"What is Linux Missing?" 720
three55ml writes "This is an editorial I wrote noting a few points about what Linux has to do before it moves totally mainstream. It talks about both the small and the large issues currently slowing the widespread use of Linux."
What it needs is just one killer app (Score:1)
Conflict of interest? (Score:1)
This is a serious issue that needs to be resolved, as it will seriously impact what Linux becomes -- will it remain an OS for the "l33t" only, or will the Linux community accept and promote the lowering of the Linux learning curve? In a perfect world, the majority of the computer-using population would have the time, patience and will to learn these computing skills; In reality, the majority of computer users are people who just want to do a few simple things: write papers, e-mail, cruise the internet, and do their taxes. They want their computer to be like a toaster... simple -- they just don't have the time to figure this stuff out, their lives do not revolve around the computer the way ours do... If you want Linux to be for the masses, sacrifices must be made.
Linux needs only one thing: APPS!!!! (Score:1)
The only bit of resistance I EVER run into is that
Linux won't run users' favorite apps. For example
a small business owner the other day who depends
on Print Master, or an accountant who needs his
accounting software, or a kid who wants his
favorite game.
Judge Jackson was correct in pointing out that
it is the "applications barrier to entry" that
Microsoft protects at all costs, and it is the
single thing that keeps Windows going...
The only way to make Linux more successful to the
average end user is to convince application
vendors, e.g. ID, that they can make a profit by a
Linux port... when Linux reaches critical mass in
applications, the road to World Domination will be
paved.
Mark
Here's some solutions (Score:1)
Making Ethernet Adapters Work with Linux [scrounge.org]
Using Linux with a Cable Modem [scrounge.org]
DHCP problems? Try using dhcpcd instead of pump. [scrounge.org]
Use IPWatch to monitor your IP address [scrounge.org]
Re:Why I use Windows, and not Linux (Score:1)
Re:A Standard UI (Score:1)
Re:Use the Macintosh Human Interface Guidlines (Score:1)
Driver compatibility... (Score:1)
SDL, PLIB, Clanlib, CrystalSpace (Score:1)
Good printer drivers. (Score:2)
Before Linux is going "Mainstream", the driver problems must be resolved.
IMHO,
M.Aartsen
maartsen@wildcard.demon.nl
Support problems (Score:3)
We, as a community, have no control over hardware producers willingness to support the OS with either commercial drivers, or open specs to driver developers within the community, other than as customers. The only way that we will arrive at a situation of having Linux supported is by asking, every time we go into a computer store, "Does this work with Linux? Is this app available on Linux? Can you order me this game for Linux?" Once we get the stores to recognise a demand, then we may get proper support and recognition.
A problem I see with the community is our willingness, all too often, to accept that things are the way they are. We do our research before buying hardware, so that we know beforehand if support exists. If we buy a piece of kit that doesn't work, we reprimand ourselves, not the producer. We don't make a fuss.
If you ask me, the way into the mainstream (and it's by no means a given that that's where we want to go) is through financiers pockets, combined with millions of requests to helplines, e-mails to support@supplier.com and bemused queries as to why our local tech store isn't stocking Linux commercial software, or why hardware in the store doesn't have a little tux sticker on the box.
For too long we have suffered in silence. Now it is time to ask for help.
about the slow printers... (Score:1)
i guess this is just another driver thing. those makeing personal printers make windose drivers (and mac drivers if your lucky) and linux just does it like the rest of unix, postscript. unfortuneatly, postscript is unheardof in commercial software land. if adobe was not greedy about postscript it probably would have become the standard it was supposed to be.
(1) This is expensive. all ive seen are laser printers.
(2) Sco, i dont care what your lawers think, linux is a unix.
Re:flashy educational software (Score:2)
I feel I've seen this 20 times now... (Score:2)
- New, cheap computers with Linux installed. Setting up X is horrible, and that's just the way it is. Setting up Windows isn't that easy either. So we'll avoid it just like Windows does. The great thing is that Linux computers could be really cheap. We need an eMachine, not a VA Linux.
- Printing needs to be fixed. Printing under Un*x just isn't very good. Ideally, a whole new system could be set up, complete with nice libraries. Maybe CUPS [cups.org] would be a good alternative. I don't see any real salvation for lprd.
- Applications aren't that important. Lots of people never need anything more than the MS Works that comes with their computers -- StarOffice clearly matches that, and hopefully with a little polish the Free alternatives (KOffice, AbiWord, the GNOME efforts, etc.) will be there soon too. Polish is more important than features at this point.
- Mindshare (uck... I feel dirty for using that word). Anyway, people have friends who know Windows, and that provides an important support structure for people. This makes for a chicken and egg problem, but the Internet has already shown us the solution. Linux has a great Internet support structure, but for Internet newbies this can seem a little intimidating. Making something a little more friendly would be helpful.
- Games. Linux doesn't have lots of good games. People like games, parents buy computers with the notion that while the computer might not be for games, the kids would really like to play them. And some people just buy a computer to get games. LokiSoft [lokisoft.com] is really the best short-term model for this. Commercial companies will dominate this for a long time, and that's just fine IMHO.
- A few things need to be hidden. Like all those libraries. Debian's Apt provides a good start for this. People shouldn't need to think about what libraries they have to have installed, it should just happen.
Applications aren't that important. People like to buy a computer that can do something before they get more software. Linux distributions already typically include far, far more functionality than a normal new Windows install. This is something Linux advocates should be pushing. Shrinkwrapped retail boxes aren't where Linux will shine, and it doesn't need to.Ironically, that's why I don't use it... (Score:2)
Linux is definitely worse from a simplicity stand-point and I think the people who consider it a non-techie desktop replacement are getting way ahead of themselves. I use linux on servers and my desktop, but I'm also not going to be scared about using installing the latest Netscape (or compiling it [mozilla.org] for that matter). The reason why I consider Linux a win over Windows is that, although it may not be as simple, it's much more predictable. I've never had a Linux system where an application magically corrupts itself to the point where you need a fresh reinstall of the operating system.
What I've been eyeing as a potential desktop OS replacement is BeOS [be.com]. It's got a very slick and polished GUI and, unlike Linux, the user never needs to go near a command prompt if they don't want to; if they do, they'll find a well equipped command prompt with all the usual Unix software (lots of ported OpenSource stuff). They just aren't forced to use it. It's got very solid multitasking and handles a heavy load much better than Windows (it's a tossup between BeOS and Linux on my machine as far as scalability goes) with very well developed multimedia support, too.
The best part, however, is the speed. Installing it took something like 15 minutes, which consisted of telling it which partition to use and most of the time was spent copying files off of the CD. A 5 second reboot later, everything was up and running - no plug and pray, no hardware configuration dialogs. The only config I had to do was providing my IP address to the network driver. Everything else was auto-detected without more than a second's delay.
The bottom line: BeOS is fast, extremely stable and it's got a very easy GUI. My mother could install this and use it without training even though it's also got enough neat technology that I find it interesting as well.
(A few weeks ago, I summarized my experiences installing stuff on my new PC [improbable.org].)
Shareware hell? (Score:2)
Re:Ironically, that's why I don't use it... (Score:2)
Of course, in real life, you get Close Enough using the filesystem under /etc if you enforce a sane directory/file structure. We just need to civilize those remaining programs (e.g. sendmail) which use arcane mutant config syntax...
One thing missing: Games games games games games (Score:3)
Most everything else is important, but game software needs to be taken serious to get Linux onto the average Joe desktops.
Re:A Standard UI (Score:2)
Consider for a moment that people keep finding it necessary to create new UIs; this is evidence that the elements needed to establish such a standard have not "settled down" yet.
