pointertovoid

Todavía no. I want to try it, as I've a W2k on a separate disk just for experiment purposes. So I'll have observations for i815ep and J8080, which are P-Ata and bring the equivalent of Ncq - J8080 is very efficient on Ncq, it is Excelstor's name for a licensed fabrication of Hitachi's 1680, a 160GB platter with one 80GB face used. If UniAta switches Ncq properly, I'll see it with Atto and W2k. On the other hand, I've no hardware to test it on S-Ata. As for 128GiB, my W2k installation Cd are all HfSlip'ed with Sp4 (...and more!) and the EnableBigLba switch, so the driver's effect shouldn't be observable. HfSlip is anyway the right method. I had to slip the Sp4 because W2k Sp0 installs on a Raid-0 only with certain stripe sizes... Real luck that the support at Silicon Image knew it!

Hi everybody! - I'm not quite sure the Bios associated with an Apollo 133 can use a Usb keyboard, even connected to the mobo's Usb. Did you check in the doc? Maybe your Usb port works. - With a Usb keyboard and a mobo of this age, you're looking for trouble. Why shouldn't you buy a PS/2 keyboard? They're so cheap at eBay for instance! - You i440bx is way better (much faster, no bugs) than the Via. Unless you want >256MB Ram, keep the i440bx. It runs over 133MHz. Some people can run 512MB on it. Just improve the Northbridge cooling. - The Apollo 133 has never and will never work properly. Throw it away. Replace it with an Intel i815ep for instance : quick, no bug, Tualatin-capable (but Ram 512MB max). Such a mobo costs about 5€ +shipment on eBay now. Twice the throughput on Pci, Ram. But avoid the Tusl2-c from Asus (yes!) as it forces Cas and Ras to 3 and Agp to 4x when you increase the Fsb - regardless of your hardware's capability. Take a Msi 815ept or a Ga 6oxt: they don't fool the user as Asus does with this chipset.

Do you mean that .NET Framework installs Vb and Vc runtimes (and J++ and and and) as well? Then, I'd put .net 1.1sp1 and 2.0sp1 (or sp2) on W2k and have all newest Msvcrt and Mfc at once. Nice. But for W95b, even .net 1.0 isn't supported officially. Does it work properly?

Hello everybody! I'm having a look at my next configuration, on which I will run my W2k (don't even try to convince me). At least from documentation, it seems that Intel doesn't provide Ahci nor Matrix Storage drivers for W2k on any IOH (southbridge) newer than ICH7R, but hopefully some people have practical experience, so please bring it here. (1) First, is it correct that no Ahci nor Raid drivers exist from Intel for W2k on Ich8-8r-9-9r-10-10r? Nor from the motherboard manufacturers? (Seen some for the Sata & Raid adapters added on the mobo, but none for Intel's IOH) (2) I've read that W2k works on these southbridges without any added driver, only Raid is impossible. Is that correct? What is the observed performance then: <133MB/s buffered read and no Ncq? <300MB/s buffered read and no Ncq? Or <300MB/s buffered read and Ncq as good (measured by Atto or IOMeter for instance) as with Ahci driver? And does hot plug work for eSata without the Ahci driver? (2b) For the non-Raid versions of Ioh, that is Ich8 Ich9 Ich10 without R, do I need an Ahci driver to get Ncq and >133MB/s and hotplug? Future Hdd and current Ssd need high transfer speeds. (3) I considered misusing Xp drivers on W2k, but saw that IaStor.sys for instance calls KeAcquireInStackQueueSpinLock and KeReleaseInStackQueueSpinLock which don't exist in W2k's Hal.dll; more absences in W2k's Ntoskrnl.exe (the very same sys file is installed on an existing OS or by the F6 floppy). Can someone confirm it doesn't work? (4) Or can I misuse on the Ich10r the W2k driver intended for the Ich7r? I guess it involves editing hardware identities in an Inf file of the Ich7r driver, doesn't it? Where do I find a description of this operation? (5) What would the driverpacks http://driverpacks.net bring to this case? Their ichaahci.cat (=Ahci for Ich10) cites Win 5.0 5.1 5.2, but I just fear they did the type (3) operation. (6) Does Uniata http://alter.org.ua/en/soft/win/uni_ata work on Ich10r? Their site tells Ich8 and the Inf file tells Ich9. Would a change in an Inf file suffice? It won't bring Raid up to now, but at least Ncq. (7) (8) (9) I've certainly forgotten options...

