn’t know (nor want to know) anything about programming nor device driver or window managers subtilities, would be reassured a lot whenever they start a Linux box they could find the same base of tools even if they are very basic or incomplete at first installation. It’ll be a wise choice to let the end user choose what software package he/she really needs and not provide them with all what the community has produced (make sense no ?). What is more annoying than to find out that your favourite OS is bundled with many preinstalled editors, browsers or terminals ! Also, a good graphical software installer/uninstaller system along with a reliable OS installer would be a really major step forward in getting Linux OS adopted by far more end users than today. Most installers today fail at some point, you still have to dig in config files to correct things when it is still possible.
I hope Linux to start to win this battle or others will soon,
PS: “chaos theory”, it’s like my dad talking 🙂
Open Source programmers are not a willing mass of work force which can be thrown at the project of _your_ liking. They do what _they_ like, unless they work for a Linux distributor. That’s why Linux _needs_ commercial distributors and big company support from IBM and Co.
Regardless of what Joe A. Programmer writes, Red Hat, SuSE and Mandrake are by no means required to use them. Red Hat, SuSE and Mandrake (and other controlled distributions) can choose what to include. There is no reason for them to include Galeon, Phoenix, Mozilla, Netscape and Konqueror, when one does the job well enough for inclusion. This (part of the) article is more aimed towards that, I would say.
Have you taken a look at Mandrake 9.0? Its software installation tools are excellent. It includes a “what to do?” menu that offers simple choices to newbies — “enjoy music and video,” “play games,” etc., that choose a *single* application for the task.
You said you didn’t want to start a “KDE vs. Gnome” debate, but that’s exactly what you’ve done with your recommendation. I *prefer* KDE because it’s better than Gnome, at least at present. Religious considerations about the purity of its open-source lineage don’t interest me.
One example: while both environments have multiple desktops, KDE’s are much easier to navigate. Just my opinion, but I’m speaking as a long-time Windows users who made the switch a little over a year ago.
There is also very little cross-platform development going on, anyway.
I’m also puzzled about your comment that the courts will force vendors to ship barebones PCs with no OS. There would be such a public outcry that there’d be legislation within 30 minutes of any such decision. Could you share the source of your omniscience?
Finally, read the comments here about choice. I *like* choices. One of things I like about Mandrake is that it’s very easy for the Newbie, but doesn’t cripple itself for more advanced users.
If you need proof of this, I installed a Mandrake machine at work a few weeks for one of our employees. She’d never seen Linux before. She absolutely loved it and whined when I temporarily replaced it with a Windows machine (we were making some changes in the system).
I think that if the LSB says that Gnome (or Galeon or Evolution or TomartoPaste) should be used, it won’t be the end for KDE (or Konqueror or Kmail or TomaytoPuree). It just means that these won’t be as obvious to the end user.
If the LSB would standardize on Gnome, KDE developers would create a competing standard, thus making LSB worthless. Distributions would still have to decide which standard they try to support, or whether they want to support both (which would be the only way to get 100% of all users).
KDE and Gnome would still compete.
I couldn’t agree more with Marco, Linux big failure was not being able to deliver a consistent OS to end users.
The OS is consistent if you stick with a single distribution. It doesnt make sense, from an economical perspective, for distributions to offer exactly the same software. A distributor can only make money by offering more value than the competition (and the competition, like Debian, is free). And if all distibutions would offer the same software, someone would create a new better distribution that would supplant those distributions that stopped innovation.
>Open Source programmers are not a willing mass of work
>force which can be thrown at the project of _your_ liking.
>They do what _they_ like, unless they work for a Linux
>distributor. That’s why Linux _needs_ commercial
>distributors and big company support from IBM and Co.
I agree 100% with this. Why can’t they see it? Thats also the reason I greatly respect Red Hat and IBM.
And as for going for the Joe users, like my dad, I think that those people should be the last on the list. Its far more logical to first get a lot more technical users on Linux. Go bottom-up.
Not even 10% of the people in my Computer Science department boot into Linux now at the university (it has dual boot configured machines). I think thats a sad state.
bah
find setup.exe :: 50 items found
[installer with not a standard name] :: You aren’t administrator, I can’t do anything about
I think that have 6 command that the avarage use can cut& paste is way simpler and less fault prone…
My opinion against your obviously
(still you have to find something easyer that type emerge [theprogramiwant] to install packages)
>I strongly believe this would be a fair price to pay for >seeing Linux rule in the desktop, don’t you?
NO! i will not change my freedom just to see Linux running everywhere. This is not the correct way to thing. Maybe you dont get the “kernel” of GNU (and in part Open Source) phylosofie…. Free software (again OS, too) is just about freedom…
What you say is like: “Let’s put chains insome of us, who is completaly free, to almost free others and get everyone almost free”.
No thanks.. i will not put chains in my hands, sorry.
” Programmers waste their efforts (and ability, time, money) in creating similar projects, and related configuration tools, documentation, translations.”
You might think so… but consider this (a fact that not many people know) you can grow trees much faster by growing them in a clump where they have a lot of direct competition than growing them on their own… You can infact intesify this phenomenon by wrapping the clump of trees in black plastic reducing the amount of resources being competed for. This is just the way that trees work. I think that software works in a simular way.
“if you have 3 or 4 widely used window managers, this becomes pretty impossible for the vendor, and thus Linux will always be seen as “more difficult than Windows”.”
Every window manager out there is out there for a reason… purhaps linux systems should be referred to by their desktop environment as a different desktop environment are as good as being different systems. The rule with WindowManagers is USE THE ONE THAT IS BEST FOR YOU. Since you can choose your desktop environment when you login this is hardley an issue at all. It only becomes a problem when the WM/Desktop environment you use is not installed.
” Even if one is not interested in recompiling from sources, there are a lot of other problems to face: various package formats, dependencies, files location on disk.”
