Open Ballot: is Graham Morrison wrong?

TuxRadar

Our kid Graham has had a rough time of it on the internet recently. His article for our sister site TechRadar, "The trouble with Linux: there's too much choice", sparked off a few flamewars. Most notably, Caitlyn Martin over on the O'Reilly blog delivered a no-minced-words response: "Are you intimidated by breakfast cereal?".

We want to know what you think, for the podcast we're about to record. Read both sides of the argument and let us know. Is Graham on the right track, and the vast range of options in the Linux world is confusing for newcomers? Or is he wrong, and having many choices of distros and packaging systems is like having choices of breakfast cereals? Post your thoughts below - and use a name other than Anonymous Penguin if you want to sound like an awesome person in our podcast.

You should follow us on Identi.ca or Twitter


Your comments

Surprisingly Simplistic

Frankly, I was surprised by how simplistic both arguments were, given how long this "problem" has been around in open source. Clearly Graham is right, the BSD wars proved that too much choice will kill market adoption.
And even more to the point, Ms. Martin is the obvious winner, because (if the large IT vendors have taught us anything) competition is the only way to avoid the vendor-lock-in boogey-man.

So I say "Everything in moderation!" Meaning, lets have everything, give me 100 apps for each and every mundane task and I'll decide when its safe to go spelunking through all the abandonware to find that random feature that I know in my heart I've always been missing. If you're really lucky I might even copy out some code and make a patch for a more popular program (though I suppose I'd have to be really lucky to find feature with a compatible licence that I could copy paste, but these articles weren't about licence proliferation, or duplication of effort right???).

Gifts in the box

Yes, Graham Morrison wrote something childishly simplistic and got his pants flamed, for the reason that his comments detract from the quality of available Linux offerings.

His main example, Shotwell, is like the figurine in the cereal box - nice to have, if you like that kind of thing, but not why you choose that brand of cereal. The alternative - whatever of the wealth of photography applications happens to be his chosen alternative - is still a mouse-click or a single command away in the package manager. There are very, very few applications removed from circulation - if you used it within the last twenty years, then it is probably still available.

The fact of increasing choice need never detract from choosing, installing and using the most appropriate package, or the package to which you are most attached. Most distributions of Linux ship in a pretty good functional state, and simple to configure to your own needs.

... Assuming you know your own needs, which is my guess of why Mr Morrison is suffering from such confusion.

... as long as it's .deb

Choice of package managers and choice of default applications are two different issues... aren't they?

I assume fragmentation of package managers in part explains the lack of third-party Linux support by Amazon/Audible, which reduces choice (or flexes the work-around muscles).

Whereas most users will happily look for alternatives through a well crafted software repository interface, if the success of the Apple App Store is anything to go by.

Choice is a two-edged sword

Hmm, my summary of this would go "A choice of a single app to do something is restrictive, 5 apps is good, but 50+ is just plain stupid". And at the moment, the progress that Linux could be making is being held back more by the spread of distros than that of apps. (imho of course)

Yes, there is a lot of wasted effort duplicating features in a spate of NIHishness for example - echoing Graham's favourite hobby horse - why the heck do we need any package format other than RPM or DEB? The next person to invent yet-another-package-toolset needs a damn hard slap!! On the other hand, the package nirvana of a single toolset/format is NEVER going to happen - heck, even Windows can't manage it (MSI v's InstallShield v's etc).

And I've got another example to go along with Graham's Fedora/Shotwell/FSpot one. Ubuntu replaced gqview with geegie - why? The latter takes more resources; does the same job; but crashes more often than the former. It's provably a worse app, but you've got to be quite sneaky to get "good old" gqview back.

Personally I think we need a software meritocracy - apps regarded globally as "good" survive and thrive, whereas the geegie of this world die.

Actually no, on second thoughts, the "big name" distro's should do this (yes, I'm talking about you Ubuntu and Fedora), but I've no problems with the Arch, Gentoo, etc putting any kind of app in there - because the folks they're targetted at are going to be the ones that'll quite happily patch and/or download the bleeding edge build from SVN.

How would we choose?

If we went on the majority view we would no doubt end up with Ubuntu running a Gnome desktop and some rotten colours on the desktop too. However the majority of the Linux community would not want that. Me, I like LinuxMint using Gnome and, even then I modify the desktop. I know my solution is the best but there will be a few (millions) who will disagree!

What we should do though is standardise some of the basics. The file directory layout and package management. I don't care which one it is just as long as it works efficiently and, in the case of the directory layout is simple and logical.

We will always need distribution(s) for those of a technical bent, who want to pull the system apart, break it, then hopefully put it back together again. The rest of us just want a system to run our applications upon and the system just enables this and otherwise keeps out of the way.

I remember saying this once already

The package manager is irrelevant, a distro is created by the packages themselves, the way they are configured, and the versions used. You can install any package in any distro and it can sometimes work or it can fail miserably since it was compiled against different versions of other software, or differently configured. So what you're asking for is not a single package manager, it's a goddamn single distro. And for the most part you already have that, just use ubuntu and stfu;p I like my gentoo and debian, both serve me very well and just because you want a single distro that has one perfect application for each task (just for you since different people have different needs) doesn't mean everyone else should stop developing for their needs.

Is he thinking at all?

Mr. Morrison is wrong. It is as simple as that. The wide variety of distributions allows choice and accommodation for different needs and purposes. Old hardware need not be thrown away. The choice gives incentive and means to create and innovate.

Graham is...

...a beautiful specimen of Humanity and cannot possibly be wrong about anything.

6 of one and 0110 of the other.

As usual with these kinds of arguments neither side is right. Both have taken a black or white stance on it and Caitlyn seems far too fired up about it for any rational debate.

Choice is the cause of many problems in Linux and many of its successes. Standardisation on a single package format would help but you wouldn't want to prevent future ideas being developed by disallowing other projects.

The choice aspect of Linux (and FOSS in general) allows for a great deal of cutting edge work being available to the general public very quickly. We're starting to see things appear first on Linux (USB3) and many ideas from Linux are starting to appear in other OSes.

