Home » Daynotes Home » Week of 19 July 2004

Photograph of Robert Bruce Thompson Daynotes Journal

Week of 19 July 2004

Latest Update: Sat, 21 May 2011 8:09 -0700

Click Here to Subscribe Buy PC Hardware in a Nutshell, 3rd Edition:
Buy Building the Perfect PC:
[Amazon] [Barnes & Noble] [Bookpool]
[Amazon] [Barnes & Noble] [Bookpool]
Visit Barbara's Journal Page

Monday, 19 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

11:24 - Back from the dentist. Ugh.

Still doing clean-up this week on Building the Perfect PC. It's the usual final rush of trying to get everything as right as possible before the camera-ready copy goes to the printer. I think that happens on 5 August, so there's not much time left. There's still the QC2 pass to do. Ordinarily, that's in-house only, but I asked O'Reilly to send me the QC2 PDFs. Even if I have only a day or so to go through them I may be able to catch something. At that point, though, they're nearly cast in stone, so it'd have to be a pretty minor edit or a pretty major problem for them to make any changes at that point.

Meanwhile, I have lots to do around here to get ready for the book I'm starting on now. There are also housekeeping/administrative things to be done. One of them may be my backup methods.

For years, I've been doing a full (data-only) backup every weekend to DDS-3 tapes. Lately, that's been getting to be more of a problem. DDS-3 in theory stores 12 GB native and 24 GB compressed. In fact, for my backups, I can seldom fit more than the 12 GB native amount of data. Much of my stuff isn't very compressible--zip archives of older material, images, and so on.

I have DDS-4 drives and tapes, but I'm beginning to wonder if it's worthwhile migrating to DDS-4 for my primary backups. A 67% increase in storage capacity just isn't going to cut it in the long term. I could move to AIT, DLT, or LTO/Ultrium, but any of those would require a significant cost for tapes. Off the top of my head, I think what I'll probably do instead is segment my data into chunks that will fit on a DVD+R/RW disc, and back it up as follows:

Working data - Grandfather - father - son rotation scheme to DVD+RW (daily/weekly/monthly full backups); sweep Working data to Inactive data monthly.

Inactive data - Monthly full backup to DVD+R; weekly differential backup to Working Weekly disc; sweep Inactive data to Archive data quarterly.

Archive data - Quarterly full backup to DVD+R

I'll also set up rsync on a couple systems and configure it to copy changes throughout the day to other network volumes.

I've resisted abandoning tape for optical disc backups because tape has superior error detection and correction. However, simply by using redundancy I can overcome that objection. For the working data, for example, I'll have five or six Daily +RW discs, five Weekly +RW discs, and a dozen Monthly discs, which I'll probably write to +R discs and store permanently. It would require an incredibly unlucky set of coincidences for me to lose data with that much redundancy.

The nice thing about using optical discs is that I can "restore" a particular file simply by loading the disc and copying the file, rather than having to go through an actual restore operation. We'll see how it goes.



Tuesday, 20 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

9:34 - Sixty years ago today, a very brave man attempted to assassinate Adolph Hitler. Oberst (Colonel) Graf (Count) Klaus Philip Schenk von Stauffenberg left a bomb in a briefcase near Hitler during a military briefing about the Russian front at Führerhauptquartier (Hitler's headquarters) Wolfschanze, near Rastenburg, East Prussia.

Colonel von Stauffenberg left the meeting after placing the briefcase under the table near Hitler's feet. Another meeting participant found the briefcase to be in the way of his own feet, so he moved the briefcase from Hitler's side of the solid pier upon which the table rested to the other side of the pier. That action saved Hitler's life, and probably extended the war by several months. Hitler suffered only minor injuries when the bomb detonated.

Hitler's fury at the conspirators was unbounded. Colonel von Stauffenberg and scores of others, many of whom were not actually involved in the conspiracy, were hanged from meat hooks with piano wire and allowed to strangle to death slowly. That was filmed, and films of those executions became popular evening entertainment at Hitler's headquarters.