A Universal package manager for linux would be a huge boon… Such a program would automatically take into account different file locations, dependencies different distrobution formats and would even try to compile packages from source if neccessary.. I think that Mandrake, Debian and others are doing a good job sorting this out.
“Is there any point – for the majority of us – in providing multiple Text Editors? (KDE – for instance – installs KEdit, KWrite and Kate!)”
One of the problems with KDE and Gnome is that they don’t allow you to install or uninstall individual apps.
Yet another ‘what Linux needs article’. These must appear at least once a week around here.
– ” Programmers waste their efforts (and ability, time, money) in creating similar projects, and related configuration tools, documentation, translations.” –
…they really don’t!
Even with the vareity of applcations doing nearly the same, one can allways be sure that any function an application offers , really are needed by someone – that’s why they were made.
Some programmer felt the itch, needed a function, and made a program to do the job he needed done. Other people needed that very same function, but weren’t quite happy with other parts, so they wrote a little correction – or a new function they needed or thought could be neat to someone – or they dicovered a bug and found the solution.
The application evolves, gets better and more usable to all the people needing that function that the other apps can’t provide as well.
That’s the very essence of open source programming. The userbase are the maintainers best tool – even those that can not program themselves. They can come with suggestions, bug-reports, and in that way be just as valuable.
Now, that being the case, all the various applications out there are needed by someone. I don’t know about you, but I fell I’m not at all competent to judge in the matter of which program all users should or should not use, all the time I know that they are all needed by someone.
Trying to put all functions into one great “superapp” would fail simply because there is so many functions that people think is neat and they would miss if they dissapeared. An application containing it all would be immpossible to configure for your use.
</nalle>
First of all, I applaud all the great work done by the people at linuxprinting.com and gphoto etc. These guys need to get peripheral devices working with out the support of the manufacturer. Unfortunately, things don’t always work right and the manufacturers aren’t always willing to lend support to linux. In Windows a person buys an HP printer that comes with HP software and driver (plug and play, most of the time) or an Olympus digital camera that comes with the right software and drivers.
Linux requires more tweaking to get hardware to work. Any one can be told how to correct software problems. But, joe-average user isn’t going to brake-out the logic scope and oscilloscope to reverse-engineer their new HP printer so they can get the drivers to work correctly with linux.
The user can be told about gnome vs. KDE or evolution vs. Kmail and how to select various options. I was able to quickly show my wife (who is an average PC user) to effectively use Linux. However, eventhough I have an MSEE it took me hours to get my HP printer to work correctly.
Ultimately the hardware must be rock-solid.
Therefore, the immediate need for Linux is for peripheral manufacturers to start supporting linux with their dollars and engineering staff. They will start doing this once there is a larger market to be won.
Bad analogy:
You might think so… but consider this (a fact that not many people know) you can grow trees much faster by growing them in a clump where they have a lot of direct competition than growing them on their own… You can infact intesify this phenomenon by wrapping the clump of trees in black plastic reducing the amount of resources being competed for. This is just the way that trees work. I think that software works in a simular way.
Yes, the trees grows faster but they also become weaker (due to the resource limitations) than a tree being allowed to grow at it’s own pace without competing for resources. In nature competition works by ONE coming out on top if there are competing individuals.
Free Software/GNU/BSD is about choice. The choices would not be removed just because you standardize. Whenever I sit down at a Windows NT desktop I know that I have at my fingertips IE, NotePad, WordPad (usually) etc and I can get to work without fiddling with different menus etc. On my own desktop I don’t use NoptePad, IE etc as there’s choice even (gasp and oogle) on a Windows desktop. I download Phoenix, Mozilla, EditPad whatever I feel is working for me.
Now don’t say that duplicated work is always good. The neat thing about the idea of open software is that anyone can make their own “favorite” app. Now the bad thing about open software is that far too often everyone does that. For many hackers it is their 15 minutes of fame. I respect and totally understand that. Working in a team is for many far too anonymous. We really do need more thoughthrough apps to the free NIXes and that comes first and foremost through team efforts.
//I do not agree with Linus Torvalds when he says that Linux will be ready for the desktop in 2006. In my opinion this could happen in a couple of years. //
So … you think it will be sometime in 2005. Linus says 2006.
Big difference, that.
//They are all crap and make a human being do far too much work to get anything done. None of these operating systems handles a lot of information, many disparate types of information, or does any chores for the user automatically. //
Come again? What the f&sck do you want a computer to do? Brew your coffee and make your bed?
Do you even realize how much today’s PCs do, compared to just 5-6 years ago?
Goof.
…is love, sweet love?
Hal David would be proud!
I don’t see what’s wrong with simply abstracting the underlying differences away, and it certainly steps on a lot fewer toes.
Take the whole KDE/GNOME debate. What’s wrong with making GNOME apps look and feel like KDE apps when they run under KDE, and vice versa? It’s already done partially by BlueCurve and [K|G]eramik, it just needs some tidying up. That way Red Hat, Sun et al can stick with the interface they have chosen and make it a Unique Selling Point. SuSE, Mandrake, Lindows, Xandros, Lycoris and so forth can stick with the interface they chose. The end-user is none the wiser. Each distro may have its own particular look based on KDE or GNOME, but every app looks and works right for that particular distro. The distros have to distinguish themselves from one another somehow, don’t they?
Windows has had multiple development frameworks (MFC, OWL, CLX, Qt, .NET and more) for years and it certainly hasn’t hurt its chances – they all look approximately the same and pick up the same colours, fonts and so forth. I don’t see why the same thing can’t happen on Linux. It might not be elegant or memory-efficient to have multiple toolkits, but get over it – memory and disk space is cheap and elegance is far less important than making sure Linux doesn’t disintegrate into warring factions. It happened before with Unix in the 80’s, it could happen to Linux too, with disastrous consequences for Linux’s chances as a Windows competitor.
