Drugwash

I think legacy support for USB means assigning permanent resources for USB so that it can be "seen" in DOS for devices such as USB keyboard and USB mouse (and also - or more important - for booting from USB devices). This could mean that these resources are fixed and cannot be changed (and possibly accessed normally) when a plug & play OS kicks into action. This may be the reason for no/slow USB access. When that setting is disabled, the resources are free to be reassigned as the OS sees fit. Or I may be wrong.

Your particular setup is probably seldom used, hence the lack of replies. Reading from floppy is most likely triggered by a request not only for enumerating all the present drives but also for their total/used/free capacity. The fact that this happens only sometimes may indicate that the check is performed only when at least one of the drives' used/free space has been modified since the last enumeration. Unfortunately I can't tell exactly who is doing this, when, why or how to disable the behavior. What I can tell is that many, many years ago I had disabled one drive in that list of drives presented by TweakUI - it may have been drive B but I'm not sure - and subsequently I experienced a weird behavior which unfortunately I do not remember what consisted of but was bad enough to carve into my mind "never disable a drive in TweakUI". So please do keep this in mind in case your system starts behaving abnormally. Off-topic but related to your last remark: depending on your BIOS brand/version and your skills you may extract/save the current BIOS image to a file, edit it with a dedicated application to change the defaults exactly as you want them to be, save it as a modified file and flash it to the CMOS. This way your settings will always be the desired ones even when battery fails or while it's being replaced (although if you're quick enough the settings won't get lost). I did that to a couple motherboards in the past. The tricky part is finding the best and most compatible version of the BIOS editor for your particular brand/version. And please do keep in mind that (re)flashing the BIOS is always risky if power failure occurs, floppy drive/disk errors out or a bad/wrong image is being flashed.

Instantly, this line from the "I, robot" movie popped into mind: "Now, this is the right question!"

Merry Christmas and Happy Holidays to everyone!

FlashGet (1.9.6) can navigate/download FTP, Total Commander has a built-in FTP client…

Yes, it's OK to install it on any number of computers as long as the installation is done through the official installer. I've skimmed through the EULA just earlier while installing v5.4 and that's what it read. I find that license quite large for an application of that type. Probably they're trying to protect the 'paid mode' files (and/or registry keys?) from being copied over to other computers. Dunno, I'm not familiar with such legal stuff nor am I affiliated with Brightfort or anybody else - just stating my findings, hopefully it would help others not to get into trouble.

Well, technically this may well be possible, but I was talking about their EULA or whatever it's called. I read about that some time ago, can't remember when or where, but it did strike me that they explicitly prohibit "migration" of the update files. I haven't tried such a migrated update from XP to my 98SE so can't tell if it works or not.

According to front page the range of supported OS is Windows 2000, XP, Vista, 7, 8, 10. (32/64-bit) Latest version compatible with Windows 9x is 4.6. I've been using this for many, many years. Sadly, their policy appears to be extremely restrictive in that the updates cannot be ported from one system to another (say from XP to 98SE). Too bad. Still, it's one valuable piece of security that everyone should have installed and up-to-date (at least on supported systems). Thanks for the heads up!

I like this phrasing. Looks like Microsoft are already dead but they don't know it yet. Why don't we help them cross over 'Ghost whisperer' style…? JorgeA above has a point, for one.

I've got usbccgp.sys v5.1.2600.5585 on my trusty old 98SE machine. There's also usbport.sys v5.0.2195.5652 and usbehci.sys v5.0.2195.6882. The hub driver is usbhub.sys v4.90.3002.1. Related may also be usbntmap.sys 4.90.3000, usbstor v4.90.3000.1, usbd.sys v4.90.3000.1, usbauth.sys v4.90.3000.1 and a shady usbu2a.sys v1.42. Used to have MS usbhub20.sys v5.0.2195.6891 which I recently replaced with VIA's 4.90.3000.11. They both work(ed) fine with the built-in VIA Tech 3038 USB 1.1 controller and the VIA VT6202 USB 2.0 add-in card. But maybe my SOYO SY-6VBA133 motherboard is more resiliant to drivers than other boards, especially the new ones.

Guess this topic could be nicknamed The (Windows) X files.

Someone at their high quarters must definitely be a Highlander series fan: "There can be only One!!!" Of them all above, OneShot seems the most appropriate for Micro$oft. In the head!

