osRe

It could auto-flush after a few seconds with no activity, or when there are no open handles (though at least on XP there seem to be remnant handles for unclear reasons). I suppose it does that anyway, or else power outages or crashes would be more catastrophic. A little tray icon to indicate the write-behind status, and there, just like waiting for the floppy LED to turn off. The Windows commandline one, the DOS, or the GUI one that's part of the driver? You can get info on the chip using their DOS based UPDFLASH; in its cryptic menu mode, one of the options (don't remember which) outputs a cryptic number which is the flash chip ID. Then a web search to find what it is. Instead, possibly flashrom (http://flashrom.org/) can show the actual chip name with even less hassle. If it is flash you can probably update the BIOS regardless of the type. The chip on the card I have is of the exact same family as some of those listed in the SI Readme, just with less capacity. I had to edit the 128K update to 64K (didn't try writing the 128K but I assume it wouldn't work right). Afterwards I flashed it with UPDFLASH choosing the closest chip type. I don't think it affects anything but the BIOS. Could be different for cards that do their own processing, but I believe this one delegates everything to the drivers.

What reading speed are you getting with the uPD720101? Only tried on XP so far, but it seems limited to 15-16MB/sec (tried some flash drives and an SATA enclosure), which is unexpectedly low, especially the HDD.

I think that's just how the SI drivers work. Is that a bad thing? I never got this "safely remove". It's like technology regression; you didn't have to tell the OS before removing diskettes, now you have to notify it in advance? I think if you just give it a few seconds to write any delayed writes it should be okay. How did you try upgrading and what BIOS version do you have? I think you might still be able to update it. (I also think the BIOS doesn't matter for anything but DOS/int13 access).

50 pound! Why so expensive? I think I read about similar 1.5/3.0 problems with SI (though I don't remember clearly now). Perhaps the connected devices are also part of the problem. As for 1.5Gbit being a limitation, in this case the PCI bus is going to limit you first (and, well, only recently drives started really reaching faster speeds). Solution to the rareness problem: eBay. Quite a few sellers, mostly from Hong Kong and China, sell new controllers.

Thanks for the all info. There are more reports of trouble with VIA so I'm going to try to get a NEC 101 (too bad that unknown NEC 10X is more common). The NEC 100 was faster than VT6202 in a review I read, so I assume 101 would be even better, maybe even faster than VT6212. I don't know for sure but my guess would be that unless the card has real electrical issues the only factor would be the chip. There's not much on a USB card more than the controller chip. Audio usually has an analog part which would need a more careful design.

XP on that CPU can't be fun. (256MB is also not ideal, but seems much more workable than the CPU if you're okay with occasional swapping sprees.)

Found a few mentions of problems with VIA USB chips, e.g. this thread and others on the same forum. Some things mentioned are PCI bus parking being a possible problem, and the number of IRQs used by multi-port cards. I intend to run this on a 440BX loaded with multiple other devices (SBLive 4ch, NIC, SATA). Anyone encountered problems with too many IRQs in use? Even if it works I wonder if too much contention could be a problem for some devices.

sp193: So it was stable at first and then trouble started? The IRQ issues being mysteriously resolved could be because you used a different slot.

Thanks. So far it looks like VT6202 is more common, so I might still go after a NEC. But here D72010XF1 is more common. Any idea how it is? I can't find info on the NEC/Renesas site and the chip looks unusual, so not sure how compatibility/performance would be.

madechidna: The best way to see what works for you might be to try multiple CMS options and see which one suits you. There was also site, don't recall its name now, that allowed you to experiment with the administrator-side of various server softwares (CMSes, forums, etc). But I think you'd find more advice on that in webmaster-specific forums. BenoitRen: It's not a bad idea making DB-based sites or using CMSes. It makes management much easier and global changes simple. There's also nothing wrong with regenerating pages if the site is responsive enough (but anyway, caching is quite common).

Are there any preferable USB2 chips for Win98? One advantage to VIA is that they offer drivers directly in their site. No such luck with NEC, where I understand it's about scrounging drivers from random generic card manufacturers or generic USB driver collections. On the other hand, VIA's VT6202 seems to be the worst performer. Not sure about VT6212 vs. NEC uPD720101. Any idea how/what's that odd looking NEC D72010XF1?

I ended up erasing and recreating the partitions normally since I didn't find anything. Also it seems actual data rearranging would be needed as turning Primary into Logical would need an extra sector at the beginning for an EMBR (or more than that if head alignment is to be kept). Even if some software has a simple solution to that with minimal data rearranging, it seemed less likely and less straightforward, so I went ahead with the standard destructive approach. Still curious how I temporarily got the primary partition to get the last drive letter and not one in the middle...

