“8” Down and One to Go

I see a train wreck about to happen. Normally, that would be a bad thing but this train wreck will be a joyful event for me. M$’s “8” looks to be the biggest train wreck since Vista (18% at its peak, dead for a monopoly).

  • Here’s what a fan of M$ wrote recently:“It doesn’t make computers easier to navigate and understand, it makes them more difficult, paradoxically by making the interface so brain-dead simple that it can’t do anything someone with a brain might actually want. Want to close an application without using Alt-F4? Forget it. Want the menus and settings intelligently organized? No chance. Want to just display two windows on the screen at the same time? Good luck with that.”
  • SJVN says it’s DOA and another supporter says people should stay away from “8”‘s Metro UI for a couple of years as folks usually do for M$’s usual release:
    Windows 8 showdown: Face-off on whether Windows still matters

The silver lining in all this is that M$ is wasting its time as Android/Linux races along actually selling units and GNU/Linux is emerging on ARM in the consumer market.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology. Bookmark the permalink.

27 Responses to “8” Down and One to Go

  1. In a swap file the stuff you really need to read and write in a hurry is fluffy with all kinds of spaces and file-fragments interspersed. This increases access times a random amount but it could be bad as a drive fills with fragments.

    In a swap partition, an OS/file-system can optimize the data as it sees fit with no competition.

    When swap runs out, which I have seen a couple of times on this server…, Linux kills offending processes and usually gets it right if, as was the case, my system was spawning processes like mad that never finished. In that other OS, I suppose M$ might create a swap-file but that would be unwelcome and further degrade the swapping performance.

    I prefer a swap partition because you can do neat things with it like put it on RAID 0 or whatever to speed up performance. Swap-files are much more clumsy. It’s a small thing because swap should rarely if ever be used but M$ likes to use it a lot. In these days of huge RAM a client or server machine should never need to swap.

  2. aardvark says:

    Hmm. Interesting.

    Is there any functional difference between a swap file and a swap partition? The only one I can see is that the OS can extend the swap file where necessary, but is limited by the original definition of the swap partition.

    I’ve no idea what any OS does these days when it runs out of swap space, tbh. Which algorithm does any of the major three (TOO$, Mac, Linux) use to figure out what the offending processes are, and how to resume normal operations?

  3. Ted replied to his own demand for a technical explanation so I don’t have to give one. Thanks, Ted.

    M$’s OS does not deal with out of memory conditions. It slows down to be unusable while Linux kills the offending process and resumes normal operations.

  4. Ted says:

    “Well. That other OS just goes belly-up and crashes.”

    I’ve already quite clearly indicated that it doesn’t.

    Windows comes across a low memory condition – and JUST DEALS WITH IT.

    “A swapfile is a fragmented piece of trash compared to a swap partition.”

    I’d like to see your technical reasons for this supposition. But I expect I’ll be waiting a long time.

    Swap files only fragment (in terms of file contiguity) if they have to expand due to a low memory condition and the contiguous blocks of the disk are not available. A slight loss of performance in the page file vs. running out of memory, by far the the lesser of two evils.

    After you find your memory-leaking application and deal with it (or find out the correct settings for paging for your usage), you turn off paging, reboot, defrag, turn paging back on with appropriate settings, reboot. You’re then back to a contiguous page file.

    Or you could put the paging file on a separate partition or physical hard disk, or multiples thereof.

  5. Well. That other OS just goes belly-up and crashes. A swapfile is a fragmented piece of trash compared to a swap partition.

  6. Ted says:

    “If you try that with GNU/Linux it just keeps plugging away as long as it does not run out of swap and then Linux kills some processes and it carries on.”

    It does WHAT?? It arbitrarily kills processes?

    What gives the OS the right to kill any running processes – especially user processes – without the user having a say in it?

    What happens if Linux decides to take down something important, or something that would fix the problem or help stop the problem happening again? Apache? No-mark! Get rid! Logging? Can’t see the use of it – out it goes. SSH? Who needs that? Bye!

    If Windows did anything as frankly brain-dead as this, at least the OS can detect a service stopping unexpectedly and restart it, or run another program, or even schedule a reboot.

    And “run out of swap”? Here’s a choice quote; “Windows is increasing the size of your virtual memory paging file.” – Windows Server 2003 from you guessed it, 2003.

  7. I’ve had hundreds of freezes on that other OS and I cannot remember one in GNU/Linux. OK, occasionally I can get a system that doesn’t boot the way I want but I am using beta software these days, Debian Testing.