To my best knowledge, Xp can use Lba48 independently of the Bios. At least W2k can. One way would be to install the "Intel Application Accelerator" driver after Xp is installed. It just means that the volume where Xp resides will be smaller than 128GiB - this is good practice anyway. Another way would be to integrate the registry value EnableBigLba in the installation CD. Explained here for W2k Sp4 : http://home.arcor.de/jterlinden/w2ksp4_128.htm Even better, use HfSlip to integrate both the Sp2 or Sp3 and the EnableBigLba on your installation CD. Little to do by yourself.

Random access time and contiguous read speed aren't the whole picture in disk performance. One good reason for this is that accesses aren't random at all - Windows 98 and above make big efforts to avoid random file placing, Xp improves it further, Vista a bit more. So solutions like CF cards on Pata adapters which may look fantastic can be - and generally are - very disappointing. If you measure read and write speed at about 8kB chunk size with Atto (forget HdTach and HdTune) the figures match far better the speed you'll experience from a Flash storage. Notice that mechanical drives are better at Atto than most Flash - especially multilayer cells MLC must be banned as OS disks - and installing an OS on them confirms it.

Yes, some useful hardware has no official driver for W2k. Noticeably, the Ahci and Raid drivers for Intel's Ich8, Ich8r, Ich9, Ich9r, Ich10, Ich10r aren't available from Microsoft. I didn't try Xp's drivers for Ich8+ but had a look at them: IaStor.sys calls KeAcquireInStackQueueSpinLock KeReleaseInStackQueueSpinLock in Hal.dll which aren't available in W2k. More in Ntoskrnl.exe. This gives me a very bad feeling. Some hope would be to use on Ich8+ the W2k drivers intended for Ich7r. I haven't tried yet. Somebody did it ? Also, the very latest video cards (dX10.1) may lack drivers for W2k - but they're sometimes available from Asus when not from nVidia or Ati.

I seriously put in doubt that every bad opinion expressed about Win ME bases on personal experience. Look here, 2/3 of all votes tell ME is the worst one, but as ME has never been very common, I even doubt 2/3 of all people have ever used it. My personal experience covers Dos, 95a, 95b, 98fe, 98se, ME and W2k (marginally Xp), and ME isn't that bad: - It has working USB 2.0 drivers, while these plagued 98fe and 98se. - It has working Internet protocol programs and Internet connection sharing, while 98fe has none. - It ships with Dll newer than 98-98se, meaning fewer are replaced by applications, leading to fewer blue screens. - It boots a bit faster than 98-98se do, if you have enough Ram. - Its Defrag is way better than 95-98-98se, its FDisk has one bug less. And it has the pinball! - Most of its other "improvements" must first be switched off, granted. Blue screens were often provoked by authentic hardware faults at the time of Via chipsets. On a Via, my 98se was as stable as a piece of soap. Now, I maintain a ME on a Sis chipset, it's been stable for 3 years. All together, I got huge amounts of anger with 98-98se and little with ME - though none of these is as stable and swift (but difficult to install) as 95. And I will replace my W2k as soon as Microsoft gives up Wga and activation, meaning that my next one could well be a Linux. So I fear the horrible reputation of ME relies little on direct personal experience and much on rumour.

