by Craig Zacker and Kyle Bryant
One of the oldest relics of DOS that has been retained in Windows 98 is the way that the operating system handles its local disk storage. The file allocation table (FAT) file system and the use of drive letters to represent the disk drives installed in the computer are unchanged from earlier versions of DOS.
However, Windows 98 has not stood idle in this respect. The capacity of hard disk drives has grown faster than virtually any other area of computing technology. Also, there is now a whole new breed of storage devices, such as CD-ROM and removable cartridge drives, that are rapidly becoming standard equipment on new systems. Windows 98 supports today's large capacity drives, as well as a wide range of other storage devices.
The steady rise in capacity and sophistication of the local storage devices in the typical PC has forced operating systems to change in order to keep up with the technology. Most new systems today are equipped with a CD-ROM drive in addition to the standard floppy and hard disk drives. Many are also shipping with a removable cartridge drive, usually an Iomega Zip drive, as standard equipment.
In addition, the capacity of the typical hard disk drive has skyrocketed to the point at which drives holding from two to eight gigabytes or more are common, even on entry-level machines. Windows 98 supports a wide array of storage devices, and yet it maintains the familiar drive access paradigms at the Windows command line and in the graphical interface.
The floppy disk drive was the original storage medium for the personal computer. At one time, most users ran their applications directly from floppies, and if they were lucky, they had a second floppy drive to store their data. Today, as a software distribution medium, the floppy disk is virtually obsolete because of its limited capacity and the size of current applications.
As a case in point, the core Windows 98 installation files would require more than 70 floppy disks, and even that number would be higher if Microsoft didn't use a special format (CAB) to store 1.7 megabytes on each disk. When compared to a single CD-ROM that can store much more data and be produced much more cheaply, the choice is obvious.
As demonstrated by Microsoft, the physical disk can be formatted to a greater capacity, but the format is a proprietary read-only one that is designed primarily to inhibit the copying of disks. IBM introduced a 2.88 megabyte floppy disk drive standard in 1991, but like some other IBM innovations, it never caught on with other manufacturers to any significant extent.
All DOS and Windows operating systems reserve the drive letters A: and B: for floppy disk drives. Today, floppy drives serve primarily as an alternative boot device in case of a hard drive failure. Most systems are configured to boot from the floppy if a disk is present in the drive. This is why the Windows 98 installation process strongly recommends that you create an emergency startup disk.
Hard disk drives are, of course, the core of a PC's data storage subsystem. Hard drives usually offer the greatest capacity and the highest speed of any storage medium in the machine. To keep up with ever larger applications and data files, hard drives have had to become faster and more capacious at an incredible rate. At the same time, the price of hard disk storage has plummeted: A one gigabyte drive that sold for $2,000 in the early 1990s can now be had for less than $150.
This need for increasingly larger hard drives has resulted in serious inefficiencies when they are used with legacy file systems such as FAT16, which was originally developed for use with DOS. Windows 98 addresses these problems with a new file system, FAT32, that makes it possible to use large hard disk drives more efficiently.
The large capacity of today's drives and the limitations imposed by existing file systems often make it necessary for users to divide hard disk drives into separate partitions. A partition is a portion of a hard disk that is devoted to exclusive use by a specific file system. When you create multiple partitions on a single disk and all are compatible with Windows 98, the operating system mounts each one using a different drive letter.
NOTE: A hard disk storage device consists of the disk itself, which takes the form of a stack of platters, and the drive that spins the platters and moves the heads to the appropriate position. However, common terminology among DOS and Windows users also uses the word drive to refer to a logical drive, represented by a letter that the operating system assigns to a partition as it is mounted. Thus, you may hear the physical device referred to as a hard disk or a hard drive, but although there is only one mechanical drive in the unit, there may in fact be several partitions and, therefore, several logical drives. The first logical drive designation in a sequence of drives is usually referred to as Drive 0, followed by Drive 1, Drive 2, and so on.
You can also create partitions for different file systems on the same disk. You can, for example, create one FAT32 partition for use with Windows 98 and another NTFS partition for use with Windows NT. However, to access a partition, you must load an operating system that supports its file system. In this example, each operating system will be able to access its own partition, but not the other partition because Windows NT doesn't support FAT32 and Windows 98 doesn't support NTFS. If you used FAT16 instead of FAT32, however, Windows NT could access both partitions.
NOTE: Windows NT 5 is expected to provide support for the FAT32 file system.
When a partition is not supported by an operating system, it is effectively invisible. In the example just presented, each operating system mounts its own partition as the C: drive, no matter where each partition is physically located on the disk. Each operating system has its own function or utility for creating partitions on hard drives. NetWare, OS/2, and other operating systems all have their own file systems.
In a very short time, the CD-ROM drive has become a common part of the average PC configuration. Because most applications and operating systems are now distributed on CD, a CD-ROM drive is essential on a standalone system. On business networks, CD-ROM drives tend to be less prevalent because software can be installed and distributed over the network.
Originally, CD-ROM drives required SCSI (Small Computer System Interface) subsystems, which made them more costly and difficult to configure. However, the adoption of the EIDE (Enhanced Integrated Drive Electronics) interface has enabled CD-ROM drives to share the same host adapter used by the EIDE hard drives that are typically used in desktop PCs.
For a CD-ROM drive to operate on an EIDE interface, the device must be supported by the system BIOS, just like the hard disk and floppy drives. This support makes it possible for the CD-ROM to be used as a boot device, simplifying the installation of operating systems onto new systems.
NOTE: For older systems that do not have BIOS support for EIDE devices, you can purchase an expansion card that effectively replaces the system BIOS and permits the connect of additional devices such as CD-ROM drives.
When Windows 98 detects a CD-ROM drive on the system during the installation process, it mounts the disk using the next available drive letter after the floppy drive and hard disk partitions. You can access files on the CD-ROM using Windows Explorer or the My Computer window just as you would those on a floppy or a hard drive--except of course that the CD-ROM is a read-only medium.
NOTE: Writeable CD-ROM drives are now available at reasonable prices, along with software that enables you to write to the disk by dragging and dropping files in Windows Explorer. These are known as write-once/read-many, or WORM, devices. In addition, a new breed of rewriteable CD-ROM drives are now on the market, but the disks that they create generally are not readable by standard read-only drives.
Although the technology has been available for many years, rewriteable, removable, magneto-optical cartridge drives are now becoming increasingly popular options on desktop PCs. Many desktop systems aimed at the home consumer market now include a cartridge drive as standard equipment.
The Iomega Zip drive is rapidly becoming a de facto standard for removable data storage. To their advantage, the drive and the media are both inexpensive, their proliferation makes it easy for users to exchange files, and their 100 megabyte capacity makes them a more practical alternative than floppy disks.
Iomega also produces higher capacity cartridge drives, such as the Jaz drive, available in both one gigabyte and two gigabyte versions. The capabilities of these devices rival those of hard drives in speed and flexibility. Cartridge drives can use either the EIDE or SCSI interface to connect to the system. There are also slower models that connect through the computer's parallel port, which makes it very easy to move the drive from system to system.
Like the other drive types discussed in this section, Windows 98 mounts cartridge drives using a standard drive letter. Larger capacity drives, such as the Jaz, even support the creation of multiple partitions, enabling you to use them just like hard disk drives.
