It seems Java is not enabled in your browser. Please use these links instead.
  [ System] [Cache] [Memory] [Hard Disk] [Display] [Printer] [Scanner] [Modem] [Ports] [Backup]

The 'system' is the first thing we see. It's the outer shell which attracts us. That sleek black case with the rounded contours that 'would go so well next to the stereo'. If this is what appeals to you, be sure that the cosmetics don't come at the expense of performance, growth, and expandability. Direct your focus to what's inside. As with people, beauty is only skin deep. The most important things you should care about are the microprocessor,   motherboard, main memory, graphics performance, hard drive size and speed, case size, and adapter slots available. Today's choices on the market are  Pentium 'Celeron', Pentium III, and Pentium IV. There are non-Pentium (read, non-Intel) processors quite worthy of consideration, most prominent of which is AMD (Advanced Micro Devices).  There is really no reason to opt for Intel's Pentium if you trust the manufacturer who supplies the AMD system. The entry level systems sold today are largely adequate for the web designer. If I were to profile a minimum system for web design activity, I would look for the following: Pentium III processor or equivalent AMD processor running at 800 MHz  or more. That's not high-end today, as technology has raced beyond the 'gigahertz' boundary, but it is still very, very fast. My 400 MHz Pentium II is still very adequate but I become impatient when I'm working with graphics.   Memory requirements seem boundless, so get as much as you can afford. 64 megabytes is no longer adequate for graphics programs, so probably 128MB would be considered a good target to start with. All the new systems are expandable beyond that. Parity checking memory, also called "ECC" memory is a good thing to have but not widely available.   Graphics adapters should be configured for 16MB to 32MB for performance. I'd stay away from integrated graphics adapters since they tend to be low-end and not expandable.   Hard drives today are 7.5GB and up. I would want 10GB to 20GB to be comfortable. That's quite easy to achieve today since 40GB drives are selling for $200. It's smart to have 2 large drives enabling you to back up the main drive on the second. Tape backup systems for drives of these capacities are more pricey than adding a large hard drive.   Adapter slots are disappearing as cases are shrinking. I would want at least 4 adapter slots available after the vendor's configuration requirements are met. That would likely mean at least 7 adapter slots total, include the advance graphics port.   Finally, I'd want a case that's as large as possible. Seems like it's 'reaching today to find a case with more than 2 3.5in external bays and more than 2 5.25in bays. I'd never settle for those mini machines that can only take a CD-ROM and a ZIP drive, in addition to the standard diskette. Things to be added include additional CD drives like CD-RW and DVD and tape backup drives. The power supply should be 200 watts or greater and probably 2 cooling fans would be a minimum requirement. A few other thoughts: Be wary of high-styled cases. Smoked plastic, see-through pastels, and the like. Ask yourself whether you'd be able to replace or upgrade the disk drive or cd-rom supplied by the manufacturer, and at a fair price. Beige doesn't look to good with translucent magenta, does it?   If you have dogs or cats, think about the fur in your air, just waiting to clog your every computer orifice. Consider placing your tower on your desk, away from the floor and rug level. Consider, too, an optical mouse or trackball, unless you enjoy taking the mechanical mouse apart daily and pulling out hairs and other gook. A cover for the keyboard could be your best friend if you have cats.   Consider a case with a flat top. The surface will frequently come in handy for a modem, external tape or zip drive, cd-rom, or all of the above in a nice tidy stack. In lieu of that, how about your morning coffee, a couple books ... you name it. All these things are at risk on many of today's  'designer' cases.   Don't forget a surge suppressor or two. Be sure it protects your phone line as well. A favorite of mine is the APC Surge Arrest. It is D-shaped with 8 outlets. 3 outlets are on the circular part of the 'D' so 3 transformer can fit side by side. Also, 2 outlets remain hot when the suppressor is turned off, great for lamps and telephones.

 

Cache memory (pronounced 'cash') is special memory used by the microprocessor, called Level 2 cache, or just L2. (OK, there's also a Level 1 cache but all Pentiums have some and I won't talk about that here.) The processor can take advantage of L2 cache to store data prior its being needed. This ability to do two things at once, process current instructions and data and get new data in advance can be very beneficial should that new data be needed. It's already there! Access to this cache is very, very fast compared to accessing the data from system memory. (When needed data is already in the L2 cache, we have what's called a 'cache hit'.) By the way, the graphic on the left shows a microprocessor. Can't show you just the L2 cache because it's integrated and hidden away. Benchmark results have shown most emphatically that this technique can improve performance of many applications by as much as 20% or more. Too much cache is not necessarily a good thing but no cache at all is certainly a bad thing. The older 486 processors had small amounts of Level 1 cache (8KB) but I don't believe any Level 2 cache. Pentium processors and their non-Intel equivalents introduced Level 2 cache, originally optional, and then up to 256KB (usually standard), and the newer Pentiums have as much as  512KB, except the 'Celeron' version. The original Celeron Pentium had no L2 cache and suffered in the popular benchmarke. Today's Celeron Pentiums generally have 128KB and they are pretty fast. Bottom line if you're buying a used system, ask if it has Level 2 cache. If you're buying a new system, see how much L2 cache it has. Here's a link to a related article in PC Magazine which you might find interesting. I hope its location doesn't change for a while.