Hello everybody! I've just bought a couple of UDMA multi-card readers, they are both fast and cheap, so I felt other people may be interested. The Digisol 47200 52in1 takes 52 Flash Cards formats (including the two I need), plugs to USB 2.0, bases on the chip UT335, and is fast: according to http://www.hjreggel.de , only more specialized card readers (offering just CF on FireWire) are significantly faster. Bought with Paypal at http://www.digisol-online.com/ on Monday evening, received in a neighbour EU country on Friday noon, and they were cheaper than eBay competitors. I measured the speed with Atto v2.02 available there http://members.home.nl/rvandesanden/ATTO%20benchmark.html http://www.hugesystems.com/supportspace/bench32.exe http://www.pocketpc-blog.de/download/bench32.exe http://www.techpowerup.com/downloads/123/A...mark_v2.02.html because Atto gives measures near 16kB which represent the speed experienced by a user on typical computer files. Neither sustained read speed nor access time, as measured by HdTach or HdTune, come close to this experienced speed; sustained speed is useful for ~1MB pictures only. The setup is: - A motherboard fast enough (Pci measured at 124MB/s) - A D-Link Usb 2.0 card on PCI, with chip Nec - An Adata CF card, 8GB 266x. It's Slc and Udma5 and fast and has among the shortest reaction times. So this is the speed of the CF at various IO chunk sizes on: (Left) a P-Ata adapter directly on the i815ep chipset (a Silicon Image 680 would be marginally better); (Middle) the Digisol 47200 52in1; (Right) an old Hama 9in1 reader, without Udma. ----------P-Ata--------47200--------9in1 ---------W-----R------W-----R------W-----R --1kB----2.5---7.7----1.6---2.7----1.0---1.2---MB/s --4kB----3.4--21------2.9---8.3----1.7---3.3---MB/s -16kB---25----36-----15----18------6.0---6.3---MB/s -64kB---36----45-----22----25------7.7---8.1---MB/s 256kB---40----47-----22----25------7.6---8.1---MB/s I didn't see the 44MB/s said to be achievable with a TI chip at the Usb/Pci card, but these are the best measures I've made up to now. And similar measures on an SD card are quite favourable as well. Bye!

Hello everybody! Just like many people, I'd like to have my personal preferences integrated to the installation of Win right from the beginning by modifications in the registry. Though, using some kind of .bat file isn't very convenient: - Some modifications are best done very early, for instance enabling BigLba on W2ksp4 (at best before installation begins) - Modifying setupreg.hiv can be better, but the commands for Reg.exe aren't as clear as they could be - One may want to change many keys, for instance for timezones - Higher risk of errors which can't be seen quickly So the easiest way would be a program that takes a normal .reg file and injects the changes in setupreg.hiv: we could export .reg files from a Windows tuned as one likes it and duplicate the tuning in the next installation. - Does such a program already exist? - Or is it feasible, and would someone want to write it? - Maybe simpler: a program that takes a list of .reg files and turns them into a .bat file understandable by Reg.exe? Thanks!

Hello everybody! Sure, applications should bring their runtime libraries with them - but older ones don't always bring them, and I love old applications. Also, I want to have the newest safest runtime libraries on my computer - but I'm not willing to search for application security updates, and many of their software editors have disappeared anyway. Sometimes I want to bring the newest libraries back on my computer after a badly designed application installer messed it. For all these reasons, I'd like to have something like an installer (or a few ones) for all the newest Visual Studio Runtime libraries. To be accurate, I'm not integrating them in my application, but providing them to existing applications - I expect to get the best answers from this programmers' forum, though. So far, I've found: - vbrun60sp6.exe and its KB957924, plus vcredist_x86.exe (2008sp1) - this latter seems to request W2k at least, as it needs Msi3.0 - Or mmenu.exe and msmenu.exe which I converted to 44 Msi files using msm2msi.exe: not very convenient, and I doubt mmenu.exe (=6sp5) has VID6, VFP6 and VJ++6 as announced, since mmenu.exe only brings Mdac.msm that msmenu.exe (=6sp6) lacks. - For DotNet, I believe to have everything. Not useless to know: I want to install the runtime libraries on W2k and on W95b, yessir. Though W95b probably won't accept even DotNet1.0sp3, or will it? Maybe I can forget VID6, VFP6 and VJ++6, since I'm not aware of any of my applications using them. Suggestions? Limit myself to VB and VC? Then, what is (and where do I find) the latest vcredist_x86.exe that fits on W95b? Thanks a lot!