We're at a turning point in computing where we need to leave behind a lot of old ideas and start thinking about how we really want to use computers and the best ways to make that happen. Linux is better placed than the others to move with the times are to offer something truly new.

Also this argument so often relies on whether or not you consider Linux in commercial terms. Some wish Linux to supplant Windows and, effectively, destroy it. If you want Linux as a single entity to replace Windows as the publics OS of choice then you do need everything to be simple and the same everywhere. However, one of the main problems with Windows is that fact that it has become the standard OS worldwide. If one version of Linux was on 95% of computers globally then this would only be marginally better and definitely not the same as the Linux we know and love now (in all its manifestations).

So, to sit firmly on the fence, there are things that can be improved by standardisation and increased co-operation and there are those that benefit from competition and also provide for differing scenarios. In many ways the introduction of newcomers to Linux relies more on available information and support than the nature of the OS itself.

Poorly made, but a good point is a good point

Here we have a person who is willing to stick his neck on the line and tell us what he thinks. The fact that we are prepared to give him a hard time for doing so is despicable.

On the point itself, I believe that the Linux Foundation and the Linux Standards base should do a lot more to standarise the key parts of the OS. Yes, let's keep our myriad of choice for those who want it, but surely it would be better for those of use who want to get things done to have a unified face for Linux. So yes, I do agree with Graham Morrison

"The lady doth protest too much, methinks"

This little debacle is in the end, a great bit of journalistic hyperbole especially on the part of Caitlyn Martin.

And to answer her question "Are You Intimidated By Breakfast Cereal?" I personally am not.

But then again, I don't see what f*&^%$ flour and corn based products has to do with open source software and the "too much choice question"?
They happen in two very different contexts making the "argument" unbelievable weak.

Yes... always

...but Mike is never wrong!

P.S.

I look forward to reading "The trouble with Linux: there's Just too much to argue about"

Thank something that people aren’t committing this shit to paper.

Right Slash Wrong. The Doublethink is king

Everyone is wrong

and everyone is right

'Are you intimidated by the breakfast cereal aisle in his supermarket' - no, however when all the breakfast cereals are hidden all over the place and require you to have different bowls, then yes.

For a new comer to linux, who has just been introduced to linux then yes the huge choice and different ways of doing things is realllllllllly bad. Its like being buried in breakfast cereal, quite scary.

However when you have dug your way though the cereal and found the one you like, then all is good, and you often don't care how other distros do things and you stick with what you know.

However when you climb a bit more up the linux knowledge ladder and start to experiment with different distros, it can be confusing to start with, but then you get used to the annoyance and get on with things. and possibly moan a little.

So all the choice at the beginning is bad from the new users point of view, however after a while (most) will come to love the differences.

So Graham isn't wrong, and neither is Caitlyn.
Unless Mike agrees with one. Then they are wrong.
Mike is always wrong.

If your confused

Just choose what ever application starts with a 'K':

KDE
Kmail
Koffice
Konqueror
Kopete
K...

Seriously, does that mean we need two different versions just so the naming conventions are less silly? Many versions of some applications consist of smaller differences.

Quick, the maintainer forgot to put a smiley in his email, let's fork!

More choices will win out in the long run

I can see there are valid points with that 'too many distros' argument but taking all the good and bad aspects into consideration I still think more is better than less. There's never going to be an absolute winner in this argument, both options have varying and evolving strengths and weaknesses. I tend to see diversity as being a primary factor in breeding inspiration, experimentation, innovation, and competition. Monocultural thinking is more often than not the antithesis of that. Yes there are a confusing number of Linux distros, but there's an overwhelming amount of good and bad information out there, along with a huge mix of bias and balanced opinions, plus the FUD that corporations like Microsoft seed the media with. Researching which distro that will suit someone's needs shouldn't be such a mess that it is, but that's the problem, NOT the fact that there are numerous options to choose from. Restricting the number of Linux distros may look good as a superficial fix but I can't see that as being a viable solution to the actual problem. A basic search on any topic online can turn up an overwhelming amount of information and sources, and I don't think blaming the number of Linux distros is valid -- even a new computer user thinking about either one of two choices, Windows or OS X, has to sort through an overwhelming amount of FUD by the fans of each.
So I think the assortment of Linux distributions is a great thing, providing the public with good information is a more realistic matter that needs to be addressed.

Graham isn't cemplaining about the choices available

From the original FA: "Most of us are utterly confused by the options, and we'd just like a decision to be made on which is the best."

Graham isn't complaining about how many choices there are out there, he's complaining that the choice hasn't been made for him.

But also, when Fedora (and Ubuntu) drop F-Spot in favor of Shotwell, he's mad because the choice being made for him wasn't the one he wanted, which, apparently amounts to a loss of freedom.

So, which is it, buddy? Do you want to pick from one of eleventy decent image editors and STFU when the distribution managers don't read your mind about which you personally think is best? Or do you want to sift through the Googles to see what people like, try a few out (at no cost but the time you feel like spending on it), and make the choice yourself?

What bugs me most about the "choice is bad" argument is that it's a total red herring. What image editor does Windows include? OSX? Are those best of breed, or do you go do some legwork yourself to obtain and install software to do what you want to do? Does that legwork drive people away from either operating system?

PS, eff F-Spot. It came with my mom's Ubuntu 9.04 desktop, and I might as well have told her to manage her images with Visual Basic in Wine, the interface was so convoluted and inconsistent. She's happy with GThumb now, maybe I'll check out Shotwell in the next LTS release.

Make your first choice...

... and the rest follow. Say you pick Ubuntu, that means you'll get GNOME, debs, etc. If you went for Kubuntu, it's KDE and debs. No other choices are necessary. Pick Mandriva and it's KDE and rpms. Well supported distros offer a wide variety of packages, so you'll likely find the package you want in a native format and not have to worry about using Alien with debs on Mandriva or rpms with Ubuntu.
Now, one can get a lot more sophisticated in one's choices and pick other desktop environments than the standard, but that's pretty much true of other OSes, there are alternative desktop shells for Windows and a good many installers for software, some of which won't work across all varieties of Windows.
There are alternative application menus for OSX and different package formats there too, but I don't recall OSX and Windows users complaining about choice.

