Sharing files and printers are the primary reasons for networking computers in the first place, and whatever other services you plan to provide, you will almost certainly want to implement these basic capabilities. This chapter examines the procedures for sharing the file and print resources of a Windows 98 system, both in theory and in practice.

The theory behind device sharing is as important as the physical procedure, because you will want to create a network environment that both protects your resources and makes them easy to access. You should carefully plan where your users are going to store their data files and how you are going to protect them from accidental or deliberate tampering. In the same way, you should plan where your shared printers should be located, how you are going to connect them to the network, and who is to be permitted to access them.

To give network users access to system resources, you must create what are known as shares. A share is a logical entity that represents a physical resource, such as disk drive, a directory, or a printer. Shares appear in the network hierarchy displayed by Windows 98 file management utilities like the Network Neighborhood and Windows 98 Explorer as subordinates to the computers that host them, as shown in Figure 7-1.

File system sharing enables users to access files on any computer anywhere on the network. Even on a small network, there should be an administrator who controls access to files. The administrator is ultimately responsible for seeing to it that the files are protected and yet available when needed. This may involve a number of different tasks, from setting password security to performing regular backups.

In Chapter 2, Networking Basics, you learned about the differences between client/server and peer-to-peer networks. Strictly speaking, your Windows 98 network will use the peer-to-peer model, because each system will be responsible for controlling access to its own resources, but from an organizational standpoint, the decision about which systems implement the client and server functions is up to you.

Every system on a peer-to-peer network can function both as a client and a server simultaneously. You will have to decide which files your users will need to share and where they should be stored. For example, you may want to implement what amounts to a logical client/server arrangement by selecting one computer to serve as the respository for all of the network's shared files. Even though this one system runs Windows 98 like all the others, logically it would be functioning as a file server.

Alternatively, you may want to distribute the shared files on systems throughout the network. This prevents one system from carrying the entire file sharing burden for the network, but it can also make it more difficult for users to locate the files they need. In either case, it is critically important to share only the files that your users need to access. Granting everyone full access to every disk drive on the network is easier for the network administrator in the short run, but it is also asking for trouble.

The most obvious network sharing strategy is simply to share the root of each disk drive on every networked computer. There are two main reasons why this is a bad idea:

Confusion: When users are faced with a dozen different shares representing the disk drives on a dozen different machines, it can be difficult for them to figure out where their files are located. They might have to pore through the directory structure of several different systems before they locate what they need

Security: Sharing an entire drive, particularly a drive on which the operating system is installed, grants users to a great many files and directories that they would be better off not seeing. Users typically do not need access to operating system and application files on other machines, and may cause damage if they were to inadvertently delete a required file or directory.

The solution to this problem is to create shares from specific directories and not from entire drives. The files that users are most likely to access over the network are documents and data files; sharing applications and running them directly from network drives is another issue entirely that is discussed in Chapter 9, Network Applications. Therefore, you should organize the directory structure of your systems in such a way as to store documents and other shared files in logically-named directories, and then create shares out of those directories.

One of the most common traits of computer novices is the failure to recognize the need for an organized directory structure, or in some cases any directory structure at all. After creating document files, these users save them in whatever directory the application uses by default. Depending on the application, this can be a logical place such as the \My Documents folder or an inappropriate spot like the root directory.

While the use of a standardized directory like \My Documents is a step in the right direction, it is not enough to create a suitable network storage strategy. A dozen machines with all kinds of data files dumped together in a single directory is not much of an improvement over storing them in the root. As the administrator, you must exercise a greater degree of control over each user's computer than you would if they were not networked.

This need for centralized control may be the biggest adjustment you have to make when you switch to administering a network from a group of standalone PCs. Each user's actions can have an effect on the network's other users, and somebody has to develop a set of policies that everyone must follow.

Before you create file system shares, you should plan where your users' data files will be stored. This is an important decision for several reasons, including the following:

• Users must be able to easily locate and access the files they need.
• Files must be protected against accidental or deliberate tampering and deletion.
• Network traffic should be distributed between systems relatively evenly.

The strategy you develop for storing your users' data files will depend largely on the efficiency of your network and the sophistication of your users. Let's consider a few storage scenarios and see how they affect your network and its users.