And now, an updated version of the table, with the missing KB829759 and more indications. The table is plain text to help you copy it. In order to display your copy properly, you'll need a fixed width font (like Courier new) where you paste it, for instance an Html document or a text editor (Metapad does it). Again, 123456 means KB123456 and 1234 means 5.0.2195.1234 PAY is shorthand for "not publicly available", WGA for "no caution warning" and BSOD for "better keep away". ------------------Sp4,R1,------------------------PAY-----WGA----------BSOD ----------------Security---829759-883528-838771-838921-838989-843503-838417-836111-823086-817765-841880-890202-- ================================================================================================================ hidclass.sys----2195.6655---6824------------------------6882---6882 (hidusb.sys)----2142.1 openhci.sys-----2195.6675---6824------------------------6882---6940---- uhcd.sys--------2195.6655---6824------------------------6882---6882---- usbarw.sys------2195.1620 usbcamd.sys-----2135.1-------------------------------------------------------6883-- usbd.sys--------2195.6658---6824-----------------6882---6935---6935---7008-- usbehci.sys-----2195.6709---6824------------------------6882---6882-- usbhub.sys------2195.6689-----------------6883---6884-----------------7006-- usbhub20.sys----2195.6655---6824------------------------6891---6891-- usbintel.sys----2134.1 usbport.sys-----2195.6681---6824------------------------6926---6941-- usbprint.sys----2195.6655----------6968-- usbscan.sys-----2195.6655 usbser.sys------2195.6655---------------------------------------------7006-- usbstor.sys-----2195.6655-----------------------------------------------------------6773---6871---6934---7009-- ================================================================================================================ ---------------------------829759-883528-838771-838921-838989-843503-838417-836111-823086-817765-841880-890202-- NB (by dencorso): When requesting KB838417, make sure to select the so-called "SP5" release, not so-called "latest" release, which contains earlier files and omits usbser.sys, as pointed out by Bristols here. I corrected the tables to reflect the contents of the "SP5"release, but poinertovoid tested the "latest" release, instead, because the existence of two different versions was not known at the time, so perhaps the "SP5" version is useful and does not cause BSODs... This, clearly, warrants testing.

I've tried separately the various functional patches for Usb. In short: - Most work without obvious flaws, but what they bring doesn't correspond to my computer - KB838417 is bugged and should better be avoided. Not only does KB838417 produce a blue screen each time I disconnect my Adsl modem (even with proper ejection before), it also lets my Cdrom reader make from time to time the same noise as during power up, lets the mouse scroll go mad, and partially closes Explorer plus TClockEx (a better clock for W95-98-98se-Me-2k, very nice). To produce such varied nonsense on a computer, the most probable way is to write at irrelevant locations in the Ram - something a driver is probably allowed to do. I also tried the Usbhub.sys v7006, which is announced in the knowledge base KB838417 and is integrated in Gurgelmeyer's unofficial Sp5 (the KB838417 patch brings v6935, and also Usbd.sys v6935 which KB838989 and KB843503 already bring). For that, recreating with Cabarc a modified Drivers.cab was necessary. It has similar bugs. So to users of uSp5 who would experience similar mishaps, I would suggest to bring Usbhub.sys back to v6689 in Drivers.cab, DllCache and Drivers. By the way, I find a bit disappointing that several editions of Usbhub.sys display the same patent bug. The bug probably explains why the patch doesn't contain the file versions announced in the Knowledge Base - but it didn't help. And as the Sp5 was suppressed, there was logically no money left to correct Usbhub.sys. Such bugs explain why Microsoft tell to use the patches with caution and only if experiencing the related problem. I couldn't find a better version of Usbhub.sys : v6883 from KB838771 doesn't address the issue of installing W2k from a Cdrom reader on Usb, and v6884 from KB838921 isn't publicly available. No other patch for Usbhub.sys seems to exist. So to people willing to install W2k from a Cdrom reader on Usb, the suggestion would be to find a P-Ata Cdrom reader. Has anyone tried from a S-Ata Cdrom reader?