Similarly packaging systems are not going to just disappear. Like it or not, Debian is not going to change to the technically-inferior RPM because of market concerns. It’s not a commercial entity, why should they change? Debian can already install RPMs. They’ll take the Debian-based distros – Lindows, Xandros, Knoppix – with them too. Gentoo will do their own thing. It’s probably not possible to standardize here – but none of the existing solutions are all that great anyway. Instead, why not coax everyone into using a single new system, that learns from the mistakes of the old systems and does things better? I think everyone would be interested in that.
Forcing standards will only further polarize the Linux community and make it even harder to create a viable single platform. What Linux really needs is more tolerance and understanding between developers and more inclusive, outside-the-box thinking.
I don’t see what’s wrong with articles that point out weaknesses of Linux, how else would it get this far. If some of these young “wizards” remembered the days when we switched from DOS to Win GUI there were “experts” predicting that the real computing was going to be done only in command line – guess they were more than wrong. Nemesis of Linux has reached this far by listening to the user body 90% of the time. I don’t trust and never will people who think that they don’t need outside control, suggestions and good will guidance.
Some of you see KDE vs. GNOME as problem, not me, thanks to SuSE and RH I feel comfortable with both and switching from one to another was a 15 minute browsing through and I plan to try out some of those light one GUIs on my laptop. Numerous apps are also beneficial, why is this going to be a deterrence – average Joe is not a computer wizard but neither is he a complete dumbass.
All of this I’m writing as 90% of the time Win XP user and the rest RH 8.0 because for some apps I have no alternative. When Tyson was the reining champ I hoped that somebody would come along and kick his ass, that happened and I’m glad. MS is today lean-mean-muscle killing machine, let us users help Linux learn how to throw devastating left/right hooks and uppercuts.
I like, Stephen Poole, are puzzled by this statement? You say you are from Italy; is there talk of such legislation in Italy for this? In the US, there is no basis to think the courts would ever do this-it is like saying automakers can not sell a car with the engine installed. The courts have talked about preloading some programs with the OS-is this what you are talking about? -Dpme
——————–
You wrote-
Moment will soon come when Courts will forbid vendors to preload an operating system when selling a pc. Linux distributors must be ready to take advantage of it and offer a stable, secure, usable and uniform product, while differentiating in documentation, technical assistance, presence on the territory, courses, partnerships. Maybe there could even be points of sale for Linux-related hardware and software (like Apple’s ones).
that’s what distros are all about, i mean you can always builkd your “linux from scratch” and use the software you like, but if you stick with one distribution you’ll have a small range of choices and you’ll be sure that that software will work on your distro. Commercial software developers shouldn’t support directly each distribution, but instead each distribution should decide the software to ship and care about making it works in it’s environment. That’s my point of view and that’s why when i type: “emerge foo” it always works, because somebody in the gentoo staff mantains the package and make sure that that application works on gentoo, so the developers can develop and not mantain distro compatibility.
I’m happy using WindowMaker right now for my desktop manager. Fast and minimalistic. I’m running Slackware current ( not 8.1 ) right now and its really nice to have XFree86 4.2.1 on the desktop and some of the other updated apps. One was or another – as perfect or imperfect – that the Linux distros and apps are I’m happy to use it.
The Linux desktop might not be perfect right now. The fact that the programmers have – there is the exception – been doing this all on free and unpaid time and have accomplished this much is spectacular. Gnome, KDE, Mozilla, X, and all the distros are just painstaking amounts of work and need a lot of collaboration. My thanks to all the everyone involved.
I’d like to see a Desktop Publishing app either Quark or PageMaker. It will take time and convincing to get these out but I will then I’ll switch permanently to Linux.
just a few comments
1. I don’t think that Linux should have a consistent look and feel. Linux is a kernel. Distributions package the kernel into usable forms. Asking all of the distributions to package the same stuff would essentially eliminate the reason for their existence. Without distributions driving Linux (and the resulting competition among distributions), Linux would be little more than a hobbyist OS.
2. I agree with some of the author’s commnets about the Gnome license vs. the KDE license. This is simply my opinion and is based off of my belief that Linux needs support from commercial non-free apps to flourish. However, I think that Gnome and KDE can peacefully coexist. KDE apps can currently run on Gnome systems and Gnome apps currently run on KDE systems. I think the two projects could cooperate more towards making it seamless. A linked themeing engine for Gnome (Metacity & Sawfish) and KDE would help here.
3. One thing that Linux REALLY needs is a common method of distributing applications. I’m not talking about package management (RPM vs. DEB). The Autopackage project looks promising (at least in philosophy). This is again based in my belief that Linux needs commercial non-free applications (that don’t come with the OS).
I’d like to see a Desktop Publishing app either Quark or PageMaker. It will take time and convincing to get these out but I will then I’ll switch permanently to Linux.
http://web2.altmuehlnet.de/fschmid/ – Scribus homepage
I think Linux Journal is going to switch to this desktop publishing package (either them or Linux Magazine though I’m pretty sure it’s LJ) once it becomes mature enough to suit their needs. At least, that’s what I recall in reading one of the editor’s notes sections. I think they use PageMaker right now, but everything else is done in Linux.
I didn’t know about GNUpdate, but reading the FAQ they seem to be doing different things. My conclusion is we will need both packaging systems.
GNUpdate will allow you to install .debs on an RPM based distrobution using a consistent interface. Autopackage will work on a system without any packaging management. Autopackages resolve dependencies, update any packaging databases (rpm, deb etc.) and will give a consistent GUI or CLI interface.
Autopackages are packaged applications while GNUpdate is a package management layer.
Autopackages have better commercial prospects (IMHO) since they are effectively the same as Windows installers, but consistent in interface and they don’t require a package management layer (although they do require Autopackage to be installed, but Autopackage will be fetched for you when you excute an autopackage file).