What Linux could do with is a more rigid set of standards, so that those who want to make non-standard choices can do so more easily.

Obviously

To be fair, I actually do find the cereal line in my local store intimidating. But that is not a good example, because it is, like, ridiculously extensive.
ANYway: The product is a result of the mindset, and I feel that differences in distributions - sometimes subtle, sometimes not so subtle - are a mirror of the people from the community and vice versa.
As a Slackware user, I would hate to be locked up in some automagical hell of a distribution making decisions for me. For many people, a "let's just have some sensible defaults and let people set up the rest" would be a painful experience.

Choice is good

I wouldn't like to be stuck with Gnome or KDE desktop/window managers.

Choice, Evolution and Darwin

Choice and freedom is a cornerstone of Linux. However you'd think that with these wide variety of distros, just as in nature, the strongest would thrive and the weaker would become extinct.

I'm an Ubuntu user and so wish everything would be a derivative of debian and use sudo and apt-get but I'm only 4 years in Linux-land. Older more influential users may perhaps more justifiably for the years they've put in want yum and su.

Good luck to them

Graham is correct

... in principle at least.

Having dabbled with Debian many years ago & decided it was not usable as a desktop operating system I have returned to Linux over the past 12 months, using Ubuntu exclusively since May for running my business plus all home use & streaming media etc...

The biggest problem I had was the vast choice of applications for any particular task with no real guidance available as to which would suit me the best. Following the links in the package manager sometimes lead to a fantastic website full of useful information. After the application was installed it would become apparent that all the effort had gone into the website & little into the application itself. The opposite was also sometimes true.

Although I welcome choice, some sort of ranking system would be helpful.

Standardising on file system locations and packages (doesn't really matter which ones although I personally prefer .deb) would also help.

And before I get flamed for using Ubuntu I use it because the computers have a job to do. I do enjoy (and am quite capable of) tinkering with the system to get it working but Ubuntu does mostly just work.

What choice?

I thought there was Ubuntu, Fedora, and SuSE? There are others <eyeroll>.

There are many choices for Windows:
Windows CE 4
Windows CE 5
Windows XP Embedded
XP SP3 Home
XP SP3 Pro
XP SP3 Tablet Edition
XP SP3 Media Center Edition
XP SP3 Media Center Edition 2005
2003 Web Server
2003 Basic Server
2003 Standard Server
2003 Advanced Server
2003 Enterprise Server
2003 Datacenter Server
Vista Basic
Vista Enterprise
Vista Home
Vista Ultimate
2008 Web Server
2008 Basic Server
2008 Advanced Server
2008 Enterprise Server
2008 Datacenter Server
Windows 7 Starter
Windows 7 Home Premium
Windows 7 Professional
Windows 7 Ultimate
Small Business Server 2000
Small Business Server 2003
Small Business Server 2008
Windows Home Server
Windows Home Server Vail

I think a choice of Ubuntu, Fedora, and SuSE (along with respins or derivatives is not too much choice). Also, linux can scale like Windows from an ARM or MIPS processor for embedded devices to a datacenter server with 64 CPUs and Terabytes of RAM. For windows you pay for the server, the CPU licenses, and the Client Access Licenses.

timon37, I really see no

timon37, I really see no reason for the use of profanity in your response.

It's all personal

I don't think Graham is either wrong or right. His opinion just differs from others. Some may agree with him, some may not.

Personally I don't think there is too much choice. There never can be, but the call for 'standardisation' is what will unify our differences.

As others have pointed out, we need a standardised menu, file system layout, package management, etc. I think these 3 are the major points, but hang on, these is a standard file system layout specification, and a standard menu layout specification. the problem is getting the distros to adopt these standards.

Once these standards are applied, distros then separate themselves by the software they choose to provide as default and support on their system etc.

If you get at least the file system standard applied, then package management is not going to be an issue as you could quite simply install anything on any distro and it 'would' just work (am I being optimistic?).

The only other standardisation I would like to see is package naming conventions. Regardless of a package being a .deb or .rpm etc, it should have the same name. I know this is how it currently is with 3rd party apps, so it again comes down to distros. Why is it the same package could be called foo-1.2 in distro 1 yet be called bar-2.3 in distro 2?

So to put it all simply, standardise the file system layout, menus and package naming conventions, then distro's, go for your life in providing the packages and look that defines your philosophy.

Cheers

Clean-up on aisle 4.

So Caitlyn's posed this question to Mr Morrison: "are you indeed intimidated by the breakfast cereal aisle in your supermarket?"

Can you send Graham to his local Morrison's to find out? And whilst there can he pick me up a pint of milk at the latest copy of Nuts magazine? Ta.

Graham Morrison? Is he the

Graham Morrison? Is he the guy that uses KDE?
If yes, then he's always correct :D

Too much choice?
Simple, just remove all the *buntu distro... LOL

There is too much choice because there is no standard

The fact there there are thousands of choices for things doesn't matter if there were a standard that everyone looked at. In the Microsoft world the standard currently is Windows 7 and Microsoft Office. In the Apple world it is OS X and Microsoft Office. In Linux there is no "standard" to abide by and therefore too much choice is confusing. If Ubuntu, .deb, and Gnome were the standard then it would be fine to have other options such as LXDE, KDE, rpm, Suse etc. but the truth is they aren't standard and having this many different options only makes things difficult for someone to narrow down what they want.
I used Windows XP with GeoShell for years and there were plenty of other explorer shell replacements out there. The fact that the explorer shell was the "normal" shell made me know what to start with and then look for options from there.

How bout KDE/GNOME or Emac/vi while were at it?

Being wrong or right isnt the issue, its rehashing the same old stories, over and over and over again.

How many times do we have to hear a BSD/GPL debate before we scream ENOUGH!!!!!!!!!!!!!!?

Are tech writers so bereft of ideas that they cant seem to do anything but hit the old classics whenever they are close to deadline?