 

Generally speaking, more memory is better than less memory, up to a point. Any system with Windows 95/98 should have at least 32MB, a new system at least 64MB. and most do except a few leftovers and low-end laptops. What really drives the need for more system memory is the larger programs being written today, especially higher end graphics programs. I would say that today 32MB is marginal for most systems and if you're seriously involved with web design, you should consider 128MB, if your budget can withstand it. Memory has not come down in price as much as most other things, making it less affordable than it was previously. Systems today can support upwards of 512MB. (My first IBM PC in 1981 had 16KB and 256KB was the maximum ... wow!) Hopefully the newer versions of Windows will better utilize large memory than Win95/98 did, by eliminating other resource bottlenecks. There are many different memory architectures like VRAM, DRAM, EDO, SDRAM, etc. You can read about them on-line if you wish. Speed is a primary variable. SDRAM running at 100MHz was recently state of the art, now being replaced by RDRAM  in vogue on the new high-end systems. If you click the Kingston logo, look for the listing of Educational Tools on the left navigation panel. There is some very good material there. An explanation of 'parity checking memory' might be in order. Parity memory has the ability to check itself for errors and would halt the computer if it found a single error. This might sound a bit harsh, but otherwise, a memory error could show up as lost data, a system crash, or other unpleasantries. Older computers had parity checking memory but manufacturers realized they could lower costs by eliminating it. So ... they pretty much all did. The exception has been PCs used for corporate servers. They have had parity memory all along and the function of that memory has been improved as well. Finally we've gone full circle and parity checking memory for the consumer is back, albeit very quietly, from certain manufacturers only, and as a cost  option. When I saw that Dell offered it, and at a very fair price, I grabbed it. This was a major reason I bought their system and I thought it was such a shame that most people didn't have an awareness of this type of memory. The memory I purchased is the more advanced type called 'ECC', like most servers use today. Older parity checking memory could detect an error in a single bit of data (there are 8 bits to store a single character, like the letter 'A'), but it couldn't do anything about it other than stop the computer and basically suggest you should buy some new memory. ECC memory can actually detect 2 bits in error, and best of all, if only a single bit error occurred, it has the smarts to correct that 'on the fly' and let you keep on working. An important note to all this is that the system's motherboard must be able to support ECC memory and the Setup screen when you boot will ask whether you have ECC memory or not. I hope I've raised your level of awareness of this important memory feature and that you'll consider it when making a new system purchase.

 

Hard disk drives basically come in sizes and speeds, and the two are usually related ('balanced performance'?). The ways that they attach to computers today are SCSI (Small Computer System Interface) and EIDE (Enhanced Intelligent Device Electronics). SCSI drives can be faster and are usually the province of network servers. They are quite pricey even today and require their own adapter. They are the 'premium' technology and are rarely offered in sales to consumers. SCSI technology also allows the additional connection of other SCSI devices (hard drives, scanners, tape backup ...) without the need for additional adapters and without consuming a very scarce system resource called an IRQ (interrupt request line). Most EIDE drives available today are made by only a handful of companies. Price competition in this area is most intense, causing considerable financial stress and mergers among the companies. I hate to mention names because they might not be here tomorrow but Western Digital, Seagate, IBM and Maxtor come to mind readily. It used to be uncommon for a hard drive to be greater than 2GB (gigabytes, more on that later). Now, a small drive is 8GB and some are as large as 80GB. Most of you who do web design will benefit from 6GB or greater, but that's what they're selling anyway so it's not a major decision for you. (If you store a lot of graphic files or MP3 music files, their requirement would be added to my 6GB recommendation.) Upgrade pricing is really low, as you've seen if you shop the ads. Speed of a hard drive has 3 determinants: the speed of its rotation (called latency or rotational delay), the time it takes for the read-write heads to traverse the drive (called access time), and the transfer rate between the drive and the system bus. Generally a drive with a faster rotation will be faster in the other ways as well ('balanced performance' again). Most offerings today are still in the 5400 to 7200 RPM category. The drive I bought for my new system was made by IBM, runs at 7200 RPM, and really flies.  10000 RPM drives are now becoming popular on the newer and more expensive systems. To conclude this, I'll treat you to an explanation of megabytes (MB). (From the ads, you would think the drive manufacturers don't even understand this.) What aroused my interest was the advertised capacity of my IBM drive as 10MB. Imagine my surprise to learn that it was really 9.4MB. I'm not really that naive but I hadn't thought deeply about this until Windows 98 called it to my attention. A kilobyte (1KB) is not 1000 bytes, but rather 1024 bytes! Big deal you say? Well, a megabyte is not 1000 kilobytes, but rather a 'kilobyte of kilobytes' (1024x1024), or 1,048,576 bytes. See what's happening; there is now a 48,576 byte 'bonus'. If you had a choice of winning a lottery for a 'million bucks' or for a 'megabuck', by all means take the megabuck. The final extension of this is that a gigabyte is not 1000 megabytes, but 1024 megabytes, or 1,073,742,000 bytes. Therefore, my "advertised" 10GB disk drive should have been 10 gigabytes in computerspeak and 10.73 billion in decimal. Basically the ad misrepresented the capacity of my drive by NOT saying 9.4GB in computerspeak. The difference was 730 million less than advertised. Now that's a big number! If you've being following this discussion, you're doing just great! Web design is so much easier than understanding megabytes and gigabytes. So, you've now become an informed consumer and when you see a drive advertised as 20GB, you'll know the truth and that you're being "shorted" by almost 1.5GB. Some of you with a mathematical bent probably appreciate this exercise, and the others probably think I'm a bit strange, which of course I probably am.

 

When we discuss a display (a.k.a. monitor) we don't tend to think of its adapter card, but we must. No matter how capable the display in terms of, say, resolution and refresh rates, it's not going to realize its potential unless we have a fast and capable video adapter with adequate memory. Yes, video adapters have their own memory and more of it means better performance and more colors. Most of today's systems will provide an acceptable adapter with 4MB or more of memory, but you should investigate this aspect of the 'video' that comes with your system, or if you upgrade your adapter. The newest wave of Pentium motherboards have a special 'slot' for the video adapter called AGP (advanced graphics port). It's basically a PCI slot but reserved just for video. It has a speed of 66MHz, which measures the rate its data flows across the system bus. The previous speed was 33MHz, so it just doubled ... very good for performance. As for the display size itself, that's somewhat subjective, though bigger is better, up to a point. 15" displays were the standard until relatively recently, and 17" displays were quite pricey. Now, the 17" display is the standard and 19" displays are being configured with the higher end systems. 21" displays are still pricey and are overkill for most people. Technical considerations with any display must include its dot pitch (smaller is better here), and also its maximum refresh rate, which is its ability to sustain its image, especially at higher resolutions, without flicker. The higher the display's resolution, the higher must be the refresh rate to avoid flicker. The display's resolution, along with the size of the screen, contribute to what you think of as pleasing to the eye, fun to use, and ultimately contributes to your productivity, especially when doing web page design. Screen images are composed of pixels (picture elements). These pixels are horizontal and vertical. The display's 'mode' or 'resolution' determines how many pixels will be fitted to the screen at one time; you choose the value when you configure the display adapter settings. They are expressed as 640x480, 800x600, 1024x768, and higher. The higher number of each pair is horizontal. Higher resolution values mean more (and smaller) pixels. You'll find that large screens actually look bad at low resolutions because the pixels are so large that the characters become jagged. Conversely, small screens at high resolutions require eagle-eye vision to avoid discomfort. Fortunately, most video drivers supplied by the adapter manufacturer will make it easy for you to change resolutions on the fly, no reboot required. The suggestion is that experiment with the different modes on whatever size display you have and pick the largest mode you're comfortable with. On average, 15" displays are good with 640x480 and 800x600. 17" displays match up well with 800x600 and 1024x768. The bottom line here is that you would want to use the highest resolution mode your display is capable of, consistent with your eyes being comfortable. That way, you can fit more windows on your screen at once and more easily navigate between related applications, multiple browser windows, and the like. Regarding web design, that's a different story. Here, you have to consider the display that your audience is using, not your own as above. You should restrict your design to some reasonable average like 800x600 so that a majority of users can read your web pages with their present hardware and with a minimum of scrolling. The same with colors. When you select your screen mode, you have a choice of 16, 256 (8-bit), 64K ('16-bit' or 'high color') or 16M ('24-bit' or 'true color'). Pick 8-bit or 16-bit to appeal to the greatest audience, and be sure to test your web pages early in their development cycle so you can backtrack with a minimum of time loss. Be sure to do your development work at the screen mode you are targeting. A friend used a background color which was a pleasing tan on his own system, set to 24-bit color, and came out a putrid pink on another friend's system. There's a nice program I use called BrowserSizer which lets me set my Netscape window for any of the above resolutions. This enables me to see how my pages fit at the various resolutions. My own monitor is 19" and I have it set at 1024x768 resolution. If I target a web site to a design point of 640x480, BrowserSizer will model that for me. Otherwise I would have to change my physical screen resolution to 640x480 and that would be very ugly and unproductive. One word of caution. Be careful about buying a display based on its size alone. Some companies will entice you with a cheap 17" or 19" display, trying to attract you to its size alone. Definitely see a demonstration of the display set at 1024x768 and see if it's readable. A better display should have a dot pitch of .25mm to .28mm, with .26mm being a good average. I've seen some cheap displays with a dot pitch exceeding .40mm. You will certainly see the difference and in fact, probably won't be able to use such a display at higher resolutions, even if you have a good video adapter. The characters will be muddy because the dot pitch is so wide that the display hardware can't separate the pixels enough for the image or text to be clear.