As you can tell, I favor Autopackage, but I see both as useful.
So the author started using linux in 1988? Funny, it wasn’t even out in 0.01 for until the end of 1991.
“When I first got in touch with the Linux world – it was back in 1998…”
I believe your eye examination is overdue
Let’s look at it the way our good friend CEO Steve Ballmer does. It’s not about removing the users choice, we just want to ease them along, which is why Microsoft provides what he calls, “default choices” for the user.
I personally think that the future of the UI belongs not to the OS creator but to 3rd party companies. Look at Stardock for example. I can take my Windows XP and make it look and function very much like OSX. So now it doesn’t matter to me what OS i’m using.. I can make it look however I want. So UI isn’t going to be an issue in the near future… so what will matter? IMO software installation and uninstallation. In windows and mac it’s rather easy to install and uninstall apps… in linux it’s a bit more difficult (although it’s not *that* hard). When linux gets better install/unistall methods it will surely pick up marketshare in the workstation market, and that will lead to more software for linux which will compel other companies to also release their software on linux.. and maybe we can finally be done with the monopoly of Microsoft. Thank god linux is here though, because chooosing between Microsoft and Apple is really just picking the lesser of 2 evils. Apple isn’t much better than M$, and atleast M$ doesn’t force you to use their hardware.
He didn’t say you have to remove choice. He’s not pushing and end to all of your favorite programs. He’s just saying that LSB need to agree on one to be standard, and make the others optional.
Someone earlier made a comparison that there are lots and lots of email clients for OS X. This is true, but only one comes bundled with the Operating System. Only Mail.app is included. All the others must be bought/downloaded for yourself.
Someone was advocating using the command line for package installation. That’s a perfectly horrible idea for end users. My dad became a Mac user in the early days because he was terrified of the command line in DOS. My aunt still lauds Windows XP because she doesn’t have to use “that dreadful command line” like she did in DOS. Average users are afraid of the command line. Deal with it.
This is not a proposed end to freedom. You’re still free to install what you wish. It’s just a proposal as to what should be included[/i] and what should be user’s option only. In all likelihood, Joe User doesn’t care whether he uses KDE or Gnome, so making it a major choice for him is merely confusing. Stick with one as the standard, and let those who really love the other install it themselves.
Someone also said something about “it’s already consistent within a distro.” That’s not the point. Linux is recognized as a name, not SuSE or Mandrake or Lycoris. Joe User decides he wants to run Linux. He should be able to go out and buy any distro, and get a system that will be largely compatible with any other distro. If he uses his friend’s Linux system, they ought to be similar enough for him to know what he’s doing. People coming from the Windows-desktop world expect a system to have a certain degree of uniformity. Yes, it is possible to create a highly customized form of Linux, but that shouldn’t be the default.
This is one of the most useless “What Linux needs to rule the desktop” articles I have ever read. It completely discounts what made Linux so attractive in the first place, and wants to turn it into a non-commercial version of Microsoft.
What a shame, all those developes wasting their efforts on projects you deem useless. What about all those writers and painters, who squander their energies on books and paintings you never see? Wouldn’t it be better if they could just pool their efforts into one or two great books? Or all those appliances from different brands that keep coming out. Damn, that gets confusing! How am I supposed to keep up with them all? Couldn’t they work together and come up with one really good blender and be done with it?
For that matter, do we really need two operating systems on PCs? Isn’t one easier for everybody? And since WIndows owns 90%, doesn’t it make sense to stick with that and improve it?
The fact is, I have alternated between KDE and GNOME for years because they have competitively gone back and forth with the quality of their features. I used Gnome until KDE hit 2.0, and have been using that ever since (because of anti-aliasing). But I suspect I may go back to Gnome 2.0 once it becomes more mature. Or, maybe I go to a BeOS on top of Linux desktop environment someday, or whatever else is new and better that comes along.
So this is really pointless since distributions already take care of this. Go with Xandros or Lindows or Lycoris. My apoogies that that they don’t support Gnome, but that leaves an opening for a Distribution that does. I like that Red Hat, Mandrake and Suse support both (and many other window managers). If they didn’t, I would have to go somewhere else.
When I created a KDE REDHAT 8.0 environment for my girlfriend’s computer, I went through all the set ups and configurations to make everything work and simple for her to use. THIS is what needs fixing! Simple Wizard tools that handhold new users through the process. That’s why K3B has become my new favorte CD Burning App. It’s so damn simple. But — I keep an eye out for Arson and others, fo the day when those Apps are better. Hence — the greatness of competition.
Enough time wasted on this. The packaging problem is true and needs resolving. But otherwise…
Holy smoke, some people are just *dense*. No one ever said kwrite, kedit, gedit, nedit, pico, vi, nano and bluefish should bow to emacs. No one said Sylpheed, kmail and mozilla mail are out of the question since there’s evolution. Who the hell was the first to start the whole “people are stoopid for wanting to take away our choice of software!” crusade?! No one ever said anything like that. What’s only ever been (seriously) suggested is not putting 4 ASCII editors side-by-side in a menu by default. I didn’t think anyone would be zealous enough to miss it, but it bears emphasis: KEEP THE DAMN CHOICES, but keep them as “optional” instead of “default.” So we keep our choice, but at the same time we spare the indecisive among us (and also those of us who care how inane it is to have 4 of the same thing cluttering our desktop and /usr/bin).
To those who think there’s a single answer to all of our packaging problems- explain to me why people are still having problems with it, and don’t start the arguement with “well cuz they’re stupid for not switching to <insert system I’m used to>”. And emerge isn’t a package-dependancy-panacea, no matter how simple it is.