Martin is right, choice doenst hurt and choice is not hard (were in the middle of painting our place and there are 100 shades of yellow!) but people are afraid of choice.
I think we need to look no further than american politics where choice is a two party system. Is this more or less democratic than Russia or Italy with 120+ parties and where proportional representation means that porn stars can be parlementarians (it is the WILL of the people after all).
Yet we act like having a third or fourth party would be too much choice (Nader is taking away democratic votes and helping Republicans stuff).

I could waste my time and talk about the dangers of monoculture, about not being able to stop people from making their own distro (our IT dept has its own distro and I even had my own) because the same license that allows me to start my own distro is the same reason Mageia and LibreOffice can do what they did. But it would be a waste.
The link whore has won.
Heck, he's getting a podcast about it to boot!!!

You think he will stop rehashing old flamewars?
Are you kidding? You see how well it worked last time, why the hell would he stop if it brings the magical page hits?

I see him next week coming out with a ESR quote about the GPL and flinging it like a gorilla would with its poo and see how it works out.

Were not mad at your nitwit per se, were just mad because hes just the latest in a long list of nitwits.

You know how you feel after your 67th blog where some excited new Linux users tells you that "Ubuntu is teh bestest Linnux evar...OMG!!!:-):-):-)"?
That's how reading this warmed over poo felt like.

At least the hyper Buntu newbie is doing it because he thinks that he discovered something no one else has.
Why is Graham rehashing the old classics?
To get hits, controversy... anything to be noticed.

So yes, the cereal analogy was dead on because it conveys the silliness of too much choice but also the exasperation of Linux sites running this kind of rehash.

We know the benefits and disadvantages, so if youre gonna bother, at least have the decency to not recopy word for word the previous two thousand articles written about this.

As for the 'installing Linux apps is hard' intro, I could suggest that both my long retired parents and inlaws give Morrison a primer. My dad installed every game available in the repos one day when he had nothing to do.
I asked him where they were. "I just unchecked the boxes that I had checked off to install the games and kept only the ones I liked."
Dad has been retired for two decades so he has time to help newbies... and he has more patience than I do.

My father in law noticed that he and my dad were both running PCLinuxOS and that I was running Mandriva and my sons were running both Kubuntu and he asked me if they were different, why did they look exactly the same?
He realized what many dont: distros that use the same desktop environment really all look the same apart from a few cosmetic changes like wallpaper, icons and themes.
Put those three distro I mentioned next to my brothers OpenSuse and I dare you to tell me that a newbie would get lost or confused (if themes is enough for you to consider a distro different, then we have nothing else to talk about).
My inlaw has been using Linux for under 2 yrs and thinks that all Linux distros look the same.
Right now he seems more astute about Linux than your writer, ... maybe you could use some of his writing skills since he was a colllege english professor.

Next week: Minix vs Linux, which is better?

Next month: Indent styles: Kernighan/Ritchie or BSD/Allman, who you got?

I cant wait,....

I guess I'm late to the flame party

My fellow commenter is on the money I think, except of course, people who think Allman indent style is a viable option are clearly dense (or have really tall screens).

It's been pointed out already but given that there is no way of achieving your proposed unity unless you become Dictator of the World makes the point moot. You just can't stop us customising our system when we have all the sources - you'd have to take them away first. You'd have to deal with RMS first though and I heard he fights dirty.

As for the theoretical value of the proposal - would it be beneficial if it was achieved? Beneficial to whom?

See, a lot of arguments like this stem from the diverging opinions about who is "linux" written for (linux being a catch all word here). Some people think linux is for human beings, competing on the consumer market. Which happens mostly by accident. When there's money in it - someone wants consumers to use it, buntus being an example - then they make it user friendly and unified. But in fact, "linux" (again used to mean the whole userland) isn't being written for anyone in particular. Most OSS devs are mostly scratching their own itch or producing production software for and collaborating with other devs. Knowing this makes a lot of design decisions more intelligible. All the theming and unification is often done afterwards by someone else who has an interest in making it a attractive and usable to end users. You are welcome to do it but you have no way of forcing original devs to conform to your ideas - they're nobody's people. Unless you hire them :D

Gee wouldn't it awesome?

Wouldn't it be awesome if everyone just wanted to work on existing projects, moving past the fact that their opinions hold 0 weight. After all who would think making their own project exactly how they want it made would be an attractive prospect?

This discussion is meaningless and even if we reached a consensus, reality isn't going to change. Contributing to something greater then yourself in the end can make you feel very small.

Sorry about the length...

Both Morrison and Martin are talking complete rubbish.

Morrison because his column does not make sense as a piece, and Martin because she is exagerating a point that Morrison made rather than pointing out he is talking bollocks.

His headline is "The trouble with Linux: there's too much choice", but he doesn't show whether any problems actually arise. He claims that choice "goes hand-in-hand with redundancy and duplicated effort" but doesn't show that it does.

Morrison's example of Fedora changing the default photograph management tool shows that there are differences between the two: so in the minds of the developers at least this is not a case of duplicated effort but rather two different solutions to a similar problem.

Morrison also seems to think that the people behind Fedora should not have changed the default photo manager because the former default had more features. Surely it is for them to decide where there product is aimed and whether to use complex, simple or simplistic software as the default, and to make their own decisions about what to put on and leave off the CD?

His comments about package management do not really make sense: in essence, people do not care about the package manager, they use the software management tool that comes with the distribution and can effectively ignore whether they are installing .deb or .rpm files. Most distributions seem to have very little software "missing".

Morrison's most stupid point, though, must be "If we could all agree on what should be the default photo manager for a certain distro..." Each distribution has chosen a default, and certainly shouldn't be forced to keep that default for ever, no matter what other options come available and the focus of the distribution changes over time. It isn't as though Fedora changed the default with every release, he says that F-Spot was the "long-standing" default.

The point of the column was intended to be that package management is a mess in Linux but his solution of each distribution choosing their own is the current situation.

There are two problems to the idea of "universal binaries" in Linux (by universal I mean will work on the vast majority of systems, not necessarily all.

Firstly, the big distributions, ie Debian, Red Hat and Suse need to standardise on names, in order for dependancies to work properly.

Secondly, dpkg and rpm need to be able to read packages in the other format. There is, I believe (but I have no actual knowledge of packaging), no great obstacle to this: both package formats essentially contain the same information.