Every year at this time, I wonder what might have happened had Colonel von Stauffenberg succeeded. With Hitler dead, the course of the war would have been shortened, certainly. But its outcome might also have changed the future political landscape. It was only at Hitler's insistence that the Western Front was given equal or higher priority than the Eastern Front. Nearly all of Hitler's advisors understood by this time that the war was over, and favored throwing all available resources at the Russians to stop them in their tracks. The Americans and Brits would have had a much easier time of things had that happened, and it's likely that the eventual demarcation line would have been a lot farther east. There would probably never have been a divided Germany, and even Poland might have remained in the Allied sphere of influence.

Many of the German political and military leaders hoped to negotiate a separate peace with the Western Allies and join them in attacking the Soviet Union. Certainly Roosevelt deeply distrusted Stalin, and many Allied leaders, including Churchill and Patton, spoke publicly about the desirability of turning Allied and German forces against the Red Army. We might actually have seen Patton's 3rd Army penetrating deep into Soviet territory with Sepp Dietrich's 1st SS Panzers on its flank. Stalin might have been deposed, and the Soviet Union might have collapsed in 1945 instead of 1991. There might never have been a Cold War.

None of that happened, but it might have had only one man not decided to move a briefcase.

And some comments on my plan to shift to optical discs for backup.

-------- Original Message --------
Subject: backup
Date: Mon, 19 Jul 2004 21:38:46 -0500
From: Chris Christensen
To: Robert Bruce Thompson

I've misgivings about optical backup media based on some past experience: you can't tell if your media is truly readable without doing an md5sum on the individual files. Don't believe a directory listing! For archival storage I'd consider several hard drives. External, or in a linux server (you can unmount and spin them down when you don't need them). You probably have enough lying around. An unused hard drive will last a long time. I've recovered data from 12 year old (scsi) disks without difficulty, that had been out of service for at least 8 years.

You were fortunate. The magnetic domains on hard drives, as with all other magnetic media, begin fading as soon as they're written. As Jerry Pournelle puts it, "all the 1's and 0's gradually become 1/2's." In that respect, a hard drive is no more reliable than a tape. I'm sure I have 12-year-old tapes around here that are readable, but I sure wouldn't count on it.

As far as writing DVDs, I'd do that under Windows with Nero Burning ROM, so there's no need to run MD5 checksums on the individual files. Simply marking a checkbox in the burn dialog causes Nero to do a verify pass after the write pass, comparing the source data with the data written to the DVD.



Wednesday, 21 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

9:22 - I'm working on the index today. O'Reilly did their usual excellent job on the index, but I have to go through it item by item. Finding errors is easy enough, but they also want me to come up with items that need to be indexed and weren't. I can never think of anything. Oh, well.

More on the reliability of hard drives versus optical discs for archiving data.

-------- Original Message --------
Subject: backup
Date: Tue, 20 Jul 2004 22:50:10 -0500
From: Chris Christensen
To: Robert Bruce Thompson

I'll reiterate: Modern software burning error free optical media (CD's) has resulted, for me, in lost unrecoverable data. Data that was written with Nero, as well as Adaptec software. I wouldn't trust optical media for more than half a year: it has a temperature/uv/time/humidity/etc.. based decay that makes it problematic. Hard drives are well sealed. I don't know what the decay rate is for magnetic domains in hard magnetic ferrites (damn freaking low!), but they look pretty permanent compared to optical dyes! Consider the oldest hard drive you use: if you go to a file from OS installation (oldest file), do you have any problems reading it? In my case, either the hard drive works (spins up, etc..) or it doesn't, and if it does, then all the files are accessible. Don't confuse optical media permanence from pressed media (OS distribution disks from MSoft), with R or RW media!