 

This is a pleasant topic to discuss because printers have come down in price so much and there's a great selection available. Just about everyone will want a color printer because we're living in a world of color and the web is so colorful. The utility value of color and the sheer fun it makes available speaks for itself. A black and white greeting card??? The major manufacturers have color ink-jet printers in the $100-$300 range which are excellent. The primary determinants of price are speed and resolution, rather than quality and durability. If you do a lot of photo-image printing, you will want a printer in the upper end of this range because they can offer a special capability which uses smaller ink droplets and can actually drop different ink colors on top of each other (layering). This provides a colored image (even on plain paper) which compares well against an actual photograph. The coloring and resolution are superb. I've bought an several HPs with this technology have been very happy with them. The best thing I can suggest is that you see demonstrations of your favorite manufacturer, if you have a favorite, and maybe compare their  test sheets. There are printers from these same companies priced considerably higher than $300 and generally that is not because they're any better, but rather they may be faster, accommodate large paper sizes, have larger ink cartridges for network sharing among many users, etc. Ink-jet printers are not the only printers worth considering, despite their color advantage. Laser printers have come down in price as well and their resolutions can compare quite favorably to or exceed the ink-jet printers I've discussed. However, they are black and white, with color laser printers still very pricey. You'll find that the cost of printing on a B&W laser is very much less than ink-jet, on the order of maybe 2-3 cents a sheet compared to 15-20 cents. I have both ink-jet and laser, and pretty much use the laser for personal correspondence, business writing and documentation, and quick snapshots of web pages. I use my ink-jet for labels, greeting cards, and generally for fun. For some of you this might appear obvious, but here are two cost-saving tips for those who only have ink-jet printers, or who are considering buying one. If you check your printer driver's settings, you should find an option to use 'grayscale'. This will save you considerable drain on your somewhat more expensive color cartridge, and is yet quite adequate for printing web pages which have a thirst for color. You'll also find an option to lower the print intensity, great for rough drafts or for printing web pages, whether in color or not. I print lots of things at 50% ink intensity and am quite satisfied with their clarity and contrast ... and they print so much faster. Finally, here's a caveat! Before you buy any vendor's ink-jet printer, investigate the cost of their ink cartridges relative to its capacity. What some companies are doing is offering a 'choice' of an economy cartridge and a larger size cartridge on their newer printers. They also may only offer the smaller cartridge. The economy cartridge costs very close to what the larger cartridge sold for.  The bottom line here is that they're going to charge you more for ink so they can recover the costs of selling you a heavily discounted printer. (We're not talking small change here, either. The economy size may have 21 ml. vs the full-size of 42 ml., and for $5 less.) Certainly reminds me of Polaroid almost giving their camera away for $20 to sell you film for $12. If you're a careful shopper, you'll find certain models which use the higher capacity cost-effective cartridge, but the manufacturer is not going to be of any help.