Yes and no (Score:2)
When you need to repair your car, do you go back to the dealer you bought it from? To one specific garage, three streets down on the left from the supermarket, every time? Do you go there to check your oil level? Add gas? Clean the windows? Top up the anti-freeze?
When you are cooking, do you call the shop you bought the food from, for the recipe? If the cake is falling, do you call the people who made the flour for a fix?
If your house is burning down, do you call the real estate agents or the fire brigade?
The fact is, the ONLY facet of society people call a single source for EVERY single problem is in computing. It's simply not natural for people to do this. People naturally call the people most likely to be able to help, not a sales clerk.
So why not break the habit, and do what WORKS, in computing, as well as everywhere else?
Pump vs. dhcpcd (Score:2)
Windows easy? You haven't been at it long enough (Score:2)
Judging from what you've said it looks like you've never been through DLL Hell [desaware.com].
There may be a lot of things that are easy in Windows but software installation management is definitely not one of them. Jaded Windows users may be impressed by one-click self-extracting installs and Install-Shield uninstalls, but that's only because they've never seen a real package manager like RPM or dpkg in action.
I will concede that software installation in Windows is easy if you can show me in Windows how to:
- List the files that Word97 installed on the system,
- Tell me which programs require ctl3d32.dll or any other particular dll,
- Show me what programs on my system will break if I upgrade to Office2000.
These are all utterly fundamental chores for a system administrator. You cannot argue that Windows software management is easy until the day comes when Windows can do these tasks just like RPM and dpkg can today.Comment removed (Score:3)
Re:A Standard UI (Score:2)
Applications and drivers. (Score:3)
While KDE and Gnome and the likes are not perfect, they definitely give the mainstream user a comfortable, usable GUI.
Installation procedures have been improving a lot too, just look at Caldera or Corel. Besides, there are tons of books such as Easy Linux that start with explaining drag-and-drop, so even the new computer users can probably learn Linux without a lot of obstacles.
What Linux continues to need though, is all the latest applications and drivers. The diversity of reasons not to use Linux I hear has shrinked to "I cannot run FavProg97" or "my DVD/webcam/whatever isn't supported".
There are many viable alternatives and workarounds for this problem, but to go mainstream Linux should no longer require workarounds. Native support, mainstream.
My mother thinks the computer at work is running WordPerfect. Since the has no interest in the OS and mostly types papers and letters on the computer, she cares about WordPerfect. I set up a Linux machine at home with KDE and WordPerfect and she continued exclaiming: "yes I run this at work too!", even when I showed some screensavers.
Linux was ready for my mother because it supported our hardware and ran the programs she wants to use in a solid, clean interface (KDE). Linux will be ready for mainstream use as soon as it supports most or all hardware and programs.
Re:How about a DHCP client... (Score:2)
First, make sure your network card is detected. You can use modprobe, but I have PCI NE2000 support compiled right into my kernel, and in my opinion, that's the way to go.
In
# Connecting RoadRunner Services (replace ethX with your ethernet device.)
dhcpcd ethX
RoadRunner no longer requires a Login, so it's very easy to get it working in Linux. Once you add the dhcpcd line, you should be all set.
Also note, this is for Time Warner RoadRunner. Your cable service may require a login, in which case you would have to use the OS that the login program was written for, or get the Linux equivilant if it exists. I know several RoadRunner login programs were written for Linux before Time Warner eliminated the login, so you may have some luck.
(NOTE: I'm not responsible if you screw up your system, people. This is just how I got my connection working, and working quite well I may add.)
-- Give him Head? Be a Beacon?
KDat: tape backup! (Score:2)
Thinking more about CD-R backup, it seems it wouldn't be hard to adapt KDat to burn CD-R disks by frontending existing burning tools. I figure the resulting tool would work as follows:
Font rendering? How about font management! (Score:5)
I'd like to take the complaint over font handling a step further. I agree that TrueType support should be part of the standard fontserver. Mandrake already does this, replacing xfs with xfstt, and unless there are license conflicts, other distros should follow suit.
But further hobbling things is the horror of installing and managing fonts under Linux, and under Unix in general.
More fonts, Postscript and TrueType alike, with their accompanying config file additions, should be packaged as relocatable noarch RPMs.
There should also be simple command-line tools and GNOME and KDE frontends for installing and inspecting not just these font RPMs, but also for one-step, jargonless, click-and-drool installation of arbitray Windows and (binhexed) Macintosh fonts and font bundles. Such tools should transparently perform all of the steps necessary to install such fonts under *nix, perhaps even creating the aforementioned font RPMs as an interim stage. Added bonus points if the same fonts were automatically made available to TeX, not because I think TeX is going to supplant word processors, but because it should be pushed hard as a Free Software answer to the Crystal Reports layout engine.
A failure to click. (Score:2)
Well, I guess it is informative in a way. I mean, I certainly didn't know that about Geri Halliwell's hair.
Printing, fonts and printing and fonts (Score:2)
One problem is the basic design of the printing model -- you basically just dump postscript files to a spooler, or possibly dump them to a post script filter, then to a spooler. The problem is that the finer features like choosing paper trays and possibly stapling documents have no place in this primitive printing model. CUPS does a lot to address this, but the applications are usually developed for the lowest common denominator.
Another problem ( partly related ) is font management. So, you've installed fonts into the X server. Well, that's nice, but using them from your applications is still a very nontrivial task. The reason for this is that most applications that need to print install their own minituare font system which properly handles display *and* printing of fonts. because the bitmapped fonts that show up when you use xlsfonts are not printable. Under linux , the true type fonts aren't printable either unless you have one of the newer post script printers.
What linux distributions need to do is
(a) move towards a better printing system,
(b) find a way to render true type fonts to a printer, maybe code some kind of printing layer
(c) possibly supply some kind of API much like the one Star Office uses to handle the problems of WYSIWYG printing.
(d) Ship some decent fonts. It looks like Corel is making some steps anyway ( they're shipping 200 bitstream fonts )
Well, part of the fun (Score:2)
Raw unix isn't for everyone, it's user friendly but picky about it's friends, etc etc.
But seriously, what they'll be talking about after all these IPO's is market penetrataion which means custom tailoring the system to the markets existing level of expertise or a tolerable learning curve to expand the user base. It'll probably be up to the developers paid by RHAT and VA to do all the boring 'user friendly' extensions necessary to make Linux accessible to the average person in the mall; a lot of the developers who 'do it for fun' and share their code get their kicks in other ways, usually pushing the technology in front of them and probably couldn't care less about coding in proactive help and dancing paper clips - it's hard enough just to get release notes typed up! Of course some developers may get a charge just out of eroding the mighty snake god's empire enough to help with the friendifying extensions sometimes requires cunning trickery.
Maybe if Video card mfgrs distribute a linux CD already configured for their card. Or rhat sponsors a hardware testing/certification program so vendors can put a "Linux OK!" sticker on their box and an easy to find rpm just has to be loaded to use it.
Boojum
that got mangled! (Score:2)
Missing paragraph - Windows isn't always a dream itself, they've got their share of tech support calls and upgrades that require an expert and cunning trickery to get runnings.
One of my hosed boxes at work has a win95 display properties windows that is 50 screens wide - sure takes a long time to get to the "OK" button, heee.
Boojum
Linux needs standard API's (Score:3)
Imagining everyone agreeing to a single standard for both text mode and graphical interface mode. That way, we can optimize the operating system for MUCH improved ease of installation and use, not to mention better hardware support and much improved printing capabilities.
