Well, now. This is interesting: A study of corporate Linux adoption polling 1,275 IT professionals says:
Linux desktop roll out is easier than expected for properly targeted end-user groups
Those with experience are much more likely to regard non-technical users as primary targets for Linux. The message here is that in practice, Linux is easier to deploy to end users than many imagine before they try it.
It’s become fashionable lately to be pessimistic about Linux’s future on the desktop, but I have to say this matches my experience pretty well. The handful of Ubuntu deployments I’ve done in the last couple years for end-users have indeed been easier than one might have expected.
The study further notes that most respondents have only deployed Linux to approximately 20% of their users, so the pressure on Microsoft is likely to increase as planned rollouts shift a cohort of the userbase literally 5 times larger than today’s. The study also notes “Desktop Linux adoption is primarily driven by cost reduction”, which is especially interesting given our second news item…
Here it is: Linux to regain 50% market share. And cost reduction is what the oncoming shift to low-power ARM processors in netbooks is about, a move that will lock Windows 7 out of the fastest-growing segment of the hardware market.
All this perfectly fits the prediction I made in January that Linux would regain netbook market share from Windows as purchasers seek out small, cheap, light machines and netbook makers drop Windows in order to claw back profit margin. Microsoft can delay the latter trend only by kicking back what would otherwise be profits to the netbook makers…and when your earnings are dropping for the first time ever it is a very bad time to be giving up profits. As ARM chips displace Intel processors, even the bribery option won’t keep Microsoft from hemhorraging market share.
When the final collapse of the monopoly happens, it’s going to happen fast…
> When the final collapse of the monopoly happens, it’s going to happen fast…
You know the ‘too big to fail’ mantra? Somehow I think Microsoft can scare members of congress to bail it out if they just babble about “essential internet infrastructure” enough.
I find that Linux is incredibly easy to install. In some sense, it has to be because it doesn’t usually come pre-installed.
As far as the rest of the deployment goes, I’ve noticed that many users can’t even really tell the difference between Microsoft Office and OpenOffice with the exception of powerpoint or complicated graphing in the spreadsheet and they deal with firefox and thunderbird just fine.
I am a bit skeptical of the last sentence, though I suppose it depends on the meaning of the word “fast” and when you no longer deem it a “monopoly” (because it’s debatable if it’s even a monopoly right now).
There have been rumors of an ARM port of Windows 7 precisely for the uprise in low-power long-battery-life netbooks, although the obvious problem with such a port is lack of x86 compatibility in the hardware. Sure, they could make an ARM port, but running all your favorite Windows applications would be no simple task. One possibility, that was actually employed in the Alpha version of Windows NT 4, is to implement an x86 emulator into the operating system and have compatibility, albeit at drastic performance costs (I don’t even want to think of how much power such an act would take, I wouldn’t be surprised if emulation negates any benefits of ARM in the first place), with all your favorite x86 programs. To put it blunt, an ARM port of Windows 7 is doomed unless everyone is running open source software compiled for Windows/ARM (well then, why don’t you just run an open source operating system?).
Dallas: There’s enough competitors to Microsoft in the big arena that I doubt it’d be as simple as that… at least I hope not, but don’t count me as surprised if it does happen.
Maybe a tiny bit off-topic: I’m probably biased because I work at Google, but I think that Android (atop Linux, of course) has the potential to really take off and take a huge bite out of the low-end Microsoft market.
Actually, I’d like to clarify my point of “unless everyone is running open source software compiled for Windows/ARM” as I’m sure there will be the obvious counters that there’s no reason proprietary software can’t be ported as well: Proprietary software is driven by whether the effort of a specific task will be beneficial to their business. If there were an x86 emulator built inside Windows/ARM, why port it to the platform specifically, possibly having to rid their software of architecture-dependent or endian-dependent code (or worse, an #ifdef to split up code between x86-specific and ARM-specific portions)? That’s more effort, and even though the software might run slower, why is the Netbook market so important for them? Really– I’m sure people, somewhere, actually use Photoshop on a Netbook; Adobe (and most other proprietary vendors) can’t even be assed to port to x86_64, the now leading architecture on personal computers, why should they bother with a third architecture, even if that means Photoshop will be slower than ever? Open source has a unique benefit — the developers do it for the benefit of the community at large. For the most part, OSS developers don’t seek retribution for their work, they are happy enough making the world a better place, and having a native Windows/ARM port would indeed be better for users running such an operating system.
Anyway, that’s enough speculation (for me) for what is only a rumor.
>Somehow I think Microsoft can scare members of congress to bail it out if they just babble about “essential internet infrastructure†enough.
Not with IBM and Google fighting the other corner they won’t.
>Sure, they could make an ARM port, but running all your favorite Windows applications would be no simple task.
Indeed. Add the footprint from the emulator to the massive bloat typical of Windows apps and you’ve got a no-win situation for Microsoft. Oh, and there’s the device-driver issue…the havoc wrought on binary drivers by the 64-bit transition was nothing by comparison…
Let me take a role of advocatus diaboli: I wonder how portable the whole CLR of .NET platform; would be enough to port CLR (which as I understand is equivalent of Java’s virtual machine) for all closed-source .NET applications to run automatically and without problems on ARM? Or do I not understand CLR?
>I think that Android (atop Linux, of course) has the potential to really take off and take a huge bite out of the low-end Microsoft market.
Um…doing what?
That’s not a hostile question, I have an Android phone and I like it a lot. But I can’t see Android devices as replacements even for netbooks, let alone PCs.
Actually you bring up a good point, that not all is lost for Microsoft; it’s a minority of programs, but at least some of them would be able to run on ARM without recompilation or x86 emulation. I don’t know how well-coded .NET itself is, but Mono (an open source CLR implementation) has been ported to ARM.
I have a suspicion he’s referring to a netbook-tailored version of Android…
> I can’t see Android devices as replacements even for netbooks, let alone PCs.
