Skip to content
  • Categories
  • Recent
  • Tags
  • Popular
  • World
  • Users
  • Groups
Skins
  • Light
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • Dark
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • Default (No Skin)
  • No Skin
Collapse
Code Project
  1. Home
  2. The Lounge
  3. Execution Drift

Execution Drift

Scheduled Pinned Locked Moved The Lounge
helpquestion
22 Posts 15 Posters 0 Views 1 Watching
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • N NowIsForever

    Call this "execution drift" for want of a better term. Everyone who uses a PC knows about this problem: A computer must be rebooted periodically or "things go wrong." Some software in the system (frequently many programs) will begin behaving erratically, or not function at all. Mainly, I'm looking for more information on the phenomena, and in particular, is there a standard name for this? I tried several topics with Google, pc drift, program drift, execution drift, etc. and nothing valuable came up. Presumably, the culprits are mostly applications and add on code rather than system code, drift in system code is perhaps not uncommon.

    :doh: Eschew Obfuscation :doh:

    J Offline
    J Offline
    Jorgen Andersson
    wrote on last edited by
    #8

    To me this is something that happens to others, not everyone. My work computers and servers have the programs they need, nothing more. And they don't need restarts except for windows update once a month or if there's a new version of a necessary program. If I need to test a program I do it in a virtual pc. BTW, a better name is "Attention Deficit Ooh Shiny".

    My postings are a natural product. The slight variations in spelling and grammar enhance their individual character and beauty and are in no way to be considered flaws or defects.

    1 Reply Last reply
    0
    • D DaveAuld

      I don't think you need any fancy term. Its just bad programming. Either at the OS or Application Level or firmware. If everything released resources correctly, cleaned up after itself correctly when being uninstalled, dumped temporary files correctly etc. etc. etc. then things should work for ever. Look at mission critical systems where money is no object and every level of detail is looked into. They don't get powered off and on all the time do they?

      Dave Don't forget to rate messages!
      Find Me On: Web|Facebook|Twitter|LinkedIn
      Waving? dave.m.auld[at]googlewave.com

      R Offline
      R Offline
      Rob Graham
      wrote on last edited by
      #9

      daveauld wrote:

      Its just bad programming. Either at the OS or Application Level or firmware.

      I don't think that's the case. memory leaks are indeed bad programming, but they don't explain the issue daveauld complains of. More likely it's heap fragmentation, which will eventually kill any OS that implements dynamic memory allocation to an arbitrary collections of applications. Sooner or later, an APP that needs a large block of contiguous memory will fail to get it, then things begin to unwind. In your example of mission critical systems (particularly the 24/7 process control type applications) the memrory is is tightly controlled (often preallocated and dedicated), so leaks and fragmentation are by design not an issue.

      I 1 Reply Last reply
      0
      • R Rob Graham

        daveauld wrote:

        Its just bad programming. Either at the OS or Application Level or firmware.

        I don't think that's the case. memory leaks are indeed bad programming, but they don't explain the issue daveauld complains of. More likely it's heap fragmentation, which will eventually kill any OS that implements dynamic memory allocation to an arbitrary collections of applications. Sooner or later, an APP that needs a large block of contiguous memory will fail to get it, then things begin to unwind. In your example of mission critical systems (particularly the 24/7 process control type applications) the memrory is is tightly controlled (often preallocated and dedicated), so leaks and fragmentation are by design not an issue.

        I Offline
        I Offline
        ian dennis 0
        wrote on last edited by
        #10

        If an application has memory leaks which ties up most of the available memory, and then you attempt to launch another application which needs more memory than is free, it would explain some of the symptoms that daveauld is describing. I'm sure that MS.Word is riddled with memory leaks and have found that the occasional reboot is necessary to "clean" up the memory.

        1 Reply Last reply
        0
        • N NowIsForever

          Call this "execution drift" for want of a better term. Everyone who uses a PC knows about this problem: A computer must be rebooted periodically or "things go wrong." Some software in the system (frequently many programs) will begin behaving erratically, or not function at all. Mainly, I'm looking for more information on the phenomena, and in particular, is there a standard name for this? I tried several topics with Google, pc drift, program drift, execution drift, etc. and nothing valuable came up. Presumably, the culprits are mostly applications and add on code rather than system code, drift in system code is perhaps not uncommon.

          :doh: Eschew Obfuscation :doh:

          S Offline
          S Offline
          SimulationofSai
          wrote on last edited by
          #11

          I thought that this was because of memory fragmentation. This will be more evident if the OS relies on a time based paging algorithm to determine which blocks goes to the page file. Oh wait, while this explains why applications will be slow when reactivated after being left open for a long time, it does not explain why anything may appear erratic. :doh:

          1 Reply Last reply
          0
          • D Dan Neely

            Or even just fragmentation of the heap.