Intimidating cereals...

According to mysupermarket.com, Tesco has 163 breakfast cereals, increasing to 249 if one includes porridge and meusli.

I am not "intimidated", but I certainly haven't tried very many of these; and it is very possible and even probable, that I may prefer one of the ones I haven't tried.

I get through about one box of cereal a month, so it will take about 20 years before I have tried them all and can make an informed decision.

ENOUGH

Enough with the stupid tick for tack arguments. This comes down to how pragmatic vs idealistic you are. You (the posters)
think that you are perfect in what you think and do in life and when it comes to Linux. In my humble opinion Caitlyn Martin just needed to keep her mouth shut.

She is as much to blame for bringing this flame war back. She adds nothing new and uses very lacking comparisons to try and prove a point that ultimately does not matter. I think that for the most part distributions do look similar and that is where the confusion can come. you think they install packages the same and they don't and so you get frustrated.

I am the first to admit I am more pragmatic when it comes to Linux. Choice is good but just like everything to much and it becomes a bad thing. Most people just want to use there computer to do a simple task and they DO NOT CARE ABOUT HOW A OS WORKS. They want to get on with life. So distributions should just pick software that they think works the best but give a few common alternatives so if someone does not like it they can easily find one that does.

I unlike most of the posters above realize my opinion means absolutely nothing to anyone but me. Graham Morrison is in every right his opinion just like I am to mine. I think He is right in his general point. The reason why these arguments come up it because nothing has changed.

I am from America. Do you think this country with all its freedoms would have been formed if someone did not take over and just come to a agreement between two sides and make middle ground and enforce it? If it was left for everyone in the country to come to a agreement we would still not have a government.

I also want to point out how many of you follow Kernel development? I bet not to many of you well do you think core Kernel changes comes to a vote? Think again Linus has to approve of it , and you want to know why that is? It is that sometimes a strong leader that controls how something is done works. a standard base would help and from there packages can fork all they want.

P.S. how many customized packages (that are changes to a main stream package or a fork) that server a small (less then or around 1000 users) could just be merged with there parent package?

choice is our choice

when you choose foss, you choose choice.

how can you have it any other way, different people think in ways... and that can bring some exciting development.

Too much choice is a bad thing... for new users

Look at Windows users. How many of them know and care what Explorer is? Not many. Now imagine if Windows prompted the user to choose a desktop environment, window manager and file manager when you install it or log in. They would freak out and give up, labelling the operating system as not user-friendly and for technical users only. And that's one reason why Windows is popular. Most users are too dumb to want choice - they just want something that works.

For experienced users choice is a good thing, provided the options have different benefits and aren't all trash.

Right and wrong

Graham is certainly right about package managers - both for the developers' and for the users' sake a standard package management system would be an incredible step forward. How can you explain the average gaming company that, after porting their game for a tiny market share to Linux they should also put some effort in distributing it in at least 5 ways to reach an audience? And how can you explain a completely inexperienced user that he can't use that .rpm file on his Ubuntu installation?

However, choice in applications is good. I, for example, always hated F-Spot as it is a bloated, constantly crashing mess. Therefore, I am happy it's not the single standard photo manager under Linux! Let applications vary - the good ones will survive, the bad ones will slowly die.

What's the boundaries of open and free software?

Frankly I don't find Graham Morrison's article coherent enough. Selected examples don't back up his claim very well. A unified package manager wouldn't change how businesses view customer information as a useful source of making profit. Hence it in my view becomes a mish-mash of problems and causes.

I did buy Linux Format for a time but stopped because of Graham Morrison's articles. My main concern about his agenda is that it's influenced by restricting corporate ideas, probably unintentional, not respecting the nature of how free software come to be. As of now the opportunity to create whatever piece of code you like without facing judgement about whether it fits into the decided conformity, is in need or surplus, works as an incitement for creativity. Mathematics is art, and you don't tell an artist what he can and cannot do; and from mathematics the step to code isn't far. Some creations become beautiful and useful, others not, directed by users' selection.

That said I'm not sure I'm in agreement with all said by his critics.

Choice is a good thing that some people aren't used to

I'll start by saying I think Caitlyn is maybe slightly missing some of the points Graham is making - Graham is saying more (as I see it) that excessive diversity can be not such a good thing as it does mean that effort is wasted and duplicated and some apps maybe don't get the attention they deserve. Personally, I think we need a good balance - package managers is one area where the choice seems rather redundant. I know I'm repeating what has been said in the podcasts but I don't give a rat's arse which one I use, so long as it work. However for the majority of software I think the choice is great - it's one of the things I will tout about Linux to non-Linux users - I'll open up the package manager and show them the vast array of software I can download at the touch of a button _and for free_ and ask them if they can do that on Windows. I also point out that because it's free they've got nothing to lose but a bit of time to try them out.
One way that the amount of choice _is_ bad, however, is in the culture-shock of coming from another OS where there isn't that choice - people aren't used to having lots of apps to look through. The main problem with this isn't the choice itself, but rather the fact that a lot of people today can't be bothered to think for themselves and put the effort in to find what's best for them - the "middle ground" in this case has got to be getting a good set of "fits-most-people" apps on the default install.
With Linux, as with most of the good things in life, you get out more if you are prepared to put a bit of effort in, and the choice in Linux enables people to do that on a great level
twitter.com/aSheepie

Choice is mostly bad...

Sure its nice to have options but the main proplem with Linux is the lack of standards.

This is why all the hints etc are using commands because the only thing which seems to be standardized is Bash.

I think Graham is correct in his opinion and he has my support. We need standards and to stop messing about with the "choice is great attitude".

meh

You're just trying to generate clicks by inciting flamewars or quasi flamewars over non-issues, and then resuscitating these stories when they die out.

We seem to be having an inordinate amount of empty debates in the Linux world. And an increasing amount of ridiculous, shallow blog posts, and this here isn't even the worst.

There are better ways to generate web traffic you know. I'll- leave you with that.

Well...

For OEMs, choice is bad.

Example:
Dell should be offering PCs with Mint, not Ubuntu, simply because Mint is full-featured and easy-to-use out-of-the-box.

@ Anonymouse Penguine

We're just passing time before we shuffle off this world.

...and the best way to generate web traffic it seems is to piss off 4chan. Insert piracy (or copyright in general) into your 'just a minute' knock-off wheel of fortune thingy.

Standards are wonderful

...but who gets to make them? The Linux ecosystem isn't owned or controlled by any one person or group. If some group did try to impose standards, there would be some group of rebellious mother forkers who who would just make a new distro their way, and the chaos of the bazaar would begin again. Cathedrals are built to standards, bazaars just happen. ESR may not be as "in" as he once was, but his metaphor is still apt. Linux, by its very nature, cannot be a cathedral, no matter how useful it might be in so many ways... So, while I sympathize with Mr. Morrison, it is a pipe dream that Linux can be that way. Stop smoking the happy weed or else join a cathedral-building empire like the Church of Apple.

For the consumer, choice IS

For the consumer, choice IS confusing and overwhelming. We shouldn't shove 15 photo management apps at them, just three really polished ones. It's nice for us geeks to have choice, we have hours to waste away comparing them all and actually understanding what the differences are. The general person just doesn't care, they just want a polished app that does what it says on the tin.

choice, to be or not to be

choice is good, but too much choice like too much beer or too much anything else is not so good. the question is whether we have too much choice, or who is to judge that?

you can choose not to choose

If you want to limit choice, look at #!CrunchBang with the OpenBox desktop: it gives you the basics and manually configuring OpenBox for the novice is as simple as working on your dishwasher while it's running! Just let the n00bs try to figure out that one...

Past that, I end up using pretty much the same programs in whatever distro I'm using; gedit, GIMP, Inkscape, terminal, etc.

The real difference is the community, wiki, and forums. The reason I use #! is because they don't make me feel like an idiot when I sometimes ask a dumb question. Patted on the head like an ignorant child, maybe, but I don't feel burnt from the flames.

So...we have a choice.

yay. I choose to be the chosen for the formidable task of choosing which applications and standards will be used in the following releases of every linux based distribution, and, my first choice is that you have no choice.
Don't you like my choosing??? But I do so live with it :P.

Good news really

Reading through this it is good to see folks talking about consistent structure and packages. Sure, a proliferation of distros can only be good, provided they are built on a consistent structure with a single package format.

I think Graham is on the

I think Graham is on the right lines although his generalisation-engine appears to have run over the sanity check occasionally. Are most of us really that confused? Dear oh dear.

But on his specific point, concerning Fedora and the photo imaging app switch, I agree -- and I've never used either of them.
I know Fedora is intended to be experimental, and it's fine to introduce new packages, but there should be some thought given to users who happened to be using the old one. Ms Martin responds to this by saying that
"Fedora didn't drop F-Spot. It is still in the repository."
True, but what does the user do to continue to use it? Turn to the F13 release notes of course - and they say
"Shotwell replaces Gthumb and F-Spot as default photo organizer. ... Gthumb and F-Spot ... available in the Fedora repository. They are not installed by default anymore."
I think those notes should say
"To revert to using F-Spot, do these steps ..."

Another similar one from back a while - let's call it "Rythmic-throbbing-sound" to avoid triggering more incendiaries -
Fedora 8 Release notes :
"Rythmic-throbbing-sound is now installed and enabled by default. ..."
And did they go on to say how to turn this off?

Choices abound

Choice can be a good thing or a bad thing, it goes both ways. When Microsoft released Vista, they had around 6 or so (don't really remember) versions for users to choose from. That frustrated many people because they didn't know what they needed and the person selling it wasn't sure what the user really needed, so in the end they ended up with the wrong version.

Linux choice is astounding. With Distrowatch listing all available versions of Linux, that is quite overwhelming. Choice is a very good thing, but when a Sanskrit Ubuntu is released instead of being another Ubuntu version, its counted a whole new version of Linux.

Instead of making each individual version of Ubuntu count as a single different version of Linux, make it count as only Ubuntu with multiple desktops. The overwhelming number of Linux's (linuxes, linuxi, ???) is part of the stumbling block in widespread adoption, but also makes it one of the great things about the OS. An individual user, given time can find the Distro that is right for Him/Her.

Graham is close but wrong

The real problem with Linux is that the big distros don't have the courage of their convictions

Another perspective?

One thing I never hear is that, perhaps, just maybe, Linux as a whole SHOULDN't be mainstream. In terms of an analogy, perhaps Linux already succeeds as a specialist tool. Lots of people use knives, but no one would say that everyone needs and should use a surgeon's kit or a filet knife when making a pasta sauce. Breakfast cereal is overly simplistic, but so is Graham's argument. Some tools are made for a certain job, excel at it, and don't work as well elsewhere. The world is so rarely as it should be.

But what's the market share?

Personally, I like choice, but I suspect that many of us are just making the same choices, so that there are a few key players in each category . . . and then a load of also-rans for the adventurous, curious and just damn awkward. So, deb and rpm probably account for about 90% of software packages actually in use on Linux; *buntu, Fedora, Suse, Debian and possibly Mandriva probably account for the vast bulk of desktop Linux use; OpenOffice and Gnome Office probably account for the vast bulk of office software in use on Linux, etc. So, there's a reasonably common core and a wide, varied periphery.

Plus 1 for me...

Sorry, but I'm in the GM was wrong camp. Anonymous Penguin (not verified) - October 4, 2010 @ 5:15pm is pretty close to how I feel on the matter.

worker

Of course there's too much Linux choice. All you need is Red Hat/Fedora, which is the distribution I've been using since 1994. Pity other people can't see this, and just think and do as I do. Come to think of it, I have this problem in many other areas of my life.

well yes, but...

I'm not sure it's reasonable to compare software or distros to breakfast cereals. Most people already inherently know how to operate a spoon. The mechanics involved in bringing granola to one’s mouth happen automatically. Most people do not inherently understand what an RPM is, or whether Mono is at the root of OS software selection decisions.