Hmmm. I don't think so. Hard drives fail quite regularly, and the usual reason has nothing to do with the electronics or mechanicals. It's because they begin experiencing read failures because the magnetic domains laid down by the original low-level format are becoming unreadable. In the old, old days, we used to low-level format hard drives quite regularly. That's not possible with modern ATA drives, which are low-level formatted at the factory, with servo data embedded in the data tracks. There's no way to refresh that low-level format short of having the very expensive piece of equipment needed to do the job. The so-called low-level format utilities simply write zeros to the existing tracks. When those tracks themselves begin to degrade, the drive is good only for the scrap heap.

Optical dyes like those used in CD-R and DVD-R are at least a full order of magnitude more stable than magnetic domains. If stored properly, which is to say at room temperature or lower, with moderate humidity, and in the dark, good-quality CD/DVD write-once discs should be readable without errors for at least a couple decades, and perhaps a century. No one knows exactly how long they'll last, but even the least stable of the current dyes (metal-stabilized cyanine) should be good for 20 to 50 years if stored properly. Pthalocyanine is more stable still, and azo dyes are probably twice as stable or more than the cyanine-based dyes.

The metal used also has an effect. The best is gold, which is essentially incorruptible. The late lamented Kodak Gold CD-Rs were wonderful. I once burned two identical discs on Kodak CD-Rs and left them sitting exposed to raw sunlight for a couple of years. I did binary compares on them periodically, and after two years of sitting in the sun the binary compares showed no differences. Stored in the dark, I don't doubt they'd be good for 100 years.

CD/DVD rewritable discs are probably even more stable. Instead of depending on dyes, they use a phase-change data layer. Changing a bit on a rewritable disc requires heating the data layer to its phase-change or melting point, which is unlikely to happen other than in a CD/DVD writer. I have no concern at all about data I've written to rewritable discs.

The bit error rate is a different issue. Typical optical drives have an unrecoverable bit error rate around 10e-12, compared to a hard drive's typical 10e-14 and a tape drive's 10e-15 to 10e-19. All that means is that it's 100 to 10 million times more likely that an unrecoverable bit error will occur on a writable optical disc than on a hard drive or data tape. That's easy enough to address, though. All you need do is make two identical copies of the disc or, alternatively, do a binary compare on the one disc you write against the original data.



Thursday, 22 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

10:40 - Someone commented the other day that Microsoft must really hate the continuing flood of viruses, Trojans and worms that afflict Windows and Windows applications. I don't think so. I think Microsoft really loves all these exploits, and I'll tell you why.

Consider this. Microsoft has two main goals. First, of course, is to generate lots and lots of revenue. Second is to shift users in the directions Microsoft wants them to go. As an example of the latter, take DRM, which Microsoft clearly hopes to monetize to their benefit. Bill Gates probably daydreams about getting a cut of every transaction processed through his DRM-laden software.

With regard to their first goal, Microsoft has a problem. They have a 95% market share in a saturated market. Revenues from software bundled with new systems is grossly inadequate to keep Microsoft in the style to which they've been accustomed. In such a mature market, there is only one way to produce the levels of revenue Microsoft requires. Current users have to upgrade. But the simple fact is that most users don't upgrade voluntarily.

Millions upon millions of desktops, both business and home, are running old versions of Microsoft operating systems and applications. Until recently, we were standardized on Windows 2000 and Office 2000 here, and we're not alone. The uptake in businesses of Windows XP, not counting XP bundled on new systems or installed under volume purchase agreements, has been pathetic. Simply put, almost no businesses have upgraded to XP voluntarily. Nor have home users. There are uncounted millions of desktop systems out there still running not just Windows 2000, but Windows NT 4 and Windows 98. Similarly, Office XP and Office 2003 haven't made any dent at all in Office 2000 installations. Even Office 97 remains in widespread use.

From Microsoft's point of view, that's intolerable. How dare these people continue to use old software that they've already paid for, when Microsoft needs them to upgrade? But how to get them to upgrade is the problem. They don't need or want the new features promised by the new versions, and left to themselves they'd simply continue using that paid-for software forever.