    GNU/Linux is a rugged system and I don’t need M$ or anyone like M$ to keep it running. GNU/Linux systems were running for years without failures back in the day when that other OS could barely manage a few hours. I was in one place I explored Lose ’98 because I had never used it before. I could crash it in a few minutes just opening a few windows in browser and word-processor. At one school I had a student who figured out how to crash XP. I spied on him to see how he did it… He selected every icon on the desktop and asked for them all to be “opened”. Strangely, XP would grind away for minutes and not crash but as soon as you closed any window it would freeze up… That was about 2001 when XP was new and the school division shipped us a bunch of machines. If you try that with GNU/Linux it just keeps plugging away as long as it does not run out of swap and then Linux kills some processes and it carries on.

  8. Ted says:

    “My computer locked up mid-mouse click solid. It turned out I had some bad sectors on the hard drive.”

    Are you saying that LINUX locks up? For a single hard disk error? On a disk that hasn’t otherwise failed? Really? So a single duff sector on a hard drive could potentially bring a Linux server to its knees and need the engineer’s finger to fix it?

    Quite a revelation, that. Thanks.

    I’ll tell you this – Windows does NOT behave that way with bad sectors. It’d time-out the operation after retrying, not hard lock-up. You’re more likely to get a BSOD if the file affected is a system file, but a hard lockup? Nope.

    “The computer wouldn’t even boot when I pressed the reset button.”

    Reset button? What decade was this? I’ve not even had a tower case _with a reset button_ for six years. It’s 10 seconds press-and-hold the power button on these new-fangled ones, BTW.

    “If I was using using Windows and NTFS that would have meant a complete reinstall. ”

    No it wouldn’t.

    Boot from Windows DVD or a bootable USB key with the Windows install media on it, Start Windows Setup. “Repair your computer”, SHIFT-F10, “chkdsk c: /r /x”. If the disk isn’t actually failed as in damaged of defective, you’re sorted. The bad sectors will be written into the FAT as unusable.

    Remarkable how that procedure is similar to the Linux one, isn’t it? It’s almost like it could happen to any operating system.

    My computer booted fine with no loss of data. Same thing happened a few months later with the same results.”

    ANOTHER HARD LOCKUP? ON LINUX? OVER A CORRECTABLE FILESYSTEM/IO ERROR? Oh dear!

    “You can keep NTFS, I’ll keep ext4 with the “maybe loss of data” and the proven saved-my-ass-no-loss-of-data.”

    I will keep NTFS, or maybe I’ll use exFAT. Depends on usage scenario. Haven’t lost data on exFAT, either.

    By the way, there’s NO “maybe” about it; EXT4 LOST DATA. FOR NO REASON. AND IT WAS RELEASED THAT WAY. I’ve already proved that, with a “Canonical” source. (I’m here all week, or until Mr. Pogson bans me again. Remember to tip your waitress.)

    And before you drag out my “random lockups” from earlier (Even I’m selectively quoting me now) please remember the context – I said this happened in progressive hard disk FAILURE, not a couple of bad sectors.

    “If you weren’t so in love with it you should be condemned to use it.”

    You’re a weak troll, Koz. Try harder.

  9. Kozmcrae says:

    Dear Ted:

    You can use NTFS and be happy with it. I use ext4 and I’m very happy with it for good reason.

    My computer locked up mid-mouse click solid. It turned out I had some bad sectors on the hard drive. The computer wouldn’t even boot when I pressed the reset button. If I was using using Windows and NTFS that would have meant a complete reinstall. Not so with GNU/Linux. I booted the computer with a Gparted disk and ran the disk repair utility (fsck). My computer booted fine with no loss of data. Same thing happened a few months later with the same results.

    You can keep NTFS, I’ll keep ext4 with the “maybe loss of data” and the proven saved-my-ass-no-loss-of-data.

    Simply put, NTFS is a piece of crap compared to ext3 and ext4. If you weren’t so in love with it you should be condemned to use it.

  10. Ted wrote, Did you bother to enquire?

    Nope. No time. No phone. No budget. I was a classroom teacher with no access to a telephone during class time, because I was teaching and no access to a telephone outside of class time because the office was locked and teachers did not have a key… I suspect Fat32 was chosen for reasons of familiarity. Did I want to keep 80 copies of installation CDs in my crowded cabinet? Nope. GNU/Linux is so much more convenient. The licence permits me to do everything I needed to do with no fussing over installation CDs. I used USB drives and/or PXE-loaded networked installers. No CDs needed.

  11. Ted says:

    “When I arrived there in 2009, all of the XP systems were installed with XP SP1.”