However, using multiple partitions on removable media can cause a rearrangement of the system's assigned drive letters, depending on how many partitions exist on a particular cartridge. The drive letter of a CD-ROM mounted after a removable drive, for example, changes depending on the number of DOS partitions on the cartridge that is currently loaded. You can avoid this problem by configuring Windows 98 to permanently assign drive letters to specific devices. To do this, open the System Control Panel and select the Device Manager tab. Open the Properties dialog box for the removable device, click the Settings tab, and select a range of drive letters. After you reboot the system, those letters will not be assigned to another device.
Various manufacturers continue to develop new types of storage devices for use with PCs, hoping to see them come into popular use and become a standard in the way the Zip drive has. The most promising of these is the DVD drive.
Originally named the Digital Video Disk, the acronym is now often called the Digital Versatile Disk because of its adaptation to use on computers. The DVD medium looks just like a CD-ROM except that it is double-sided, and each side can have two separate layers. By adjusting the focus of the laser that reads the disk, the device can read a separate data stream on each of the layers on each side. This increases the potential capacity of the disk enormously.
DVD was originally intended as a read-only video transport medium. It used MPEG-2 compression to store more than two hours of video, plus high-quality audio and other material on a single disk. Movies are already becoming available on DVD, and many PC vendors are offering the drives on their high-end systems.
As a data storage medium, DVD has great potential for use in the PC world, but it's still too early to gauge its success. There is not yet a ratified standard for the technology, and a recordable version of DVD is currently in development, which could end up replacing the read-only version in general use. Investing in DVD technology today might be akin to the old days of investing in 1X CD-ROM drives.
To view, create, and manage a disk drive's DOS partition table in Windows 98, you still use FDISK. FDISK is a command-line utility inherited from DOS that has changed little in appearance over the years. The primary difference in the Windows 98 version of FDISK is its capability to create partitions by using the new FAT32 file system.
NOTE: FAT32 was actually introduced in the OEM Service Release 2 of Windows 95, but Windows 98 fully integrates it into the operating system by enabling you to convert existing FAT16 partitions to FAT32.
Under Windows 98, no matter which of the FAT file systems you use, the partitions that FDISK creates on a disk drive are referred to as DOS partitions. FDISK creates a partition table on the drive in the master partition boot sector, which is the first sector on the drive. This table lists the locations of the other partition boot sectors on the drive. Each partition boot sector precedes the section of the disk that has been allocated for that partition.
Each partition boot sector contains information that defines the size and nature of a particular partition. The partition boot sectors and the master partition boot sector are strictly DOS conventions. Other non-DOS file systems create their own partitions and have their own methods of allocating space on a drive.
NOTE: All Windows 98 hard disks and some types of removable disks have a master partition boot sector and one or more partition boot sectors. Floppy disks, however, do not have these and, as a result, cannot contain multiple partitions.
Since DOS 3.3, it has been possible to create multiple DOS partitions on a single hard disk. The first partition that you create is called the primary DOS partition, and any others are called extended DOS partitions. In contrast, some other file systems (such as NetWare's) allow you to create only one partition of that type per disk.
As long as it has sufficient capacity, you can use one hard disk drive to boot several different operating systems, each of which uses its own file system. You can create DOS partitions of any size, leaving unallocated space on the disk, and then use the partitioning utility from another operating system to create partitions for a different file system in that empty space.
Each area of a hard disk that is allocated as a DOS partition by the partition boot sectors begins with a DOS volume boot sector. After you make a particular DOS partition active (or bootable) with FDISK and you install the DOS, Windows 95, or Windows 98 system files on it, the system BIOS passes control of the machine to that partition's DOS volume boot sector each time the system restarts. When this happens, the code in the DOS volume boot sector runs and attempts to load the system file IO.SYS from the root directory of the partition. If IO.SYS cannot be found, the boot fails. To view a complete summary of a hard disk's allocated file structure including numbers of files and directories, use the DOS Directory (DIR) command with the /A /S and /V command-line switches.
NOTE: Because floppy disks have no partition boot sectors, the DOS volume boot sector is always placed on the first sectors of the disk. That enables the floppy to function as a boot device just like a hard drive. It is the system BIOS that determines whether control of the machine should pass to the floppy drive or to the active partition on the hard disk.
To create DOS partitions on a hard disk, you must run the FDISK utility from the DOS command prompt. You can run FDISK from a Windows 98 DOS session, but only if you do not intend to work with the currently active partition. Obviously, if you destroy the partition containing the system files and operating system the computer is currently using, the system will halt as soon as you exit FDISK.
After you create a partition, exit FDISK, and reboot the system, you will find that the system has assigned the new partition a drive letter. You can switch to that drive letter from the DOS prompt, but you cannot read from it or write to it because the drive has not yet been formatted. You must use the FORMAT utility from the command line, the Windows Explorer, or My Computer to format the drive before you can use it to store data.
In most cases, it is best to run FDISK from a boot floppy, so that you can continue to access the machine even if there are no partitions on the hard disk. You can create a Windows 98 startup disk by launching Add/Remove Programs from the Control Panel and selecting the Startup Disk tab (see Figure 9.1).
FIG. 9.1 A Windows 98 startup disk contains all the tools you need to boot the system and partition and to format its hard disks.
The Windows 98 startup disk includes the system files needed to boot the computer to the DOS prompt, as well as the FDISK and FORMAT utilities you need to prepare the hard disk for a full Windows 98 installation.
If your system contains a hard disk drive with a capacity greater than 512 megabytes, you see the following message when you run FDISK:
Your computer has a disk larger than 512 MB. This version of Windows includes improved support for large disks, resulting in more efficient use of disk space on large drives, and allowing disks over 2 GB to be formatted as a single drive.
IMPORTANT: If you enable large disk support and create any new drives on this disk, you will not be able to access the new drive(s) using other operating systems, including some versions of Windows 95 and Windows NT, as well as earlier versions of Windows and MS-DOS. In addition, disk utilities that were not designed explicitly for the FAT32 file system will not be able to work with this disk. If you need to access this disk with other operating systems or older disk utilities, do not enable large drive support.
Do you wish to enable large disk support (Y/N)...........? [N]
Your response to this question determines whether or not FDISK will create FAT32 partitions on your hard disks. Under normal circumstances, drives with a capacity of less than 512 megabytes cannot use FAT32. On the other hand, if you have a large capacity drive and want to create a partition larger than two gigabytes in size, you must use FAT32. Otherwise, see "Choosing a File System," later in this chapter, for information that can help you decide whether or not you should use FAT32. This initial screen is the only indicator that the Windows 98 version of FDISK provides support for FAT32. The screens that follow are the same, whether or not you enable large disk support.
The main FDISK screen appears, displaying the following information:
){kind=link}
Microsoft Windows 98
Fixed Disk Setup Program
(C)Copyright Microsoft Corp. 1983 - 1998
FDISK Options
Current fixed disk drive: 1
Choose one of the following:
1. Create DOS partition or Logical DOS Drive 2. Set active partition 3. Delete partition or Logical DOS Drive 4. Display partition information 5. Change current fixed disk drive
Enter choice: [1]
The following sections provide details about each of FDISK's available options.