No. The motivation was and is seamless integration of advertising into everything, confusing users and luring them into clicking at random, thus rising the chances of getting more revenue by means of click count, referred purchases or whatever. Why does Windows 10 need constant updates? Because the advertising elements embedded in the code have to be replaced every now and then. Simple! It's all about the money, it has always been about the money. The user doesn't count for sh.. in this equation.

You're welcome!

Now we're talkin'! There's one particular quote în the last link's article that I'd like to emphasize: Considering in this particular situation Win98SE can do that resolution I'd say we're out of the woods regarding a bad video BIOS, but for anyone else having a similar issue with another chipset/video this may be useful information. That article has a lot of technical information that's quite hard to hack together even by a knowledgeable person. I wonder what happened to old-school programmers that had the WinAPI at the tip of their fingers…

I wouldn't call them very clever as long as they use .NET, but that's just a personal opinion. The image in the 7-zip archive is clearly 1400x1050. So definitely something interfered with the image size attached to the board but that's not even important anymore. Problem was and remains making XP allow a non-standard resolution of 1400x1050 in full screen, which requires manually overriding the EDID. The tool referred to by jaclaz may do the job but I, for one, wouldn't install .NET just to run that thing. To each their own though. Good luck!

Well, I'm sure something tampered with your first screenshot because what I saved here on my HDD is 93,538 bytes while the size seen in one of your later screenshots (in MS Paint) shows 132,886 bytes. If the board software does this automatically, this is very bad. OK, now that we figured this out and also seeing the EDID report, I can say that Win98SE is much more tolerant with respect to unsupported resolutions, while XP is strict to a T and will not allow any unsupported resolution other than in pan-and-scan mode. So I guess you'll have to find an XP driver that can force a non-standard full screen resolution. Unfortunately I have no experience with Intel drivers so can't help here. Hopefully someone else can.

There's a multiple-model manual here containing technical details towards the end. Not of much use, in my opinion. 1400 x1050 reported by the control panel is a clean 4/3 ratio. 1200 x 900 of the real screenshot is also a clean 4/3 ratio. The fact that the control panel reports a resolution while in reality it is a different one could be anybody's fault but I'm inclined to point to the drivers who may use a built-in flat panel scaling capability that I've seen in NVIDIA's drivers. If such option exists, it should be set to use display's built-in scaling, not driver's.

The picture you uploaded is 1200 x 900, despite the control panel reporting 1400 x 1050. Hm, this is getting weird. According to the manual, the monitor has both DVI and D-SUB input - which one are you actually using? Regardless, try the other video output/input if your videocard allows. If you have DVI-out on the motherboard you may also try using an adapter with a D-SUB cable to connect to monitor's VGA input. Please make sure the video cables are in perfect shape - if possible, replace them at least temporarily. All this is to test if any of your monitor's inputs fails to report the correct EDID data to the driver. There's been some talk about the EDID issue even recently, you may search the board. There's also at least a free application that can build an inf from the monitor's EDID. The inf can then be manually edited if anything's wrong. Of course, there's always the possibility that some fault lies in the video BIOS If the motherboard has an AGP or PCI-e slot you may try with an external videocard. If it works then you know it's the built-in video's fault. Let's not forget the drivers - both 98 and XP ones - which may be the latest but not the best fit. Try older driver versions, if possible.

I've been into electronics for about 37 years. At this point I'd rather replace the microswitches a dozen times than replace my favorite mice. Which are all ball-based, with Logitech chips. One of them is my first ever mouse, more than 15 years old and it's got a twin brother: model M-S35. Wouldn't trade them for golden mice the same weight. Just sayin'.

I just got proof that, after all, we still are human… Glad things got sorted out in the end. And… enjoy your favorite music!

You may be right, I just looked at the Activex/OCX tab offered by the FileInfo plug-in in Total Commander, without actually checking whether registration is available. My bad! However, I see: dispinterface ITextDocument dispinterface ITextRange dispinterface ITextSelection dispinterface ITextFont dispinterface ITextPara dispinterface ITextStoryRanges as references and typedefs for a tom library, that would suggest a COM interface. I'm not familiar with COM though.

Before moving/renaming riched20.dll you probably should've unregistered it (regsvr32 /u riched20.dll at a command prompt in C:\WINDOWS\SYSTEM). If you decide to put back the originals don't forget to manually register the original riched20.dll. A reboot may also be in order after unregistering/moving/registering the files but it's only a hunch.

Might be an incompatible LOGFONT structure being passed to CreateFont().

Well, what's the actual aim (pun intended): communicate over a certain IM network or just have a certain software work?