            3x12=36 2x12=24 1x12=12 0x12=18

            L Offline
            L Offline
            Lost User
            wrote on last edited by
            #12

            I know how it feels (on a Monday). :sigh:

            Join the cool kids - Come fold with us[^]

            D 1 Reply Last reply
            0
            • L Lost User

              I know how it feels (on a Monday). :sigh:

              Join the cool kids - Come fold with us[^]

              D Offline
              D Offline
              Dan Neely
              wrote on last edited by
              #13

              But it's Tuesday.

              3x12=36 2x12=24 1x12=12 0x12=18

              L 1 Reply Last reply
              0
              • N NowIsForever

                Call this "execution drift" for want of a better term. Everyone who uses a PC knows about this problem: A computer must be rebooted periodically or "things go wrong." Some software in the system (frequently many programs) will begin behaving erratically, or not function at all. Mainly, I'm looking for more information on the phenomena, and in particular, is there a standard name for this? I tried several topics with Google, pc drift, program drift, execution drift, etc. and nothing valuable came up. Presumably, the culprits are mostly applications and add on code rather than system code, drift in system code is perhaps not uncommon.

                :doh: Eschew Obfuscation :doh:

                L Offline
                L Offline
                Lost User
                wrote on last edited by
                #14

                You may find this[^] paper interesting. It talks about DRAM errors as they occur in the wild. These are errors from hardware rather than software and I suspect, after reading the paper, that they may come into play a lot more than people realize. Cheers, Drew.

                1 Reply Last reply
                0
                • P PIEBALDconsult

                  Sooo... poor management then. :-D Bad managers hire bad developers and give them bad deadlines. Plus, in an environment where having a system crash affects everybody (mainframe, mini-computer), more effort is made to deliver a quality product. But when each user has his own system and a crash only affects one user, quality tends to go down the drain. X| (5)

                  Richard Andrew x64R Offline
                  Richard Andrew x64R Offline
                  Richard Andrew x64
                  wrote on last edited by
                  #15

                  Not necessarily poor management. Just management. There is a cost/benefit analysis that goes into deciding how much development and debugging time goes into a product. So if a management decision is made that says, "OK, for this $100 product, X number of bugs is acceptable," that's not poor management, that's just management.

                  P 1 Reply Last reply
                  0
                  • N NowIsForever

                    Call this "execution drift" for want of a better term. Everyone who uses a PC knows about this problem: A computer must be rebooted periodically or "things go wrong." Some software in the system (frequently many programs) will begin behaving erratically, or not function at all. Mainly, I'm looking for more information on the phenomena, and in particular, is there a standard name for this? I tried several topics with Google, pc drift, program drift, execution drift, etc. and nothing valuable came up. Presumably, the culprits are mostly applications and add on code rather than system code, drift in system code is perhaps not uncommon.

                    :doh: Eschew Obfuscation :doh:

                    D Offline
                    D Offline
                    Dr Walt Fair PE
                    wrote on last edited by
                    #16

                    I call it Windows grunge. I've had a Linux server that was only rebooted to upgrade the operating system. It went for 3 years once, with no problem at all.

                    CQ de W5ALT

                    Walt Fair, Jr., P. E. Comport Computing Specializing in Technical Engineering Software

                    M 1 Reply Last reply
                    0
                    • Richard Andrew x64R Richard Andrew x64

                      Not necessarily poor management. Just management. There is a cost/benefit analysis that goes into deciding how much development and debugging time goes into a product. So if a management decision is made that says, "OK, for this $100 product, X number of bugs is acceptable," that's not poor management, that's just management.

                      P Offline
                      P Offline
                      PIEBALDconsult
                      wrote on last edited by
                      #17

                      "Good enough" isn't. :-D

                      Richard Andrew x64R 1 Reply Last reply
                      0
                      • P PIEBALDconsult

                        "Good enough" isn't. :-D

                        Richard Andrew x64R Offline
                        Richard Andrew x64R Offline
                        Richard Andrew x64
                        wrote on last edited by
                        #18

                        Not when I'm the one laying out the money for the product, it sure ain't. :thumbsup:

                        1 Reply Last reply
                        0
                        • D Dan Neely

                          But it's Tuesday.

                          3x12=36 2x12=24 1x12=12 0x12=18

                          L Offline
                          L Offline
                          Lost User
                          wrote on last edited by
                          #19

                          It's software - always delivered late.

                          Join the cool kids - Come fold with us[^]

                          1 Reply Last reply
                          0
                          • D Dr Walt Fair PE

                            I call it Windows grunge. I've had a Linux server that was only rebooted to upgrade the operating system. It went for 3 years once, with no problem at all.