When you select the Create DOS Partition or Logical DOS Drive option, FDISK presents the following options for creating a new partition on the selected hard disk:
Create DOS Partition or Logical DOS Drive
Current fixed disk drive: 1
Choose one of the following:
1. Create Primary DOS Partition 2. Create Extended DOS Partition 3. Create Logical DOS Drive(s) in the Extended DOS Partition
Enter choice: [1]
If the disk currently has no DOS partitions, you must first create a primary DOS partition. If the disk does already have a primary DOS partition and there is still unallocated space, you can create an extended DOS partition. Finally, if the disk already has an extended DOS partition and you want to divide it into two drive letters, you can create logical DOS drives in the extended partition.
When you create a new partition, FDISK first asks you if you want to use all of the space available on the disk and make the partition active. This enables you to create the most common partition configuration in one step. If you say no, you are prompted to specify the size of the new partition in megabytes. Alternatively, you can specify a size using a percentage of the free space available by including the percent symbol ("%") after the value.
FDISK also prompts you to specify a volume label for the partition. This label can contain up to eleven numbers or letters used to identify the partition. The Windows Explorer and the My Computer window display the volume label along with the drive letter for each local drive on the system.
After you create a partition on your system's first or only hard disk, you will most likely want to make that partition active so the system can boot from it. When you choose the Set Active Partition option, you are prompted to select the partition that will hold the system boot files. Only one DOS partition on your computer can be active at any given time.
Making a partition active does not in itself make the disk bootable. It only means that the BIOS will turn over control of the system to that partition at boot time. To make the disk bootable, you must also place the system boot files on that partition by formatting the drive with the system files option or by using the SYS utility.
If you select the Delete Partition or Logical DOS Drive option from FDISK's main screen, you'll see a list of the partitions and logical drives on the selected disk. This list includes all the DOS partition types, as well as partitions created by other file systems (which are identified as non-DOS partitions).
Deleting a partition is always a big step that should cause you to stop and carefully consider your actions. Deleting a partition destroys all the data that is currently stored there and, in the case of an active partition, the system boot files as well. FDISK repeatedly prompts you to confirm your actions before it deletes the partition, including a requirement that you enter the volume name of the partition to be deleted. This virtually ensures that entering a random series of accidental keystrokes cannot cause you to lose your data.
Obviously, you must back up any important data before you delete the partition where it resides. It is also important to be sure that you have a means of booting your system after the partition is gone. If you delete the active partition, you must have a boot floppy disk, such as a Windows 98 startup disk, to restart the system.
If you find yourself in a situation in which you must boot from the startup disk, keep in mind that you might still have access to your CD-ROM drive. Whether or not you do depends upon your CD-ROM drive and the limited driver support provided by Windows 98. Review the CD-ROM drivers on your startup disk and try a test startup disk boot to determine whether or not your hardware is compatible. Note that Windows 98 loads a RAM drive as part of the startup disk boot sequence. This might cause the usual drive designation letter of your CD-ROM to change.
CAUTION: Many of the personal computers sold today use a small non-DOS partition to store a configuration program for the system BIOS. This program is activated when you press a particular key combination during the system boot process (the exact key combination varies depending on the BIOS manufacturer). Be sure not to delete this partition unless you are certain that you have the configuration program on some other medium. Otherwise, you might be unable to access the information in the BIOS.
When you choose the Display Partition Information option, you see a list of all the partitions on the selected disk. The list should look similar to this one:
Display Partition Information
Current fixed disk drive: 1
Partition Status Type Volume Label Mbytes System Usage C: 1 A PRI DOS DRIVE1 2014 FAT32 100%
Total disk space is 2014 Mbytes (1 Mbyte = 1048576 bytes)
This example shows that fixed disk drive 1 contains a single 2,014 megabyte partition that constitutes 100% of the space on the disk. It is a primary DOS partition that has been set as active and will use the FAT32 file system. FDISK has assigned this partition the drive letter C: and the volume label DRIVE1.
FDISK is a program that many people look upon with fear and trepidation because of the catastrophic damage that can result from accidental or improper use. The Display Partition Information option is totally safe, however, because this screen contains no controls that can affect the state of the disk.
FDISK offers the Change Current Fixed Disk Drive option on its main screen only when it detects more than one supported disk drive in the system. FDISK can address only one drive at a time and defaults to the first drive in the system BIOS. To manage the partitions on the system's other disks, you must first use this option to select the desired drive.
Selecting the option displays a list similar to the following that outlines the disk drives in the system along with their disk numbers, partitions, drive letters, sizes, and the percentage of the disk occupied by each:
Change Current Fixed Disk Drive
Disk Drv Mbytes Free Usage
1 2014 100%
C: 2014
2 2014 2 100%
D: 2012
To change to another disk, you select it by specifying the disk number shown in the Disk column. Once you have selected another disk drive, it becomes the default for all of FDISK's other functions until you change it or exit the program. Most of the other screens contain a line that specifies the number of the disk that the program is currently addressing.
When you create one or more DOS partitions on a hard disk, you must format them before you can use them to store data. The formatting process divides a partition into clusters. A cluster is a logical unit that represents the smallest amount of disk space that can be allocated at one time. A disk formatting program determines the size of the clusters to be created based on the size of the partition and creates an entry in the FAT for each cluster. Windows 98 includes the character-based real-mode FORMAT program inherited from DOS, as well as protected-mode GUI alternatives in the Windows Explorer and the My Computer window.
TIP: You must reboot your system after you create a partition but before you format it. The Windows 98 version of FDISK does not reboot the system automatically after it creates a partition.
The Windows 98 FORMAT.EXE program operates from the DOS command line and provides more options than the GUI formatting utility. The program includes all the switches from the previous versions of the utility included with the DOS operating system, although many of them are obsolete unless you use the older floppy disk types. Despite its similarity to previous versions, however, it is imperative that you use the Windows 98 FORMAT program on disks with FAT32 partitions.
The syntax for FORMAT is as follows:
FORMAT drive: [/V:label][/S][/B][/Q][/C]
These are the switches:
| drive: | The variable drive is replaced by the drive letter of the disk to be formatted. |
| /V[:label] | The /V switch specifies the volume label that is to be assigned to the disk being formatted, where label is replaced by a string of up to eleven characters. |
| /S | The /S switch causes the program to write the system boot files to the specified disk after the formatting process. |
| /B | The /B switch causes the program to reserve sufficient space on the formatted disk for the later addition of system boot files. |
| /Q | The /Q switch causes the program to perform a quick format on the selected disk by overwriting its FAT table. This removes all of the existing files on the disk, but does not check the clusters for damage. You can perform a quick format only on a disk that has already been formatted. |
| /C | The /C switch causes the program to test the clusters on the disk that have already been marked as "bad" by a previous format. |
[/F:size][/T:tracks][/N:sectors][/1][/4][/8]
Those switches function as described here:
| /F:size | The /F switch specifies capacity of the floppy disk to be formatted, where size is replaced by a value in kilobytes or megabytes, such as 160, 180, 320, 360, 720, 1.2, 1.44, or 2.88. This switch is not needed when formatting standard 1.44 megabyte 3.5-inch floppies. |
| /T:tracks | The /T switch enables you to specify the number of tracks on each side of the floppy disk to be formatted. |
| /N:sectors | The /N switch enables you to specify the number of sectors on each track of the floppy disk to be formatted. |
| /1 | The /1 switch causes the program to format one side of a floppy disk. |
| /4 | The /4 switch causes the program to format a 5.25-inch 360K floppy disk in a 1.2 megabyte high-density drive. |
| /8 | The /8 switch causes the program to format a floppy disk with eight sectors per track. |
CAUTION: The /Z switch is a powerful option that is not recommended for use on production systems without extensive testing. The switch does not override the 65,536 cluster limit on FAT16 drives, so it should generally be used only with FAT32. Specifying a smaller than normal cluster size may lessen the amount of disk space wasted by partially filled clusters, but it can also create enormously large FATs that severely affect the performance of the file system.