Then you’re looking in the wrong direction.
Android has a (nice) browser. Many ‘desktop’ apps can run in … a browser.
One can write apps for Android, so its certainly possible to write custom software for such a machine.
Android runs a linux kernel with a ‘java’ stack over the top. Technically, it is even possible to run ‘native apps’ (binaries) on Android,
though these require a re-compile due to the different libc.
Perhaps you are unaware that Android already runs on several of the (x86-based) netbooks. In another 18-24 months, I think it possible, and even likely that an Android-based netbook would be on-par with (or even superior to) Ubuntu (or Fedora) on the same hardware. Perhaps not for developer desktops, but certainly for the types of deployments cited by your article:
And none of this touches on the real issue. Google is a household brand. Linspire, Cannonical, etc are not.
Novell is better known, but Aunt Tille doesn’t know who Novell is.
She does, however, use “Google” as a verb.
Fortune 1000 companies see a “big company” behind Android, and (with the exception of IBM) only small companies behind linux.
Android really could unseat the linux ‘desktop’ on its way to destroying Windows on portable devices.
>Then you’re looking in the wrong direction.
Don’t assume your conclusion. There may be a niche for a tablet-like device that is, in effect, an always-on browser with WiFi, and certainly Android could fill that – but this doesn’t span anything like the entire range of what people buy netbooks for and would not replace them. Even for the transaction/light-professional workers there are some minimal requirements that can’t be really be met through a browser; presentation software, for example
Accordingly, I must consider talk of Android-based netbooks mere handwaving from anyone who is not presenting a clear functional profile of what a netbook is and how it differs from a cellphone or a browser tablet.
(Mind you, I might be talked into buying an Android browser tablet at the right pricepoint, below $100 or so. But I wouldn’t think of it as a netbook and wouldn’t use it for the same things.)
Oh, one more thing.
On the other end of the scale (massive multi-processor machines, Intel has an 8 core part coming out soon, AMD a 6-core part. Both of these are destined for motherboards with 4 (or more) sockets.), linux doesn’t scale that well above 8 cores.
( but it scales better than windows.)
Even if linux were to (somehow) scale, the applications (other than virtualization) won’t, because programming with side-effects does not make for stable parallel programs.
And here we can again look towards Google and functional programming idioms such as map reduce.
not that they invented the idea, but they did make its use… popular.
> Even for the transaction/light-professional workers there are some minimal requirements that can’t be really be met through a
> browser; presentation software, for example
First, presentation software does run in the browser. See: http://googleblog.blogspot.com/2007/04/were-expecting.html amoung others.
Otherwise, NeoOffice is a version of OpenOffice written in Java. Given its existence, can a port to Android be far behind?
And again, its possible to run compiled (non-Java) code on Android, so a native OpenOffice could be run, though the GDI portion would need porting.
Similar to Linux vs Windows, Open Source apps would be ported and fixed on the ARM faster than closed source applications. This is a good sign for not only Linux, but rather for the Open Source eco-system and movement as a whole.
>When the final collapse of the monopoly happens, it’s going to happen fast…
That read my mind too. Very true.
>I’m sure people, somewhere, actually use Photoshop on a Netbook;
People certainly want to, if they don’t already do. Scaling down GIMP or InkScape should be easier and faster (than PhotoShop or Illustrator) as the source is open and is not subject to business strategy decisions.
I have slightly “edge-on” view on this.
Microsoft will own the desktop as long as it can perform adequatly. And that is for pretty low values of “adequate”.
But the desktop will, very slowly, fade into irrelevance. Moore’s law has made a twist that will see to this.
The chip producers have finally met a brick wall in terms of single processor, classical computer performance. This happened around three years ago. So, we have to live with 3 GHz Xeon’s, or their equivalents, as the ultimate mainstream desktop processor. So, the machine will need to work well on that.
We will have machines with multiple processors, but they are slow to penetrate corporate systems. And the multiprocessor systems eat resources, like power, like crazy. On the other end, the desktop environment feels like a stifling straightjacket. Except for the gamers, there is nothing exiting, new happeing there. Microsoft cannot handle such excitement.
However, below this we see a whole underforest of interesting things happen. Moore’s law is still in effect, even if the top end speed corrolary has ceased to function. Nowadays we can have very decent 1GHz machines that have a _lot_ smaller footprint, and 200Mhz machines that can fit in your shirt pocket and run a week on AA batteries. This has only just started.
It will be different from the desktop. Very different. And Microsoft will not be the monopolist here. Sure, they can claw onto some market share, but not as the ruler of the universe. They simply haven’t got the products for this. Meanwhile we will see a lot of open and closed applications, like the iPhone, the EEE, and scores of others yet to come.
This will not be the desktop. Just like the first EEE that arrived in this house was confiscated by the wife, to be used as the mobile unit of choice. Linux is irreleant for that. It may be the enabler, but the product is the key. We may see a lot of “application as product” machines.
Which will leave to us, the professionals, to have the servers scale to handle this application load.
— mrr
@Bret — I’d have to say I agree. The only major issue I really have with Ubuntu is that it doesn’t like my wireless internet (it happily takes being wired up to the router, though, for some mysterious reason). Other than that, it’s a good OS, easy to work with, easy to install. (Though I will admit that uninstallation, if you use the grub, was more difficult! But that’s because I was doing it for the first time, so… and yes, I dual boot Windows and Linux.)
Open Office, I actually run in place of Microsoft Office. I like being free from Clippy. *g* It’s easy enough to work with, and I like that it saves to various formats, and you can still open your stuff in MS Office.
Firefox…. I’ve had Firefox for a while, and it’s my preferred browser, but I’m having problems with it lately. I don’t know if anyone else notices that it seems to have gotten bloaty in the last few releases? It takes forever to start up, and even then, it tends to freeze up for the first minute or two after loading, which annoys me to no end.
This said: I abhor Gimp. It’s actually harder to work with than Photoshop!
mikec > Maybe a tiny bit off-topic: I’m probably biased because I work at Google
I know Google has its reasons not to make this data available, but it has to be said: could you get the powers that be over there to publish user agent statistics for Google search and other apps again? I think the sample is large enough that it should give a reasonable estimation of the Linux user base and the growth trend, if any.
I recently saw a blog post by one of the Fedora developers reporting that they had used the Smolt unique ID data to estimate what proportion of the IPs accessing the update downloads are actually unique systems. They found that the number of unique IPs is a underestimate and should be multiplied by 1.15, which gives about 16 million Fedora systems at the moment. I was wondering if the other major distros could also count their unique IPs and perhaps give the list in confidence to some third party that could analyse how much the installed bases of the distros overlap and come up with some sort of a total.
I use Linux as a matter of course now. There’s no longer the X-factor surrounding it. 8 years of Debian and Debian-based distributions have had their effect.
I use Vista very very occasionally for Skype (64-bit Skype apparently hasn’t been released – another thumbs down for proprietary code) because people seem very keen on Skype and not on alternatives like Ekiga.
I must say though that so long as the popular commercial applications support only Windows, there’s very little chance of Linux completely overthrowing the Microsoft platform.
Clarification: when I said 64-bit Skype, I meant 64-bit Linux version of Skype.
Regarding ease of deployment. I have a bit of experience getting Windows and Linux to talk to each other. I’ve also done my fair share of clean installs with both operating systems. This is my recent home network experience.
A couple years ago I bought a low end laptop (Compaq C501NR) with Vista pre-installed. The wireless networking never really worked right (Broadcom something or other). The short version is that once the machine went to sleep, the wireless connection could not be re-established. And browsing the local Windows network was hit or miss even when I did have a connection. The “solution” was to reboot the machine. I experienced the same problems on the same hardware with Vista, XP and the 7 beta.
Ubuntu, on the other actually, actually works on this hardware. The install was easier, connecting it to the printers on my wife’s XP machine was much easier than with Windows, the updates are easier, finding software that I need for basic tasks is easier, everything so far is easier. And the wireless connection, after the machine goes to sleep and wakes up, re-establishes itself automagically. Prior to Ubuntu I used Slackware more often than not and was always hesitant to recommend it to Windows users, fearing that I would become their “support guy”. I don’t worry about such things anymore and life is good. I hope I haven’t rambled on too long.
JC:
linux doesn’t scale that well above 8 cores. ( but it scales better than windows.)
I’m not sure that’s true in most meaningful cases; the truth, AFAIKT, is that pretty much any OS scales fine to reasonable levels as long as most of what you’re doing is user-space compute and not I/O intensive… the number of applications that need both massive tightly-coupled compute power and have high I/O to compute ratios just isn’t that large.
I do agree that 2010 is the point where the high-end is really taken over by x86.
esr:
Mind you, I might be talked into buying an Android browser tablet at the right pricepoint, below $100 or so. But I wouldn’t think of it as a netbook and wouldn’t use it for the same things.
I’m not sure that’s true for the general populace, though. It may in fact be that what they want is compute applicances with a powerful browser and some simple canned, sandboxed applications rather than personal computers. Apple is having remarkable success along these lines with the iPhone/iPod-Touch application store, and Android is targeted at similar lines… it’s no longer obvious to me that that sort of device, moving into a tablet or netbook format, won’t take a big chunk out of the comsumer PC market. Most people may be willing to trade away that flexibility for simplicity.
Eric,
Well, several cellphone manufacturers targeting the Chinese market are replacing their Windows Mobile kit with Android-based kit…
And what is a clear functional profile of netbooks? The original marketing profile for them was not much greater than browser tablets: space- and compute-limited devices that run all your apps in The Cloud(tm). Even Mary Lou Jepsen came out and said that the personal computer would shrivel into a screen with an always-on wireless connection.
That’s not what customers saw, however. They saw “tiny cheap laptops that can fit in a purse or satchel”. In other words, PCs, with a bit of CPU zorch traded for low price and effortless portability. That’s how Microsoft pwnz0red the netbook space in about a year and a half: most netbooks are Windows devices, same as the rest of the PC market. So if you accept the “browser tablet” profile of netbooks, then Android becomes a viable solution indeed (that is, until Apple releases their much-rumored iTablet). If you accept the “tiny PC” profile, no version of Linux stands a chance of being much more than a geeky niche platform.
> Most people may be willing to trade away that flexibility for simplicity.
For most people the flexibility you are reffering to simply doesn’t exist. You have to be a user who uses that flexibility. And this is what? Can you name an example for that flexibility that a member of the most people group would care about?
On the other hand: simplicity. Do you mean I can have an internets that runs with lesser problems and my boss let’s me use it too?
@fiona — Ubuntu works fine wired because you don’t need a driver for the router only the NIC. Try ndiswrapper. It lets you use Windows NDIS NIC drivers in userspace.
Lawl! The x86 as high-end CPU. Even ten years ago it was still considered consumer-grade trash.
“RISC architecture is gonna change everything.”
“Yeah, RISC is good.”
I for one welcome our new low-power load-store overlords. But need I remind you that the NT kernel once ran happily on Alpha and PowerPC ISAs? If you think that Microsoft won’t rip out and rewrite the ASM parts of Windows, and hit F5 on their cross-compiler for the rest, you’re profoundly short-sighted. And if the market is there for these things, the apps and drivers will follow Microsoft in short order.