Because Microsoft uses the WIN32 API (which has been improved a lot of the years), it can do a very good job of optimizing the operating system to support system hardware, and MS was able to develop a very good printing system (which works with just about every printer out there). I mean, look at the output from the 1000 dpi-plus color inkjet printers--most of the output is generated by Windows-based computers, unless it's one of the newer models with USB connections, then you can use a newer-generation Macintosh to generate the stunning output. Sadly, that type of output is quite yet available to Linux users.
Re:Ironically, that's why I don't use it... (Score:2)
That happens with Windows because it uses a centralised registry. For one thing, it's too easy to interrupt processes while they are in the middle of updating it. For another, even if writing to the registry isn't interrupted it is still possible for the contents of the file system to get out of step with what's in the registry. In a more general sense, it makes the system less transparent and puts all your eggs in one vulnerable basket.
On Linux, applications never interferes with each other or with the system as a whole (system-wide libraries are always shipped separately).
Although things can break (say if the power goes off suddenly and fsck can't retrieve all the files that were open at the time) the worst you'd ever have to go through is to reinitialise the affected package's
There are a lot of know-nothings raised on Windows who are desparate to see a registry on Linux. Occasionally they even pop up on Slashdot. Don't listen to them, fight them with every breath in your body. In their ignorance they'd destroy Linux's greatest asset - stability.
Consciousness is not what it thinks it is
Thought exists only as an abstraction
Re:Too many apps! (Score:2)
Er, no you can't; you can only pick what is there and that's the whole point. You can't pick a video player to play the latest Sorenson-encoded QuickTime movie trailer, and you can't pick a finance package to do your tax return. There might be though if those with the talent and free time to code stuff didn't waste their time re-inventing the wheel. I mean, 30 versions of the same program, that's just plain silly.
Consciousness is not what it thinks it is
Thought exists only as an abstraction
Re:Math is the foundation of CompSci (Score:2)
That's a nice way of putting it.
Consciousness is not what it thinks it is
Thought exists only as an abstraction
Re:What does linux do that windows doesn't?? (Score:2)
I think I can answer that with a degree of authority as I'm one of those people who uses linux and Windows about 50/50 at home. The answer is stability.
I run Windows 98SE and when I finally got it installed with a few apps(after a dozen abortive attempts) I thought: wow, how slick.
But by the time I'd installed a few more applications and, more to the point, all the recommended updates from windowsupdate.com, the system had become so flaky and unstable that it was regularly sending my blood pressure through the roof.
For example, one of the most recent recommended updates for IE5 made the special IE5 icon disappear completely from the desktop (there's no way to make it come back except by reinstalling... and that means re-downloading half of the updates again).
For another example, ACPI support just plain doesn't work; having reinstalled from scratch with plain old APM I find that doesn't work either; the machine frequently goes to sleep and will not wake up unless I hit the reset button despite my having standby mode completely disabled!
For another, the system frequently hangs partway through the shutdown sequence, with one of those kernel32.dll errors. You click on OK and it comes back again straight away. And you can't get the task menu up with Ctrl-Alt-Del as long as it's there, which is to say: ad infinitum. So again the only recourse is to hit the reset button.
Oh, and often enough rundll hangs on shutdown instead.
Don't imagine this is anything like a substantial summary. It's only the tip of the iceberg.
The fact is, IME Win98SE is a mess. Period. It may have the gloss when it's working, but the frequency with which it stops working completely destroys the experience for me. Often enough as a result I just switch over to Linux for several days at a time because I'm so damn disillusioned with using Windows.
Consciousness is not what it thinks it is
Thought exists only as an abstraction
Re:[Off-Topic] Registry (Score:2)
Seriously, you shouldn't write me off as an anti-Microsoft bigot. I've bought two versions of Windows in the last two years even though I already run Linux. When Microsoft comes out with a new OS I'm ready to part with the GBP70.00 to GBP140.00 it costs here, even though they've bitten me in the past. So to imply unreasoning bias is a bit uncalled for.
The thing is, every time I go through this I end up bitterly disappointed. My experiences with Win98SE have been appalling and despite hours downloading patches from Windowsupdate.com things are getting no better. The most expensive failures I've encountered on Win95 and Win98SE have been those which corrupt the registry basly enough to require a reformat and reinstallation of the OS and all applications and patches. This takes days of my time, and happens so frequently that it has cost me thousands of pounds in lost time this year alone.
Now, about the other matter.
With regard to your assertion that a registry would work better if written by someone else, try to understand this simple concept: a centralised registry is unworkable in the real world. It is therefore a bad idea, period.
It creates a single point of failure for the whole system. Remember a chain is only as strong as its weakest link. If you have an entity upon which the whole system depends, and which is written to by every program on the system (and thus vulnerable to bugs and other mishaps ocurring in every program in the system), then that entity will inevitably get corrupted and the system will be unrecoverable. As with windows, you might get it to work with some manual fixes but you probably won't be able to find all the damage and the system will never be quite right again.
A registry might look cool from a theoretical architectural point of view. But this type of model hardly ever takes into account the fact that software contains bugs and that hardware sometimes fails. It would work if these things never happened. But they do, all the time, and until we see infallible hardware and guaranteed bug-free code a centralised registry will always be a liability.
Consciousness is not what it thinks it is
Thought exists only as an abstraction
Pretty close to the point. (Score:4)
Hardware... almost everything on the shelves is Linux compatable except for the cheap junk.(Win products) but is there a linux sticker anywhere? no way! We hear of these companies saying they support linux, but they refuse to put a works with linux sticker on the box (Creative Labs... get that sticker on every box you dummies!)
But then linux isn't made for the general computer user. windows9x is great for the clueless... it's great for the un-knowing... it's geared for the person that has very little in their heads or very little desire/time to do more than push a button. Linux didnt become popular because it was effortless, it became popular with us because it is hard. ls -al is dang difficult to the windows user. they cannot understand keyboard commands (before you win-lovers start flaming... I support windows95/NT for 120 users at my office... they cannot understand anything that is more difficult than a doorknob. yes there are a few that have a clue but it is very few) Unix has been around for eon's longer than wondows/dos even was a fleeting idea. Unix was "shunned" for being complex. Dos came around... it too was "shunned" for being complex! Only the geeks had 8088 or a 80286 (-rich geeks) computer at home, and everyone else said "oh god no! no pc for me! I cant even set my new-fangled clock on the VCR!"
well now it's 1999... and these same sheep cant set their VCR clock, but they think they need a computer!
I dont want them using linux. Actually, the best thing for linux is to stay "fringe" or in the "power user" domain.. it will become de-facto in business (It's replaced most of my servers) because it can be made to easily run what the sales dept., accounting, etc... needs and will keep the "I know something about computers" people from screwing it up! A linux box properly configured for business is 100% bullitproof, idiot proof, and makes an admin's job almost trivial.
So please linux, dont go mainstream... stay in the realm of power hobbiests, and business users.
Leave the fodder of the "home user" to Microsoft.
Too many apps! (Score:5)
Doing a search for a Linux ICQ client on Freshmeat produces (get ready) 30 matches! That's right, there are 30 seperate ICQ clients for Linux. They are all open-source, and many have all the usefull features, but people, DO WE REALLY NEED 30 OF THEM?
The linux kernel programming community is a pretty loosely knit bazaar--a lot of programmers, but at least they have a single focus. There aren't 30 versions of the kernel!
Although the application programming community also has alot of programmers, they are all duplicating each other's work!
We don't need more applications... We need better applications. Instead of saying, "I want to write my own web browser," why not contribute to Mozilla instead? Instead of writing "Yet Another MP3 Player," contribute code to one of the more established ones.
________________________________
Re:Why I use Windows, and not Linux (Score:2)
Unix is a huge pile of software that's evolved under the direction of a large number of people. Getting it to work involves getting lots of different layers of software to work together, and this is both a strength and a weakness. For example, Unix users get to choose between window managers, and having chosen, they have a lot of choice about how to configure their window managers. That's a whole class of software that doesn't even exist as far as the typical Windows user is concerned.
And the problem is, if anything getting worse, with more layers being added to the system... Pretend you're a naive user, and then consider Gnome. Everyone's talking about it, but what the hell is it? It's not a window manager. It's not an application. It's not a C library. Do you need it? How are you supposed to know?
You might hope that it would be the job of the distribution packagers to work out these issues, but as far as I can tell, they haven't been that much help. And there's a limit to how much they *can* help, because one of the virtues of the Unix world is the amount of choice it offers the user. Freedom is what it's about, and freedom can be confusing.
Re:Question (Score:2)
C2 is the first, and least restrictive certification. It is in reference to the red and orange books, which I'd have to look up the actual reference numbers of. Most commercial workstation vendors have had this for some time now. NT now has it. It is granted after a very entensive code review by a branch of the NSA.
FIPs is actually a set of a few hundred Federal security standards. There are some for password security, encryption, system control, etc...
Re:NT isn't C2 either... (Score:2)
Certifications for gov use (Score:3)
I've gotten Linux into a lot of government shops and applications, but its always an uphill battle because it isn't C2, or FIPs, or POSIX (early on for that one). Now with NT being C2 and FIPs (I don't remember the actual FIPS, maybe 140 or 180) certified this arguement has only gotten stronger. Put on top os this of not supporting ATM (yes I know it sucks for LAN apps, but tell the government that) in most distributions and pre built hardware and a large hurdle exists to get into the government and their contractors shops.
Maybe RedHat or VA should use some of their newfound richs towards certification. C2 would be a good start and clear most of the problems.
How about a DHCP client... (Score:2)
Re:Direct X and Visual Studio (Score:2)
As for VB, have you tried tcl?
Re:How about a DHCP client... (Score:2)
Re:Direct X and Visual Studio (Score:2)
Re:Admins have to care about the average user (Score:2)
People at work don't have the time to learn awk. They don't have the time to grapple the finer points of command lines. As a network admin, I can vouch for the fact that most users don't even have the time to go into "Tools, Options" in Microsoft Office when they're trying to get their work done. They race around, looking for a button that says, "Analyze All My Data And Print The Good Stuff In Arial 12, Landscape." If it's not that easy, then they don't have the time to use it. Of course you don't care about the average user, because not only are you brighter than one, but you're also not tasked with helping them on their computers.
Those of us who actually have to deal with the average user wish, beg, pray for a super-usable version of Linux and applications to match. I'm right there with you, man - I *know* how stable Linux is, I know how powerful it is, I know how flexible it is. But you and I have taken literally years to play around with Unix & Linux, and I can't ask that from a secretary when her next sentence to me is, "I don't know why I'm learning this anyway. I'm looking for another job over at Company Y where they pay another fifty cents an hour."
I care about using my computer, too. But I also care about using my VCR, and I was really happy when VCR Plus came out. I'm not stupid, but wow, it's great to just push five numbers and hit GO, and know that it's definitely going to tape Futurama in its entirety. Sure, I know how to program it, but I don't even have to think when I use VCR Plus. It just works, bam, done.
Is vi as fast as VCR Plus? Sure. But it ain't easy for a secretary to walk in and create a two-column report with a bar graph in vi. It can be done in a matter of seconds with Word without ever having to read a single word of text.
I'm not slamming the usability of Linux - I'm just saying that yes, it is good enough for you as a mathematician familiar with awk and vi, but it's not good enough yet for the secretaries. And there's a lot more secretaries than mathematicians.
[Off-Topic] Registry (Score:2)
I would like you to separate the concept of a registry from its implementation on Windows 9x. What is "wrong" with the concept? The idea of the registry is a good idea, I think. One place to change settings throughout the system instead of the zillions of unsimilar configuration formats in a files that are usually in
The implementation is another thing. Microsoft made an operating system that is less than stable. That doesn't mean all operating systems are unstable. IIRC, the GNOME people are developing a registry program called GConf but it seems to be entirely unlike the Windows implementation. Instead of a binary database, GCong will be based on XML. [This is all AFAIK]
The truth is, users *want* one place to change settings throughout the system. Microsoft has achieved this (even though it leads to problems with security (easy to access registry on any Windows 9x system and unlockout the control panel, etc.) and most of the information is undecipherable to humans). Registries are a good idea if done right.
Not everything Microsoft comes up with is a bad idea.
Re:A Standard UI (Score:3)
Re:The Post/Ghost-script model (Score:2)
The same method worked with the graphic card manufacurs.
I think a other approach is the best option. The postscript method can't handle the options the modern printers have to offer.
The problem isn't anything to do with drivers at all - it's to do with the fact that 99% of Linux apps print by spooling out Postscript code DIRECTLY to a file, and then forcing you to go through Ghostscript to print it out - which frankly, is ludicrous.
Windows has good printing technology which makes all this a none-issue. Rendering to the screen is IDENTICAL to rendering to the printer for 99% of all jobs. It makes it easy - which is surely the whole idea?
Of course, then you *would* need printer drivers. Catch-22 I guess.
Simon
Re:Windows does NOT have "good printing technology (Score:3)
Bzzzt! Sorry - thanks for playing.
Who mentioned WinPrinters? You did. I didn't.
I'm talking about the mechanism which means that to print, you can re-use your graphics code. You don't have to touch Postscript - your printer driver sits one level *up* from that under Windows.
It goes something like this:
Create Device Context (Printer in this case)
Draw to Printer Device Context
Finish.
That's it - no "how do I write something which generates the correct postscript to do X, Y and Z?" - the GDI subsystem and the printer drivers handle all of the nasty stuff so that the printer gets the data in a format which it can understand.
Winprinters... well, I wouldn't buy one, that's for sure. But given that some of them a while back cost $50, and were being bundled with NEC machines, I can't really knock them - if it were a choice of winprinter or no printer at all (and the printer was effectively "free" in this bundle), I'd take the winprinter. And I did do so.
However, if I'm in the market for buying a printer, I'd buy one that could be used on any machine, just for future-proofing. (Same reason I only buy external modems if I can help it - although that criterion is going now that ADSL is picking up).
Anyway, to sum up - don't go off half-cocked shouting and screaming about something when in fact you're not talking about the same thing at all. It makes you look stupid.
Not only that, but I noticed that for all your flame, you didn't have the guts to stand behind a real alias - you posted as an anonymous coward. Tsk tsk. Try getting the cahones to stand behind your beliefs - or don't post them.
Simon
Complexity versus consistency (Score:2)
Windows is perceived as easy, because it is perceived as not terribly complex, which it isn't if all you do is run minesweeper and sol.exe. But, if you get under the hood, or even if you do any mucking around in the passenger seat, you'll find it terribly inconsistent. From the UI to the interminable crashes, Windows just doesn't behave the way it is supposed to (or ought to).
Linux, on the other hand, is perceived as complex, and frankly it is: there are many options presented to the user, and having to choose among many options is something that requires prescience or at least forethought. Linux, however, is also incredibly consistent: things don't crash when they're not supposed to (I'll leave it to the reader as an exercise to figure out when it is ever the proper time for something to crash). Programs, when they do exist, do their thing and nothing more. To me, that is much easier than is windows -- if I spend some time investigating an issue under Linux, I'm certain to be rewarded with some insight to bring to the table in the future if the circumstances arise again. Under Windows, I just have to hope that whatever happened before won't happen again, and if it does, that some magical combination of reformats and reinstalls will keep it at bay for a while.