Windows 98 also includes a GUI-based formatting utility that's accessible as a drive letter under Windows Explorer or My Computer. Figure 9.2 shows the GUI display.
FIG. 9.2 The Windows 98 GUI disk formatting utility provides access to the same major options found in the command line FORMAT program.
In the Format dialog box, you can select the capacity of the disk to be formatted and the type of format (quick, full, full with system files, or system files only). You can also specify a volume label for the disk. The utility will not allow you to format the hard disk drive where the Windows 98 operating system files are stored and will prompt you to confirm your actions before it destroys any files on a disk that has already been formatted.
When you elect to include the system files during a format by using the /S switch with the FORMAT program or by activating the Copy System Files check box in the Format dialog box, the following boot files are written to the disk after the formatting process is complete:
The first three files are flagged with the system attribute, which you must remove before you can delete them. COMMAND.COM functions as a normal executable and is launched whenever you open a DOS session from the Windows 98 GUI. Together, these four files occupy 393,728 bytes, which is a substantial amount of space on a floppy disk. Unlike the boot sector and the FATs, which reside outside the disk area that is visible to the operating system, the system boot files are visible and can be manipulated by conventional means. However, such manipulation can damage the disk's capability to boot the system.
During boot time, the system BIOS in nearly all PCs is configured to check the first floppy drive for the presence of a disk containing system boot files. Therefore, you can use any floppy disk containing those files to boot the system by inserting it into the A: drive. On a hard disk drive, however, the mere presence of the system boot files on the disk is not sufficient to boot the system. You must also configure the partition containing those files to be active, as detailed in the section, "Making a Partition Active," earlier in this chapter.
In addition to its partition boot sector and its DOS volume boot sector, each DOS partition also contains two copies of its file allocation table, or FAT. The FAT is a matrix that correlates the files and folders stored in the partition with their physical locations on the hard disk. Two consecutive copies of the FAT are stored in the disk area before each partition. Like the boot sectors, the FATs are stored outside the disk area that is visible to the file system.
NOTE: Although the data structures that comprise a partition's boot sectors and FAT tables cannot be displayed or manipulated by the standard Windows 98 file management utilities like Windows Explorer, they are not inaccessible. Many products on the market address disk devices directly instead of through the file system. These products, called disk editors or sector editors, enable the direct manipulation of data in a disk's boot sectors and FATs. In many cases, professionals use them in a last ditch effort to recover the data on a drive that has already been damaged by an attempt to manually repair a FAT.
When files are written to a disk, they do not necessarily occupy a contiguous space equivalent to the size of the file. Instead, the files are broken up into clusters of a given size (sometimes called allocation units), which may be scattered all over the partition. As a result, the FAT is not a list of files and their locations as much as it is a list of the clusters in the partition and their contents. Whenever you access a file on a given partition, Windows 98 accesses the directory entry for that file, which contains the starting cluster. This cluster points to the entry in the FAT representing the next cluster in the chain. Each successive cluster then points to the next until the entire file is read.
Each entry in the FAT is a 12-, 16-, or 32-bit hexadecimal number, the size of which is determined by the FDISK program--even though FORMAT actually creates the entries. All floppy disks and hard disk volumes under 16 megabytes in size use 12-bit FAT entries. Hard disks and removables with volumes that are 16 megabytes or larger typically use 16-bit entries, but a volume larger than 512 megabytes can now use Windows 98's FAT32 file system and 32-bit FAT entries.
The size of the clusters on a given partition is determined during its creation by the size of the partition that you specify in the FDISK program. The intent is to strike a balance between the number of clusters listed in the FAT and the size of those clusters. This need for balance is the source of the FAT file system's biggest problem--a problem that is addressed by the new FAT32 file system. FAT32 does not completely resolve the problem, but it does provide a greater flexibility of solutions.
Obviously, the smaller the clusters used on a partition, the more clusters that partition can hold. The more clusters there are on a partition, the larger the FAT table must be and the longer it takes for the operating system to search the table for the information it needs to access a file. On the contrary, when you create larger clusters, the FAT table becomes smaller and more manageable, but you also increase the amount of disk space wasted by clusters that are only partially filled.
A file can be of almost any size, but when Windows 98 writes it to disk, it splits the file into a series of uniform clusters. The result is that the last cluster is almost never an exact fit. The end of the file is written to the last cluster and the remainder of that cluster stays empty (see Figure 9.3). This slack space is wasted as long as the file remains on the disk.
FIG. 9.3 Because the FAT file system divides the disk into clusters of a uniform size, disk space is wasted whenever a cluster is not completely filled with data.
How much partition space is wasted is determined by the relationship between the size of the clusters and the size of the files. If, for example, you have large files and small clusters, relatively little space is wasted. However, the combination of small files and large clusters (a more common occurrence) produces the exact opposite effect. When you store a large number of 10KB files on a partition with 32KB clusters, more than two thirds of the allocated disk space is wasted.
Although the problem of disk space being wasted by partially filled clusters has existed as long as the FAT file system has, it was exacerbated to an alarming degree by the ever-increasing capacity of the typical hard disk drive. As disks and the partitions on them grow larger, the cluster size must also grow to prevent the FAT from becoming too big. FDISK uses the following cluster sizes for FAT16 partitions of various sizes:
| Partition Size | Cluster Size |
| 128MB | 2KB |
| 256MB | 4KB |
| 512MB | 8KB |
| 1GB | 16KB |
| 2GB | 32KB |
CAUTION: FDISK uses the literal definitions of kilobyte and megabyte. That is, one kilobyte equals 1,024 bytes, and one megabyte equals 1,024 kilobytes or 1,048,576 bytes. Some hardware vendors have been known to make their drive capacities seem larger by defining a kilobyte as 1,000 bytes and a megabyte as 1,000 kilobytes or 1,000,000 bytes. Be sure to take this into account when you are trying to create partitions with a particular cluster size.
When you purchase a new PC today, it is not uncommon for the machine to be equipped with a hard disk drive holding six or eight gigabytes or more. The primary reason for FAT32's existence is to prevent users from having to divide a device that large into three or more separate partitions. FAT32 can support drives holding up to two terabytes of data (1 terabyte equals 1,024 gigabytes) and single files up to nearly four gigabytes, so it is likely to remain a viable tool for some time to come.