Once again, I am concerned that the discussion we’re having will be limited to people who know what grep is. Choice in computing is fantastic, of course, but not to my Mum. Perhaps Graham’s article should have been titled, "The trouble with Linux: too much choice creates difficult collective marketing."

Hmmm

Interesting as these kind of debates are - I ask myself are they relevant?
Nokia and others might have asked in the beginning if what Apple did was removing choice from the mobile phone users. Well now they are not asking but trying to catch up.
Humans are by default quite lazy - it is in our nature - we follow OSPF by default. The same comes with computers and technology - how many times do we say RTFM, however we never seem to ask why do we need to say RTFM. With Apple users say you do not need the manual as it is sooo intuitive.
I have never used Apple myself, I read manuals and I dabble with computers a bit (as I like it) however most people don't and hence what do they choose.
I hope they will choose Linux and that they will choose what suits them, just like I do. However like with all things first impressions make a lasting impression. So if the choices shown to them are confusing and not to their liking are the likely to continue or give up? History has shown that they give up and choose the OSPF method...
So Graham has some points as does Caitlyn arguing about it does not seem to get us anywhere though.
Driving FOSS forward we need to listen to users and present simplicity while allowing for choice. I could not live without the choice anymore. I am often so confused when I use windows now adays and when I realize that the things I take for granted are things I have to pay for if I am to use in Windows. I can do more at home on Linux than I now can on my office computer and I work for a large corporation. That is diversity and something to celebrate and build on not argue about.

Real Life Relfections

how many of us actually eat more than 1 (or maybe 2) types of cereal? yes, there are a lot of choices (for linux or cereal). However, once one picks a flavor, it rarely changes. I think what helps us in choosing a cereal are the pictures (and the nutritional info panel) on the boxes (taste really comes after the cereal is bought) So, with linux "flavors" I suggest that the "boxes" could be a bit more descriptive. There are so many flavors of linux trying to do it all. Maybe each distro could promote the focus and strengths of their product, so that the consumer can make the best choice. oh, and 1 last thing no anti-competitor distro stuff. that's not in the spirit of linux.

mostly not

Linux distros agree on a common kernel. It would be nice to have that expanded. It would not be unreasonable to have a standard file system format, configuration utilities, shell, package manager and desktop environment.

It is unfortunate that we are divided by some of these issues. We are not talking about a few less popular distros creating these divisions. The top five or ten distros have many differences on fundamental levels.

Distros could have their "standard" version. That wouldn't prevent them from making a few remixes (ubuntu, kubuntu, etc...)

You don't have to take away choice. Users who want to change their environments, applications, etc... could do so at will.

I do think it is important for developers to have a standard environment to target. Otherwise it will not be worth their time (in a business sense).

As much as I enjoy my custom system and open source applications and even finding creative ways to solve use cases that are trivial on other OS'es... if we are talking about growing the user base then you have to get to things that "just work".

Having the choice to use commercial applications on a standard platform would be important in that matter.

Until people can get things to just work the platform won't be taken seriously in the home user space.

Businesses are largely the same. Small business seems to be a large segment where MS is dominant. SUSE Enterprise seems to have taken a stab at it with some success. In the end I don't think they offer enough of a benefit over MS servers for MS shops to justify the change.

All this aside there seems to be another issue... for example, something like Netflix. Cases where it seems the platform is deliberately isolated puzzle me. I think the man is just trying to keep the penguin down.

Linux has advanced far enough that a great desktop (as defined by the other 98%) is a realistic goal. If the community can establish a standard toolset and some solidarity the user space is there to be had.

Solidarity!!!

p.s. I appreciate the work the Linux Format crew does. keep up the good work.

cereal killer

by the way the cereal analogy is a terrible one. you are just changing what you are pouring into the bowl. changing the os seems like it would be more akin to swapping out the bowl and spoon rather than the cereal.

since the analogy is terrible why not cling to it here...
the bowl is the linux kernel.
the spoon is your distro--the desktop environment, shell, utilities that what you use to do things or enjoy content. the cereal would be the content and maybe content specific apps.

while some people may eat their cereal with a claw arm, a stick, or just drink from the bowl it is nice for those less adventurous types to have a spoon to resort to.

note to self: wasting time on flame wars is a waste of time. civ V ftw!

if ( grahamsArgument.isTrue() ){ "it's a matter of scope" }

Linux is so much bigger than this...

Graham, in a way you are correct. If the number one priority for a group of Linux distributions is to become a Windoze replacement for home desktops making use of commercial software... then standardization is important.

Many distibutions are so much more than just desktop environments for home users though. So, really your argument is... to prevent fragmentation of Linux you should create a new fragment. If a group of distributions decide to specifically target a market segment and work together to create a standard then it could be a good idea.

However, really, eventually users will just pick a distribution and suddenly there's your standardization. It's all just an abstraction.

mrEvans: some proprietary software products involve an exchange of funds for usage rights. active sync is a great example of this, silverlight may be too. a free distibution may not want to pay for the usage rights. or it could just be the man geting you down.

no

No

No but yes but no but yes

see above

Graham Morrison is not wrong.

Great podcast guys. I wanted to wait to hear it before drawing a final conclusion on this, and it's clear to me that Graham's point was taken drastically out of context.

Regardless of that, these flame wars are doing no one any good, and certainly not helping the cause of Linux. If the flamers could learn to curtail their 'knee jerk reactions' and discuss the topic instead of responding emotionally and angrily right away, it would be vastly more constructive.

That said, I'd like Graham to know that I and my rather large family appreciate his efforts and opinions and we say, keep on truckin'!

I disagree with Morrison.

I disagree with Morrison. However, I think I somewhat understand his agitation. Linux is trying to come into a world where, for over 20 years, there's really only been ONE choice.

Windows.

A lot of people think, and incorrectly, that having a lot of flavors of a particular operating system seems to make compatibility between all the choices an issue. They're COMPLETELY oblivious to the nature of POSIX and what it allows a Linux distribution to do. You *can* and you *will* consistently run the same software across multiple distributions, be it Ubuntu, Debian, Arch, SuSe, or Red Hat.