I tend to correlate ease with the absence of frustration more than the absence of complexity. Unfortunately, others don't seem to think that way.
emphemera/conf dialgs/backup/stock kernels (Score:2)
2. Richard Stallman needs to explain to my wife why she must edit
3. Where is the GUI tape/CDR backup program?
4. The sign of bad a bad Windows helpdesk is "reinstall Windows". The sign of bad Linux help is "recompile the kernel". Either make more things available in binary form for mainstream distributions or (shudder) make kernel recompilation graphical and bulletproof. Until a small cult believes having a stock kernel is a virtue this isn't going to get fixed.
Direct X and Visual Studio (Score:2)
Also, like so many others, I make a good living writing Visual Basic apps. you can knock together some very nice database apps in no time at all. If I had this available under Linux, I would happily tell all my customers to switch.
Linux needs AOL and the bunch (Score:2)
--
Phase 1: Collect Underpants
Phase 2:
Phase 3: Profit
Preinstalled Linux is the Answer. (Score:2)
He's like 90% of "real world" computer users.
He has all this MS software because it was installed that way out of the box. He'll never upgrade to Windows 2000 (why would he?). And he'll never upgrade to Linux, even if it was easier, cheaper, and seemless.
Linux advocates should think about the development and marketing of low-cost PCs that have pre-installed Linux and it's associated application tools. People buy computers these days regardless of the specs. A reliable Linux-based computer for 20% less than the competion will be (I speculate) quite successful.
I think the Linux community can learn some from (gasp!) Apple Computer. "Be different". An alternative OS in an alternative-looking box - and with amazing sales. I like that strategy for linux. Put the OS in a low-cost 4" x 6" x 10" black box, and sell it as "cheaper & faster & better & cooler".
And like Apple, if you start selling your hardware with MS-Windows on it, you loose your "different" status, and you lose the whole game.
So how can "we" get all this done? Well, the existing PC manufacturers (HP, compaq, apple, etc) won't be much help - the current Linux market isn't worth a zillion dollars. So perhaps only a smaller organization, along with technical advice and assistance from the Linux community, can get the ball rolling
If sales get big enough, retailers and the big boys (Dell, Compaq) will have to take notice and will need to start actively marketing Linux for the desktop. All in the name of profit, of course.
Make Linux idiot proof? Would it be Linux then? (Score:2)
1) Simplified installation. Preferably you drop the CD into the computer, boot the computer, type in your name and the name of the computer, and the thing sets itself up. (Of course that requires a degree of plug-and-play which isn't currently supported in older hardware PC models.)
2) Standardized and simplified user interface. First, the interface must be totally standardized so that my mother doesn't get confused when one button is shown in grey and another is shown in black and white. Second, the interface must be totally transparent--that means totally eliminate things like "right-click" unless there is some standard graphical element that says "you can right click on me," or unless we adopt a convention such as "all icons are right clickable."
3) Restructure the file system so that adding device drivers, file system support, and other elements to the kernel is a matter of dragging and dropping a file to a special magic folder and rebooting the system. For example, we could create a
4) Rename the folders so that they make sense to my mother. For example, instead of
And so forth. I think you can see that I'm going in the direction of pulling in Macintosh-isms and Window-isms onto Linux in order to create a consumer ready system.
There's the problem, though: by pulling in all of these consumer-friendly features such as simplifying the boot process or restructuring the file system, Linux is no longer Linux. Further, I fear that in order to accomodate the user, you wind up having to weaken a number of things such as weaken security (my mother tends to forget her password--and frankly doesn't understand why she needs one in the first place).
Perhaps we shouldn't ask the question "what do we need to change about Linux to make it mainstream," as I think from the above that if we did that, we wouldn't have Linux anymore.
Instead, perhaps the question should be "if we were to design and build an open-source consumer level operating system from the ground up, what features should it have?"
Re:Make Linux idiot proof? Would it be Linux then? (Score:2)
You know about end-users; they're the ones we insult as idiots while we extract money from their wallets. I'd like to think we shouldn't bite the hand that feeds us--but apparently, at least in the computer industry, that puts me in the (very small) minority...
Re:Make Linux idiot proof? Would it be Linux then? (Score:2)
>why would we want to reboot? Thats a winblows thing- rebooting for installing things.
Beats rebuilding the kernel and then rebooting, as is done with some Unix variants.
>>Rename the folders so that they make sense to my mother.
>Maybe your mother needs to make sense of the existing names for herself, they make sense to me.
Quantuum mechanics, C++ and VLSI design makes sense to me, but that doesn't mean the entire world should be forced to learn QM, C++ or VLSI either.
>> Perhaps we shouldn't ask the question "what do we need to change about Linux to make it mainstream,"
>no, perhaps we need to ask "what do we need to teach people about linux to make it mainstream", if even mainstream is a good thing.
My mother is an architect who spends 12 hours a day building houses in California for folks of all income ranges, from the extremely rich to housing for folks on welfare. Are you suggesting that she should drop her job (losing a ton of money in the process) because you and I as computer programmers are too damned lazy to do our job?
I'm constantly amazed at the attitude of the software development industry: "I'm too lazy to make it easy to use, so I'm demanding that the rest of the world learn something that I only spent a few years of high school and college learning."
Re:Closer than you think.. (Score:2)
This is one of the reasons why Windows apps are so bloated in download / cd size. Almost every application you can run will come with the shared components it needs to allow itself to be run. When was the last time you had to go online to download a shared library for Windows? Short of downloading the vbrunXXX.dll's, I don't remember a single time.
That's another reason why Windows is so unstable. One app may require component A, and install component A. Another app may require component B, and install component B. Another app may require both component A AND component B, but install predated versions of A & B than what was previously installed. Thus, when those two components are called by the first two programs, they may perform executions in a manner different than what was intended for the original applications, therefore - CRASH. The install programs are getting better with shared components, but it's a dangerous assumption to know that an installer will know what's best for all applications on your computer.
Re:Carpel Tunnel Syndrome. (Score:2)
Re:Fgnyyzna'f qevccvat phz. (Score:2)
Ubjrire, gur Freinagf bs Fgnyyzna jvyy cebonoyl fzbxr hf bhg. :-(
Re:Here's an idea... (Score:2)
Hm... what about Tivoli? Have they managed a port to any Linux yet?
Re:A Standard UI (Score:2)
Usually simple make program suffices, or better yet, just make. Imagine mk were aliased to make. Now, learn about completion. So you often need to type no more than
to get build that program, and probably less than this.And with reasonable programmable completion, you can just hit the tab after the command, and it will tell you valid arguments. For example
Don't knock what you don't understand.Re:Fgnyyzna'f qevccvat phz. (Score:2)
#!/bin/sh
tr a-zA-Z n-za-mN-ZA-M
or simply
$ perl -pe y/a-zA-Z/n-za-mN-ZA-M/
works as well.
Re:How about a DHCP client... (Score:2)
cablemodem from castel@home, running suse linux 6.1 (I think. I'm at work now and forgot the version. That's what you get from free stuff(as in beer...)
//rdj
Re:A Standard UI (Score:2)
Not that I need help compiling new libraries
Re:My $0.00 (Score:2)
Let me know if I left anything out.
Re:My $0.00 (Score:2)
My $0.02 (Score:3)
I don't think so... The main attraction of Linux is that you can get anything for free. What needs to be developed is an easier way of getting these things, something like an online store without money... now I have no problem with the current system, "Search Engines" (hehe), but the average person would like a place he can go online to just browse what's available and say "Ooh that looks nice, I think I'll get that..." rather than knowing beforehand what he needs.
