To: William T. Katz who wrote (7436 ) 1/29/1998 5:13:00 PM From: batskinner Read Replies (1) | Respond to of 64865
Which of those things will an NT box not do? Your never-crash-in-2-years scenario seems of little use to many engineers, particularly those in areas with poor electrical infrastructure. I think the point that was being made is that Solaris boxes basically do not crash. I am also an engineer and have used Sun workstations of various sizes and shapes for almost 10 years now. In that time I have seen one system crash (an old 3/60) and one have to be rebooted due to a program that went nuts and spawned off thousands of processes. (I'm not talking about scheduled and intentional reboots here. More below.) I have used NT for about 2 years and have had it crash several times (probably about 8 times total) during that period. And by crash I do mean the blue screen of death, not a single process crashing. The interesting thing is that I use the NT-PCs for wordprocessing essentially and some occasional light computation. On the other hand, I regularly run computational and resource intensive software on the Suns I use (Sparc 10, 20, Ultra). And I'm not referring to simulations that will run in under a minute on a PC. I'm referring to computational solid mechanics problems that run for days/weeks on an Ultra. I also run 3D visualization codes that eat memory for lunch and really push the systems to the limit in terms of resource hogging. Under these conditions, I have never seen a Solaris box crash. As for the "two years without a crash" thing, the point isn't that anyone runs a system for 2 years without rebooting (even Unix needs to be rebooted occasionally to cleanup zombie processes, etc.). The point is that the system doesn't crash at random or simply because it's being heavily loaded. I've done Unix network sys-admin for more than 5 years so I'm not just some engineer who uses the boxes for compiling and never sees the "other" things that will force reboots. Yes, there are rare times when the easiest way to fix a problem is to reboot. But the point is, you can choose when to schedule this sort of thing. It's not the same as my NT box crashing right before my eyes.I think engineers will continue to be split on NT vs Unix for a while, but NT will continue to gain mind share and market share. Yes, NT is taking hold for some engineering applications that a few years ago would have required a workstation but which will now run on PCs. But Sun has recognized that and introduced the Darwin line for the types of applications that require lower end machines. And I too think that NT will gain some ground in engineering but IMHO, it's not there yet. I'm not in love with Sun. If NT would let me do my work better/quicker, I'd have switched to it already. As a final note, let me just say that I realize that the entire world does not view the great workstation debate the way engineers do. Not everyone needs to do simulations that take hours/days/weeks. I sure don't when I'm at home. But the debate here seemed to be focused on engineering applications so I thought I'd throw in my two cents. Clearly though, there's more to the whole picture than just engineering workstations. --Batskinner