At the same time, FAT32 eliminates the file allocation table's 65,536 cluster limit, making it possible to use smaller clusters than FAT16 partitions of the same size can use. This results in many more clusters and a larger FAT, but FAT32 can also dramatically reduce the amount of space wasted by partially filled clusters. FAT32 partitions of various sizes use the following default cluster sizes:
| Partition Size | Cluster Size |
| less than 260MB | 512 bytes |
| 260MB-8GB | 4KB |
| 8GB-16GB | 8KB |
| 16GB-32GB | 16KB |
| greater than 32GB | 32KB |
FAT32 also includes other improvements besides larger partitions and smaller clusters. Since its inception, the FAT file system has always used two identical file allocation tables for each partition. The two FATs are written consecutively in a contiguous area of the disk just before the beginning of the partition itself. During normal operations, the operating system updates the second FAT by copying data from the first on a regular basis.
FAT16, however, takes very little advantage of this redundancy. A FAT16 file system will use the second copy of the FAT if damaged disk sectors prevent access to the first copy, but it does not access the second copy of the data if the first simply becomes corrupted. In fact, the file system is far more likely to continue to update the second FAT, overwriting it with the corrupted data from the first copy. Some third-party disk utilities can take advantage of the second FAT and use it to repair the first, but only if they perform the repair operation before the second FAT has been damaged, too.
FAT32 takes greater advantage of the two FAT copies. In any situation in which the file system finds the data in the primary copy of the FAT to be unreadable, it switches to the secondary FAT, which then becomes the primary. This is in direct contrast to FAT16, in which the first FAT is always the primary. In addition, FAT32 provides greater control over the FAT mirroring process. The file system can temporarily disable the process by which the data from one FAT is replicated to the other.
The modification in the behavior of the FATs results in a greater degree of fault tolerance--without the need to interrupt system operations and apply third-party utilities.
Despite the first entry in the cluster size table (shown earlier, in the section "Introducing FAT32"), the Windows 98 FDISK utility in its default configuration will not create FAT32 partitions smaller than 512 megabytes. Indeed, when it detects a drive smaller than 512MB, it does not even present the option to enable support for large drives.
However, certain third-party products on the market, such as Partition Magic 3.0 (www.powerquest.com), make it possible to convert FAT16 partitions smaller than 512MB to FAT32. In addition, an undocumented switch for FDISK enables you to create new FAT32 partitions smaller than 512MB. By running FDISK with the /FPRMT parameter, you can specify a FAT32 partition size smaller than 512MB.
CAUTION: FDISK's /FPRMT parameter is undocumented. Experiment with it at your own risk. It should not generally be used on disks containing vital data. Because hard disk drives smaller than 512 megabytes are becoming an increasingly rare sight, the compatibility of these drives with FAT32 should not be a major issue. In addition, because a 512MB drive with a FAT16 partition already uses 8KB clusters, the savings realized by FAT32 will not be significant.
After you enable large disk support in FDISK, the rest of the program functions identically with both FAT16 and FAT32 partitions. To verify that a partition you have created is using FAT32, you can either use the Display Partition Information option on FDISK's main screen or right-click a drive letter in Windows Explorer or My Computer window and choose Properties from the context menu. In either case, the display should show the selected drive as using the FAT32 file system (see Figure 9.4).
FIG. 9.4 The Properties dialog box for a hard disk drive identifies the file system that has been used to create the partition.
Because FAT32 operates at the file system level, it does not have compatibility problems with most applications (whether they are shrink-wrapped or custom written). The sole exception to this is a type of applications that address the storage devices directly, such as disk utilities and some antivirus products.
CAUTION: Disk tools designed for use with the FAT16 file system, such as sector editors, disk repair tools, and defragmenters, should absolutely not be used on FAT32 drives. They can potentially cause serious damage to the partition, including data loss.
Windows 98 includes new versions of ScanDisk and Defrag, in both real and protected mode, that are designed to work with FAT32. If you are familiar with the functionality of ScanDisk, you'll easily understand how converting a FAT16 drive to FAT32 affects the program's diagnostic process. Whereas the examination of the FATs on a FAT16 drive takes only a few seconds, the larger tables on a FAT32 drive of the same size take longer because of the greater number of entries.
While the Windows 98 disk utilities are certainly adequate, many users have come to rely on the greater functionality provided by third-party disk repair and utility products. Most manufacturers of these tools have by now released versions of their products that support FAT32 drives. Be sure, however, to verify that a disk utility supports FAT32 before you actually use it. The damage you could cause might be irreparable.
The primary drawback of FAT32 is unquestionably its incompatibility with previous versions of the DOS and Windows operating systems. FAT32 was originally released as part of the OEM Service Release 2 of Windows 95, but the original Windows 95 release cannot access FAT32 partitions. Likewise, all previous DOS versions and all current versions of Windows NT are incompatible with FAT32. A FAT32 drive will not be able to serve as a boot device for a Windows NT system.
Because of this incompatibility, you cannot configure a system to dual boot Windows 98 and Windows NT or Windows 3.1 using a FAT32 drive to store both operating systems. You can, however, create separate partitions for the two file systems and boot each one from its own drive.
Here is a list of other issues you should consider before making the decision to convert to FAT32:
CAUTION: Using Drive Converter to format a removable disk as FAT32 might void your removable disk's manufacturer warranty. Check with your removable disk manufacturer before proceeding.
When Microsoft first released the FAT32 file system as part of the Windows 95 OEM Service Release 2 product, the only way to use it was to delete the existing partitions on a disk drive and create new ones. Obviously, this resulted in the loss of all the data and, therefore, the need for a complete backup and restore procedure. Windows 98 remedies this problem by providing a FAT32 conversion wizard, known as Drive Converter (FAT32), that enables you to preserve your data while gaining the advantages of the new file system. To access the wizard, you can launch the Drive Converter program from the Start menu's Programs, Accessories, System Tools program group, or you can execute CVT1.EXE from the Run dialog box (Start, Run).
The wizard provides you with information about the new file system and the possible consequences of the conversion. After you select a drive to be converted (see Figure 9.5), the wizard scans the drive for applications that might cause problems with FAT32 and then gives you the chance to remove them before proceeding. The wizard also gives you the opportunity to back up your data with Microsoft Backup, and it warns you repeatedly about the consequences of your actions, requesting your confirmation before proceeding.
FIG. 9.5 Windows 98's FAT32 converter implements the new file system on selected drives without the need for repartitioning or reformatting.
The actual conversion is a DOS process. After you complete the wizard's information and configuration screens, it reboots your system to the DOS prompt. Because the converter must manipulate the existing data on the drive as it creates new clusters, the process can take many times longer than it would to partition and format an empty drive. The length of the conversion process is, of course, dependent on the number and size of the files on your drive, as well as the speed of the device itself. It is not uncommon for a FAT32 conversion to take a considerable amount of time.
CAUTION: You should take whatever steps are possible to ensure that the conversion process runs from start to finish without interruption. A power failure during the operation might not be avoidable, but you can use common-sense precautions (like connecting the system to a UPS if possible and not converting your disks during a thunderstorm) to lessen the chances.
When you have partitioned and formatted your disks, you can use the standard Windows 98 disk tools to store and manage your files, regardless of which file system you have elected to use. From the drive letter icons in Windows Explorer or My Computer, you can copy files, directories, and entire disks by dragging and dropping them onto other drives or directories.
By default, when you drag a file or directory to another location on the same drive, Windows 98 moves it, deleting the file or directory from its original location. When you drag a file to a location on another drive, Windows 98 copies it, leaving the original in place.
