I love it when there are file server problems at work and Windows grinds to a halt when you do something advanced like select “My Computer”.
In Vista when that happens you get a little wav file that says, “I’m sorry Dave. I can’t do that.”
Upgrade now home boy.
I recall a blog entry by Raymond Chen who said that Explorer would ping all drives in the network when you open it, and they didn’t use a background thread for this job. So if the network doesn’t reply Explorer simply locks up. Brilliant!
That sounds about right.
Windows networking is the reason why I’m still happy to recommend Macs for my company’s 60-host network. Departments that don’t want to pay for Macs get thin clients.
The best part was when you’d go to My Computer, and Explorer would freeze, because of course, you’ve got mapped network drives.
Apparently one of the other servers staged a coup against the PDCs, calling itself the browse domain master. I’ve seen this happen once before – when a 95 box rebelled against it’s NT overloads, but you know what… it’s not supposed to happen at all!
Whenever I find myself messing with any kind of windows networking issue, I always think, in the back of my mind, that I am doing something very wrong. It is almost impossible to think that something could have so many caveats and dont’s and so many ways to break it without trying.
The explorer issue is the most amazing, thrown-in-your-face unaddressed bug in the history of computers: map a network drive, unplug your network cable, immediately open the drive, then on the white, empty window that appears, just try to close it normally. Voila. They don’t fix it because, in the Windows philosophical approach, it’s not even a bug: you’re doing something you shouldn’t, and explorer only breaks if something else (namely your network) is broken.
In other words, its a turd they can’t even be bothered to polish.
I wonder if they’re fixing this for Vista.
Honestly, I can see the cost-benefit analysis. Explorer currently works very poorly over networks with problems. However, fixing it would be a near-complete rewrite, and require upwards of a year to have the same functionality.
If I was in Microsoft shoes, I wouldn’t fix it either, and concentrate on fixing it for the next generation. I’ve got a similar issue at work – I’ve got something now that “mostly kind-of” works, with known workarounds and ways to fix it, and have a replacement for the entire shebang in the works. Do I want to spend resources on the current stuff? Only in the more dire of situations.