Proprietary software is not even that much of a concern even if POSIX governs binary compatibility by no means (Which is why the Linux Standard Base came about.). Yet, again, the same proprietary apps still consistently find themselves working generally the same and just as well as any other distribution.

And distributions competing amongst each other is VERY good for end users. This competition can and does spurn on a LOT of advancement and upstream contributions to individual open source projects.

Bottom line: Linux (Far more than even GNU) has done so much to bring open source to the top. Because of its popularity and attention from millions of developers worldwide, you run a useful open source project, you'll find its hard to keep up with what your users do to improve your software.

Choice is Fun

I personally enjoy trying out different distros. To find the one that suits my needs. Especially if a distro becomes too buggy and bloated and no longer works for me.

However isn't great in the GNULinux world people are free to have their own point of view on GNUlinux like Mr Morrison.

Not enough choice!

When you visit your local PC World (Or whatever other shop) there are Windows PCs & Macs. Generally there's nothing with Linux pre-installed.

To carry on the dubious car analogy put forward in the podcast - you're basically taking your new car home and fitting a different engine that you haven't test driven and don't know will work properly in this car. (OK so probably most readers of this will have installed an operating system but this is a big deal for most people who have never needed or wanted to do this.)

Until machines with Linux pre-installed are visible and easily available on the high street I suspect the market share for Linux is unlikely to exceed 2%.

Choice or newbies, they don't go together

Two questions are (a)What is the target market and (b)is there a need for Linux to dominate?
Having tried lots of distros, from early slackware on, I like choices and sometimes swap apps along with distros. I frequently disagree with some of the choices made (GIMP I find awful, Evolution is too heavyweight, the open java clones are not as good as the Sun/Oracle implementation etc). I know others disagree and they're right to. For us more techie types, we can easily find/replace the things we want as long as there is choice, so choice is good. But what about new users?
My wife uses one of my machines. If I change the email client on it I know I will be in trouble. She can't find things and has no idea how to get back and re-install the one she uses. Choice is bad for new or nervous users. As someone who taught IT in a school for many years I know the easiest way to confuse teenagers with a computer is to change the version of (say) Word so that menu items move. That's all it takes. If your target market is new users, choice is bad - and look at the criticism over all the versions of Windows 7.
But does it matter? Unless we want to dominate the world then as long as there are enough users (and there are) Linux will stay healthy and prosper.

Choice for the experienced, recommendations for new users

New users want advice, not just choice.

I would suggest the best way forward would be to have a rating system for applications. In any one category, a user can select the application with the best rating. The choice is still there though.

We all benefit if linux is more popular, if only in that more hardware manufacturers will provide drivers.

My cereal doesn't last that long

The cereal analogy is weak at best. When I select a breakfast cereal, I am making a choice that I will live with for what, a week?
Graham is a long term Linux user, enthusiast, writer, etc. He wants Linux to grow past the 2% uptake that the current linux model has succeeded in attracting. We've tried status quo for more than 10 years. Perhaps it is time to try another approach.

Im a newbie.

Before a year or so I had spent a whole hour in an non dos/windows environment. But I had a pretty good idea what they entitled.

I tested Knoppix to its full extent, or rather what I believed was its full extent. When I gave up I grew more and more discontent with how windows operated. It seemed to slow and buggy. But with enough usage you learn how to make it bearable. Now seven years after I first got interested in linux i have tried two versions of K/Ubuntu, OpenSuse and Fedora. Iv never actually used it more then a week in a row. The reason for which I thought I would share here now.

When new to linux you have steep learning curve. Which of course it true if you switch every time you switch an operating system. But once you have learned the basics they are there, and the changes really are trivial enough that you don't have to get confused, etc. On linux on the other hand the basics is the same but how you do the basics differ. It ranges from everything from .deb to .rpm to tar.gz and even less technical things like how to change your settings.
I could easily learn how to do the simplest of things using the terminal in Ubuntu, trying it in OpenSuse didn't work to well and took some reading to figure out. I even got the wrong package format a few times.

While little things like this can seem kinda trivial as really it only takes a bit of reading to figure out. I really believe it to be what keeps linux out from the masses(besides the obvious lack of "professional" software and games).
I really do believe that everyone would be better of with a uniformed control panel, one package manager front and to go with that one package architecture. That way no matter which system someone uses they could get help from virtually anyone, regardless of their system of choice.
I'm not saying that taking away distortions or even killing of software projects because there is already one that does that is the solution on the other hand. I believe that to be a strength.

While I think it won't happen as there isn't one governing organ, it makes me kinda sad the fix for, to me, the biggest issue with linux would be overcome by simply working together a bit more. Even then it doesn't have to be the same program(referring to the control panel), just uniform.

The answer was in what i believe was a quiet rude fashion, that alone makes me inclined to agree with him besides the points said.

Give me a universal distro when it works universally.

Not all versions of Windows work on all machines. Try running Windows 7 on a 256M Windows XP machine, even if you can find the drivers. Try running OS X on a Mac Plus with 128M RAM.

We can't even standardize on one type of computer or automobile, why should we expect to have a one-size-fits-all distro?

The arguments in both articles are inane. Automobile drivers are expected to go from a Mini Cooper to a Cadillac Escalade at the drop of a credit card at a car rental kiosk, why can't knowledgeable computer operators be expected to go from a Windows 7 to a Mac Mini to a Fedora laptop with the same facility?

Its all about choice

6 billion people in the world, 7 billion opinions (schizophrenia, you know). Why not adopt the signature of a linuxoid friend of mine - "Linux, the lifetime learning experience." (I am a linuxoid also but not as brainy as he is.)

In the last 10 years, I have installed at least 14 different distributions, some of them in several versions. I am still looking for the perfect distribution. If every distro becomes exactly the same, it will still not be perfect and the bar to innovation would be much elevated. Bad deal.

BSD is also a very good OS but there are very few versions compared to Linux. The ones I have tried have focussed on parts of the OS that are less addressed by the Linux community. IOW, BSD'rs are a great but different community from the Linux community. Let's keep the choice.

Post new comment

CAPTCHA
We can't accept links (unless you obfuscate them). You also need to negotiate the following CAPTCHA...

Username:   Password:
Create Account | About TuxRadar