But then comes the threat of viruses, Trojans and worms. "Ah," Microsoft thinks to themselves, "if we only provide updates for our current products, we can force these people who are still using older products to upgrade if they want to be protected." So you end up with situations like the current one, where XP SP2 is the Holy Grail. It will fix the recently exploited security flaws in Internet Explorer. It would have been easy enough for Microsoft to distribute the fixed version of Internet Explorer for earlier versions of the OS, but they're not going to do that. IE, you see, is now tied inextricably to the specific version of the operating system in question, or so says Microsoft. As much as they'd like to distribute the fixed IE for Windows 2000, they just can't do it.

So what are the millions upon millions of people who continue to use Windows 2000 to do? Well, Microsoft's answer is that they should upgrade to Windows XP. And, oh, by the way, they'll have to pay Microsoft's upgrade fee to do that. Frankly, I'm cynical enough not to be surprised that Microsoft will be allowed to get away with that.

Which is why I don't think Microsoft is in the least bit upset by the continuing flood of viruses, Trojans and worms. From Microsoft's point of view, the more of these things that hit and the more severe the consequences, the better. Every new exploit makes it more likely that more people will upgrade to Windows XP, and pay Microsoft for the privilege of doing so. The only way Microsoft would consider the flood of exploits to be bad is if they started to effect Microsoft's market share, which is why the fact that these most recent exploits have caused IE to lose share to Mozilla are interesting.

I'm not putting up with it any more. July 4th was our personal Independence Day. We officially declared our switch to Xandros Desktop Linux. It'll be a while before we've fully migrated away from Windows, but it's happening right now..

More on using writable DVD for archiving and backup.

-------- Original Message --------
Subject: format
Date: Wed, 21 Jul 2004 21:49:33 -0000 (UTC)
From: Bo Leuf
To: Robert Bruce Thompson

> ... The so-called low-level format utilities simply write zeros to the existing tracks. ...

Actually, you need a special wipe utility to do that. Windows "low-level" format mostly does read test for all its lengthy trundling. FAT or NTFS. A tiny "random" selection of sectors are test-written, most are not. Fully 90-something percent of the data is physically untouched in the full format process, 99-something in the quick format or in partitioning. Easily readable directory structures are lost but most of the pointers do remain.

No, you're thinking of a high-level (also called a logical or DOS) format, which is all Windows provides. I was referring to the so-called pseudo-low-level format utilities that are available from the drive makers. In fact, all these utilities do is write all zeros to the user-accessible tracks. They don't refresh the low-level format itself, which includes the embedded servo data.

-------- Original Message --------
Subject: DVD+R/W Reliability
Date: Wed, 21 Jul 2004 22:42:15 -0700
From: JHR
To: Robert Bruce Thompson

Bob -

You wrote: "CD/DVD rewritable discs are probably even more stable. Instead of depending on dyes, they use a phase-change data layer. Changing a bit on a rewritable disc requires heating the data layer to its phase-change or melting point, which is unlikely to happen other than in a CD/DVD writer. I have no concern at all about data I've written to rewritable discs."

Not necessarily so. I have a Sony brand 4.7 Gig DVD+RW disc, formatted packet-writing UDF using BHA SW via a Sony DRU-510 drive, that failed miserably. So far, it is the /only/ one to do so. I have had no problems with other Sony brand DVD+RW discs that I regularly use as "belt & suspenders" backup.

I had almost given up on the damn' thing. Fortunately, I discovered CDROLLER (http://www.cdroller.com/). It retrieved, /very/ tediously and slowly, all the files on the disc. It did not/could not repair the disc.

The peculiar thing is that this failure occurred between uses. One day it was accepting files just fine, like all my other Sony discs. Then when I again attempted to use it, I got an error message. There was no apparent damage to the disc, and gently washing it (which had restored some other CDs on occasion) had no effect Clicking on the "Properties" option showed the disc to be 100% full. It was not - as the retrieved files are <650MB.