And finally, someone mentioned that “dumbing down” isn’t the way to go, but educating the user *is*. First off, no one’s dumbing anything down. When ya start with something like Linux that has a community like this one backing it, don’t worry, it’s always gonna be technical. Second, the user’s aim is not to have to learn a new skill in order to send mail and look up air-fare on the Internet. They often don’t WANT to be educated. Techies are supposed to bring the service to them, not they bring themselves to the services. That’s not what a good service sounds like! and if you don’t like catoring to the needs of users, stay out of discussions on Linux for users.
“Awa ! Good ideas !
One email client, one browser, one desktop …
one government, one plannet ??!!
The point of free software is about choices.
Do not pretend to be God, and decide what is good
for everybody.”
The author was NOT deciding what is good for everybody, nor was he recommending we actually DO AWAY with all the apps. What he recommended was only INCLUDING one of each app with a distro. Read the damned article before commenting.
“Please not another “what Linux needs” article… it does not change anything anyway. Why? Because the only way to make a change in the Free Software world is to contribute.”
Not true at all. Blind contribution, without any sort of common goal (and don’t come back with some etherial garbage like “the goal is choice!”) is what has led to so much fragmentation and duplication. And if you don’t like reading articles on this subject, why the hell do you bother reading it and commenting on it? Moron.
I, for one, agree wholeheartedly with the author. Good job. Yes, adopting standards does decrease the absolute measure of choice, but so what? It increases compatibility and productivity. And it in no way will do away with any of the fifteen thousand text editors out there; it will simply standardize THE WAY Linux operates, not WHAT will operate on it. So suck it up and march on.
Not true at all. Blind contribution, without any sort of common goal (and don’t come back with some etherial garbage like “the goal is choice!”) is what has led to so much fragmentation and duplication.
Yes, and this is why free software works (and free markets as well, BTW, and you are proposing the equivalent of planned economy). There is no real leadership. Everybody is free to fork any project, or to start a new one. If it is better it will supersede the old one. If not, it will die, sooner or later.
As somebody else already pointed out, there is no willing mass of work force. The reason why there are two major desktop environments is that there are two large groups of developers who fundamentally disagree about what is the best approach, and you can’t change this.
You can make the decisions for yourself, or you can pick a distribution that made the choices for you (like Lycoris and similar distributions did). But how do you want to change the choice of other people? What you are saying is like “Xandros and Lindows should stop selling their distributions, so we can standardize on Lycoris”. Or “General Motors should stop making cars, so we can standardize on Lexus – everything else leads to fragmentation and duplication”.
It’s completely useless whining, and do indeed regret that I spent so much time arguing about it instead of working on free software.
Why should the FSF tell me what to call it, when people recognise the name ‘Linux’? GNU won’t do anything as a desktop either, for that matter. Hmm… XFree86 isn’t GNU. KDE isn’t GNU. Mozilla isn’t GNU. ls and date are GNU, but they aren’t desktop tools… GNU/Linux and Linux are as much a desktop as the other.
You seem to be confused, let me put some things into perspective for ya (taken from the kde.org website):
1. Each and every line of KDE code is made available under the LGPL/GPL licenses.
2. The KDE CVS source code repository holds currently about 2.6 million lines of code. (To put things into perspective: The GNU/Linux kernel version 2.5.29 consists in about 3.1 million lines of code.)
This brings me to another point. People keep talking of Linux on the desktop, but a lot, if not most, of what they talk about applies equally to (say) FreeBSD. Really, it could be ‘X on the desktop’. But I guess one letter isn’t as user-friendly or as topical as ‘Linux’.
Ever hear of directfb? I think its GNU, even supports the X protocol, but might still be alpha, I dunno. The point is this software is almost a feature complete desktop lacking only a couple minor things that I’m sure will be filled in within the next couple years. GNU is the license covering 80-90% of the software that makes up any typical Linux install. Its not like Linus developed this kick ass kernel and all of a sudden GNU came around and started making demands. If it weren’t for the GNU license Linus chose its quite possible we’d equate Linux to that Minix operating system you keep hearing about. But the Hurd would still be where it is today, possibly further. GNU is a force to be respected. I’ve never seen anything like it, have you? BSDs aside…
As a newbie who has gone back and forth between Windows 98 and SuSE 7.3 (with Win4Lin) for the past year or so, I have found both KDE 2 and Gnome 1 to be valuable, with KDE 2 more mature and stable. The problem is that despite the availabilty of regular (and, apparently, excellent) upgrades to my software and OS components, I have had limited success upgrading.
I was, after a couple of tries, able to upgrade to Mozilla 1 and it is the best browser/mail app I’ve ever used. However, I’ve been completely unable to upgrade either KDE or Gnome due to seemingly insolvable problems with dependancies and other arcane infrastructure issues. For example, I’ve upgraded Qt, which is required to upgrade KDE, but have been unable to convince the system (or package manager, or whatever) that I have done so, despite many alterations to configuration files. I have posted my problem on a newbie site and received friendly but ineffective guidance. As a busy professional, I have decided to move on and live with my existing installation as-is until I upgrade by purchasing a new version of the SuSe package. (Here is another upgrade issue–I can’t just upgrade easily SuSE 8.1 from 7.3: they say that it MAY work, but I MAY have to just reinstall the whole system if the upgrade fails. So, I’m not jumping at the opportunity to upgrade this way just yet either.)
In order to get folks like me, who have influence over institutional desktop choices, to recommend linux for a broad range of users, there needs to be a simpler way to install and upgrade software and OS components. One of the strengths of this community is that the software is always being reworked and improved. However, this strength is lost on many new users because of the complexities of managing installation and upgrades. From what I’ve seen in my attempts to get help with my own problems, more experienced users also have ongoing trouble with installation and upgrade, too.
I can’t recommend a solution. However, I can identify this as really the only place that the OS (or perhaps just my distro) falls down, in my experience.
I had the same worry whn I wanted to try and update my Mandrake 9.0 to 9.1 Beta3. I figured I might have to reinstall everything and lose my preferences etc.