In addition to the Windows 98 graphical file management tools, you can also use the traditional DOS COPY and XCOPY programs. COPY is an internal DOS command that's designed to copy one or more files to a new location on the same or a different drive. XCOPY is also a holdover from DOS, but it is an external program that's used to copy entire directory trees.
The syntax for COPY is as follows:
COPY [/A:[[-]rhsda] /C /E /H /K /M /N /P /Q /R /S /T /U /V /X /Z] ¬source[+]...[destination]
These are the parameters and switches:
| source | The source variable is replaced with one or more filenames. Standard DOS wildcards are permitted. If the filenames are separated by plus signs (+), the files are joined into a single destination file. |
| destination | The destination variable is replaced with the name of the file, directory, or DOS device to which the source is to be copied. When the destination is omitted, the source is copied to the current default directory. |
| /A:[-]rhsda | Copies the source files flagged with the attributes specified after the /A:. If /A is used but no attributes are specified, all files are copied, regardless of attributes. |
| /C | Copies a file only if a duplicate file in the destination directory is older than the source file. |
| /E | Suppresses the display of non-fatal error messages during the copy process. |
| /H | Copies files flagged with the hidden and system attributes. Normally, COPY ignores files with these attributes. |
| /K | Causes the copied files to retain the DOS read-only attribute if it is present in the source files. |
| /M | Copies only files that are flagged with the archive attribute. |
| /N | Parses the COPY command for testing purposes without actually performing the copy. |
| /P | Prompts the user for confirmation of each file copied. |
| /Q | Suppresses the display of filenames and totals during the copy process. |
| /R | Prompts the user for a confirmation before overwriting files in the destination directory. (By default, COPY silently overwrites files.) |
| /S | Parses the subdirectories of the source for files to be copied, creating identical subdirectories at the destination as needed. |
| /T | Displays the total number of files that have been copied, but suppresses the display of individual filenames during the copy process |
| /U | Copies files that are newer than files of the same name in the destination directory as well as files that don't exist at the destination at all. The /C switch does not copy files that don't exist at the destination. |
| /V | Verifies that the data copied to the destination directory is readable. |
| /X | Removes the archive attribute from the source files after they are successfully copied. |
| /Z | Causes COPY to overwrite a read-only file of the same name in the destination directory. |
XCOPY source [destination] [/A /C /D[:date] /E /F /H /I /K /L /M /N /P /Q /R /S /¬T /U /W /Y /-Y
| source | The source variable is replaced with the name of the file or directory to be copied. Standard DOS wildcards are permitted. |
| destination | The destination variable is replaced with the name of the file or directory to which the source is to be copied. |
| /A | Copies only the source files that are flagged with the archive attribute, without changing the attribute at the source. |
| /C | Continues the copy process after encountering errors. |
| /D[:date] | Copies only the files with a date more recent than that specified by the date variable. If no date is specified, only source files newer than the destination files are copied. |
| /E | Creates all subdirectories found at the source to the destination directory, even empty ones. |
| /F | Displays the complete path names of files and directories as they are copied. |
| /H | Copies files flagged with the hidden and system attributes, in addition to all other files. |
| /I | Assumes that a nonexistent destination is a directory when copying multiple files. By default, XCOPY prompts the user to identify a nonexistent destination as a file or directory. |
| /K | Maintains the attributes of the source files in the destination directory. |
| /L | Parses the XCOPY command and displays the files to be copied without actually performing the operation. |
| /M | Copies only the source files that are flagged with the archive attribute, and then removes the attribute at the source. |
| /N | Copies files to the destination using DOS 8.3 filenames. |
| /P | Prompts the user for a confirmation before copying each file. |
| /Q | Suppresses the display of filenames during the copy process. |
| /R | Overwrites destination files flagged with the read-only attribute. |
| /S | Copies all subdirectories of the source to the destination, except empty ones. |
| /T | Creates the directory structure of the source in the destination directory but does not copy files or empty directories. |
| /U | Overwrites older files in the destination directory. |
| /W | Prompts the user to press a key before beginning the copy process. |
| /Y | Overwrites files that already exist in the destination directory without confirmation. |
| /-Y | Prompts the user for a confirmation before overwriting files that already exist in the destination directory. |
Windows 98 includes DriveSpace 3, a disk compression product that can increase the capacity of your disk drives by 50% to 100%. As with many of Windows 98's optional features, however, you should carefully consider whether using DriveSpace on your system is a good idea.
Unlike static compression programs such as PKZIP, DriveSpace compresses files automatically as they are saved to your drive and then decompresses them on-the-fly when you request access to them. Naturally, this introduces an additional level of processing overhead that can diminish the overall performance of your system.
Compression programs function by scanning files for redundant bit patterns and replacing them with codes that take up less space than the patterns themselves. The degree of compression the program achieves is based on the nature of the files being compressed. If, for example, you apply DriveSpace to a drive that contains mostly ZIP and GIF files, you will see little gain because these file formats are already compressed and cannot be reduced any further.
Bitmap and database files, on the other hand, contain a large number of redundant bit patterns and often can be reduced to one-fifth their original size or smaller. Executables and dynamic link libraries typically fall between these two extremes, compressing at a ratio of approximately 2:1.
DriveSpace also can save disk space, even when you choose to create a new drive without actually compressing the files. Instead of using the standard FAT16 cluster size, which can be as large as 32KB, DriveSpace breaks files down into 512-byte pieces in order to minimize the wasted space caused by partially filled clusters.
DriveSpace 3 is a new version of the DriveSpace program that was included in Windows 95. Originally released as part of the Windows 95 Plus! product, this new version can create compressed volumes up to two gigabytes in size, while the Windows 95 version of DriveSpace was limited to 512 megabyte volumes.
Dynamic compression technology has been around for several years, and at one time, it was a sound method for improving the economy of PC data storage. Today its value has lessened--for several reasons. First, the price of hard disk storage has plummeted in recent years, to the point at which a new two gigabyte drive can be had for one-tenth the price of a slower model five years ago. Therefore, it might be more economical to simply purchase additional storage for your system instead of reducing your productivity by adding compression.
Second, you can improve the storage efficiency of your hard disk drives without the aid of compression by using the FAT32 file system included with Windows 98. The additional available disk space you realize with FAT32 will not be as great as what you could achieve with DriveSpace, but even a fast system will exhibit degraded performance under DriveSpace.
If system performance is of paramount importance to you, you might find the performance penalty suffered under DriveSpace to be unacceptable. Carefully consider your other alternatives (converting to FAT32 or adding a new hard drive to your system) before you make the decision to use compression.
CAUTION: DriveSpace compression is not compatible with the FAT32 file system. If you have any plans to use FAT32, you should not apply DriveSpace compression to your drives.
Finally, like FAT32, DriveSpace renders the data on your compressed drives inaccessible to operating systems that do not support the technology. If you ever plan to boot your system with other operating systems, carefully consider whether you will need access to the files on the drives you plan to compress.
NOTE: DriveSpace 3 can read disks that have been compressed with the earlier DriveSpace versions included with MS-DOS 6.22 and Windows 95, as well as disks that use the DoubleSpace technology that was part of earlier MS-DOS versions. If you want, you can upgrade these older drives to DriveSpace 3 in order to take advantage of its additional features. However, it's a one-way compatibility: The earlier compression products cannot read disks compressed with DriveSpace 3.