Even though there are numerous applications for Linux available, there really aren't that many of comparable quality and usability to their Windows counterparts in areas which the average user needs. For example, there is nothing for Linux that is comparable to something like Quicken for Windows, a popular financial application. While there are small applications being developed, there aren't any commercial applications developed that serve that purpose. Some Linux users seem to fear the commercialization of software, but in a sense, it is required for the further advancement and acceptance of Linux.
Picky picky picky... Yes financial software is one area where linux is deficient, the other (more prominent) area being web browsers. We've just about solved the web browser problem, with Opera and Konquerer on the way, and especially Mozilla. M12 looks great, and it seems to be almost as powerful as netscape 4.7 (and it IS more stable
For the most part, I agree with everything else in the article... I've felt the wrath of printer support, especially with the HP 722c's. Back in RH6.0 I was able to get mine working... slightly... After that, I haven't been able to. I'm now on Slackware 7 and I still can't. (The driver is available, at PPA for the masses [httptech.com].) BUT... I am extremely impressed in the amount of drivers available for linux, in the most recent kernel there is a working driver for the SB1000 Cable Modem [linuxpower.cx], a Hybrid Cable Modem card used in some areas (like mine) where fiber optic cable is not yet available. This is such an obscure device I didn't think there would be a driver, but lo and behold there was a driver and a HOWTO!!! You don't need to go looking for drivers in linux, almost everything is included with the kernel. Recompiling a kernel is not hard, either. Try it some time!
yebyen@adelphia.net
How quaint! (Score:2)
"You can have any car you want as long as it's black." - Henry Ford
One of the charms of Linux and all other Unixen is the _freedom_ to have any interface you want.
What we need to standardise is the plumbing under _all_ interfaces, not the user interface itself.
If you want an OS with an enforced standard look and feel, go play with a Mac.
Merry Christmas!
-M
Re:Open source is hard to be accepted by managers (Score:2)
To equate a software product's value to counting the hours someone paid to have it written is seriously bent. Even equating its value to the hours spent writing it is absurd. The value is much better measured as the satisfaction of the user without regard to the method of production of the software, or its cost. Measuring it this way naturally includes such factors such as support and accountability.
BTW, the managers are correct that support and accountability are the playing field, that is exactly what RedHat, VA Linux, etc. need to provide and *market* to succeed. It's only high cost as indicator of quality that is a ridiculous criterion, and I suspect most smart managers know that. How large that subset is...
"Do one thing *well*" (Score:2)
I think it important that we do whatever possible to help anyone who *wants* to try it have a successful experience. I think this is pretty easy in business and academia, where there is usually a paid guru around to help anyone with problems. It's still chancy for a home user with a low patience threshold, or with needs for specific applications.
While we should be friendly to the curious who want to try Linux we shouldn't do anything to alienate it's core constituency: software and web developers, students, scientists, and engineers. If Linux can capture most of that desktop market, and a good share of the general server market, then application and driver scarcity will be history, and more desktop marketshare will follow naturally with time, as will the polished desktops and pricey application sets. World domination is a nice, fun hobby. Keeping the core developers and users of Linux happy is a matter of survival.
In other words, Linux should always be the choice of those who value computing power and are willing to learn to have it. Those who value being able to tinker and share knowledge, rather than buying off-the-shelf proprietary solutions. It very likely will also be made useful to nearly anyone, but that must remain a secondary goal. It is not elitist to say that "Do one thing *well*" is a fundamental philosophy of Unix.
A trick from the musicians' book (Score:3)
Obviously, this is not the case: the piano is a very popular compositional tool, even amongst people for whom it is not their primary instrument. Simplicity of music doesn't even exclude it from greatness (although "great music" is a subjective thing): I personally like using Louie, Louie as a proof for the existence of God :)
There are some people who have a piano in their living room simply for family sing-alongs. They'll never play a tune more complex than Louie, Louie, Blue Moon or Chopsticks and will probably never compose their own music. Others with pianos will produce hundreds of original compositions. Both groups, if we ignore the type and quality of the piano, use essentially the same instrument.
You never hear of a great piano maestro looking down his or her nose at "newbie" piano players (the same applies to any musician, regardless of instrument). A maestro would never say "We don't want your kind playing this instrument. Perhaps you should play something more suited to your abilities...like the radio. Hur hur hur." It's not part of their nature to scoff at someone because they only play other people's music and never write their own compositions, or that they never playing anything more complex than a three-chord pop tune. You never hear anyone jump on someone's case because they don't know that the frequency of middle "A" is 440Hz or becuase they can't tell their canons from their fugues. Musicians understand that not everyone is going to be hardcore, but that many people amongs the non-hardcore still want (and deserve) the joy that comes from playing music. Many musicians encourage people to take up an instrument -- one argument they use is that these people should break away from music provided to them by corporate radio and TV.
Unfortunately, we don't always have this attitude in the Linux community. Can't write your own program, or even put together a simple Perl/Python script or C program? Stay away. You're not really interested in doing more than writing an essay, playing games or surfing the web? Go away. You're an artist who wants to use the computer to create art but can't be bothered to RTFM so you know what "ps auxww | grep -vv root" means? Fuck right off.
What we have here is a clique that's no better than the clique of popular kids at school, who are "better" only by some arbitrary measure. Ours is the only industry where this kind of arrogance is tolerated by the market, and that tolerance won't last forever. We'll probably never have a Columbine-esque incident (forgive the Katzism), but the desertion of the non-hardcore for other OSs will be just as devastating, and we will have earned it. What Linux needs more than anything is not hardware nor software; there's always a ready solution or workaround for tech problems. The hard thing will be a solution for its human element. As long as members of the high priesthood continue to act as they do, Linux will be relegated to a niche. We should borrow a trick or two from the musicians' book.
Re:my 10 cents worth (Score:2)
I was thinking about this the other day -- idiot-proofing Linux.
The only way you're going to do it is by the following:
Making sure that every last need that your end-user can have is covered by the GUI configuration tool.
Locking down all the scripts so they can't ever touch them, and auditing every last package to make sure it doesn't do anything ``unexpected''.
Let's face it, no computer is idiot-proof. (How many times have I, in my past life as a T/S whipping boy, had to help someone clean up a mess that they or another program made by doing something M$ never predicted?) And I don't think anyone is going to be able to anticipate every one of the end-user's needs -- that is what set-top boxes are for :-)
And in any event, once you idiot-proofed it (assuming you could), who would buy (into) it? How many people are swayed by promises of power when their needs are probably much better met with a lack of power?
Re:A Standard UI (Score:3)
I don't think there needs to be a standard user interface per se, but there is a need for standards on user interface behavior.
Common functions of all user interfaces (for example, generating program launch icons or registering file types with a file manager) could be abstracted to an API, or better yet a program that could be called, so that software vendors could release a program with an installer that would work cross-UI and cross-platform, registering itself with package managers and adding itself to launch menus along the way.
Just remember -- you don't need an rpm package to register your package with the rpm system. The FreeBSD ports system handles this rather well, registering programs with the package system as it goes yet never actually creating a package per se. (Note: I'm not claiming FreeBSD is the solution to this problem) :-)
Re:Why I use Windows, and not Linux (Score:2)
"I don't study information technology
This is a perfectly valid statement to position the user prior to their comments on Windows vs. Linux.
PS. How do you know they're English at all? They may be Russian trying to write English, young (perfectly valid) and/or just too lazy to use caps (or on a laptop or PDA where its hard). Get off it.
Re:Pretty close to the point. (Score:2)
no! we can't leave them in the clutches of the evil empire .... home users are people too .... if you want windows to go away (or go open source) you have to starve it's creator ....