Turned out to be pretty simple: booted CD 1 selected Update and 30 min later everything was updated. Now I’m not sure that this would work in SuSe or if it will work between major releases but I was very happy on how esy it was.
Go ahead, it normally works OK. The “maybe” clause is there because Linux users tend to do awkward things like build custom kernels, upgrade the desktop manually and so on. If you haven’t heavily tweaked the setup it’ll be alright most likely.
… is that developers want to work on glamorous software and will only give a half-assed effort to the less interesting stuff like good documentation, user friendlyness, polish (just look at ANY MS app), stability, reliability, and robustness (not that I blame them, I understand, I’m a developer myself). How do all these things get done? Someone has to pay them to do it. That’s the problem. The result is that you end up with developers working on the same thing lots of other people find interesting and then you have several text editors, or several GUIs, and none of which is really complete or will ever be for the reason noted.
BTW Some people need to be reminded that an OS is a tool to accomplish a goal. While you may like to tinker with command lines, that is not what the majority of people want to do. They want to accomplish work at minimal cost (I’m talking about time, purchase price is meaningless).
Steve Q.
I agree and disagree on several points. First off, MS apps are often not-polished, are a usability nightmare, and are often inconsistant interface-wise with MS’s own other apps. People keep thinking “Microsoft is a billion dollar company. They must know how to design UI’s, or else people would never buy their software.” Kind of like some people in the business world think “Microsoft is a billion dollar company. The must know how to make software secure, or else people would never buy it.” Microsoft is a textbook example of how not to do things.
But I completely and totally agree that linux’s number one problem are the linux developers themsleves. They do not have a developer culture that values the end-user experience, and efforts to put linux on anything other than a server fail miserably because of this. If the linux developer community had put a fraction of the energy and skill into making a usable desktop system that they put into making the linux kernel and Apache, linux would have been a desktop-ready OS by 1995. If linux developers are completely unwilling to understand what they have to do to put linux on the desktop, they should quit whining about Microsoft and traipse back to the server closets they came from. Given the way they have acted, they have shown that linux in its current form is clearly not deserving of a place on grandma’s desktop.
The solution is not to get corporations with lots of money to force these developers to clean up their mess. If you have to save a desktop OS by using money to overcome the developer community’s apathy towards the end-user experience, that desktop OS never really had an ass worth saving. Only OS’s that have a developer community that does care about the user experience should be put anywhere near desktop computers.
The solution, IMHO, is to found a new linux developer community, completely devoid of the old guard linux developers who bring that apathy with them. This new linux developer community would value things like good documentation, usability, and general polish as much as they would value technical stuff. This is the way to create a viable alternative to Microsoft.
thats the point of open source software, to allow freedom of choice. if a newbie WANTS to use OSS then he takes the pain of learning it. you had to learn windows at one point didnt you ? its not got the easiest conventions, but its the most DOMINANT one and therefore people set standards by it.
i dont see why people get sore about linux not having the major share of the market. let people use what they want, how they want.
everyone has their own prefs and you cant say “but that sucks, its lame etc etc” cause that person will just turn around and say “but it lets me check my mail and irc”
i agree that we could make it easier, but it cant be a forced change. OSS is about the people that make the software and the people that use it as well.
>1988?
> By Kon (IP: —.dsl.lsan03.pacbell.net) – Posted on >2003-02-17 17:28:38
>So the author started using linux in 1988? Funny, it wasn’t >even out in 0.01 for until the end of 1991.
Kon needs to learn to read. It says 1998.
>That’s why K3B has become my new favorte CD Burning App. It’s >so damn simple.
Shoot. You blew your argument. CDRECORD at the command line is easy. If you get confused then fire up cdrecord –help. You don’t need all that fancy front end crap.
I was about to respond to some of what’s been written before but maybe I should ask someone to clear up some of what is confusing me.
Maybe the original author is right. Less choice for the beginners.
Using a 56k connection. In the Caribbean. This means hours of downloads or paying someone who has a fast connection and therefore I won’t be trying too many different distributions.
What should I choose: Which Distribution? What Software? What can I do with it?
Will I still be able to open and edit MS Office documents? How can I essentialy replace MS without too much work or too many problems?
These are the things that keep me from switching to Linux and until these questions are answered, I guess the switch won’t be too soon.
I liked the idea of Lycoris — Install everything all at once and work with it. I tried it but had a problem. I got lost in the basics, I did not understand the basic level of the OS, (DOS-like stuff, how to create and delete files and folders. I don’t think I could even find them. How to set up software. Where to get it from etc. )
I installed Linux on the slave drive of a computer running MS Windows, then the Linux drive crashed. I could not access the master drive at all, I had to resort to FDISK and reinstall MS Windows. Yep, you guessed it, I’m no expert.
I know some people resent helping newbies ‘tie their shoelaces’ but that’s the stage some of us are at and maybe we need a little help and advice at first. Thanks to MS we already know where to find the slip-ons.
Please. No insults, If I wanted insults I got a mirror.
Just list the stuff that can get me started on the trip out of MS
Whilst I agree on most of the points in the article, this :-
“Moment will soon come when Courts will forbid vendors to preload an operating system when selling a pc.”
I cannot agree with.
There is no way that any court will be able to pass a ruling like that and in fact, it could be damaging to Linux, as it will forbid the pre-installation of Linux too.
Microsoft gained the majority of it’s market share from pre-installed computers.
They will try to do it again with home entertainment systems and will probably succeed in doing so. (The XBOX is just the beggining of this revolution)
You can’t expect Average Joe Public to install thier own Operating system, unless it’s as easy as inserting a CD and pressing one button.
Pre-installed computers is an area where Linux can reach the masses simply because it brings the price of an operable PC right down.
Anyway, PC Sales have declined and will continue doing so.
The next revolution will be Home Entertainment systems and embedded Operating systems (linux is making good inroads here)
The final choice is obviously between KDE and Gnome, the most mature projects up- to-now. I would choose the latter.