 

Just a few years ago scanners were the province of the graphic artist and the printing & publishing industry. With the popularity of the Web and the birth all its 'budding' web publishers (that's you and me), scanners have become an everyday tool. Their prices have come down very dramatically as well, making them affordable to everyone. It's hard to believe that there are so many quality scanners available for under $100, like several Visioneer & Microtek models. I can't imagine not having a scanner just for convenience copying and faxing, let alone its important contribution to web design.  Having seen some of the clip-art collections available, where there might be 100 worthwhile images in a collection of 125,000 pieces, there's a feeling of power having my scanner at the ready. Now,  web graphics are available to me in every magazine, newspaper, and photo album. In my topic Using Text Fonts I needed a way to show an example of a font manager's printout. It didn't take me long to realize that I should just print it, scan it, and show it that way. That's the magic of having a scanner. We let it sit idle, usually switched off for days on end, until suddenly it calls to us to solve an immediate requirement. Possibly the most difficult thing about using a scanner is learning the software and techniques essential to manipulate the scanned image and turn it into something useful, even profound, to adorn our web page. Editing an image was overwhelming to me. It had a jargon all of its own with its lassos, histograms, bezier curves, unsharpening masks, color indexes, and on and on, a language with which I was not familiar. I needed a kickstart,    a way to feel comfortable with the different graphic formats (BMP, GIF, JPG, rasters and vectors, et cetera) and gain an understanding of video and printer resolutions as they applied to scanning requirements. In my quest to get my arms around this, I ran into a fellow on the web named Wayne Fulton. He not only had expertise I needed, but he is an excellent teacher and is very generous with his time. One of the biggest favors I can do for you, worth the entire cost of admission to my own site, is to pass you along to Wayne. I am providing his link here and again in Sites to Visit. Wayne is also modest and calls his site 'A Few Scanning Tips', a dramatic understatement of what he offers us. He has his own domain now, ScanTips.Com. He's so good that Microtek refers to him from their website. Do visit him, and if you like the information on his site, he places all that tutorial material in a book he prepares himself for something around $20. I don't want to get involved with product recommendations since I have only used a couple of Microteks. Wayne Fulton liked them and that was enough for me. I have since bought a couple more for my children and they took to it quite easily. Their ScanWizard software driver is widely recognized in the industry. Microtek shows a lot on their web site and you probably would want to visit it, not just for their product information but for the educational material and links they provide. So ... now that I just lied and made a product recommendation, I'd like to share a few opinions and tips. Definitely consider a few features you want in a scanner while you're shopping. Speed is very important and varies from one scanner to another. Parallel port scanners were the most prevalent until the USB port came along on the Pentium II computers. Parallel port speed is very fast in the EPP or ECP mode of the modern computers (since 1995). USB is possibly even faster. Some scanners have both a parallel and USB connection and all things being equal, I would opt for that.   The characteristic called Optical Resolution can be important as well. That's the 300-600 dpi or 600-1200 dpi you read in the ads. 300-600 dpi is nearly obsolete now and 600-1200 is very adequate for most scanning requirements. Scanners sold today should all have that resolution. As I explain in other topics of this site, I scan most material at the higher 600 resolution and then resample the image smaller using PaintShop Pro. Web sites only use 75-92 resolution but I would prefer to capture more data using the scanner and then have PaintShop discard what it doesn't need when it resamples.   Be aware of whether your scanner has an on/off switch. Many do not and that's something you may not realize until it's unboxed at your home. Since you will not be using your scanner that often, it's nice to be able to leave the cord in the electrical outlet and shut it off from the switch. Some scanners will go into standby mode and turn their lamp off after say 15 minutes, but to me, it's still ON and I don't like that.   Parallel port crowding. USB ports take some pressure off the single parallel port which comes with most computers. It's common however to have a need for multiple parallel ports for things like scanners, zip drives, tape backups, and of course multiple printers. Parallel scanners provide a 'pass-thru' port to connect your printer to the scanner in order to connect it to the single parallel port. This can be problematic,  and with the setups I've seen require your scanner to be on all the time for the printer to work. Look at my item here called Ports  for my recommendation of a terrific space-age printer switch from 'Belkin' I have three of them in my family and we all love it.   Finally, understand how the case comes apart so that you can clean the glass. My scanner had dust under the glass when I bought it and when I asked Microtek how I could clean it, their response was 'you don't, it will void your warranty'. Well, for $50 or so, I'm not going to worry about voiding my warranty. I found it to be very simple and safe to remove the upper glass and clean both sides. 90% isopropyl alcohol and a soft lint-free cloth is generally recommended.   You may have realized that I never mentioned "single sheet vs flatbed". I'm a flatbed bigot and don't want a scanner that has to move the document perfectly in defiance of gravity, and I don't want a scanner that limits me to single sheets of paper, and not books and other 3D objects. Also, they're not very convenient for repeated scans of the same item and they're not really less expensive any longer. They do take up considerably less space, however.