Re:How about a DHCP client... (Score:2)
---
Re:Who Cares About the Average User? (Score:2)
Are you nuts? The amazing "beeping and not beeping" vi, and "ctrl+alt+right shift+F12+esc+del" emacs... Then pipe the stuff through ispell? For people who can barely remember the shorcut to spell check in Word97, you expect these to work? Corel Wordperfect -- now there is a usefull alternative.
"Ah, the joys of awk." Ah, the screams of end users. GNUmeric, perhaps. I myself prefer to avoid using things like awk by hand. Scripts are more productive with them, true, but... They are components, and most users don't like manually stringing together components.
"GNU/Linux is a tribute to the best and brightest in the world from the best and brightest in the world. If the average user wants GNU/Linux, fine. Let him learn how to use his computer."
It's that kind of elitism which is not welcomed in my Linux community. If an average user wants Linux, show them everything you do during the install (they likely won't see if it again, and it's just a primer to their Linux life really). Once that's done, help them setup the GUI or CLI login of choice. It's very easy to make simple scripts that let them do some trivial admin tasks without needing to know about multiuserness. Want to change hardware conf? A simple setup tool does it.
Something like the Windows Control panel which lets people to admin tasks (once a password is entered, of course), and setups which actually update things like GNOME and KDE menus by default would do a lot towards getting Linux "mainstream" as it were. From there, they're bound to pick up some knowledge. But please, don't patronize them just because they don't know everything you do. True, they may be missguided, but you'll get a lot more of them to come over and learn if you talk to them rather than scream from afar "LEARN, STUPID!"
---
Open source is exposing the interfaces (Score:4)
On the other hand, Linux and open source in general is designed to be an environment that empowers programmers. We are the ultimate power users. We want control of everything we can see. And we have the training and skills to see a lot and know what we could do with it. Open source exposes the interfaces to any programmer or user who wants to see them. Communication protocols and file formats are documented and often come from standards. Those standards and documentation are usually available for free. That is how open source attracts programmer/users who may be potential contributers. It is not something we can abandon, or want to.
As an example, there is nothing that frustrates me more when using a piece of software than when a common task requires a a lengthy sequence of mouse clicks, and there is nothing I can do to put it on the keyboard or shorten it. I don't care how obvious the icons and menu items are when I have to click six times to accomplish something that I do a hundred time a day.
The reason for this difference in preferences and usage styles is a difference in usage patterns. Programmers tend to be deep users of a few tools. We know their features completely. And we want to tailor them to our exact needs. What could be better than open source for that. It is a paradigm that can let us find even a couple dozen people scattered around the globe who want some very low demand feature. And it allows us to collaborate to build it.
Does my mother want to be able to optimize her usage this way? No. She wants to finish placing an order at an e-commerce web site so that she can go play golf. She doesn't use any given piece of software often enough to make a long learning curve worthwhile to her.
Most of the Windows vs. Linux flaming over the years has boiled down to "Your interface sucks and if you try to make me switch you're an idiot." It leaves out the most important phrase: "for my usage pattern". The real fight should be to reach a point where those of use whose productivity profits from using Linux can use it, at home, at work, with any server on the net we want to reach, period. Windows and the software under it doesn't give me what I need to make my computing experience pleasant and productive because it limits the control I can exercise with the considerable knowledge I bring to it. On the other hand, Linux is limiting to people who don't have the knowledge to leverage and never want to spend the time to acquire it.
Can one environment satisfy both camps? Can you simultaneously expose interfaces to ann potential developers and programmer/power users while hiding those interfaces from Grandma? I think it can be done, but only if the importance of both user communities is balanced. If we ever bury the interfaces so deep that new geeks can never get to them, we will lose the open source collaboration and the software will stagnate. If we don't make it possible to use the system without seeing them, we are limited in the potential audience. We can do both, but it must be intensional, or at least one of the two goals will fail.
Re:How about a DHCP client... (Score:2)
He was completely blown away because he expected it to be another fiasco like when he tried to configure his winmodem.
numb
Re:Why I use Windows, and not Linux (Score:2)
This is an excellent point which I forgot to mention. There are several programs make installation and un-installation easier (like RedHat Package Manager.) But no one seems to have agreed on a standard for this kind of thing. Most software is still distributed as gzipped tape archives...which is a great format for many reasons.
tar xvfz package.tar.gz
cd package
./configure
make install
^ is pretty easy when you know how it works, but there's no reason that this couldn't somehow be condensed into a double-click.
numb
Re:Why I use Windows, and not Linux (Score:2)
Actually I thought about it a little more and it maybe a 'meta-configure' script could be set up to pop up a Tk GUI interface asking for options such as --prefix= for configure, etc.
The standard make targets defined in the GNU Coding Standards (here [gnu.org]) include most of the work that needs to be done by the Makefile. Using a meta-configure program and make doesn't mean that you can't distribute the binary as part of the package. In fact I'm pretty sure you doesn't even have to include the source code as part of the package but it's recommended. (obviously you would need at least one or the other.)
Some of the stuff people release for Unix needs all kinds of wierd input to get it properly installed. But A LOT of it only needs stuff as simple as an installation path if you choose not to use the default.
The way I see it, autoconf and make already simplify installs and uninstalls tremendously. It seems like a relatively small step to make an interface to them that end-users can use as easily as InstallShield.
numb
Linux needs newbie stuff (Score:5)
OK, disk partitioning is sort of a problem. I really don't like disk druid. It picks the name and order of partitions for me which would be nice for a new user, but it's not cool when they force it on you. Fortunately for me fdisk suits my needs just fine for partitioning--however disk druid would probably be the way to go for a new user. It's good that RedHat provides both. However, disk druid needs some work.
I recently installed Debian 2.1r4--liked that a lot despite some wierd problems, but it would still be a little much to ask a newbie. I don't think newbies are Debian's target audience anyway.
There are little nits here and there with Gnome and Enlightenment, but nothing that won't be fixed soon I'm sure.
Newbie documentation. Linux distributions include tons of in depth documentation about the operating system and all things related. If you bought Windows, you'd have to spend $50-$100 on books after buying the operating system and you'd still have nowhere near as much depth of technical information. The X/Gnome/E help systems available for apps look great also--however a lot of the apps don't use them. There is a good reason for this. So much of this software is still being developed and improved--documenting works in progress for newbies would take away from the time that could be put into the apps themselves.
I'm pretty sure these things are being worked on as we complain
ObWinDis: Installing any Windows operating system on a machine that didn't come with it is no joy either. In comparison to installing Linux, it's a long convoluted process that takes several reboots. In fact, installing RedHat 6.1 on a system that does not already have Windows on it, and has fully supported hardware, would probably be easier for a newbie than installing Windows. It would certainly take less time.
numb
Why I use Windows, and not Linux (Score:5)
I use Windows 98, even though i know its got dismal hardware efficiency, in that the equivalent machine on a RedHat installation could work alot better than on Windows. I use Windows knowing full well (and being disgusted with) the marketing practices of Microsoft. I use Windows knowing that my love of tweaking how my machine looks, works and runs could be fulfilled much easier on a linux box. I use Windows knowing that my data is about as secure on a WinBox than my personal possesions if I were to have the lock on my backdoor installed by a thief.
The single reason I personally use Windows and don't use my set of four RedHat v6 disks sitting right here is becuase one thing, Gloss.
It's shallow i know, but im an average user, im a student, i don't study IT, im more interested in the social context of the internet, it's effects upon communication psychology, but im the user that Linux needs to convert to have any hope of ever fighting back against the "Windows on every desk" mantra.