                            CQ de W5ALT

                            Walt Fair, Jr., P. E. Comport Computing Specializing in Technical Engineering Software

                            M Offline
                            M Offline
                            Mark_Wallace
                            wrote on last edited by
                            #20

                            Why do people associate servers with work computers? They might be the same hardware, but they're completely different animals. It doesn't suirprise me in the slightest if a Linux, Mac, Windows, or any other server stays up for a decade or two with no errors. A work machine, that's used to write and build code, to create and modify graphics, to use God-knows how many rich text editors, watch youtube videos, open PDFs, operate 600 background processes, fill in web forms, act as a telephone, compile and manage documents for printing, etc, etc, etc. might not be quite as stable. I point and laugh at idiots who think that Linux is stable because it can run a server.

                            I wanna be a eunuchs developer! Pass me a bread knife!

                            D 1 Reply Last reply
                            0
                            • M Mark_Wallace

                              Why do people associate servers with work computers? They might be the same hardware, but they're completely different animals. It doesn't suirprise me in the slightest if a Linux, Mac, Windows, or any other server stays up for a decade or two with no errors. A work machine, that's used to write and build code, to create and modify graphics, to use God-knows how many rich text editors, watch youtube videos, open PDFs, operate 600 background processes, fill in web forms, act as a telephone, compile and manage documents for printing, etc, etc, etc. might not be quite as stable. I point and laugh at idiots who think that Linux is stable because it can run a server.

                              I wanna be a eunuchs developer! Pass me a bread knife!

                              D Offline
                              D Offline
                              Dr Walt Fair PE
                              wrote on last edited by
                              #21

                              Mark Wallace wrote:

                              I point and laugh at idiots who think that Linux is stable because it can run a server.

                              I point and laugh at idiots who generalize too much. My Linux work machine, which is used for most all the things you mentioned, ran 5 or 6 months without rebooting. I've been testing the waters to see if I could totally get rid of all Microsoft products for about 20 years or so. In stability and reliability, Linux wins hands down. In user friendliness, it comes in pretty bad. In office products software and interchanging with other companies, it is hurting. But in system stability, it wins hands down. On the other hand, I've never had a Windows server that could go a whole year without rebooting. Maybe the later versions are better? Maybe Windows 7 will finally catch up?

                              CQ de W5ALT

                              Walt Fair, Jr., P. E. Comport Computing Specializing in Technical Engineering Software

                              M 1 Reply Last reply
                              0
                              • D Dr Walt Fair PE

                                Mark Wallace wrote:

                                I point and laugh at idiots who think that Linux is stable because it can run a server.

                                I point and laugh at idiots who generalize too much. My Linux work machine, which is used for most all the things you mentioned, ran 5 or 6 months without rebooting. I've been testing the waters to see if I could totally get rid of all Microsoft products for about 20 years or so. In stability and reliability, Linux wins hands down. In user friendliness, it comes in pretty bad. In office products software and interchanging with other companies, it is hurting. But in system stability, it wins hands down. On the other hand, I've never had a Windows server that could go a whole year without rebooting. Maybe the later versions are better? Maybe Windows 7 will finally catch up?

                                CQ de W5ALT

                                Walt Fair, Jr., P. E. Comport Computing Specializing in Technical Engineering Software

                                M Offline
                                M Offline
                                Mark_Wallace
                                wrote on last edited by
                                #22

                                Walt Fair, Jr. wrote:

                                I point and laugh at idiots who generalize too much.

                                Pointing and laughing is so much fun! I think the two main things that helps Linux maintain its stability rep are: 1. That no self-respecting Linux user will buy anything from Adobe. 2. That open source apps, while they might be a pain in the backside, user-unfriendly, and often lack polish, are generally coded a lot better than much of the freeware and shareware available for Windows. That said, and even though my machines are loaded with Adobe malware, they never crash, and I never shut them down unless a really annoying update demands it (and demands it and demands it -- I've left "restart now?" boxes floating around for weeks). I don't think that Linux is really more stable at all; it's just that people spend more time setting it up to be stable, and don't install so much broken code on it.

                                I wanna be a eunuchs developer! Pass me a bread knife!

                                1 Reply Last reply
                                0
                                Reply
                                • Reply as topic
                                Log in to reply
                                • Oldest to Newest
                                • Newest to Oldest
                                • Most Votes


                                • Login

                                • Don't have an account? Register

                                • Login or register to search.
                                • First post
                                  Last post
                                0
                                • Categories
                                • Recent
                                • Tags
                                • Popular
                                • World
                                • Users
                                • Groups