 

Obviously all of you have modems, but perhaps some are considering an upgrade to a newer and faster one. I'll share some opinions I have on this subject, and a few tips. If you can avoid it, don't buy a Winmodem. I believe this is U.S. Robotics primarily, but anything advertised as a Winmodem or Windows modem is not 'pure'. It requires Windows software support in order to operate. This is not a good thing since performance may be impaired based on other things taking place in your computer. Other programs in your computer can more easily break your connection as well. It is always better to have hardware do as much as possible without assistance from the computer. Finally, some  software won't work with such a modem when it's written to interface directly to the hardware modem. An example of this is 'ModemShare 32' by Artisoft. If you're on the market for a new modem,  56K is the only option these days. But if you can pick up a 33k modem from a friend, consider taking it. In many cases you won't notice the difference between it and a 56K modem. Actually the 56K is really 53K and it applies to downloading, not uploading. If you already own a 28.8 or 33.6 modem, I don't see a compelling case to scrap it for a new 56K one, since the speed of the internet is so variable anyway. Don't buy an internal modem, unless you have a good reason! The only valid reason which comes to mind is the situation where you have an older 486 PC, say from around 1993 or so. These systems may have a slow serial port, designated UART8250, which doesn't support the newer fast external modems reliably (28.8K and up).  All internal 56K modems have their own serial port built-in and of course it's the fast one, called UART16550. If you want to check the speed of your serial port, go the Control Panel of WIN95/98, click on the Modem icon, click the Diagnostics tab, select or highlight your modem and click the button for 'More Info'. It will tell you the UART level, among other things. Why don't I like internal modems? That's a fair question. I can provide six reasons from my experiences.   Well, they're harder to use. You can't see what they're doing, what speed they're running at, whether they've got the telephone line 'off hook', etc. Indicator lights can be most comforting when you have a problem. It's extremely reassuring for a user to see the lights when things aren't going so well.   Also, sometimes you can't disconnect your modem from the line, and the only surefire way to reset or reinitialize your modem is to reboot your computer. With an external modem, just turn it off and back on.   Another disadvantage is that they generate some heat, and for my money any way I can keep additional heat out of my system, I'm for it.   If you live in an area prone to power surges from storms, you're much safer having a modem connected to your external port, rather than sitting right there on your motherboard where it can fry your whole system. I suffered damage to my I/O adapter on my old Gateway from a nasty storm where it jumped from the external modem to the adapter card.  I shudder to think what would have happened to the system if my modem were installed internally. (Whichever type of modem you use, you should have a surge suppressor with RJ11 phone protection.)   Another reason, perhaps a critical one, is that an internal modem takes one of your limited adapter slots. Why give up such a scarce resource when every computer has a serial port sticking out the back, just begging to be used for something?   Finally, an external modem is quite portable. You can share it with other computers you may have, easily migrate it to your next new computer, use it on your laptop if you have one, or in the case of troubleshooting a problem, you can take it to a friend's computer and see if it works OK there.

 