Is there a non-destructive software to change partition types/order? I want to change, without touching or relocating the data, from primary + logical to extended with 2 logicals. This is to solve drive letter shifting. I guess 1 extended with no primaries would be compatible/failsafe? BTW: Must partition entries and their corresponding sector ranges be ordered the same? Is it possible to have the data for the first partition physically after the 2nd partition's data? --- Besides the above, maybe someone can surmise how the following strange occurence happend... I booted Win98 with the newly added drive (on a SATA controller, unformatted primary, NTFS formatted extended/logical), expecting it to shift around drive letters. The boot actually failed with an "Abort, Retry, Fail?" sort of message, I think due to the unformatted partition taking up the letter where one driver was supposed to be. I went to format the partition in another OS and tried booting Win98 again, expecting it to load fine now but with shifted letters. But no, the new partition appeared LAST, despite the other drives each holding PRI+logicals. I attributed this fortunate surprise, with no better explanation, to the drive being SCSI under Windows. A reboot later it was still the same, all good. Or so I thought. During the format, in anticipation of problems, I was looking for a solution to drive letter shifting. One option was the software "Letter Assigner". Despite the letter assignment being okay in the end I was curious to see how that software works. I installed it, had a look (too many EXEs/VXD/no install log/not enough disclosure for my taste for this type of software), and removed it. The next Win98 boot (after a booting another OS in the middle) I was hit with the drive letter shift. I don't know if "Letter Assigner" did something or what. Trying to reassign letters in the registry (under HKLM\Enum\SCSI or ESDI) didn't do much, nor did playing with disabling/removing/reinstalling the driver. I suppose it goes based on the int13h order so the driver shouldn't have an effect, but it WAS assigned differently a few reboots earlier, so I'm totally puzzled.

I ended up getting a SiI3512 based card. Works fine so far. In the process I updated the card's BIOS (adds support for larger drives and fixes problems), which makes me wonder, are there any BIOS updates for VT6421-based cards? I thought about getting another controller for another computer and it's an opportunity to try the VT6421. But if there aren't publicly available updates and it might come with an obsolete BIOS (like this card had), that's a reason to avoid the VT6421. I only found a reference to an old BIOS file being included in older driver versions.

I hate LCD scaling. I don't want scaled resolutions but centered with 1:1 mapping, or integer scaling depending on resolution (e.g., 640x480 -> 1280x960 centered inside 1680x1050). I know current gfx drivers allow you to do that (at least centering, not sure about integer scaling), but I don't know if the ones that run on Win98 supported it. I won't necessarily keep on using the same desktop resolution but I want good 640x480, 800x600 and 1024x768 (and maybe 1280x1024/960) for various older games that run at fixed resolutions. For desktop it's less crucial although it's still a compromise as most LCDs are 100dpi natively and I prefer ~75 for non-scaled fonts (on my 19" I'd use 1152x864 if it didn't suck, instead I'm on 1024x768). LCDs should be better once there's enough native pixels to map to each logic one but such higher res monitors aren't available yet. E.g., using 1024x768 on a common 1680x1050 would get you 1.37 native pixels per logical one (1050/768). I don't remember what exactly my CRT has for physical resolution but the height is something above 1600 and below 2048. Assuming the middle ground: 1824, it works out to be 768/1824 = 2.37. Maybe improved monitor scaling algorithms could help too, but this is actually better left for the graphics card to do. I suspect the analog blur/fuzziness that results due to how images are stretched on CRTs is more pleasing than the bilinear filtering or whatever most LCDs use. But to be able to really tell I'd first have to see an LCD with enough pixels and a DPI comparable to CRTs and try to play with digital upscaling algorithms to see how it looks. Desktop LCD monitors are at around 100 DPI, 19" CRTs are 150-190. Notebook LCDs sometimes have higher DPI than desktop but they have other drawbacks and naturally aren't usable for desktop.

Any idea what/if nVIDIA/ATI Win9x drivers allow to keep a (non-native) resolution centered and non-scaled or integer-scaled? (My CRT is starting to die more and I may have to consider an LCD for the Win98 computer. I want to be able to use non-native resolutions.)

Using older versions is an option too. I use UltraVNC 1.0.2 on Win98. There could be newer versions that also work. Something I recall from years ago was Remote Administrator. I think it wasn't free, but when I tried it at the time it was much faster/smoother than VNC and worked on Win98.

Assuming graphics card assisted decoding is not possible I'm guessing it could work if your CPU is fast enough. What video format are you talking about? ffdshow might be enough for all common formats, CoreAVC is a faster non-free h264 decoder alternative (at least old versions used to work on Win9x, don't know about more current ones).

BTW, did anyone try booting from an SATA card on an old computer (to Win98 or otherwise)? It appears many of these cards aren't bootable but it seems you can integrate the needed code into (some) standard mobo BIOSes using appropriate utilities. I'm still unsure if a modified mobo BIOS is all that's needed or that's only for passing control to the card's option ROM, so a card missing that would be unbootable in all cases.

Thanks for the info. I think I saw some Sil (maybe 3512) cards but the VIA looked better as it had Win98SE specifically mentioned on their site (besides the other pros: PATA port, and eSATA). But a card that includes USB2... that's actually an interesting idea.

Soon we'd have people starting to repack Opera as "Opera Lite" w/o all the junk. It's really sad that companies can't keep it small/light for long and succumb to feature creep. Though Opera was relatively okay before v10 (I'd still get rid of the email client, BT client, IRC client, and possibly widgets thingie).

I'm thinking about getting a SATA controller card with eSATA for use also with Win98 (on an ASUS P2B mobo). Are there any specific things to look for or to avoid? I found a noname card with VIA 6421A which looks interesting. Besides Win98 driver from VIA it also has an IDE channel which could be useful in the future for having IDE (or an extra channel) on modern mobos.

Thanks. But of course I don't use "web style" but customized classic!

When typing a bad directory name into a Windows Explorer address bar it loads an IE error page instead of showing a simple error message (unlike when typing a bad directory into the Run dialog, which does the right thing). Is there a way to prevent this IE-handling and just have it show a simple error window? I know, it's time to get a decent file manager instead of explorer.exe, but it's a chore I'll leave for another day...

Are there any reliability or performance issues with PCI based SATA controllers on Win98? I'm considering going that route too for external backups.