I chose Sony brand discs on the theory that they would be compatible with a Sony brand CD/DVD burner if anything would be.

POINT: I have lost all confidence in DVD disc reliability. The next twenty discs may last and be usable for years - or the next disc may be a bummer. I hesitate to use them for anything but redundant backup any more.

Sure, discs fail, but what you describe happened during the writing process. If you'd written and verified the disc the previous day, it was almost certainly readable until you tried to write it and failed.

Also, I should mention that, although your scheme of using discs labeled with the drive maker's name sounds reasonable, it's not always the best thing to do. I haven't looked specifically at Sony DVD+RW discs, but with CD-R/RW discs Sony was notorious for putting their name on third-rate discs made by various Pacific Rim manufacturers. Oddly, Sony also actually produced discs themselves, which were excellent, and those discs were sold under both the Sony name and other brand names. So Sony was in the odd position of putting their brand name on discs made by themselves and others, while at the same time actually producing discs that were sold under their own brand name and other brand names.

As far as DVD+R and DVD+RW discs, I use Verbatim (actually made by MCC, Mitsubishi Chemical Company) and Maxell exclusively. They're of superb quality. I'd also be willing to use Taiyo-Yuden or Ricoh DVD+R/RW discs, but they're not widely available in the US.

Once again, the potential failure window is during writes. If you write and verify a disc successfully I wouldn't worry too much about it being readable later on.

13:44 - This week, I hope to finish up all the last-minute details to get Building the Perfect PC ready to go to press. I checked the index yesterday, and the full cover today. The only thing that remains is the QC2 PDFs, which I'll probably get on 7/28 or 7/29, with any comments/edits due back by 7/30. At that point, any minor final changes are made and the camera-ready copy goes to the printer. Printing, binding, and distributing the book will take a couple of weeks, so it should be in the stores by mid-August. Finally.

I have a few other administrative things to get cleaned up, but next week I start serious work on the fourth edition of PC Hardware in a Nutshell.


Friday, 23 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

9:30 - O'Reilly needs a couple chapters to get the new edition of PC Hardware in a Nutshell rolling (and, more importantly, to approve our first advance payment...), so I'm starting work on those. In line with the traditional "something old, something new, something borrowed, and something blue", I plan to submit an updated version of the Power Supplies chapter (borrowing some text I'd written for Building the Perfect PC that ended up on the cutting room floor) and a new chapter on Wireless Networking. Now I need only come up with something blue.

More on backing up and archiving to optical discs:

-------- Original Message --------
Subject: Re: format
Date: Thu, 22 Jul 2004 22:49:56 +0200
From: Bo Leuf
To: Robert Bruce Thompson

Ok, sorry, it wasn't clear that you were referring to the format/wipe tools provided by the drive makers themselves.

I've been looking but haven't found any real hard data on how long the servo tracks might remain reliable. Possibly because there may be a rather big and largely unknown spread between different technologies here. I've never had a drive fail except under severe mechanical stress (seized bearing), so I can only say that the oldest IDE drives I have running, which are something like 15 years old, on Atari Falcon systems, are still good. A few external SCSI units are even older but also fine, last I checked.

When drive makers casually and routinely hype MTBFs of half-million to million hours for some drives, and rarely under a few hundred thousand, one assumes that the servo data must be good for most of these literal decades.

Apropos CD-R and RW. Though it's true that gold-back ones should be better, I see data suggesting that the gold layer is so thin now as to give a relatively high risk for many minor defects, both from process and from later addition of label-side layer. Some whitepapers I've read even suggest that you literally can't get defect-free CD media at all, so that we depend totally on error correction and burner compensation features to give us usable data tracks.

You've been fortunate. I've had many hard drives fail as a result of media access errors, which are usually caused by degradation of the factory-applied low-level format. Using the "low-level format" utility provided by the drive maker actually just writes zeros to all user tracks and marks the unreliable areas of the drive bad. That's a very temporary fix, because the actual underlying problem is the fading low-level format, and more particularly the loss of the servo data that's embedded in the tracks.