(i’ve mentioned this before. apologies in advance for boring anyone with long memories.)
i used to be involved in the pointy-end of financial exotics software, and our stuff had sold heavily into many of the german banks and landesbanks (equivalent to the USA’s federal reserve (what you read about as “the fed” is actually only the central aggregating body of the actual fed)). thing to be aware of: these guys take risk-aversion to a near-insane degree, even by bank standards. thing to be aware of: the financial markets are so backward re technology it’s surreal. they are social animals, and BOFH would be regarded as a good employee, punishing dissenters rather than worrying about what they were trying to achieve.
i was STAGGERED to learn in ~2000 that several of our clients were swapping THE ENTIRE BANK to linux. we’re not talking a few protected special pockets of the frontoffice — we’re talking a 100% cut-over rollout over across areas as freakishly sensitive as the middle and back office. without —and i can’t emphasise this enough— without exception.
worked perfectly.
it’s quite something to walk into a half-acre trading floor and see every screen showing linux, the bustling crowd around them and at them not giving a flying f*ck because all their KEY APPS are just running fine so who cares. then walk around the endless warrens of middle and back office, and all the same.
even when you know that’s how mac died (by its market-creating graphics apps migrating to windows (trigger point: apple’s decision to screw over warnock (white and shaking at the conference where they blandly/smugly announced it, leading him to finally allow PS to be ported to windows), proving empirically that 95% of users have ~0 platform-loyalty but instead focus on end-functionality) it’s still quite something to see an overwhelming “no-brainer” market dominating platform/purchasing-decision be so utterly thrown over so swiftly with absolutely zero effect on or complaint by the users.
(they provided transparent access to key platform-specific apps (eg, solaris, windows) by server-level services such as wine. Excel, for example, is COMPULSORY for non-trivial spreadsheet use (don’t auto-scoff: excel is a declarative language with implicit/embedded database, supporting/facilitating procedural overrides where these are more appropriate/easier, with a wholly open API and hooks throughout the entire system for extension/modification of native behaviour) — the overwhelming bulk of modern financial markets’ activity is tied together with (and even driven by) Excel. and these guys used it on linux without noticing or being bothered.)
.
i walked back into one of them in 2007 on a short-term contract (did the 5 weeks in 5 days, then uprated their whole installation to near-web2.0-compliance ;) and had a… proud? slightly spun-out? moment to sit down at my borrowed PC in this ultra-riskaverse ultra-organised megabank and… log into linux.
and everything… just worked.
“Sure, they could make an ARM port, but running all your favorite Windows applications would be no simple task.”
A netbook is a netbook. It’s based on the idea that a significant part of stuff you want to do you can do in a browser alone – Google Docs, webmail, Meebo.com for messenger, etc.
Given the limited hardware and the nature of such applications, the bottleneck is JavaScript speed.
Thus, I think you may have the right answer, but I’m not sure the question itself is right. The right question: which solution will provide the fastest JavaScript on ARM: Windows+IE, Windows+Firefox, Linux+Firefox or Linux+another browser? I would bet against the first option, but I’m undecided about the rest.
“and everything… just worked.”
I’m glad, but the main risk is – how can they be ever sure that if they must upgrade to the next version of MS Office / Excel, it will too work with wine? Of course, one may ask why “must” they upgrade if the current version does the job? My answer would be that if other companies keep mailing them files made in a later version, they have to, largely for prestige reasons, they cannot afford to ask for a “save as”, to appear to use anything but the latest version of the usual, popular apps would seem like admitting to have financial problems, of being low on money, being cheap, not a good PR.
The major obstacle of the spread of Open Source in the corporate world is that it’s unacceptable to have to tell to client: please can you send it in a different format? It just sounds cheap.
> But need I remind you that the NT kernel once ran happily on Alpha and PowerPC ISAs? If you
> think that Microsoft won’t rip out and rewrite the ASM parts of Windows, and hit F5 on their cross-
> compiler for the rest, you’re profoundly short-sighted. And if the market is there for these
> things, the apps and drivers will follow Microsoft in short order.
Yes but if you take Vista (or even windows 7 which is afterall a spit polished vista) and recompile it on ARM will it have adequate performance on a 1Gig ARM processor? I know for a fact that linux will run like greased lightning on a 1 gig processor(granted thats 1 gig intel and intel performance != ARM performance) for pretty much anything a casual user would desire. I’d argue as Linus once did… “Linux wins heavily on points of being available now”
I’m not saying it can’t be done but i think it will take more than just crowbaring some arch specific code in and slapping a sticker on it.
> if they must upgrade to the next version of MS Office / Excel
Actually the takeup of the new office has been so close to non-existant that i’m no longer sure thats an issue.
As someone who actually likes the new office (The ribbon actually made sense to me as as opposed to scrubbing damn near half the menus trying to find pretty much any individual feature) I ended up translating the few new style documents we got (our biggest client happened to have made the transition).
Umm, folks, in the real world netbooks aren’t being used as anything other than mini-laptops. Key apps are the browser, iTunes and Office. Browser-based apps simply aren’t a player in that space and most users are going Windows because they want a real copy of office instead of a clone like OpenOffice and they want their iPod to just work and be able to access ITMS.
As someone who’se using a stripped-down Vista install on a Netbook type system (Fujitsu U810 minitablet), Vista runs just fine on the 800MHz A110 processor with 1GB ram once Windows Search, ReadyBoost and Themes are turned off. It’s not wickedly fast, but neither is a GNOME-based system on similar hardware.
The application that will keep Microsoft in business is Excel. As Saltation says, there’s nothing out there that even comes remotely close to it.
(Hey, Saltation, you at the point where you can grab Minimus? Oh, and I have an Excel workbook that’s for conniseurs of “Holy fsck! I didn’t know Excel could *do* that!”)
ON this thread folks may be interested in Charlie Stross’ LOGIN 2009 speech – http://www.antipope.org/charlie/blog-static/2009/05/login_2009_keynote_gaming_in_t.html
One of the things he points out (and which I already knew) is that in terms of numbers the ARM architecture has already beaten x86 by at least one order of magnitude. I believe that ARM’s partners ships about 3 billion ARMed products per year these days. I’m not sure what the x86 total is but I suspect it is in the 100M range.