The answer to the speed issue on Usb sticks between W2k, Xp, Fat, Ntfs seems well explained here: http://www.uwe-sieber.de/usbstick_e.html under "Cache or not" [EDIT: oops, sorry, already mentioned in a post above] The whole site is very interesting: http://www.uwe-sieber.de/english.html

Solved. This was not a matter of Usb patch, nor of a registry key protection or some kind of bug. It is a security policy of Windows whose default setting is too restrictive to match my needs, and probably the needs of most users. You know, "this bug is by design". ----------------- Found the explanation here (available in English through a click): http://www.uwe-sieber.de/usbtrouble.html#auswerfen The same author gives the remedy as well, Danke Kumpel! http://www.uwe-sieber.de/files/allocatedasd_2.reg REGEDIT4 [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Winlogon] "allocatedasd"="2" Or use your favourite registry editor (mine is Vilma v1.6.0, still free). On the other hand, the administrative tools don't offer the same adjustment in W2k, so this action is only for Xp: Control Panel > Administrative Tools > Local Security Policy > Local Policies > Security Options > Devices: Allowed to format and eject removable media W2k offers an adjustment specific to Ntfs removable media, which I set to "local current user" as well to avoid future annoyances. ----------------- And then an article at Microsoft: http://www.microsoft.com/technet/prodtechn...ntry/96271.mspx HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Winlogon REG_SZ allocatedasd // yes, this is really a string, to store a number 0 Only administrators of the computer 1 Only administrators and power users 2 Only administrators and the local current user ----------------- On my W2k Sp4 R1 with all security patches but few functional ones, it worked for the power user session after a few reboots. The first reboot wasn't enough. I went through allocatedasd="1", it worked at that time, and again to allocatedasd="2" - no idea if this was necessary. Xp is said to behave identically. It needs Sp2+some patches.

Jaclaz, about Usb sticks slower on Xp than W2k, check that: http://www.uwe-sieber.de/english.html Gross Technik mate in Churmany!

I've tried more or less all post-Sp4 functional patches for Usb available from Microsoft - at least the ones that bring the latest file versions. These are KB 836111 838417 838989 843503 883528 (841880 superseded by) 890202. None brought the solution. I couldn't find any task added when the administrator session ejects a Compact Flash card from the reader. If such a distinct task exists, it must be too fast to see it in the Task Manager. To eject the complete Card Reader, Rundll32.exe C:\Winnt\System32\Hotplug.dll is used. It probably can't eject just the CF card; I didn't investigate. I also tried a shortcut to Explorer.exe with administrator rights delegation... This shortcut can be created, launched, the antivirus sees something pass by, but Explorer doesn't launch. One solution I found is to make a shortcut to C:\WINNT\system32\diskmgmt.msc (the disk management!) and allow to execute it with administrator rights delegation. There, the user can eject the CF, which is displayed as a disk. Though, I'm not pleased with this workaround. First because diskmgmt is dangerous, then because I've to type the administrator password every time. So better proposals are still welcome! Lowering some key protections in the registry? Or a third-party software? With a background service having admin privilege, or even if it needs the admin password every time? Thanks!

Ciao Jaclaz! Benchmarked the FAT32 on USB sticks: no. - To go fast, use a CF on a P-Ata adapter or in an Udma reader, not a Usb stick - I've already seen three broken Usb sticks (from "innocent" users), but none of my many CF (who have experienced heavy and exotic uses). - I also use two Win95b around me, which work very well and will stay indefinitely, but don't have Usb. They access CF through P-Ata and Pcmcia. So I have no Usb stick - not a single one. I've benchmarked a lot, but on CF with a P-Ata adapter. In short: - Mlc is very slow on small files, Slc is faster there, especially when writing, and HDD are generally better. This corresponds to the experienced speed, and is best measured with Atto, not at all with HdTach nor HdTune. - Filesystem makes no difference there, PROVIDED THAT cluster size is the same - CLUSTER SIZE changes speed an awful lot on Flash memory. This favours Fat16. Have a look at Dandu.be if you read French - others exist in German and English. - Drivers don't change CF speed significantly on P-Ata. Conclusions with Usb sticks may be different. Arrivederci!

Now, I've received from MS all patches I wanted, installed normally since they had the adequate language, and tried them. A general observation is that the versions and even the number of modified files that ship with a patch (or hotfix) don't always match the description in the KB (knowledge base). The versions seem to have evolved over time without updating the description. This would explain why, for instance, Gurgelmeyer's unofficial Sp5 contains files and files combinations that don't exist now in February 2009. What does work is that patches install all the files they carry, provided these are of a more recent version than the one in place. From all the patches I tried (basically, the ones that bring the newest files +841880 as an alternative), none did what I hoped: allow to eject a CF card from a "9-in-1" reader from the user session. Only the administrator can do it. I hoped the patches for Usbstor.sys would bring this extra refinement - no, they change nothing here. Even right-click on the Eject icon from the taskbar, and the the "detailed" option in the subsequent windows, shows all places from the 9in1 and closes them all, whatever I highlight. More disturbing: with these patches, new instabilities have appeared. One blue screen when hot unplugging the Adsl modem as it was still trying to synchronize on the line; this modem as well as the "9-in-1" reader not always detected when plugged; a user session that doesn't close... Things I hadn't seen in 2 years, using W2k in an often cavalier fashion. So Microsoft's recommendation now looks reasonable: these patches "aren't fully tested" as they say (are still bugged I say), and use them only if you experience the problem they solve. And a W2k incorporating all these patches isn't necessarily better than without them all, quite the opposite. I haven't tried each one in-depth, except for ejecting the CF. MS releases KB838989 without the accompanying warnings, so this one could be of higher quality. And to install W2k from a CD drive on Usb, the relevant patches are 823086 (replaced by 890202) and 838417 (replacing 838921).

Thank you so much, James_A! You are perfectly right, these patches are sent by Microsoft for free. I believed it would require a special maintenance contract after 2005, but my Oem W2k without Wga could receive the patches without any other condition. I was disappointed that Microsoft had written patches and didn't release them with the planned Sp5, but giving the patches easily and for free improves their image quite a bit. Knowing that, incomplete testing is indeed a credible reason not to release the patches more widely. ----------------- The contents of several patches differ from the KB descriptions which were the source for my table. KB838989 has Usbd.sys v6935 and Usbport.sys v6926 - both are more recent than announced. Maybe we shall understand KB838989 as a rather general-purpose refresh for Usb on W2k, though it isn't the latest possible update; Microsoft distributes it more widely, without the "little tested" warning and under Wga. KB838417 has Usbd.sys v6935, Usbhub.sys v6935 (older than announced) and no usbser.sys (not installed anyway on my W2k). ----------------- I missed KB829759 just because it put "for W2k Sp3" and this reason was bad. The files it brings are newer than those brought by Sp4 and R1 - but older than KB843503, yes. ----------------- I've neglected KB836111 but have now the other five most recent ones. I'll try them and tell you what they've installed and if I can eject my CF cards after that.

Nabend Schadenfreude! I didn't check if they have you patch, but these are often good addresses: http://www.hpcfactor.com/support/patching/winnt4 http://www.hpcfactor.com/support/patching/winnt4svr http://nt4ref.zcm.com.au/pdfwrite.htm http://beqiraj.com/windows/index.asp Ich drücke Dir die Daumen!