We've talked a bit about serial ports in the modem section, but what about parallel ports? Talk about scarce resources, this can be a real problem if you've bought a scanner, tape backup system, Zip drive, Jazz drive, a second printer, and so on. I have 2 printers, a Zip drive, and a scanner, all clamoring for attention from my one measly parallel port. Do I see some of you smiling in acknowledgment? Well, there are 4 ways to get some relief. 1) Use the 'pass-thru' port on your scanner or zip drive to support your printer, 2) Install a small network and put a printer or Zip/Jazz drive on another system and share it, 3) Buy a parallel printer switch and place it on your LPT1 port. 4) Buy a dual printer adapter (ISA or PCI) and add 2 more ports to your system. Obviously, if any of this was a no-brainer, I wouldn't be writing this section on Ports. Let's look at the alternatives and see what may be best for you. By the way, I have implemented all 4 approaches and can give you the benefit of my direct experiences. 1. This the easy, low cost solution, but it comes at the price of limited application and inconvenience. It only applies if you have 2 devices, one of them has a pass-thru parallel port, the other device is a printer, and that pass-thru port works properly with the printer. It may not work fully, if at all, in support of a bi-directional device built to the IEEE 1284 standard, like many ink-jet printers. Also, the device doing the pass-thru may have to be switched on, like my Microtek scanner, and I definitely did not want to turn my scanner on every day, all day, just so I could print from time to time. 2. Unless you're running a home office and need several networked computers on at all times, this is a higher cost, inconvenient solution.  Also, some devices like scanners cannot be shared on a network and the bi-directional capabilities of sophisticated printers (like page duplexing) won't work across the network. By the way, the Zip drive and devices like it can be shared since they're seen as hard drives. 3. This is the choice which makes the most sense, but it is not without its problems. The most important consideration is that most switches on the market used to attach 1 computer to multiple parallel devices, just don't work with devices like bi-directional ink-jet printers, Zip drives, scanners, and similar devices which require bi-directional capability. They were largely used to support a couple simple printers, like older model laser printers and dot matrix printers, or they were intended to share a single printer with multiple computers which is much easier to do. So, you need to buy a switch, be it manual or automatic, which can support the IEEE 1284 bi-directional standard. I am going to make a suggestion here of a specific product from Belkin which has worked great for me. I have tried different switches and they did not work, including a different one from Belkin. Belkin has 2 products called Bitronics Autoswitch Kit. The first supports 2 devices from 1 computer, their part no. F1U125-KIT. The second is a follow-on  product which supports 4 devices, their part no. F1U126-KIT. Both come with a 6-foot IEEE 1284 'straight-thru' cable to connect the switch to the parallel port. They are NOT automatic like their names imply, when connecting 1 computer to 2 or 4 devices. They are  automatic when connecting 1 device to 2 or 4 computers. The switch is extremely easy to use and works with probably all parallel devices. On the 2-port model, the ports are A and B and are quickly selectable by software. They supply a handy utility program which sits on your desktop to do the switching. If you prefer, there's a pushbutton on the switch you can use, if it's within reach.  Their literature can have you believe that no power need be applied to the switch in "most cases", although there's an AC input jack. Trust me that "most cases" will likely apply to you, so plan to buy yourself a 9V Power Adapter (min 500mA). A good one I'll recommend is Radio Shack's 9V 800mA Power Adapter (#273-1770). The Belkin switch uses Radio Shack's Adaptaplug "M" ( 273-1716), included in the $14 price of the adapter. By the way, I have since bought this switch for each of my children and a neighbor (to share a scanner with an HP DeskJet) and everyone is happy. Here are links to Belkin and Radio Shack for your convenience.     4. This is also a very good choice because it eliminates the extra cable from the switch to the computer, as well as the power supply. It is also a lower cost solution than the switch. I purchased a SIIG Dual Parallel Pro (ISA), model IO1828. It's a great product, supports the IEEE 1284 devices, not just printers, and supports a very wide range of IRQs, DMA channels, and I/O port addresses. The downside here is that you must give up a motherboard expansion slot, in this case ISA, and to install the second of its 2 ports, you must give up an additional bracket opening on the back of your computer case. This is a bit nebulous, but scanners and zip drives can be real fussy about the I/O addresses they use, so you may have to diddle with the addresses to get everything working. Here's SIIG's website. There's a PCI adapter available also. Finally, I want to address the subject of IRQs. Your system will have a very, very limited number of these available, possibly none at all. Your computer's CMOS (Setup) has probably reserved IRQ 7 for the built-in printer port and normally you'd have to dedicate IRQ 5 for LPT2, and even another IRQ for LPT3 (if you need an LPT3). This requirement will be impossible for most of you to manage, so let me give you a tip. From what I've been reading and this is reflected in my actual experience, Windows 95 and 98 do NOT use IRQs to support the printers. And indeed, the Iomega Zip and Microtek scanner did not seem to require an IRQ either. I set my jumpers on the SIIG card to IRQ 7 for both LPT2 and LPT3, and my system CMOS was also setting LPT1 to IRQ 7, and everything works without conflict. Even though I had LPT2 and LPT3 jumpered for IRQ7, no IRQ shows up in device manager. Until recently,  my environment used both implementations 3) and 4), because I had 4 parallel devices. I had my scanner and laser printer on the Belkin switch on LPT2, my zip drive on LPT3, and my ink-jet on LPT1. Recently I bought a USB Zip Drive and didn't need the switch any longer, so I gave it to my son. This means that my next computer, which won't have any ISA slots will require me to buy the SIIG PCI Parallel Adapter. I will probably buy the 4 port Belkin switch instead since I never use my scanner and printers simultaneously. I hope that some of you will find these tips and revelations as interesting and informative as I did while I experienced the discomfort of setting up my own system.

