Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
Rich Braun wrote in a message to Mike Bilow: RB> I should also point out that Linus had a much narrower goal RB> than the FSF: to make his initial contribution, he needed RB> only to make his boot-loader and device interface work on RB> the 386 chip. One could argue this, but my perception was the Linus wanted to make a system that used the native capabilities of the 386-class hardware. He had been using Minix, which is written to the 286 which Bill Gates (in the OS/2 context) appropriately termed "brain dead." In retorspect, the 386 proved to be the minimum platform on which one could build a useful multitasking operating system, and the few attempts to use the 286 -- Minix, OS/2 1.x -- were extremely limited even in they were remarkable technical achievements. And, of course, the 286 model deservedly turned out to be a dead-end. RB> The FSF was hamstrung primarily by the stated RB> goal to simultaneously release the Hurd on multiple hardware RB> platforms (pretty much all Unix hardware environments except RB> the Macintosh; and in fact one of the trickier requirements RB> was that it specifically *not* run on the Mac, owing to RB> antipathy between Apple and FSF founders). This goal worked RB> well for the GNU user environment's remarkably wide global RB> adoption, but it got in the way of kernel development. I think you misstate the situation. There was certainly no requirement that the GNU product not run on the Macintosh. Rather, the basis of the dispute as I understood it was that Apple refused to document the Macintosh (or anything else) sufficiently to write system-level code, and therefore Apple hardware was simply declared unsupported. But there were obvious very good reasons to support running on the Motorola 68000-family hardware, and this was probably seen as the most Unix-friendly of all major microprocessors of the era. There is no question that system-level programmers saw the 68000 as nirvana compared to the byzantine x86, and this was duly reflected in the capabilities of the Mac OS relative to MS-DOS. Even by the mid-1990s, it was far from clear that the dominant architecture by 2000 would be x86, as PowerPC, 88000, and other RISC contenders had to be taken seriously. Linux was at first very intimately tied to the x86, but that led to Linux becoming a really useful running system far sooner than would have been possible using a more elegant, say microkernel, design. One of Linus' greatest talents has been to get things out the door in some working fashion, even if he knew they were really hacks that would have to be cleaned up later. Also, the fact that Linux was from the beginning built by college students who were short of money had a significant effect that is often forgotten. In particular, Linux ended up being designed to run in low memory and other difficult environments because of these economic necessities -- 16 MB RAM cost in the neighborhood of $1000 when Linux development started, and many machines had BIOS limitations which prevented them from supporting more than this. So elegant solutions which would have eaten time and money, such as platform independence and a microkernel architecture, were avoided from the outset. -- Mike - Subcription/unsubscription/info requests: send e-mail with "subscribe", "unsubscribe", or "info" on the first line of the message body to discuss-request at blu.org (Subject line is ignored).
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |