>Thanks for your interest, Jim.
>
>I've dug up the following information partly on faith, and that refers
>to a hope that something productive may come of it. If your intent is
>to say, "See, that's not very standard," then I revert to a previous
>statement: The future will see more diversity rather than less, and it
>doesn't matter much whether they can't fix it or they won't.
>
>So, it's an AMD K6-2 processor, 333MHz, with an MB851 motherboard. It's
>a Packard Bell Model 955, purchased new last November. Chip set is
>SiS-5598-W877F with 2Mb video memory. 128Mb Ram, 8Gb on the IDE drive
>I'm using. FAT32 without compression. OS is W98, full version number
>is 4.10.1998, using an Award Modular BIOS v4.51PG. BIOS ID string is
>06/25/63-SiS-2298-W877F-2A5IIG5CC-00. I don't carry a lot of extra
>junk on it as it's basically a work machine. No video games, and I've
>defragged, scanned, and virus-checked it frequently. VFP is the only
>software which has behaved this way since I got the machine.
>

Bob, I don't know if you'd see this as productive, but I'd strongly suspect that there's some additional software present on the drive to make the entire 8GB drive visible; Award's 4.51PG BIOS considerably predates the advent of large drives, and probably needs drive overlay software to see the full extent of the hard drive as a single partition - something like Maxtor's MaxBlast, or OnTrack's DIsk Manager is likely to be there. Neither of these would preclude the use of FAT32 or has any history of causing problems for VFP in my experience.

There are any number of pieces of software that might not be so innocuous as you'd suspect; with FAT32 present, older disk utilities might well cause problems.

If you've upgraded the processor and/or memory yourself, you might be looking at a chipset problem in any case. The Intel 430TX chipset had a limit on its L2 cache of 64MB; coupled with a processor released well after the baseline BIOS was developed, it might be a problem with cache coherency. If the configuration is one provided by Packard Bell as part of a standard package, it probably is not worth worrying about, but if you added memory and switched processors yourself, it'd be worth investigating the exact details of the motherboard in more detail. A good starting place would be Tom's Hardware Page; it has a number of links to processor, chipset and BIOS vendor's web pages that might help identify any known issues from upgrades.

The K6-2 in particular is suspect here; the processor really requires a motherboard based on a chipset using one of Via's or Acer's recent offerings, which support the 100MHz FSB, and at a minimum, upgraded BIOS code that correctly identifies the processor and clock multiplier used in conjunction with it.

>I can dig up similar information on the second machine, on which I had
>the same problem with scrambled FAT tables. If this leads to productive
>discussion I will do so eagerly.
>
>Thanks.
>
>-- Bob