The MTBFs published by drive makers can't be used as you suggest. That is, an MTBF of 100,000 hours suggests to most people that an average drive with that rating should last 100,000/8766 = 11.4 years. It's not that simple. In fact, the service life of that drive might be rated at 5 years, which is a more accurate indication of how long the drive might last. Moore's Law has largely concealed the problem, because very few drives are in use for more than four or five years. Consider that five years or so ago, a 4 GB drive was considered large and a 10 GB drive huge. Most drives are replaced nowadays not because they're reaching the end of their service lives, but because they (or the systems they're in) are too small and too slow.

As to CD-R discs, very few of them use gold nowadays. The ones that appear to be gold in fact usually use a silver alloy that has a gold appearance. You can still get true gold CD-R discs, but they aren't cheap. The last time I looked, they were a buck or more apiece, which means that almost no one uses them. The $0.15 CD-R discs have nearly driven the expensive (and good) gold-based discs from the market. But they are still available if you're willing to pay $100 to $125 for a spindle of 100.

Economics forced the shift from using gold for the reflective layer to using silver or a silver alloy. Much of the bad press that writable optical discs have gotten with regard to archival stability are a result of the early silver-based discs. Gold is essentially immutable, and so sealing it from the atmosphere wasn't very important. Silver, on the other hand, corrodes easily, particularly from exposure to sulfur dioxide in the air. Manufacturers realized what was happening, and over the last few years they've greatly improved the coatings used to seal the reflective layer. Current-production name-brand optical discs are well-sealed and quite reliable.

-------- Original Message --------
Subject: CD stability discussion--my 2 cents
Date: Thu, 22 Jul 2004 23:46:26 -0400
From: David Yerka
To: Robert Bruce Thompson

Hi Bob:

I thought I'd give my experiences to the backup discussion.

(1) I've had clients running backup on both tape and CD-R for the past 5-6 years and the ONLY problems with either have been during the INITIAL WRITING. My clients are generally doctors or dentists and are performing both daily backups to tape and CD and archiving patient records to CD. We now have CD-R disks 5 years old holding copies of patient records which are still perfectly readable. I know because I check them and make extra copies every couple of years and do bit compares to the originally burnt CD. (Copies are made from the newest copy of the original)

Of course conservative settings are used: Copy with no compression, 650meg. CD-R disks (I loaded up with a good supply), Plextor drives. Nero Burning (latest/most stable version), multiple copies stored in separate locations, verification on, store vertically, protective cases, stored in fire safe for onsite copies.

The routine is: Daily Full to tape, Daily Differential to CD. Weekly full to tape (5wk cycle), weekly full to CD-R (duplicated and stored off site). Monthly full to tape (3month cycle) and Monthly to CD-R (duplicated and stored offsite).

Recently we've started using DVD at some sites with corresponding results for the last year only. In fact, if it wasn't for size problems, I'd probably consider eliminating tape entirely but as yet there are no 30-50gig. DVD burners which are proven dependable.

(2) I've used CD-R for working copies of CD's I use on a daily basis for the last 8-10 years and the only problems I've seen are due to actual physical damage: scratches, worn spots (from rubbing in sleeve). And these are disks that live in a CD a full 40 CD binder case and travel with me everywhere in my equipment case.

The problems I have seen are from (1) bad media (not necessarily cheap in cost but cheap in production), (2) crappy drives, and (3) crappy software; in that order. Buy the best disks recommended and test, test, test, Don't be penny wise and pound foolish, buy the best drives (and get 2, one for the machine and one spare stored offsite with the backups), and ALWAYS, ALWAYS test updated software against reading older disks and if everything is OK burn a new copy after the update (this, though, may not easy for archives but is safest).

