Geek Practitioners Blog

I have long imagined a future with ubiquitous computing, such that any surface might be a video screen or be interactive, and might even change from video screen to an apparent traditional surface when not in use.

That requires a few things. One is amazing screen technology, which we’re working toward with thin LCD everywhere, and improvements in the direction of electronic paper. Another is extreme miniaturization, crossing into active nanotech. When things the size of ever-shrinking cell phones pack the power formerly found in desktop computers, we’re getting there. Finally, good wireless communication and integration between elements that might make up such ubiquitous computing or hardware interacting with same. That, too, is getting there, if not there already, including in the miniature, passive form of RFID.

My vision of the future was brought back to mind by Microsoft’s new Surface technology. I haven’t read the entire article, but the brief video is compelling and well worth watching. (Article noted via Glenn Reynolds.)

There are two drawbacks. One is all of this has to be kid-proof. You should see what the squirts do to our plain old coffee table, and what they can do to a computer given the chance. The other is that ubiquitousness meets inelasticity if you expect or create it to be excessively built in, say as part of construction materials.

Other than that, well… bring it on! Especially if I can use it to better control the housekeeping robot. Oh wait…

I’m curious whether anyone else has opinions or a body of experience indicating the quality over time of different hard drive brands.

I’ve had the worst experience with Seagates dying or becoming unusable. In fairness, a large part of the sample size was a case of 4.3 GB drives used for upgrading some Dell Pentiums for a client. Perhaps there was a particularly bad manufacturing run and in general they are fine. Still, I have seen other Seagates bite the dust, with no relation to the main group. As a result, I try to avoid Seagates.

Once upon a time, Western Digital had such a stellar reputation that I would pay a premium to buy WD rather than a different brand. We’re talking back in the days that culminated with my purchase of a 540 MB drive for the amazing price of just over $400. Eventually, though, I started to see Western Digitals fail. Not as freely as Seagates, but seemingly with increasing frequency. Renee’s computer has a WD drive that is on the verge of dying, which is what brought the topic to mind. These days I buy WD if I have no other option, but I try to make sure I buy…

Maxtor. I don’t believe I have ever seen a Maxtor fail, and I have nearly as many of them as I have Western Digitals. If any have died on me, it was so few as to not have impinged on my quality suspicion radar.

So I am quite sad that my parts supplier does not appear to be carrying Maxtor drives. That is enough to have me thinking about checking alternative sources.

Other brands? I had an early 7200x hard drive by IBM. Small sample size, so I don’t think I’d make much of it. It was always loud, like a jet taking off. Eventually it failed such that it would work, but only as a slave. Then later it failed completely.

I’ve encountered some Samsung drives, and so far, so good. In one case, overlay software was needed, and I found out that Samsung was, at least at the time, effectively Western Digital hardware under a different brand. The Western Digital overlay utility was what had to be used to make a Samsung drive above the supported size work. That drive is still going. A couple of 60 GB Samsung drives I bought last year are fine so far, but that’s not very long.

I miss my Maxtors. I’ll have to do something about that, or hope for the best from the current crop of Western Digitals.

Comment moderation is on to control spam, but I’d be interested in any other thoughts and experiences with drive brand quality and failure rates. I’ll approve legitimate comments as quickly as I am able.

I have four old computer carcasses that ended up stacked in the office because they were “dead” for some reason or another. This almost always means the power supply died, and may or may not have taken out something else with it.

My objective is to build at least one working computer out of the mess.

It’s been so long for all but the most recent, I forgot entirely what the story was, and two of them appear completely intact; no missing parts. I just plugged in the first of those. It started to fire up, had me thinking “maybe there was nothing actually wrong with this one…”, and then blew out the power supply with a bang and a puff of smoke. One of the most impressive displays of power supply failure I have ever witnessed.

Guess it really did belong in the dead computer pile. Now what will happen if I put in a new power supply? If I am lucky, nothing fried besides that. The computer may have ended up her simply because it kept refusing to turn on or something.

What is it with power supplies anyway? Really! There are dozens of ancient computers around here; P60, P100, P200… all with their original power supplies, still functional.

Every machine from P2 450 boxes to P3 800 boxes has had the power supply go at least once. Some also lost the replacement after a while. While most of the cases were of a particular brand, they varied. The brands of power supply used in them varied. The failures happened in many locations, so it’s not just the power in one building being prone to surges killing them.

Worse, I have even had to replace one in a newer machine, a P4 built by me, in a quality case with a good power supply. It’s as if the people who make the components that are used in power supplies have serious quality problems.

I have long said the same thing. Granted, on most computers now there is a switch directly on the power supply, at the back of the machine. But sometimes there isn’t, and there was a window of time when the funky switches that are controlled through the motherboard had started, but power supplies with switches were rare or didn’t exist.