    Some very serious failings of your predecessor, then. XP’s install default was NTFS on day 1 (as it was with Windows 2000), and in Sept 2009, there was simply no excuse whatsoever not to be on SP3.

    “There were no “installation media”.”

    Most OEMs will provide a new disk for a few dollars/pounds/local currency. Did you bother to enquire?

  12. Ted wrote, “why weren’t they INSTALLED as NTFS? It’s the default option.”

    When I arrived there in 2009, all of the XP systems were installed with XP SP1. I migrated them to SP3 and NTFS.

    There were no “installation media”. I eventually fixed things by migrating to Debian GNU/Linux. End of problems. Of the dozens of machines running XP that were unbootable none had a hard drive failure. It was mostly scrambled file-systems. One or two had a failed PSU.

  13. Ted says:

    “No file-system can write data before the power is cut off if the power is cut off early enough.”

    Absolutely true. However, Ext4 was still losing data even without power outages. See the bug I linked to on Launchpad.

    “In 2010, I converted a mess of XP machines to NTFS”

    I have to ask, genuinely curious – why weren’t they INSTALLED as NTFS? It’s the default option.

    “If it’s not power messing up the file-system, it’s malware or the OS itself.”

    Malware – maybe then, but not these days. There’s no extortion money to be made or spam to be sent if they screw the system up.

    The OS itself – not unknown. For _ANY_ OS. But rare for Windows or Mac OSX to be beyond a few hours work to be back up and running unless there is hardware failure.

    You leave out a significant cause of Windows being irretrievably screwed;

    Hard drive failure.

    They’re mechanical, and the part in any computer most likely to fail.

    It’s pitting anecdote to anecdote, but in more then 95% of cases where I’ve had Windows become both unbootable and unrepairable (a distinction you do not make), the click of death is not far away.

    Windows even plays up in predictable ways beforehand, if the failure is progressive. Slow boot, sluggish performance, slow file accesses, random lockups, etc. I get those, the first thing I run is a hard disk diagnostic.

  14. No file-system can write data before the power is cut off if the power is cut off early enough. The journalling file-systems may not even be able to write their journal in which cases changes are lost but the integrity of the existing files are more sure. In 2010, I converted a mess of XP machines to NTFS and things improved but machines still became unbootable on their own/under normal use. If it’s not power messing up the file-system, it’s malware or the OS itself.

  15. Ted says:

    “Interesting link to wikipedia.”

    I thought so too. It was frankly amazing to me that some distros used Ext4 as a default file-system with such show-stopping bugs.

    Or were you perhaps trying to be sarcastic?

    Is this a case of double standards again? Wikipedia is written (sorry, vandalized) by trolls when Linux gets criticized, but gospel truth when Linux gets a good write-up?

    “As far as trolls go, you’re fairly good.”

    Troll? As in someone who dares disagree with the Linux group-think?

    “I was reading the paragraph you linked to. The source?

    “^ Thoughts by Ted blog entry, March 12th, 2009 A blog posting of Theodore Ts’o on the subject””

    You have no idea who Theodore Ts’o is, do you?

    http://en.wikipedia.org/wiki/Theodore_Ts%27o

    After you’ve read that and extracted your pedal extremity from where your tongue should be, read on…

    “Sure, ext4 did lose data if you would poweroff immediately after writing.”

    And NTFS _didn’t_, unless you were enough of a dolt to turn on the additional write-caching for fixed disks without some form of UPS or battery-backed cache and then cut the power mid-write, or to yank out a removable drive that has write-caching enabled without ejecting it properly.

    And NTFS doesn’t suffer from anything like this;

    https://bugs.launchpad.net/ubuntu/+source/linux/+bug/348836

    (By the way, “Ted” is a coincidence. Me != Mr Ts’o)

  16. Dann says:

    @Ted: Interesting link to wikipedia.

    As far as trolls go, you’re fairly good.
    I was reading the paragraph you linked to. The source?

    “^ Thoughts by Ted blog entry, March 12th, 2009 A blog posting of Theodore Ts’o on the subject”

    Ted’s blog, huh. Sure sounds academic to me.
    Sure, ext4 did lose data if you would poweroff immediately after writing. This problem was, what, up until the end of 2008 until distros started cutting down on the async writes. I’ve been using ext4 for years without problems, even now on my ssd (without journaling enabled) and I have not lost any data.

    While I’m here, I’d also like to point out that Windows 7 cannot install on my 3 TB disk, but any linux-based distro can. I had to fish out a slower 80GB drive to install Windows 7. I have since removed it though.