To me, it is like saying: “The final choice is obviously between English and French, the most mature languages up-to-now. I would choose the latter.” Face it, both of these desktops, and the others as well, will exist until all of their users perish or fail to reproduce (a real possibility judging from my local LUG
.
SuSE is a good choice for desktop users with limited bandwidth. Solid commercial KDE distro with tons of software on the CD set.
“I am not asking for the moon: I would just like all the distributions to default-install some packages instead of others, and let expert users the ability to modify this default choice and tweak their environments.”
As I read this lines, I only can say YES. Many M$-users don’t want to change their OS (windows -> linux), since they are worried about the difficult install-progress. So one or perhaps two default-install-options can solve this problem (it could called “windows-like-install”). For example: What programs need M$-users?
They want a clean and simple desktop-system and an office-system, one browser including an email-client and one multimedia-viewer.
So take this:
– KDE or GNOME
– OPENOFFICE
– MOZILLA
– EVOLUTION
– ??? (it doesn’t exist an multimedia-viewer at all)
Ok, that’s the first step. The second one is to make installing new software easier like today (one way is the redhat packages). Behind these steps we have to develop the needed libraries and drivers. For printing and scanning there are already CUPS & SANE as well-developed systems including nice graphical frontends.
So we need not to “windowing” linux, but to make it possible for windows-user (or windows-like-user) to change. We can save the multiple linux-world and make linux accessible for normal desktop-users at the same time. We don’t have to mutilate linux to do that. But most users aren’t interested in the os. The only one they want is to use their applications daily. Therefor we need at first a default-install option, which installs the selected basic applications (s.a.) and installs the needed libraries and so on at the same time without any explicit action by the user.
I LOVE LINUX.
large amount of software cannot be the issue, in a negative sense; too many sing the song that Linux does not have enough “mainsteam software.” Variety of software, too many choices: let’s look at that. I believe no one minds thousands of working programs ready to run at the first click, like a file editor for example. I feel what people mind is thousands of pieces of software I must configure before I can use it:
I feel no Linux programmer should be allow to distribute software that does not have a default configuration ready to run.
Software needs the shine of usability as a public resource; too much stuff released into the public is still built-for-other-programmers. And of course other programmers, as many as they are is not the mass market.
Also, I feel each piece of software should have an email address in about encouraging users to speak up and say what they like and don’t like and what they want more of. Any piece of software that seeks to be serious need to achieve critical mass. Again it’s the public thing.
Look at SuSE on their kdm login: a nice dropdown list of window managers that were installed all configured and ready to run. It a pleasure and an education to look at different desktops available (I use blackbox for daily chores and kde for admin stuff). Also, look at how SuSE impliments YAST2: notice how you get a login box so you can login as administrator and use the feature (not that you got to know to su first and know how to get a gui up using an terminal interface.
Also look at apt4rpm-with-snanaptic, it’s real nice but can use more work to shield the users system better. For example it could use a local database so it can come up faster and update itself as a background process.
Just use me baby!!!! … that all I want to hear from my software, so to speak.
Ease of use, THEN massive power. Linux is getting there; more distros need to be more aggressive with ease of use
Hey, developers! Read this, my suggestions for a unified, well-designed installation system..
::
Here’s my suggestions:
1) The system should be unified and system/distro independent. There should be only one version of the installer software.
a) To ensure the above, the software cannot be licenced under GPL. It must be open-source, but not GPL. We don’t want a repeat of RPM.
2) The installer will not be binary based. It should be just a single configuration file that controls all aspects of the installation; making it easy to understand, program and adapt. It also ensures Linux disto-wide and even UNIX-wide compatability.
3) It should be flexible, because of the above. The user should be given options on how and where to install the software; what location, where it will be integrated (e.g. KDE Kicker, GNOME menu).
4) It should be secure. There should be some kind of verification for the installers, as users won’t have to log on as root.
5) Advanced information should be accessible. There should be an option to view the compiler used, optimisations, architecure, etc.
My e-mail address is [email protected] if you wish to contact me.
Other notes:
a) The software shouldn’t be installed in insecure locations such as home directories. Why? The user installs the software there, runs it and has full privilages over it; it can read, write, delete and modify it. Not only that, but programs run by the user can too. It leads to the user having “root privilages” over the code.
b) We should maybe design the system where you build (i.e. compile) the code. Some Linux-desktop proponents say “no way, that’s geeky!” but I don’t think it’s such a bad idea. It’d be very useful for updates and patches. 98% of the code on a Linux distro have their sources available on the web, and often the filesizes for the sources are lower than the binaries. Most importantly, it’ll solve much of the huge difficulties of having different Linux distros and versions, architectures, etc. If the system is made right, building (compiling) applications could be, for the average user, pretty normal; part of the installation procedure. On the fast 2Ghz+ processors we have today, compiling most applications doesn’t take that long.
c) This is my philosophy for installing software. Its a three-step process:
1) Copying files – files are copied to the required destination.
2) Linking files – making them work with the main system.
3) Integrate – make it work with the user interface and easily accessible by the user, e.g. have links on the KDE menu, GNOME menu (we should also have a common system for menus) and the software installation menu.
d) Learn from other installers. Forget RPM and APT. Nullsoft’s NSIS is a very efficient system, used widely for Windows applications <LI>http://www.nullsoft.com/free/nsis/</LI>. Work with the programmers even.
e) The extension? (I know KDE and GNOME identify files by MIME types, but us humans can’t, browse an FTP directory and you’ll see.) My suggestion would be [apptitle].inst.tar, and the convention for [apptitle] should be the applications title, and its version, with no dashes, spaces or dots, e.g. gnome22.inst.tar. Other information about the file should be contained in the filename, e.g. ‘b’ for betas, ‘P’ for prerequisites included, ‘m’ for minimal, and ‘s’ for the standard release, where ‘s’ is assumed where there are no extra characters; e.g. kde33b2.inst.tar for KDE 3.3 [Beta 2] and vorbis10P.inst.tar for Ogg Vorbis 1.0 [Prerequisites Included]. Others could be used too.
f) For the source code installation, you just download the tar/bz2 or tar/gz (preferably the former as it compresses better) by FTP or HTTP, and then download the inst script [appname].inst.bz2 from the projects website or mirrors (the inst script is compressed as it could be large)
g) The name? The system could be called Integrator, LSIIS – Linux Software Installation & Integration System (well maybe not that because of the ‘IIS’ in the name, and it sounds a bit too like NSIS), any ideas?