Now of course a LOT of that is cellphones, diskdrives and soho routers and very little of it is real PCs but the high end ARM designs in cellphones aren’t notably less powerful than Intel Atoms in Netbooks (and recall that the Kindle and other ebook readers use ARM). They do however cost a lot less. Linux plus ARM (including GPU bits) costs a manufacturer maybe $20 total and probably more like $10. Intel integrated Atom plus intel graphics hardware costs $25-$45 depending on quantity etc. And then there’s the microsoft tax on top. Someone like ASUStek can probably sell an ARMed netbook with linux more than $50 less than they would have to charge for a Atom/Windows combo.
PS FWIW I have run GIMP on my Asus 701 4G. It wasn’t blazingly fast but it wasn’t a total disaster either.
We can end the performance discussions right here and now with this fact: The iPhone runs Mac OS X on an ARM processor. Darwin kernel, Cocoa, Quartz, the works. The things that will kill you on Windows performance are graphics (not a problem if you have an embedded 3D chip) and useless horseshit like ReadyBoost (Microsoft can easily disable that). Address those two and Windows on ARM will become a preferred platform to Linux on anything overnight.
If you add a Rosetta-style x86 emulation layer to the mix, then ARM Windows netbooks will be able to run native Win32 versions of Office, etc. .NET apps will run from the word go. It’s practically a no-brainer.
Now this is just absolute arrant nonsense. Power management and wireless, alongside 3D accelerated graphics, have been Linux’s weakest support areas. Meanwhile, they Just Work under Windows.
>Now this is just absolute arrant nonsense. Power management and wireless, alongside 3D accelerated graphics, have been Linux’s weakest support areas. Meanwhile, they Just Work under Windows.
You’re behind the curve. Linux power management got cleaned up early in 2.6 cycle; now that the Broadcom chips have been reverse-engineered, this is actually a pretty reasonable thing for Lim to be saying.
>If you add a Rosetta-style x86 emulation layer to the mix, then ARM Windows netbooks will be able to run native Win32 versions of Office, etc. .NET apps will run from the word go.
I can’t imagine that that would be very fast, though.
Jeff, you need to try using a Nokia N810 Internet Tablet. Welcome to the 00’s!
> Linux power management got cleaned up early in 2.6 cycle
Suspend-to-RAM is still a mess by any absolute standard. Most problems come from buggy hardware, though; it’s just as bad in Windows.
http://cio.co.nz/cio.nsf/news/A185E9909EDD1D95CC2575BD006D8140
The book
Currently ranked #32,895 on Amazon, despite being available as a PDF for free.
By the way, doesn’t Windows Mobile support ARM procesor?
Yeah, I am sure a bunch of freetards got it to ‘pleasure themselves’ with. I mean, who can resist such insightful points such as these
This commentary is also insightful.
Those ‘inisghts’ were from just five pages BTW.
“Power management and wireless, alongside 3D accelerated graphics, have been Linux’s weakest support areas.”
I have an absolutely noname laptop (crufted together from a Mitac 8080C base and even more noname other parts) yet wireless worked out of the box (it required hunting for drivers under XP) on Ubuntu. Power management could use some improvement – reporting that the battery is under 45% charged when in fact it’s completely dead long ago isn’t quite accurate.
I’ve always saw the biggest weakness of Linux it’s diversity of distros: just what hardware vendor wants to test their drivers on many distros, and release multiple versions of it when the whole user base is small to begin with? Linux needed and “official”, “default” distro. Now Ubuntu became just that. Now you can just test your stuff under Ubuntu, release a package for it, plus the source, and just leave it for volunteers to make it work under all other distros, because if you covered Ubuntu, then you’ve covered the majority of the user base and you don’t really have to worry much about the rest. If you want to be really user-friendly, with support and printed manuals and no manual editing of config files and everything, you can’t develop for such a vaguely-defined platform as “GNU/Linux”. But you can develop for Ubuntu.
>Power management and wireless, alongside 3D accelerated graphics, have been Linux’s weakest support areas. Meanwhile, they Just Work under Windows.
Tell that to my Vista laptop. It’s faster for me to power-cycle it than to wait for it to wake up from sleep-mode.
Power management in windows: It just works, wrong.
Every single windows computer I have does something different, despite me setting them to supposedly do the same thing. None of them as bad as the old Presarios my office had which had to be hard-power-cycled (pull plug!) to get out of sleep mode, but that may be because I don’t allow them to “hibernate”.
Hi, Eric !
I’m sorry to have to tell you this. When I started using Linux, back in October, I had high hopes, great expectations, and an open mind. I think I’ve given this OS a fair trial.
(For the rest of you, I’m running Kubuntu Linux 8.04.1, kernel 2.6.24-14, KDE 3.5.10., on a Dell Dimension 8200 with 500 MB of memory and a 2.5 GHz Intel Pentium 4. Upgraded from a box running Windows 98 SE )
But I’ve come to the conclusion that Linux is:
annoying
frustrating
aggrevating
time-wasting
ugly
short on necessary utilities
very short on necessary documentation
difficult to get answers to simple questions about*
unreliable
In short, simply not ready for prime time
I would say that it’s a great ‘toy’ for hackers interested in messing with OS and kernel code, but I’ve little hope that the ‘bazaar’ will devote itself to ‘the user interface’.
As I’ve said elsewhere, testing Linux has been like being at the ‘speed shop’ back in ’62, and you’ve sang
“It’s ported and relieved and it’s stroked and bored
‘ll do a hundred and eighty in the top end floored.
It’s my Linux OS
Y’a don’t know what I got”
And I have to respond “Great chassis, sharp-tuned engine, but the paint is four shades of gray primer, the upholstery consists mainly of duct tape, and the steering is great for the track but a bear to get around a parking lot
(There have been suggestions that in the event of a New Deal 2.0, there’d be a WPA Federal Programmers’ Project, devoted to making OS operating systems usable. Mayhaps the economy will improve before then. )
Anyway, I’m looking for a legit copy of Win XP, and planning to get a new hard disk to put it on. Not that I’d want to erase Linux entirely – mayhaps GIMP works better in its native OS, etc… But I’m at the point of being sickntired even of looking at that desktop screen, and will be very glad not to have to deal with Linux daily, any longer.
Yours, John Desmond
* To paraphrase the old joke about Harvard: “They’ll give you an education if you can prove you don’t need one”
Shenpen/Jon B: >> if they must upgrade to the next version of MS Office / Excel
>Actually the takeup of the new office has been so close to non-existant that i’m no longer sure thats an issue.
quite.
a/ the bank is more than big enough (in every sense) to simply say “please re-send in a compatible format”.
b/ the new formats are surprisingly rare in practice (i speak from someone very much downstream on a daily basis)
c/ word has not yet progressed materially past RTF. and excel has not materially progressed past Version 5. hell, Version 2 has 99.9% of what is currently regarded as “cutting edge” functionality in excel. basically, if you don’t need Conditional Formatting and don’t need connected worksheets to travel as a unit and don’t need to create XY graphs in expected format, excel hasn’t moved in 15years.
Adam Maas, Jeff Read, et al: >As someone who’se using a stripped-down Vista install on a Netbook type system (Fujitsu U810 minitablet), Vista runs just fine on the 800MHz A110 processor with 1GB ram once Windows Search, ReadyBoost and Themes are turned off….. We can end the performance discussions right here and now with this fact: The iPhone runs Mac OS X on an ARM processor.
hell, (full) macosx runs like shit off a shovel on a 233MHz G3 on 0.75GB RAM.
tiger runs faster on an old imac than it does on a “modern” G5. now, i know that G3s are faster at integer procesing than G5s. but still, the order-of-magnitude improvement of “snappiness” side-by-side with a “modern” machine quite takes you aback.
Ken Burnside: >(Hey, Saltation, you at the point where you can grab Minimus?/i>
minimus==your ubernifty gamerules? no >:\ as per belated post on that thread, i STILL haven’t been paid. 3months late and counting…
> Oh, and I have an Excel workbook that’s for conniseurs of “Holy fsck! I didn’t know Excel could *do* that!â€)
now that i’d like to see. i have bent excel in all sorts of weird directions, to the point of intake-of-breath moments of the sheer GENIUS of the original writers as my-next-logical-step is transparently and lunaticaly extendibly supported by apparently trivial design decisions of them at very low levels, yet clearly made with full appreciation of the full user-world extrapolations and usefulness/usability. “little” things like ranges being held as strings til the very last possible moment.
my main a/c is full&jammed until i can pay for upgrade, but email me the spreadsheet at saltation_spamtrap AT fastmail dot fm (albeit eudora is still failing to register it). i’ll bing you back a rather large financial modelling workbook that uses the truly awesome capabilities of excel’s Names + Implicit Matrix capabilities to create something that has dedicated excel financial modellers go “whoa…”
> Yeah, I am sure a bunch of freetards got it to ‘pleasure themselves’ with. I mean, who can resist such insightful points such as these
I see you missed my point.
(you don’t need to know what you’re talking about to be a successful author)
Microsoft introduced a compatability pack two years ago with Office 2007 that lets older versions read the newer file formats.
They work like a charm.
And it’s amusing that Shenpen’s list of “If you don’t use X, Y and Z…” features, Excel has been frozen for 15 years lists two of the three features I use most often in Excel…and that most of the people I deal with for Excel work pay me to set up for them.
Excel is proof that even Microsoft can sometimes get out of the way of brilliant coding.
Nothing in the Open Source world comes close to Excel. I now know more about how the Open Office Calc team has inherited a mess of code with competing and differing low level programming decisions than I ever wanted to know. (A vocal subset of my users demanded an OO port of the sheet I’m sending to Saltation. I’ll note that once we had that port done…the feedback seems rather like chirping crickets. Of course, that port merely highlights the horrible responsiveness penalty that OO Calc inflicts on you…)
Descriptions of Linux being sufficiently advanced enough to work for users not inclined to spend time putzing around with uncooperative machinery appear at regular intervals. I didn’t believe them last time, and I don’t believe them this time. I’ll believe them when the default jukebox app on my new Ubuntu install doesn’t skip uncontrollably when playing MP3s. (I think PulseAudio or gstreamer don’t interact well with Intel 82801EB/ER-type sound hardware, and playing them on the command line via mplayer-to-ESD-to-Pulse works, but that’s not really the point–I shouldn’t have to care about this.) I’ll believe them when using my wireless doesn’t randomly hard-freeze my laptop. I’ll believe them when a plurality of GTK-based apps don’t shit deprecation messages to the command line. I’ll believe them when it’s no longer even money I’ll be filing one or more significant bugs against any package I haven’t used before, because there’s something seriously wrong with it.
I file bugs for these things. I test and patch where I’m not completely in over my head. I even try to answer newbies and help people out when they get stuck. (The preceding are, I believe, prerequisites to publicly complaining.) But I can’t in good conscience tell someone who doesn’t have my willingness to constantly deal with these things that they should use Linux, and that’s why I don’t recommend it to people.
Server packages, on the other hand, are much better; I don’t have many complaints, and I’d recommend it to any sysadmin… but, then again, it’s a sysadmin’s job to deal with the machinery.
Indeed. One of the problems with Linux, apart from the required willingness to tinker with things, is the lack of a really good tutorial to point users to in order to get them up to speed on how things work, to educate them to the point where they can usefully educate themselves.
Adam Maas,
“Umm, folks, in the real world netbooks aren’t being used as anything other than mini-laptops. ”
I’m not so sure. It’d be nice to see some reliable stats, or we all are just puffing hot air here, but I think the unergonomic nature (tiny keyboard, small screen) of netbooks make them inadequate for serious work or using them a lot. In my mind they are just some second-rate devices, that easy to carry around when you just want to pass the time with reading some stuff, answering mails, preparing a post (Google Docs) or chatting a bit (meebo.com) and you don’t really want to serious stuff on them.