When you compress the C: drive on your system with DriveSpace, you end up with a C: drive that can hold 50% to 100% more data. But what you are seeing is not really a drive at all. During the compression process, DriveSpace creates a file on your drive called DRVSPACE.000. This file is called a compressed volume file, or CVF. It is the CVF that actually contains your compressed data.
When you view the contents of your C: drive after the compression process is complete, you are actually viewing the contents of the CVF, which exists as a file on your physical disk drive that's flagged with the read-only, hidden, and system attributes. You will also find that DriveSpace has added another drive letter to your system. This new drive is your original disk, now called the host drive, with all of its data migrated to the DRVSPACE.000 file. In essence, DriveSpace has created a new virtual drive on your system and then switched the drive letters so that your data appears to be in the same place it always was.
You do not have to compress your entire drive when using DriveSpace. You can specify a part of the free space on the disk and create a new compressed drive out of that space. You can create several smaller compressed drives this way. Each one has its own CVF, called DRVSPPACE.001, DRVSPACE.002, and so forth.
TIP: If possible, avoid compressing your Windows 98 boot drive; compress another partition instead in order to store infrequently used data or archive files. This lessens the DriveSpace performance penalty.
You can also use DriveSpace to compress floppy disks, and the process is essentially the same as that for a hard disk drive. A new drive letter appears on the system, representing the host drive for the floppy. You continue to access the floppy disk using the A: or B: drive letter as before, but that is actually a view of the CVF. The floppy can hold more data as a result of the compression, but it cannot be read by a system that doesn't have DriveSpace3 installed.
When the compression process begins, DriveSpace creates the CVF file and begins compressing the files already stored on the disk. By moving newly compressed files to the CVF until it is full, the program clears space on the host drive, which it uses to increase the size of the CVF. The process then repeats until all the files have been compressed and migrated. The smaller the amount of free space on the drive when the compression process begins, the more times the CVF has to be expanded, and the longer it takes to complete the entire operation.
It is not possible to compress a drive that does not have any free space because DriveSpace must have room to work. A 1.44 megabyte floppy, for example, must have at least 512 kilobytes of free space in order to be compressed.
NOTE: You can compress floppy disks and removable cartridges just like any other disks. However, when you insert a compressed disk into the drive after the system has been booted, you must mount the drive before you can access its files. To mount a drive, highlight the drive letter in the main DriveSpace window and select the Advanced, Mount command. You can dismount the disk in the same way.
To compress a drive, you launch the DriveSpace application from the Start menu's Programs, Accessories, System Tools program group. Figure 9.6 shows the DriveSpace 3 screen. This window lists the drives in your system, shows their compression status, and enables you to perform most compression operations. The display also includes any host drives created by previous compression operations. If a host drive is left with less than two megabytes of uncompressed free space after the compression process is complete, it is hidden from view from Windows Explorer and My Computer.
FIG. 9.6 The DriveSpace 3 dialog box is the main control center for compression operations.
When you select a drive for compression, DriveSpace launches a wizard that walks you through the process. The wizard shows you the current state of the drive and the estimated amount of free space that will result from the compression process (see Figure 9.7).
FIG. 9.7 DriveSpace graphically displays the estimated results of performing the compression process.
If you click the Options button, you can select the drive letter that will be assigned to the host drive, the amount of free space to be left on the host drive, and whether it should be visible in the Windows 98 file management utilities.
CAUTION: When you compress an entire drive, it is recommended that you leave the host drive letter invisible. It is possible to manipulate the CVF like any other file on the disk, but in this case, you are essentially dealing with all your files in one package. If you were to damage the CVF, change its attributes, or accidentally delete it, you could lose all your data.
Because DriveSpace cannot create compressed volumes larger than two gigabytes in size, it takes a different tack when faced with a hard disk drive larger than one gigabyte. Because the program estimates a compression ratio of 2:1, it compresses one gigabyte of the space on the drive, expecting to create a compressed volume of approximately two gigabytes. The compressed volume will contain most of the existing data on the disk, and DriveSpace creates a new host drive containing free, uncompressed space.
For example, as you can see in Figure 9.8, a two gigabyte D: drive with 91.56 megabytes free is converted to a two gigabyte compressed drive (also called D:) that contains all the original data from the disk. The free space on the compressed drive only increases to 127.81 megabytes, but the host drive, X:, is left with 987.25 megabytes of free uncompressed space. You can then take that free space on the host drive and create another compressed volume that increases the total capacity of the two gigabyte hard disk drive to nearly four gigabytes.
FIG. 9.8 When compressing hard disk drives larger than one gigabyte, DriveSpace compresses as much of the data as possible and leaves the remaining space uncompressed on the host drive.
If you attempt to compress your system's boot drive or the drive on which Windows 98 is installed, DriveSpace takes steps to ensure that access to the system is still possible after the compression process and that Windows 98 performance is not too severely affected. As you pass through the screens of the compression wizard, DriveSpace gives you the opportunity to create or update a startup disk for your system because the compression process requires a reboot that might not be possible with your hard drive in its interim state.
CAUTION: Before beginning the compression process, DriveSpace also gives you the opportunity to back up your drive using Microsoft Backup. Whether you use this program or not, always be sure to have a viable backup before beginning a major disk operation.Ensure that all applications are closed before you begin the compression routine, and do not interrupt DriveSpace while the compression process is running.
DriveSpace 3 also adds a Compression tab to the Properties dialog box of every local drive on your computer (see Figure 9.9). From this dialog box, you can launch a full compression of the drive (which is the same operation discussed earlier in "Compressing an Entire Drive"), or you can create a new compressed drive from the free space left on the disk. Another way to access this feature is to highlight an uncompressed drive with free space on it in the DriveSpace program and then select the Advanced, Create Empty command.
FIG. 9.9 The Properties dialog box of every local drive has a Compression tab that displays its compression status and enables you to initiate compression operations.
The process of creating a new compressed drive is far less involved and invasive than that of compressing the existing files. In the Create New Compressed Drive dialog box (see Figure 9.10), you specify the letter to be used for the new drive, the amount of uncompressed space to use, and the source drive for that uncompressed space. DriveSpace displays the estimated capacity of the new drive and, on your approval, creates the new CVF.
FIG. 9.10 You can create new compressed volumes out of your free space by using the drive's Properties dialog box.
When you create a new compressed drive, no swapping of drive letters is necessary because the host drive still contains other data that must be accessed as before. The new CVF appears as an extra drive letter on your system, which you can access like any other disk drive.
You can view the current status of compressed drives on your system and manage their properties using the DriveSpace utility. When you double-click on a compressed drive, a dialog box like that shown in Figure 9.11 appears, displaying the disk's used space, free space, and compression ratios.
FIG. 9.11 The DriveSpace utility can display the status of any currently mounted compressed drive.
