I interacted with Andre on and off for years when running my first company. The company built hardware for web caching, which is a disk-limited task, so we were constantly on the edge of low-cost disk technology (we also built big SCSI boxes, but the low-end ATA boxes were much more popular), so we tickled lots of bugs in the ATA drivers, and shipped out lots of boxes running on his pre-release ATA drivers. It was always, always, always, a pleasure to work with him. Not because he was nice (though he was always more than civil to me, perhaps because my issue reports were concise and complete and reproduce-able), though he was nicer than some of the other folks we were working with (Han Reiser was perhaps the most cantankerous, aside from Linus). But, the important thing was that he was impressively competent. I'd send him a kernel oops and he'd fix it by the next day, sometimes after a couple of back-and-forth communications or having him login to a box where he could replicate it, usually along with an amusing explanation of why the hardware implementation was broken. I think it would be difficult to overstate how important he was during those important growth years for Linux; without reliable and fast hard disk drivers, a system is crippled, and he single-handedly built a significant portion of the ATA drivers during that period.<p>I never met him in person or knew him on a real personal level, but I considered him a great ally of the things I care most about (software freedom and reliable software, in particular). I hate to see him go, and I'm sorry his family has to go through this. I know that he inspired many developers to stand up for software freedom as well as technically sound implementations, not because he preached about it but because he lived it, every day.