| Computer Gripes | documenting the down side of computer stuff |
|
HomeSearchMerchandiseAboutMichael HorowitzMy CNET Blog
|
| Index: A B C D E F G H I J K L M N O P Q R S T U V W X Y Z ALL |
Assorted Computer GripesThese Gripes don't fit anywhere else |
 |
September 5, 2004. I know there are things that a webmaster can tweak to influence the caching behavior of web pages on a site. But I don't know the details. I'm pretty sure that the web server has options for this and so too do META tags. Rather than research this on my own and re-invent the wheel, I asked the tech support department at three web site hosting companies that I use. What a waste of time.
You would think that tricks to save bandwidth would be something a web site hosting company is expert in. Nope.
August 10, 2004. I use a 128 MB USB keychain storage device (a.k.a. USB storage key, USB flash drive, etc) for backing up files. The files that I wanted to store on the device were using up almost the entire available space so I bought a new 256MB model. Then I copied all the same files and directories that were on the old 128 MB device.
The files now consumed only 78 MB! Before the used space on the device was about 118 MB. Same files. Sure enough, adding up the total size of each directory on the old device came to 78MB. So what accounts for the extra 40MB of used space?
It was not cluster waste, both devices were using a 2K cluster. The 128MB device was using the FAT file system. The 256MB device was using FAT32. An analysis of the 128 MB device showed it needed to be defragged in the worst way. A defrag under Windows 2000 warned that there was not enough free space to do a good job so I reformatted the 128MB device as FAT32 and copied the data again. Now it consumes the same 78MB as on the new device.
No doubt the problem was internal fragmentation in the FAT file system. A word to the wise.
Just for fun, I reformatted the 128 MB device again, this time with the FAT file system. Again, after copying the data it used 78MB, so the crucial issue is not FAT vs. FAT32 but a newly built file system vs. an old beat-up fragmented one.
When is an inch not an inch? A foot not a foot? A mile not a mile? When it appears as the rated speed or size of something in the computer industry. The first instance of this that I became aware of years ago was with CRT computer monitors. A 17 inch monitor, for example, is really 16 inches. This eventually resulted in a lawsuit and now CRT specs also include the actual visible screen size. TV makers never lied about their screen sizes. Recently I purchased an external USB Drive keychain storage device that was described as "32 megabytes". Not exactly. According to Windows 2000, its actual data storage capacity is 30.9 megabytes. Same thing with an external USB hard disk from Maxtor. It is sold as a 40 gig drive, but Windows 2000 shows its total capacity as 37.2 gig. The LAN in my living room is rated at 10 megabits/second but file transfers never achieve anywhere near that speed. In the June 11, 2002 issue of PC Magazine, Bill Howard wrote a column about USB 2.0 which is rated at 480 megabits/second. The fastest speed that he actually achieved when writing to an external USB 2.0 hard disk? 50 megabits/second. The same applies to wireless WiFi networks (I don't recall where I read the details on it). Then, of course, there is the classic about the 56K modem, which has an actual maximum speed of 53K and a real world maximum speed of 50K. May 16, 2002
FYI: Why doesn’t my hard drive have the advertised amount of space? AnandTech. November 2nd, 2002
Suit: Hard drive size does matter Reuters September 18, 2003. A group of computer owners has filed a lawsuit against some of the world's biggest makers of personal computers, claiming that their advertising deceptively overstates the true capacity of their hard drives. The lawsuit says that a hard drive described as being 20 gigabytes actually has only 18.6 gigabytes of readable capacity. A 150 GB hard drive actually has 140 GB of storage space.
FYI: Wireless LAN (WiFi) speeds are in reality half the advertised speed. In Microsoft tries Wi-Fi (September 2002) John Morris of CNET said: "Though Wi-Fi is capable of speeds of up to 11Mbps, you get only about half of that in the real world." This is also true of the faster WiFi network protocols. Woody's Windows Watch newsletter (June 2, 2003) said "the theoretical rate of 54mbs works out to around half that in the real world."
FYI: In an article in the December 2003 issue of PC World magazine (page 44), Matt Lake mentioned the actual speed of networks. 802.11g WiFi networks are rated at 54 mbps, but he says they typically run at 10 to 15 mbps. A new 802.11g router from Netgear (the subject of the article) is rated at 108 mbps, but it maxed out at 24 mbps in his testing. Wired Ethernet rated at 100 mbps has an actual maximum speed of 40 mbps.