My rule of thumb has always been that ANYTHING that depends on magnetic domains for storage is suspect. Use a single hard drive for backup? I don't trust that on our servers; hardware RAID5 is a necessity AND daily backups! After all, why have we been backing up servers to tape all these years? Because experience has proven that hard drives fail, period.

Anyway, my thoughts. Feel free to use, edit or ...

Thanks. That's pretty well my thinking as well.

10:53 - Lots of suggestions on the "something blue" part, including Blu-Ray DVD technology, blue LEDs, Bluetooth, and cobalt-blue UV lights that gamers use in their cases. One that wasn't suggested, but I just thought of, is the Antec TrueBlue 550 power supply, so I may just use an image of it.



Saturday, 24 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

11:11 - I had a spare moment yesterday, so I popped the lid of the "Kick-Ass Gaming PC", which is to become the new primary Xandros desktop system in my office. I'd tried installing Xandros on it a few days ago, and ran into problems that I assumed were caused by the RADEON 9800 XT 256 MB video adapter. The system would boot and run all the initial text-mode stuff, but when it tried to go into graphics mode to start KDE it'd just start alternating between a black screen in text mode and a scrambled screen in graphics mode.

The RADEON 9800 XT is overkill for a Linux box, especially one that I won't use for gaming. So I pulled the RADEON 9800 XT and replaced it with a RADEON 9200. Xandros comes up fine with the RADEON 9200. The display is gorgeous. I was thinking about installing the free Microsoft TrueType fonts that they released some years ago. They later withdrew them from distribution, but the original license still allows anyone with a copy of them to use them on any system regardless of OS and they're still widely available for download from the Internet.

Looking at the default Linux fonts supplied with Xandros, I decided not to bother. These fonts are gorgeous, much better than I've seen with any other Linux distro, at least until I tweaked the font directory and added some MS fonts. Of course, installing Crossover Office also installs several Microsoft TrueType fonts, including the basic ones like Arial. I haven't bothered to use those on my den Xandros system. The standard Xandros fonts are more than good enough.

And the more I use Xandros the more I like it. I have a Umax Astra 3400/3450 scanner sitting here that hasn't been used for months. When I applied a service pack to Windows 2000, the scanner just stopped working. I tried reloading the drivers and everything else I could think of. Nada.

I had this scanner plugged into the Xandros box before I installed Xandros, so I moused around for a while, hoping to find that Xandros had recognized it. The only place I found the scanner listed was in the USB Devices section of the control panel. So I used Xandros Networks to search for "scanner". It found a KDE scanning application and installed it with no hassles. When I fired it up, it recognized the scanner and asked if I wanted to set it as the default. "Sure," says I, and proceed to run a couple of test scans. Everything just worked, which appears to be the norm for Xandros.

Better still, a few minutes after I stopped using the scanner, its light turned off and stayed off. That's something I was never able to make work in Windows. I actually had to unplug the scanner's power connector each time I finished using it, or the light would stay on constantly. There was a little Windows applet for the scanner that allowed me to turn the light on or off manually, which I tried using for a while. The problem with that was that even if I turned the light off manually, it'd come back on all by itself.

Hmmm. I have an HP 6200C scanner that I paid $400 for several years ago. I never was able to make it work under Windows, at all. I tried the USB interface and the SCSI interface. At one point, it kind of worked under Windows 98SE, but was completely unreliable. HP had promised Windows 2000 drives that they never delivered.

HP did eventually get around to releasing Windows 2000 drivers, something like two years after they'd promised them, but the drivers weren't available for download. HP expected people to pay for a CD, and it wasn't just the cost of shipping and handling, either. You actually had to buy the damned drivers. I never did that. I bought the Umax Astra scanner instead, and the HP 6200C has been sitting unused ever since. That was the final straw for me with regard to HP. I've never recommended an HP product since.

Just for the heck of it, I may connect that HP 6200C scanner to a USB port on the Xandros box and see what happens. I'll bet it works.



Sunday, 25 July 2004

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]