The squeeze is on -- and that's where the magic is


April 12, 1993|By MICHAEL J. HIMOWITZ

In the weeks since Microsoft released DOS 6, I've had a lot of questions about what many users see as black magic: the operating system's alleged ability to double the size of your hard disk.

Here's the truth: A hard disk is a hard disk, and it can only store as much raw information as the manufacturer designed it for. If you have a 120-megabyte drive, the standard on many PC's sold today, it can store only 120 million characters of data.

But it is possible to compress and encode data so that it takes up less space than it normally would -- like the old peanut brittle can that holds a half-dozen rubber snakes that pop out when you open it. If you can compress your programs and data files so that they only occupy half their normal space, you've effectively doubled the size of your disk drive.

Why bother? Because programs and data files are so much larger now than they were a few years ago that they threaten to overrun advances in technology that have created the big but affordable hard drives that come with most computers today.

Microsoft Windows and a standard suite of business software -- word processor, spreadsheet, data base, communications, graphics and accounting programs -- can easily occupy 90 megabytes of disk space. Desktop publishing eats up even more: A single scanned photograph can use more than a megabyte.

Now to the magic. In the mid-1980s, programmers devised ways to compress files, largely to minimize telephone and connect-time charges for people using modems.

These programs use a variety of mathematical techniques, but basically they search your files for repetitive information, such as 10 carriage returns, common words and phrases, repeating spaces or repetitive black dots in the background of a graphic. They replace this information with "tokens" that take up far less space.

Spreadsheets, graphics and data files compress best this way, sometimes shrinking to 25 percent of their former size. Program files don't compress as well, but overall, it's quite possible to achieve compression ratios of 2-to-1.

The best known of the stand-alone compression programs is PKZIP, from Pkware Inc. PKZIP, which not only compresses files but also can merge multiple files into one large file, is the standard for file exchanges on electronic bulletin boards and online information services.

The problem with PKZIP and similar programs is that you can't use the files in their compressed state. You have to decompress them first, which is awkward for files that you need every day.

A few years ago, software publishers began developing programs that run in the background and compress and decompress files on the fly. The best known of these is Stacker, from Stac Electronics. Stacker and others of its ilk (including Microsoft's DoubleSpace), essentially take all the files on your disk and compress them into one large file, along with as much empty space as you decide to allot.

To make this process transparent to you and me, these programs fool the computer into thinking that the large, compressed file is actually a disk drive with far more space than the actual drive.

But there's no free lunch. First, these programs require internal memory, typically 40K to 50K, that could be used by other software. Computers with 80386 and 80486 processors can store compression software in so-called upper memory, where they're out of the way, as long as there aren't too many other programs competing for the same space.

Second, there's a performance penalty, because programs must decompressed when you run them, and data files have to go through repeated compression and decompression as you work with them. With today's fast microprocessors, the penalty isn't a large one, but it can be noticeable with programs that frequently access your disk drive.

Finally, until Microsoft made disk compression part of the operating system itself, these programs required some "swapping" of disk drives and multiple rebooting any time you changed your system configuration files.

Although it isn't as sophisticated as Stacker, Microsoft's DoubleSpace makes the whole process painless. I let the program's express setup do the work on my 210-megabyte main drive. It took DoubleSpace three hours, but when it was finished, the capacity of my 210-megabyte hard disk had increased to 376 megabytes. DoubleSpace also created a "new," additional hard drive with about 12 megabytes of uncompressed space for files that don't work well with compression software.

Some programs, including Windows, took a second or two longer to load after the installation. A few, such as Borland's Paradox data base for Windows, showed a noticeable performance lag -- but not enough to make me give back 166 megabytes of free disk space.

One potential problem involves DoubleSpace's estimate of the remaining space on your hard disk. Because it doesn't know ahead of time how well future files will compress, the program makes its best guess based on the compression ratios it has achieved so far.

By invoking a command, you can see the current compression ratio as well as the estimated ratio that DOS uses to tell you how much free space you have. You can then adjust the estimate.

Another problem occurs if you decide you want to decompress your entire disk. While Stacker provides an "unstacking" utility that will decompress your files and remove all traces of the program, Microsoft's DoubleSpace requires that you copy all your files to a floppy disk or tape backup, delete the compressed file, and restore all your programs.

This can be a pain, I can't see any reason to remove DoubleSpace. To add another disk drive with the same capacity as DoubleSpace would cost $300 to $400. As part of the $50 DOS 6 upgrade package, DoubleSpace was a bargain.

(Michael J. Himowitz is a columnist for The Baltimore Sun.)

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.