Hello friends! I believe you could substitute most drivers for disk adapters with an F6 diskette, but... There's a far better way! If this bug has been improved by a Service Pack, then make yourself a CD with integrated SP4! It has many more advantages, especially when repairing Win. Your have two relatively simple options (for a French W2k), and there are more complicated ones. The first is to do what Microsoft has foreseen for you. Copy your W2k CD to your HDD, apply the Sp4 to this copy (option /integrate), extract the CD's boot sector and burn a new CD with the boot sector and the updated copy. You now have an authentic W2ksp4 installation CD. If you prefer French: http://www.bellamyjc.org/fr/cdbootable.html http://www.generation-nt.com/dossiers/lire...vec-sp4-integre Every patch for W2k and its original components can be integrated this way, especially the Rollup 1 and security patches for W2k - this includes ie5, Wmp6, dX7 and the others. But not ie6, Wmp9, dX9 and other pieces that didn't ship with W2k. The other simple possibility is to use Hfslip to make a CD that install everything the way you want: the CD can include ie6, Wmp9 and its codecs, dX9, Dotnet and nearly anything you want. To my taste a better solution, especially for repairs. Detailed on msfn.org. Both options are faster than writing or adapting your own F6 diskette.

Grazie - Obrigado - Merci - Thanks! I'll try it and tell you - I'm just making too many things at the same time.

Hello! Again a small worry which isn't completely new, as I already read it described - but not solved - on another forum some years ago. I have Compact Flash cards, a Usb 2.0 reader for them, a Usb 2.0 card (chip from Nec) on Pci, and a Windows 2000 that runs the computer smoothly. However, from the user sessions, Windows doesn't accept to eject the CF card. I close every window that shows the CF and ask Win to eject the CF from the Explorer's contextual menu - as it works from the Administrator session - but Win then answers that some task must still use the CF. I'm not going to give admin privileges to the user sessions, as this is bad for security, as well as for privacy on this multiuser computer. The wraparound I use is to eject the whole card reader, as the taskbar icon proposes, then pull the CF out, and connect the card reader again if needed. But this is less direct, may become impractical if some day I have a card reader integrated in the tower, and worse, this operation sometimes disconnects as well the Adsl modem which is on the same Pci-to-Usb card. So I'd like a true solution. I've observed it with Pci-to-Usb cards from various manufacturers using different chips (Nec, Via, Ali) and with different drivers (Microsoft, Nec, Via). Several card readers with chips from various manufacturers behave the same way, again with different drivers (Microsoft or not). And several W2k installations on different motherboards do the same. Can this be a matter of registry key protection? I can edit it with Vilma. Or would KB841880 solve this issue? I still haven't found how to install it - troubles with availability and language. Or eject the CF by some Bat file that I could run with an administrator shortcut? I've also seen specialized software to eject similar Usb storage, so this worry doesn't look uncommon. Suggestions for a free one?

Hello everybody! I'd like to use a patch in an attempt to improve an uncomfortable bug in Win2000's USB. However, this functional (not security) patch, issued after W2k's main lifespan, is not common, and I could only find the English version. Though, the only file it brings is already "English" in my Windows ("silent" would be more accurate), so I hoped the patch would install. But nothing, nada, niente. Himmeldonnerwetterverdammtescheisse! I already tried to run the Setup.exe or Install.exe in the expanded patch (same result), as well as install options (no such is listed). I still haven't tried to change Window's language in the registry. Sorry if this is a hackneyed topic! I didn't find a corresponding thread here, and it's new to me. For other similar patches, I could just extract the useful new files (like Openhci.sys, Uhcd.sys, Usbehci.sys, Usbhub20.sys) and replace the old ones in C:\Winnt\System32\Drivers , restart, and to my surprise it still works... ¡Ole! The driver manager shows the new version numbers, and my Usb hardware works. Reminds of good old Win95, but somehow disappointing for W2k's security. Now, such a method doesn't work with Usbstor.sys, which exists in \Drivers, \DllCache and Driver.cab: when I replace the file in \DllCache and quickly in \Drivers (as is done to replace Notepad by Metapad for instance) the old version (from Driver.cab I suppose) is restored after about one second. And neither 7zip nor IZArc can inject the new version in Driver.cab. So what's the solution? Create or modify a patch based on the English one? And by the way: would HfSlip accept to slip an English patch in a non-English Windows? Thanks!