Still, there are ways around it, and holding the switch in really will usually result in a shutdown and off.

These days there is always a trickle of power to the motherboard so long as power is plugged into the power supply and the power supply is connected to the board. Thus the green LED that is common on the motherboard, lit even when the machine is off. I’m in the habit of turning the power supply switch off before I plug the cord into it to avoid any chance of a surge happening then.

The biggest problem is when the switch or motherboard circuitry start acting up. It’s common for me to hear the complaint that a machine won’t turn on. The usual solution is to turn off the power supply switch, turn on the power supply switch, then try the front button again.

Of course, here’s where I could get into the problem of users burying their computers in places, or under/around/between stuff, where it’s barely possible to reach the front switch, never mind the back switch, or the cables on the back.

Inspired by the Jay Tea post I linked previously, McGehee has pointed out that there is a blog called Spyware Warrior. It appears to be a nice resource.

Jay Tea describes an experience cleaning up malware, that is, adware and spyware, on an extended family computer. There’s that family tech support thing again. I can’t blame him for feeling dirty, as it was a particularly bad infestation.

He took a somewhat different approach from the hardcore one here and for that matter, here. Yahoo’s toolbar? I’d never have thought to do that.

One thing though; installing Firefox is priceless. It goes most of the way toward preventing the problem from happening again. And again and again and again.

I swapped my mouse at home last night for a generic optical mouse. Everyone says they’re better, after all. My old mouse wasn’t horrible, but had seen better days and was of an all too common shape seemingly designed to hurt my hand and wrist.

Anyway, the new mouse jumps around. No smooth, predictable scrolling across the screen. It steps backwards in little blips. It deviates up or down. Sometimes it looks for all the world as if the mouse pointer is climbing over a bump. It’s hard to get fine control to click on small things, which is the problem with the old mouse, only less so. Yeesh.

At first I blamed the mouse pad. It has colors and patterns that could have explained the odd movements. The fake wood grain was no better. Plain white paper was no better. Apparently this is just the way it is.

A shame, because it is more comfortable, with a lower profile. My favorite mouse ever also had the lowest profile every. It was a Mouse Systems brand serial mouse, which I still have for the Pentium 200 I never run anymore. Heck, I suppose I could hook it to a more modern machine and see if it works. It is not rounded side to side at all. It never gets very high, with a longer flat slope on the hand end, which starts down a short slope on the button end, just before the buttons begin. Most comfortable mouse ever.

I need to start paying particular attention to my mouse and desk usage and hardware. My hand is bothering me so much it’s almost debilitating at times.

Argh. NT4 doesn’t talk to hardware interrupts and can’t see drives via the BIOS the same way DOS and Win9x can. That Dell server with serial ATA RAID, even with a FAT16 partition ready, even installing from a directory on that partition, cannot handle the drive.

I obtained a driver that had an ever so slight possibility of making the whole thing work. What happened? NT could not see the floppy drive to add a final insult to the whole debacle.

So it turns out we will need to get Windows 2000 Server to make this work. Not Windows 2003 Server, because the network remains a mix with NT4 and 2000 is okay with that.

I’m very disappointed.

Update:

Turns out it sees the floppy if you hit F6 at the right time when NT4 setup starts. However, trying both Dell’s Windows 2000 driver for SATA and one from Adaptec for a similar controller, I got “file caused an unexpected error (0) at line 1213 in d:\nt\private\ntos\boot\setup\oemdisk.c” and that ended the attempt.

I have been told by a trusted source that if the SATA RAID drive has already been created through the RAID BIOS utilities, NT4 can be installed and will see that a drive is there to the limits of its size recognition. When NT has been installed on its own partition, then service-packed to death, disk manager can then partition the rest, treating it as a normal IDE drive.

You lose the speed benefits and any utilities that would come with a driver recognized by the OS. It should get me through to a time when they can be updated to Windows 2000 or 2003 Server across the board. The main thing is that Exchange keeps humming along with no data loss.

Here goes…

I started setting up a new Dell PowerEdge SC 1420 that has CERC 6 channel SATA RAID with three drives for an effective drive size just under 150 GB. Nice.

But I have to install NT 4.0 until next year when a more complete upgrade of the whole network takes place. It appears that will be a challenge…

Perhaps not so much as I had expected, if Adaptec has a driver after all. In diving into online mentions of this and similar situations, I found both No. Way. Ever. for NT4 and SATA such as this, and apparent workarounds or references to drivers.

Promise has a driver for NT for their FastTrak TX4000/S150 available, and I saw mention of Adaptec having one too. I have to return to the fray and find out for real.

This is a lesson in what happens when you try to keep going on obsolete software and systems for too long.