But it might just be me – I’m VERY keen on ergonomy and I’m always surprised to find most people aren’t and many are happy to use a laptop on a dining table for 4-5 hours browsing them web. I mean, we live in our bodies, we use computers in most of our waking hours, what makes more sense than making the later fit to the former as closely as possible? Still, in my last job, out of 45 people I was the only whan who requisitioned an ergonomic keyboard and a trackball. (Do you still use a mouse? Do you want to get RSI or what? For a geek RSI is like broken legs for a soccer player.) I just don’t understand this, really…
If I want to read and a review a long Word document while riding a train, I can load it into Google Docs. If I actually want to write one that surely won’t be on a netbook… that will be on a desktop computer, with an ergonomic keyboard, trackball, at least 20” screen and don’t even get me started on adjustable-height desks – it should be a _must_ have in ergonomy and I never had them in any of my jobs…
>desktop computer, with an ergonomic keyboard, trackball, at least 20†screen and don’t even get me started on adjustable-height desks – it should be a _must_ have in ergonomy and I never had them in any of my jobs…
Good idea, but not nearly as easy as a netbook.
I will probably make it my New Year’s Resolution to switch to a trackball. I’ve always been driven nuts by them, and never stiffed it through the adjustment period.
As to using netbooks – I use one of them regularly. I also plug my ergonomic keyboard into it and bring a mouse. :)
Saltation – that email address rejected the file. Said you were over your limit there as well.
>I will probably make it my New Year’s Resolution to switch to a trackball.
Recommendation: The Logitech TrackMan. I’ve been a happy customer for over a decade.
Seconded. Logitech has some beautiful, beautiful balls.
I find that with a trackball for gaming and general pointing and a tablet for art, I’m set. And these two devices perform their respective tasks better than a mouse could. The mouse is an example of generality FAIL.
Further proof: DirectX.
OpenGL was amazing, and wholly appropriate back when the 3D graphics state of the art was an SGI Indigo. Radical, paradigm-busting new 3D-accelerator technologies have emerged since then; and Microsoft has kept pace whereas the ARB hasn’t.
Qualcomm releases an ARMed netbook – the smartbook – http://gigaom.com/2009/05/28/qualcomm-turns-a-netbook-into-a-smartbook/
Awesome, FrancisT.
Myself, I’d like to get my hands on one of those Touchbooks, provided they prove to be something besides vaporware. But you guys are leaving out a huge potential player: Apple.
Apple has acquired a chip fab recently. And those of you who are up on your microcomputer history understand that part of the reason why the C64 and Amiga were so revolutionary was because of Commodore’s ownership of MOS Technology, enabling Commodore to make arbitrary design decisions in the machines’ chipset and crank out the parts at high volume, without having to worry about procuring them in volume beforehand.
Apple is getting set to change the game substantively even more so than they have with the iPhone. Dare we dream of the return of PPC-based Apple devices to store shelves?
Ken Burnside: >Saltation – that email address rejected the file. Said you were over your limit there as well.
no idea what’s going on there. i can web-login fine on that a/c, but eudora can’t connect to it. and it has a grand total of 1 msg in it, according to the weblogin: “welcome to! etc”
>:\
just another thing that’s gonna hafta wait on me being paid, i guess. as of tuesday, i have >£60,000 of my share of due invoices, mostly 3+mths late. non-creditcrunch-affected companies: just HR/admin fucking around. and a blown overdraft. and rent due in 2 weeks. to someone who can’t afford to spot me a month…
Jeff Read: >> Excel is proof that even Microsoft can sometimes get out of the way of brilliant coding.
>Further proof: DirectX.
quite the opposite, if Jon Carmack is any guide. he had an explosively vehement response to DirectX as a coder (along the lines of “if i gave fuller examples you would accuse me of choosing deliberately pathological cases”), and his consequent decision to refuse to support directx (in favour of opengl) i regard as essentially canonical.
don’t mistake the potential capability of creating a result with actually facilitating a result.
Saltation, what you say was true ten years ago.
These days, everyone I know of who is involved in high-end graphics, to a man — including Carmack — thinks the Microsoft APIs and dev environment are superior.
Wikipedia has it thus:
Jeff Read: ta. you’re right: my info is old. looking at what you cite here –in particular Carmack’s later response– i accede completely. thanks for the update!
Worth noting that Carmack still uses OpenGL for id Tech. Primarily for cross-platform support. Were it not for cross-platform issues (primarily due to the Macintosh platform adopting OpenGL), OpenGL would’ve gone the way of PHIGS some time ago.
Ken Burnside: >(Hey, Saltation, you at the point where you can grab Minimus?)
>minimus==your ubernifty gamerules? no >:\ as per belated post on that thread, i STILL haven’t been paid. 3months late and counting…
Ken, i’ve finally had one payment come through. the relief is… difficult to describe. this weekend i plan to dedicate to buying things like clothes without holes in them, then throwing myself into the local festival, and then next week starting to get my fiscal house in order. not monday, obviously. monday will be dedicated to holding my head VERY CAREFULLY and moaning softly. but once i start on the list, you’re high on the list.
Eric: you still ok to forward my pay(pal)ment?
>Eric: you still ok to forward my pay(pal)ment?
Yes.
ta
i plan to get onto this ~tuesday
eric: e~@i@.org, yeah? if other, email me preferred destination at saltatoin_signups@nospammail.net, with the obvious typo corrected
will pick up email there tomorrow night — will make the payment (ONLY with a clear head!) either tomorrow or tuesday, depending on how the celebrations go
^^e~@i~.org — and accidentally but profoundly a good indication why i’m not doing it tonight
*cough*
that should of course be thyrsus, com, and esr. or email me your preferred destination @nospammail.net : “saltation_signups”. ready to roll.