FYI: Walter Mossberg in the Wall Street Journal commenting on phony printer speed ratings: "I tested the two printers using ...the default print setting of Normal. The Dell is rated faster than the H-P, at up to 17 pages per minute vs. 14 for the H-P. But these claims are always dubious because they are based on canned test pages and the lowest-quality settings. In my tests, both printers were much slower than these claims at the Normal print setting ...". March 27, 2003.
FYI: The July 2003 issue of Consumer Reports reviewed PDAs. They pointed out that for both Palm Pilots and Pocket PCs, you can not rely on the advertised memory capacity of the PDA to estimate how much room is left for new programs and/or data. The advertised memory capacity includes memory devoted to the OS and the installed applications. June 10, 2003.
FYI: USB Gadgets May Not Work Fast As Claims from the AP. October 3, 2003. USB 2.0 actually encompasses three speeds: the original USB speed of 1.5 Mbps; 12 Mbps (USB 1.1) and 480 Mbps. Some manufacturers have confused consumers by using the USB 2.0 label for products that don't necessarily support its fastest data transfer rates. Some devices, such as keyboards, only need the 1.5 Mbps data rate to work. And the flash memory used in many cameras and so-called keychain drives is, like many USB 2.0 consumer printers, far slower than USB 2.0's top speed.
When has Windows finished all it's start up (boot) time processing? It must be none of my business, because it doesn't say. Never has. There have been about 9 versions of Windows and not one of them says: O.K., I'm done starting up, feel free to run any programs. O.K. many boot time programs are not from Microsoft, so maybe Windows does not know when these programs have finished their processing. Still, it could at least tell the user when control is being given to these boot time programs. It wouldn't be a big deal for there to be a standard so that when each boot time program finished its startup processing, it could signal Windows of this fact. September 2003.
Keeping up with bug fixes for the software on your computer is all but
impossible. Different vendors have different means of telling their users about
bug fixes. Microsoft, has Windows Update and Office Update. Separate and
independent, a sure sign the company is too big. Even using both won't get you up to speed on bug fixes for other
Microsoft products (I'm pretty sure of this but not positive). Windows
Update often has to be run multiple times (with a reboot in between) just to get
all the current critical fixes installed. It does not tell you this. Macromedia
had a couple security bugs in the Flash player in July 2002 and needed to
contact almost every user of Windows. They didn't contact me. Of course, there was no way to contact every Flash player user. Until the PC industry fixes
this (and they won't), personal computers will always be unreliable.
(August 18, 2002)
Go to this web page
to get the updated Flash 6 player with the bug fixes from July 2002.
In the March 2, 2002 issue of his newsletter, Scot Finnie re-ran a rant of his from 1999. It was true then and is still true now. Computer vendors pre-load consumer PCs with tons of software all configured to run automatically at boot time. His article describes a number of reasons why this is a bad thing. The article is based on Windows 98 though and the issue of system resources is not as important with Windows XP or Windows 2000 as it was with Windows 9x. Still, it's a gripe that has been true for years.
When you uninstall a Windows program, the uninstall is normally incomplete. That is, more of ten than not, there are some leftover effects from the uninstalled program. These leftover crumbs might be files, changes to files or changes to the registry. The situation is so bad that a number of programs exist for the sole purpose of completely uninstalling programs that don't completely uninstall themselves. November 2001.
New versions of Windows are sold in two flavors, a cheap upgrade version and an expensive full version. This is in addition to other variations on the OS such as Windows XP Home version vs. Windows XP Professional version. The upgrade version requires that a previous version of Windows already exist on the computer. If you install the new version of Windows on top of the old one, fine. However, if you need to re-install the new version of Windows (a different gripe) or if you prefer to install the new version in a new hard disk partition, then you need to prove to it that you own an old version of Windows. In the old days, this was not a problem as computers used to come with a copy of Windows on a CD-ROM and this CD served as your proof. Now however, many if not most, computers come without a copy of the operating system on a CD-ROM. Instead they come with a recovery CD that is designed to restore the machine to its initial state as it was shipped from the factory. These recovery CD-ROMs are not valid proof of owning a previous version of Windows.
The Langa List newsletter from Fred Langa discussed a new wrinkle on this problem October 29, 2001. It seems Hewlett Packard computers no longer come with a recovery CD-ROM. Instead, system recovery is performed from a hidden partition on the hard disk. If the hard disk fails, the computer becomes a paper weight.
The article also notes that recovery CD-ROMs can wipe out all your data when they restore the computer to its state as it was shipped from the factory. He says "restore or recovery disks are not to be trusted." Now there's a gripe for you.
| Page last updated: September 5, 2004 |