 

A primary transgression committed by many computer users is not backing up their computers. 'Backup' implies 2 things: the system and the data, and these are often treated individually because of their relative importance. Consider what would happen, in the extreme case, should your hard drive fail. You've just lost everything! Out come the stacks of CD-ROMs, diskettes and manuals as you try to reinstall everything you 'used to have'. Of course you have to remember what you had, and you will not remember everything until you try to use it and it's not there. Once you've reinstalled all your software, you'll now find that many do not work properly because they're not customized with your various preferences. You've also lost your browser bookmarks, e-mail logs and address books, and many other valuables. It can literally take weeks to put humpty-dumpty back together again. And by the way, it doesn't take a disaster to cause you to have to reformat your hard drive. Many software vendors who are trying to support you with some problem or another just love to tell you to reformat your hard drive. In just about every case, that's unreasonable of them to insist on this but nonetheless, they're under 'orders' to ask you to do it. And generally if you don't listen to them they won't help you further. As for backing up your data, need I say more? I'm sure each of us has lost something we sure wish we hadn't. Won't make that mistake again, will we? So what do we do about this? My inclination is to buy a tape backup system, one that's as fast as possible for the money you can afford. Tape backup is slow but it is comprehensive. These tape backup systems may take hours to backup a hard drive which contains 2 or 3 gigabytes of data. And then again almost as much additional time to have it verify what it just wrote to the tape. As a consequence of this, we generally backup everything around once a week and then maybe every few days, we do an 'incremental' backup which just backs up files which have changed or been added, which takes little time at all. Tape backup systems cost very little these days but make sure you get a good one and absolutely test its software to make sure it does work thoroughly and that it backs up Windows' Registry and any files that are marked as 'hidden' or 'used by system'. There are alternatives to tape but few really good ones. High capacity removable hard disks, like Iomega Jazz, are costly and have limited capacity compared to tape. CD-ROM drives have come down in price, as has the CD-ROM media, but 700MB capacity will likely not help much, except for your data. With hard drives so inexpensive now, it's not foolish at all to have a second one to backup the first. There are programs available (e.g. from PowerQuest) that specialize in hard drive to hard drive backups. Consider theft and fire damage here, however. That's why the best idea is always a tape that you take with you, far away from the computer, maybe to the office or a friend's house. My computer has a Seagate 20GB tape drive attached to the secondary hard disk controller. It can back up at 60 or so MB per minute. This works well for me since I only back up around 3GB of system and data combined ... takes a little over an hour and a half, including a verification pass. Today's super-large drives of 45MB or so, are not going to be backed up regularly, unfortunately. They would take too long and the tape systems to handle that quantity would exhaust your budget. However, very few people are using those capacities unless they're accumulating large numbers of graphic files or MP3 files. If that's your case, just place those files in a separate disk partition (drive letter?) and back them up separately, say once a month or less often. So ... enough already ... I've probably made some of you feel a bit nervous. If you only use 5 or 6 applications, write some notes about their customizations (or backup just their customization files if you know what they're called), make sure that you have the original installation CDs and diskettes (and any updates to them), along with your Windows 95/98 CDs, and perhaps you can just take your chances. Regarding your data, of course we're not going to take any chances! I've seen much pain, grown people cry, and workers fired. 'Data' is defined as 'your stuff'. All the stuff that's no longer there if your hard drive gets replaced, or your computer has a major accident, a program misfires (and Windows is not beyond misfiring), or you just plain make a mistake and delete something you should not have. Remember now that "data" includes your browser bookmarks and especially your e-mail files and address books. And, if you're into web design, you surely want your site(s) backed up. Until recently, I was getting away with backing up my data on diskettes every time an important file changed. And to be honest with you, I wasn't faithfully backing up everything on tape every week ... I've gone as long as a couple months until my conscience bothered me. In some cases my data from just a single application had grown too large for a diskette or 2. Finally, I grew tired of changing labels on diskettes and having them scattered all over the place, spreadsheets on this one, tax forms on that one, letters on those two, etc. For me, the Iomega Zip drive was the cure, especially since I already owned a couple from my consulting activity. Its 100MB capacity was very adequate for all my data and all my websites. Because it operates like a disk drive, rather than a complicated backup program with all its startup options, it is so easy to drag and drop my data from my hard drive to my zip drive. It's a good idea to keep several 'generations' of zip disks in case the last one goes bad, you'll still have the one before it. And naturally if any of your data is extremely precious, you can easily keep a zip diskette away from your home or office in case of theft, fire, vandalism and other unpleasant events. Many computer systems today come with an option for an internal zip drive at 100MB capacity or 250 MB. It's also quite easy and inexpensive to buy a 100MB USB connected Zip drive for around $80. There's no reason why you should ever have to suffer a loss of your personal files, and I don't need to describe to you the feeling of losing the tax return you labored 20 hours on, or the digital pictures of your European vacation. I'll wrap this up now with a tip for you. If you can't see your way clear to make an additional purchase just now, at least use the Windows 95/98 'Backup' program or the WinZip program. They're both quite good and support diskette drives. They can compress your data so that you can get maybe 50 to 100% more on a diskette and it can also span diskettes automatically if a single diskette can't hold everything. At the least, you can go to sleep tonight with some of your important files protected. Believe me, you'll sleep so much better.