Ken Burnside: done
My problem with Linux (and things ‘nix in general) is that it has “Users are Losers” embedded deep into its DNA. I loathe all things ‘nix for that in much the same way the Eric loathes things proprietary. Whenever I read a post about how well various “Aunt Tillies” manage with Linux, what *I* hear is a brag about how *pleased* the slaves are with their shiny nifty new shackles. That I might in principle avoid userdom-enslavement by becoming a programmer isn’t an answer. It would take a great effort for me to become merely a third-rate hacker, and it wouldn’t be worth the results. I’d still be locked just as deeply, if more subtly, into the system and it’s “Programmers are the Masters; Users are Slaves” philosophy.
No thank you.
Deep Lurker, a negative proof by counterexample: there exists indeed a Unix distro which puts the user first.
It’s called Mac OS X.
It’s pwning Linux in the desktop space. Hard.
Jeff Read: as someone who converted from unix+vms commandline to mac in 30 minutes in 1988(iirc), i would dearly love to agree with you.
sadly, macosx is as user-dismissive as any other unix. and from a user perspective it’s a VIOLENT step backwards from macos (drag&drop OS’s anyone?)
works as one piece out of the box, but that’s just a distro-coherence issue, not a target-audience telltale.
As GNU/Linux evolves into a platform which common users could use, I’d also like to see the hardcore users to embrace the idea that more and more noobs are going to flock their IRC channel for support.
With what I currently observed, the FOSS community in general are still the same easily-annoyed individuals, shoo-ing away this new generations of users for being noobs.
This attitude has to change as the OS we love is.
The steroidal African bull elephant in the room with respect to OSes which are friendly to n00bs is security: Windoze machines and even Macs are being compromised and herded into botnets with world-shaking powers.
The only solution to this problem comes in the form of restrictions which many free software advocates consider onerous: Either restrictions on who may connect to the net or restrictions on who can run what. Personally, my money is on the latter. I think the future of end-user apps development is the iPhone and the App Store. Instead of PC’s people will increasingly browse and tweet with handheld devices, running only that software which had been preapproved and signed by a central authority. This makes security easy and convenient from an end user’s perspective and also acts as a quality filter.
I do not expect fosstards to relent in their attitudes towards closed software platforms. They will be left behind.
>The only solution to this [security] problem comes in the form of restrictions which many free software advocates consider onerous
No, the only solution to this problem is for closed-source operating systems to die.
Which is what will happen, eventually.
>No, the only solution to this problem is for closed-source operating systems to die.
>Which is what will happen, eventually.
But it won’t be unix that does it. A winning open-source OS will have to be free of the “users are losers” taint, and no version of unix will ever be able to qualify. “Users are losers” is bred too deeply into its DNA.
(The Mac isn’t a counterexample. Rather it’s an example of the “users are losers” virus being powerful enough to corrupt a formerly user-friendly OS.)
> I find that Linux is incredibly easy to install.
I have to disagree here. Linux is easy to install IF the kernel happens to support your hardware. Otherwise, it’s downright impossible. And hardware support isn’t such a trivial matter; just yesterday i had to install OpenSuse 11.1 because the kernel, yet older than that of the Kubuntu 9.04 CD i originally wanted, supported my SATA controller, while the newer one didn’t because of some Kernel bug.
Another thing that slows Linux desktop adoption is the lack of a method for EASY driver installation. I have an SMC Wifi Adapter, i have the necessary driver files, but i’m obviosuly too stupid to get the darn thing installed even though i refer to myself as a fairly competent Windows hacker.
Indeed. Open source will only “win” if it develops the same relentless user focus Apple demonstrated with the original Mac.
AROS has a shot. Linux, not a chance.
Guess what! The first ARM-based netbooks anyone will likely ever get their hands on will run… wait for it… Windows CE.
Who the hell would use Windows CE when you could run a fully capable desktop OS?
An OS which confuses end users is not fully capable: If technology doesn’t work for people, it doesn’t work.
Probably it’s right thread for it: it’s 2009 and I’m still banging my head into the same walls: hardware support, hardware support, hardware support. I mean: the lack of it. Same kind of frustration I had about 12 years ago when I first tried SUSE 6.3. I have a nice Creative Live!Voice webcam and it won’t work under Linux. (Latest Ubuntu.) The info on the UVC developer mailing list is: sorry, it’s not UVC compatible, have a nice day, bye. I have the options of buying something else or booting into XP. Guess which one will I choose?
When will this mess finally come to an end? I really don’t want much. I’m fine with inconveniences like compiling stuff from source and probably that’s more than 99% of the computer users of the world are willing to do. All I want is drivers for each and every product of major hardware brands. Is it really much to ask?…
Not sure from whom should I ask it, from the hardware vendors, from kernel hackers, from Mark Shuttleworth, from Santa or Yog-Sothoth but someone really, REALLY should make it happen if Linux is ever to take off.
There is much that’s acceptable, such as, compiling stuff from source, patching, learning new stuff, googling a lot and suchlike. But buying HW from a major brand and finding out there are no drivers for it is NOT.
Windows 7 works beautifully on Atom-based netbooks. Linux remaining a major player in this space just got a whole lot more difficult.
You forgot the pony.
Seriously, that’s why hardware vendors have been historically responsible for driver support. Of course, as long as Windows remains dominant, hardware vendors will only see business opportunities in supporting Windows systems, which will reinforce Windows dominance…
Submit Links to Social Bookmarking Websites Automatically! Get Thousands of Visitors! Increase Website Traffic and Income with Automatic Social Bookmarks Submitter. Prepare to upgrade your servers today – you will have thousands of real and highly targeted visitors who will make you the money like crazy!
Well, is 2010 already and nothing has changed. Get off your high horse, Linux is for hackers and FOSS enthusiasts. As a desktop OS, it’s a total failure. Fragmentation killed it. The lack of a viable business model killed it.