Changing Compression Ratios
When you are dealing with disk compression, statistics such as disk capacity are always estimates because there is no way to know what the compression ratio for a given file will be until it is actually compressed. The status displays in Windows 98's various file management utilities often include an indicator of a disk's remaining free space, however. DriveSpace defaults to a 2:1 ratio in its estimates of a compressed disk's capacity. You can adjust this ratio to achieve a more realistic estimate based on the actual compression ratio of the files already on the disk and your own understanding of the types of files you will be storing on the drive. Select the Change Ratio command from DriveSpace's Advanced menu, and you see a dialog box like that shown in Figure 9.12. This dialog box shows the current actual compression ratio of the files on the disk. Beneath this is a slider showing the current estimated ratio, which is 2:1 by default. If you plan to store more of the same types of files that are already on the disk, you can adjust this figure to equal the actual ratio, and Windows 98 will adjust the estimates displayed elsewhere in the operating system. If you plan to store files you know to be highly compressible or not highly compressible, you can adjust the value accordingly. Setting the ratio to 1:1 displays the actual free space that will remain available on the drive if you choose to store files that cannot be compressed at all.
FIG. 9.12 The Compression Ratio dialog box enables you to correct Windows 98's estimates of a compressed drive's free space.
NOTE: It is important to understand that the Change Ratio feature does not adjust the degree to which your files are compressed, it only enables you to make Windows 98's estimates of the compressed drive's capacity more realistic.
Adjusting Free Space
After DriveSpace creates a compressed volume, free space is usually left on both the new and the host drives. You can adjust this free space as needed, moving it to either drive, as long as the compressed drive does not exceed two gigabytes. To do so, choose a drive letter in the DriveSpace program and select Adjust Free Space from the Drive menu9. The Adjust Free Space dialog box (see Figure 9.13) contains a slider that enables you to move space from one drive to another, changing the pie charts to display the correct proportion.
FIG. 9.13 DriveSpace's Adjust Free Space feature enables you to move the remaining drive space between a compressed volume and its host drive.
Changing Drive Letters
Sometimes the additional drive letters that DriveSpace creates can be confusing, especially when they are hidden and you want to use one of those letters for a network drive or some other device. You can change the letter assigned to a host drive or an empty drive by highlighting it in the DriveSpace program and selecting Change Letter from the Advanced menu.
DriveSpace provides three levels of compression that trade off disk space for system performance. By default, DriveSpace uses Standard compression. However, you can configure the system to use HiPack compression whenever new files are written to a compressed drive. To achieve the maximum possible compression, called UltraPack, you must use another utility called the Windows 98 Compression Agent.
The Compression Agent lets you initiate compression events on-demand and choose the level of compression that it should apply to your drive. Unlike DriveSpace itself, Compression Agent does not automatically compress files on-the-fly. Instead, it takes existing compressed drives and compresses them further when you specifically instruct it to do so.
You can use Compression Agent in any of several different ways--to suit your work habits. You can, for example, set DriveSpace to use standard compression or even no compression, to minimize system performance degradation, and to run a nightly Compression Agent job to UltraPack your files after you finish working by using the Task Scheduler.
You can also configure the agent to UltraPack only the files that you have not accessed within a specified number of days. This keeps your seldom-used files at maximum compression while speeding up access to those you use frequently.
When you display the Compression page in any compressed drive's Properties dialog box, you can see a summary of the compressed files on your disk, the type of compression used, and the compression ratios achieved by each one, as shown in Figure 9.14.
FIG. 9.14 A single compressed disk can contain Standard, HighPack, and UltraPack files.
To use the Compression Agent, launch the program from the Start menu's Programs, Accessories, System Tools program group. Click the Settings button to see the dialog box shown in Figure 9.15. You can select the degree of compression desired, and you can access other screens that enable you to apply specific compression rules to individual files and execute compression events based on the current amount of free space on the disk.
FIG. 9.15 You can use the Compression Agent to customize your system's compression behavior to your needs.
DriveSpace enables you to return the files in a compressed volume to their uncompressed state and delete the CVF. To do this, you highlight a drive in the DriveSpace program and select Uncompress from the Drive menu. The program displays a screen showing how much uncompressed free space will be left on the drive after the process.
Unfortunately, you will in all likelihood have more data on the compressed drive than will fit on the same drive in its uncompressed state. The program informs you if this situation arises, letting you know how much data you must remove before the process can proceed.
Do not confuse the process of deleting a compressed drive with that of uncompressing it. When you highlight a drive in the DriveSpace program and select Delete from the Advanced menu, the program deletes the CVF from the host drive, destroying its data and freeing up all the disk space that it previously occupied. As always, the DriveSpace program warns you of the impending data loss and requires your confirmation before it proceeds. You should only use this feature if you do not need the data on the compressed volume or if you have already backed it up to another medium.
CD-ROM drives have become an important component of personal computers. Most software is now distributed on CD-ROMs, and their large capacity makes them ideal for use with games and multimedia titles that require large amounts of disk space but do not warrant permanent storage on a hard disk drive. Chapter 24, "Setting Up Windows Multimedia," provides more information on using Windows 98's multimedia features with CD-ROMs.
Windows 98's autoplay feature enables software developers to create CD-ROMs that automatically launch a program when they are mounted by the file system. The program can initiate the installation process for a new application or load a multimedia program that is designed to run directly from the CD-ROM.
Autoplay is made possible by the 32-bit, protected-mode device drivers that Windows 98 uses to support CD-ROMs. These drivers enable the operating system to detect the insertion of a disk into the drive. Real-mode drivers loaded from the CONFIG.SYS file do not have this capability.
When Windows 98 detects that you have inserted a disk into the CD-ROM drive, it mounts the disk in the file system and searches for a file called AUTORUN.INF in the root directory. This file specifies the program to be launched and the icon to be used to represent the program in Windows. A typical AUTORUN.INF file looks like this:
[autorun] open=filename.exe icon=filename.ico
The open= directive specifies an executable file on the CD-ROM, and the icon= directive specifies an icon file.
In your working environment, you might find that the autoplay feature is more of an intrusion than a help. To disable this feature, open Device Manager's System Properties dialog box for your CD-ROM drive, click the Settings tab, and clear the Auto Insert Notification check box.
In many cases, you can save disk space on your hard drives by running applications directly from a CD-ROM. Other applications, such as games and multimedia titles, require that you do this. Some applications, such as Microsoft Office, have installation options that enable you to select how much of the software is copied to the hard disk and how much should be executed directly from the CD.
Running a program from a CD-ROM is always slower than running it from a hard disk, and you are, of course, limited to running a single application in this manner (unless your system has multiple CD-ROM drives). In some cases, however, the conservation of disk space is worth the delay. This is particularly true if an application enables you to control which software components run from the CD. For example, you might be able to install the core program files of a word processor to your hard disk but run less frequently used modules (such as the spell checker) from the CD-ROM.
If your CD-ROM drive uses the EIDE interface, Windows 98 includes a feature that might be able to speed up your CD-ROM and overall system performance. In your drive's Properties dialog box in the System Control Panel, click the Settings tab and see if the DMA check box is available. If you can check this box and it remains checked after you reboot the system, you are using Windows 98's bus mastering IDE controller drivers.
Direct memory access (DMA), also called bus mastering, is a technique by which a device performs data transfers without utilizing the services of the main system processor. If you turn on this feature for your EIDE CD-ROM or hard disk drives, the system does not use any CPU time for drive access requests. While bus mastering may not produce a noticeable increase in the data transfer rate for the device, it can result in a general improvement in system performance because of the reduced load on the CPU.
© Copyright, Macmillan Computer Publishing. All rights reserved.
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}
){kind=link}