  17. kozmcrae says:

    Phenom said:

    “Pogson, I’d welcome any example of SJVN being right about any prediction about Windows.”

    SJVN is pulling your chain and laughing all the way to the bank. I read his stuff and most of the pathetic comments in reply to his articles. He’s taking you all for a ride and laughing all the way. I doubt that he really even cares that much about Linux and open source. He uses proprietary stuff as much as open source stuff.

    He gets the Cult of Microsoft all riled up. Their comments are predictable and he plays them like a flute. Taking the job at ZD was a smart move for SJVN. He’s making fools out of the entrenched Microsoft fans and money too. Way to go SJVN.

  18. Phenom says:

    Pogson, I’d welcome any example of SJVN being right about any prediction about Windows. As for Linux, probably his only article partially on the spot was this one:
    http://blogs.computerworld.com/17183/is_the_linux_desktop_dream_dead

  19. Clearly, Phenom does not read enough of SJVN’s articles. SJVN has admitted his mistakes/errors but I believe he is right far more often than he is wrong. His grammar and spelling suggests he should have a secretary, though.

  20. Phenom says:

    SJVN has been dead wrong on basically all his predictions about operating systems.

    Sweet memories, I recall how he predicted OS/2 2.1 would succeed, how it would blow Windows 3.1 out of the water. Well, we know how this story ended.

    SJVN has no technical skills, no understanding and knowledge about the market and business needs, and no honor to admit when proven wrong.

  21. SJVN has been around a long time. He’s seen many things come and go in IT and he has lots of contacts with people in IT globally. I respect his judgement even when I disagree with him. I expect “8” will ship some units and then flop worse than Vista. For several reasons, I don’t see businesses slurping up stuff with “8”. Consumers might if that’s all they find on retail shelves but I think that ends this year. There are too many other good products being produced to keep the monopoly on retail shelves.

  22. Ted says:

    “What ever happened to the new, improved file system? Did you get it in 7?”

    No. But at least Windows has a default file system that works. NTFS may have its detractors, but it doesn’t habitually lose data.

    http://en.wikipedia.org/wiki/Ext4#Delayed_allocation_and_potential_data_loss

    “SJVN says it’s DOA”

    Vaughan-Nichols says a lot of things. He’s pretty much predicted the death of Microsoft with each new OS release. Most of his predictions to do with Microsoft products don’t come true, and I don’t see him backing the right horse here, either.

  23. Ray says:

    Like I said, people are already speculating on the desktop before it’s released ;)

  24. Kozmcrae says:

    “It introduced a new driver model,…”

    What ever happened to the new, improved file system? Did you get it in 7? Are you going to get it in 8? Looks like you’re getting it in the end.

  25. dougman says:

    “It doesn’t make computers easier to navigate and understand, it makes them more difficult, paradoxically by making the interface so brain-dead simple that it can’t do anything someone with a brain might actually want. Want to close an application without using Alt-F4? Forget it. Want the menus and settings intelligently organized? No chance. Want to just display two windows on the screen at the same time? Good luck with that.”

    LOL, well this whiner could fork the entire mess and build his own. Oh, wait… Errr…. nevermind.

  26. Phenom says:

    I still don’t see why Vista was deemed as a failure. In fact, it got up to 30% marketshare (over total OS distribution, not only among Windows setups) despite all the negative turmoil.

    It introduced a new driver model, a new desktop composition subsystem, several improvements in the API, made hardware vendors stabilize their 64-bit drivers. For such major endeavours, a 30% penetration is hardly something you will call “failure”.

    Gosh, only the search box in the start menu was enough to make me never want to look back at XP.

  27. Clarence Moon says:

    The most telling argument was the one that came out first in this clip, Mr. Pogson, namely that Windows 8 is significant simply because it is Windows (stress added). That is what people have been looking for ever since Windows 3.0, a better and better Windows that will improve their experience with their PC.

    Argue all you want about some alleged failures to deliver those hopes. Windows Me and Vista are favorites of the Microsoft critics. Maybe Windows 8 will be the same, maybe not. In any case, the lesson of history is that people who are so dismayed at the lack of promise of the new Windows simply stay with the old one. They do not switch to Linux.

    History will doubtless repeat here. Either Windows 8 will catch on wildly, as has Windows 7, or people will just stick with 7. An under-delivery on Windows 8 is not going to start a revolution. Neither Millennium Edition nor Vista did that, and Windows 8 will not do that either.

    Personally, I think it will do just fine.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>