Come on, do I have to build the system myself? (And no, I don’t have programming skills.)
Everything you need to build the system is above in those design recommendations. If you want to set up a website and start developing the thing, you can contact me by e-mail at [email protected] and I’ll contibute. We can start with just publishing these guidelines on the site. I also have some ideas for desktop enviroments and GUIs. We need to start with the Linux installation system now.. let’s get this thing on the road!
Any readers may freely copy this and post it to newsgroups, mailing lists, message boards, etc., in its entirety and unmodified form. I’ll amend it when we have a website put together.
Hmm, yet another article that points at the weak-spots of the current distributions. Nice, but imho it’s time to change. Our current problem is clearly overchoice, something nearly as difficult as having not enough to choose from :o)
Working on a nice modular distro called morphix ( http://www.morphix.org ), Debian/Knoppix based, and the main principle is that users shouldn’t have multiple solutions for common problems, but instead a small number (under 10, now only 3, then again we’ve just started) of pre-installed application-modules, ready to run on a liveCD. One module, one sort of user, and no overchoice.
YALD? maybe, but at least we can try (distro overchoice, now thats a real problem ;o)
Ah, if I were to use evolution I would fall asleep everytime I executed it as it THE SLOWEST EMAIL APP EVER!
Now I’ve given my opinion in this matter =)
Hi all,
I read every comments, sorry not for replying immediately but I really have no time in this period. I can’t answer to everyone, but I’d like to better specify a couple of points.
Courts.
When I referred about Courts probably denying to preinstall operating systems, I was just making a guess. I have no information about a similar decision at the moment, not even in Italy.
Anyway, chances are that, if someone really decide to seriously investigate on the matter of preinstalled stuff, then he decides that the only way to stop Microsoft monopoly is by eliminating it (not Microsoft, I mean the preinstalled stuff
.
This is just one possible conclusion of one possible investigation, if any…
I don’t believe this would be bad for Linux: of course we couldn’t see any preinstalled one, but vendors should be obliged to explain buyers what an operating system is, pros and cons of Linux Windows OsX BSD etc., so Linux would gain more visibility to beginners too – and then they would choose what to buy!
Choices.
As many of you noticed, I don’t want to eliminate choices, I only want some default preinstalled packages for every distribution and let expert users change them – this would be very easy if a good and advanced package management system is chosen!
Someone told ironically that, since 90% of the people use Windows, then – following my article – everyone should standardize on Windows. My answer is no, because Windows is not stable, not secure and not opensource: if it was, maybe I would standardize on it – because that’s what most of us want: stability, safety and warranties.
Just one note: I was emailed that, maybe, LSB should not be concerned with it, and the stuff I wrote should be interesting for the new Linux Desktop Consortium instead. I agree. I’m just worried that it contains many Linux distributors, some who bet on KDE others on Gnome, etc., so I doubt it will come to a concrete solution… but I hope!
Thanks everyone
Like kurumin linux (www.guiadohardware.net/kurumin)in portuguese.
A complete distribution in 180 MB.
LSB needs to go for the Slackware/SuSE filesystem layout which is more in keeping with the Unix tradition. It makes absolutely no sense to shove everything into /usr like Red Hat, Mandrake, Debian do.
Here’s my view of how the filesystem should be:
/
/dev
/proc
/disks (or /mnt)
/opt
/opt/kde
/opt/gnome
/opt/gnustep
/opt/bin (aka /usr/local/bin)
(any other app not essential to the operation of the system)
/usr (stands for Unix System Resources)
/bin (system utilities)
…
I always hear complaints about the lack of “one click” software installation in Linux. What people fail to realize is, that we’ve tried OS’s without security, and guess what? They _don’t work_. Just because it was “easier” doesn’t make it better. I’d prefer having to enter my root password in every ten minutes than have a OS without any real security.
Also, we’ve all seen how well “one click” worked on Windows. It seems like half the time, the crap doesn’t uninstall afterwards. I have far fewer problems with RPM and Loki’s installer. Sure, you can’t just uninstall whatever the hell you feel like with RPM, but then again, your stuff typically works afterwards. Can’t say the same with Windows and their “well, I don’t _think_ it’s being used anymore” routine with dll’s.
Easier != better. Everything in life has a tradeoff, and I think properly resolving dependencies and having real security is worth a little hassle. It’s not that hard, you know?
-Erwos
One of the major things that I like most about linux is that I *DO* have several progs that do the same thing. It is all about choice. I should be the one that dictates what I use, not someone else!
And sometimes, it even isn’t a matter of what I like, but what suites the sitution. One VERY common one I deal with is a slow, small-RAM computer where I need to run some GUI progs, such as mozilla to view some docs for something I am doing. I could fire up KDE or GNOME, and they work fine, but they hog up a lot of resources. When I am using a P200 with 64 megs of ram, TWM, while far `cheaper’ runs far faster!!! Not that I would use it if I had a better comp, but this is just an example on how CHOICE is what makes it different. People who do not like having a choice, and like having decisions made for them, go use Micro$oft crap and let Bill Gates’s vision be yours too!
Linux is all about choice! THat is one of the mane advantages it has, not disadvantages!
-Monty Warhawk