I use Windows becuase it's easy to install a program, becuase it's easy to see what im doing, where im going and how to get there. I use it because when i want an mp3 dbasing util, i can download a single package, and install it with a mouse click. I don't need to hunt for an obscure library file, i don't need to make sure that it will run with the Window Manager i want to use, i don't need to decode version numbers of updates to work out what i need and don't need. This may sound lazy, but i don't have the time to do this, i don't have the energy to do this, i don't have the knowledge to do this. Most of the market that Linux needs to reach out and grab is like me, the computer techs are sold, the geeks and nerds are sold, the average user ? not yet, they will, but not yet.
Please don't get me wrong, i love the concept of linux, open-sourced, free, community built and driven, and i would switch back in minutes given the chance, but linux really needs that single element that MS still has, Gloss and simplicity.
Experience from the Mac OS side applicable (Score:2)
A big thing was, there is TONS of software for the Mac. Like 11,000 titles (and that was a few years ago). But no one I talked to believed it. Why? Becuase when you went into a computer store in middle America, you would see MAYBE a few titles and that's it. None of the numerous Hybrid CD's would be separately labeled or even with a sticker. Ditto with the hardware. There was a fair amount of hardware that would work just fine with a mac (external modems come to mind, with the proper adapter cable), but it wasn't labeled.
In short, it's tough to fight the perception that there isn't software. Maybe on-line shopping has picked up enough that it isn't as big an issue. But I don't think so.
It's one of those "mindshare" problems. I STILL have yet to meet ONE non-geek (i.e., not in the IT business) who has heard of Linux.
Re:IT people are not geeks (Score:2)
What I meant (translated for people who haven't gotten out of school) is: I've yet to meet one non-computer hobbyist, i.e., a person who, while they may be very computer literate, doesn't view them as a hobby, but merely a tool, who has ever heard of Linux. Is that clear enough?
Re:Pretty close to the point. (Score:2)
There is no reason for them to run Windows either. Windows in itself is too big and scary for many users, and Word and Excel are just plain awful. A really easy to use OS would, I guess, be persistent (system state stored on the disk, as in EROS [eros-os.org], all programs being always up and running. Then there would be a menu with big icons for word processor, spreadsheet, www client etc, and always a way to pop that menu back up. All the standard programs would be "running" all the time, all of them would take up all the display while being used. As easy as a PDA or cell phone (at least older Nokias are pretty simple to use, surprisingly the new ones seem more irrational to me).
No use to tell people that command line is fine, when even the freedom given by Windows is too much for many of them.
my 10 cents worth (Score:3)
i think before linux can really make it big it pretty much needs to become slightly more idiot-proof. Something along the lines of, you whack a CD in your drive, click on some icon to install and voila there it is... but then again that's prolly not linux is really all about...
think the fact that it's predominatlyly an OS for people who knows what they're doing and/or people who are interested in computers enough to learn something about it... the moajority of computer users are still people who turns on their new user-friendly iMac to just look at stuff round the web and type up the occasional word doco here and there.. they don't give a toss really about what goes on, and i doubt they could be really bothered to just to learn about linux
Let me tell you something about printing (Score:2)
I still find the combination of LaTeX and a PostScript printer to be unbeatable for all the text processing I do. No GUI app I've ever used even comes close for quality of output. And the IBM Infoprint Color 8 I've been working on at work produces very nice output from Linux and several other unices.
All that being said, I've suggested in the past that the X protocol could be used to render graphics into a printer format. This would give you the benefits of having a single unified driver which every application could use. Apparently there's a X Consortium extension to X called xprt which does that, but I've been able to find very little information on it.
In addition, your driver could be set up to read Adobe PPD files, which would let one printer driver drive practically every PostScript device on the planet. You could gain similar benefits with PCL by writing a parser for the Microsoft GPD format, which is pretty well documented. And The Printer Working Group (http://www.pwg.org) is working on an XML based printer definition format called UPDF (I sat in on one of their meetings and spread a little GNU propiganda :-)
Re:A Standard UI (Score:2)
If users were given a standard toolbox list so they knew which was used for what they could pick and choose the ones they like and still have tons of variety.
Re:How about a DHCP client... (Score:2)
I was never able to obtain ANY information from my cable modem with pump or dhcpcd. I've tried tweaking the configuration files and the command-line without success.
Then I found dhclient. It worked instantly like a charm ! No configuration file needed, no command-line: I've just started it from the command-line and it worked. I'd recommend it to anybody.
AFAIK, most (but not all) distributions come with pump or dhcpcd. I suspect this is an obscure licensing problem (dhclient is free and open-source, but not GNU, I think).
You can get dhclient from http://www.isc.org/products/DHCP/dhcp-v2.html [isc.org]. It should fix your problems instantly !
Stéphane
The Post/Ghost-script model (Score:2)
The Y2K goal for Linux seems to be the desktop. For that purpose, I'm not sure that the Post/Ghost-script model is OK : most desktop users use low-end printers, that aren't Postscript compatible.
Moreover, with Ghostscript, the applications (say the Gimp) are not aware of the limitations and abilities of the printer. Think about the paper borders, for instance. That's hard to print a precise job when your printer and your app can't communicate in both directions. Worse, only through an emulation
If desktop printers were Postscript-compatible, there would be no problem with it (for instance, Postscript does an excellent job in high-end Macintosh (pre-)press systems). But using an emulated mono-directional language, just doesn't make it for this purpose, IMHO.
Since companies (HP, Epson, Canon, Lexmark,...) don't seem to be ready to develop Linux drivers, we should encourage the projects to develop performant Linux printer drivers, not based on Postscript emulation anymore, but in a Windows-like approach, in other words in a bitmap approach. (I'm not trolling here, but the Windows approach is better for the desktop printers).
Stéphane
Re:A Standard UI (Score:4)
I think this is a "yes, and no" answer. OT1H new users would *rather* have to learn *one* UI once and for all. OTOH, experienced users do NOT want to be stuck with a "standard UI" that they cannot configure.
I think we need something of a compromise. Perhaps have a standard default UI that is uniform across all distros or something, but leave it configurable. The newbie who just bought his Linux pre-installed PC will get the "standard default" UI which is the same as the one every other newbie on his block has. The advanced user, or just the picky user, can always change the default during/after installation. There is no reason/need for imposing a "standard" that everybody has to stick to, because... there is no such thing as a one-size-fits-all. There is only such a thing as one-size-fits-few.
Remember, one of the best things about Linux is configurability, and I would add IMHO, the lack of presumptiousness. One of my favorite peeves with MS crapware is that they are extremely presumptious. Ever tried MS Word and have it completely screw up a paragraph that begins with a number? Apparently some people actually like that. Well, the whole idea behind Linux is that you should be able to opt for things like this if you wanted, but it certainly should not be forced down your throat. And if it comes as a default, it must certainly be easy to change (not buried somewhere deep in some esoteric menu or worse, unknown config file).
So, in short, we need a default UI that the majority of users can grok, but all the other options must not be eliminated. Users must be given the full power to choose. If some people wants to be spoon-fed, fine, they can choose it that way. But don't assume that there aren't people who don't want to be spoonfed. We do not need Linux apps that won't work unless you have a particular flavor of UI active.
Re:Pretty close to the point. (Score:2)
Re:Why I use Windows, and not Linux (Score:3)
The linux community really should take this comment to heart, or at least this one word: gloss. Gloss is the one thing that always seems to be missing from open-source apps. The functionality is usually much greater in open-source apps than $$$ apps, but the gloss is almost always missing. Techies don't notice because mostly all they care about is functionality, but we really need to start noticing these things and start adding the gloss to win over Mr. Consumer.
simple idiot proof install (Score:2)
My second attempt was successful(mostly) but it wasnt as easy as installing Windows and dear god nothing near as easy as installing macOS. Making the installer less ugly, less stupid, and less frustrating would also help.
(oddly i've always have more problems installing 98 than NT)
piranesi (unfortunatly an idiot)