One possibility mentioned earlier is to use one computer as the repository for all of your data files, essentially positioning it as the file server for your network. This is certainly a valid strategy, but it presents some drawbacks as well as advantages.

On the plus side, storing all of your users' data files on one machine makes it easier for the administrator to control the directory structure in which the files are stored and easier for the users to locate the data they need. By creating separate directories for various file types and sharing them, you have a relatively simple interface for users to navigate in the Network Neighborhood or Windows 98 Explorer, as shown in Figure 7-2. You can create data directories based on the types of files to be stored there (such as word processing files and spreadsheets), the users or departments generating the files, or any other organizational paradigm you wish.

The single system method also makes it easier to back up the data at regular intervals and secure it against tampering. You might choose to grant certain users read-only access to the files while others have full control.

One important question that you must consider is whether the computer functioning as a file server should also be a user's workstation, and the answer to this leads us to the potential drawbacks of this type of setup. Depending on how much data you plan to store on your network, you must consider the traffic generated by having everyone accessing a single machine. If you routinely deal with very large data files, such as graphic images or other multimedia content, you may find that a typical PC is overwhelmed by the constant file input and output. When a system is frequently occupied with serving files to other users, its performance as a workstation may become prohibitively slow.

Another drawback of the single server scenario is that storing all of your data files on one computer leaves your network with a single point of failure. Should the file server go down for any reason, then none of your users can access their files. You absolutely must perform regular backups to tape or some other medium to minimize the chance of data loss, and you may even want to mirror your data between drives on separate systems for added protection.

Storing your data files on various drives distributed throughout the network is closer to the natural working pattern of standalone systems and requires less adjustment both for your users and for the administrator. If you are going to be networking a collection of PCs that users have been running for some time as standalone systems, then the distributed scenario might make the adjustment process a bit easier.

In an existing business or organization, your users probably have already established places where the store their data. Very likely, the computers are dedicated to the tasks performed by specific workers, with the data stored on the local drives. For example, accounting spreadsheets are stored on the bookkeepers' PCs, marketing materials on the salespeoples' systems, and general correspondence on the receptionist's machine. Disrupting an already functional system is usually not recommended, unless the user is completely ignorant of proper file storage practices. (If you find that the receptionist stores the correspondence document files in the \Windows directory, for example, it's time to step in and reorganize.)

However, you can still implement an organized sharing strategy by selecting only appropriate directories to be shared. If your users have stored various types of document files in different directories all over the hard drive, you can modify their work habits only slightly by moving all of the data directories to one place on the disk. For example, creating a \Data directory on each computer and storing the various types of data files in subdirectories can greatly simplify the sharing process. You can then either share the entire \Data directory or create individual shares out of the subdirectories as needed.

Another good idea is to select one system to host home directories for all of your users. Home directories are individual user subdirectories, that the users are free to use as they will. Typically, you would create a directory called something like \Users or \Home, share it, giving everyone full access, and then create a subdirectory for each user's logon name. The idea is for users to have a place to store files that everyone else can easily locate and access. So, if Betty is telling Veronica about the new utility she downloaded from the Internet, she can simply say "it's in my home directory if you want it," rather than having to specify a path.

If you're building a network for home use or even for a very small, casual business, organizational principles can certainly be more lax. You probably have specific reasons for wanting to connect your home PCs in the first place, such as sharing a printer or Internet connection, so it can be a good idea to begin by limiting yourself to solutions for your original problems. For example, go ahead and share the printer, but don't enable file sharing until the need for it arises.

If you do want to share files, however, some of the same principles from a business network can apply to the home. On a family network, there may certainly be files on the parental computer that the kids should not access, so you may want to limit your shares to specific directories. If you don't have any need for these restrictions, and you don't have to protect your systems against malevolent or inexperienced users, then go ahead and share the entire drive from the root. This is certainly the easiest way to go, and provides full access to the file system.

When you buy a home computer, a CD-ROM drive is nearly always standard equipment, and many systems today also have removable cartridge drives like the Iomega Zip. On a network, however, this is not always the case. In a business environment, CD-ROM drives are used primarily to install software; multimedia and gaming applications are usually not part of the picture (at least during business hours). As a result, having a CD-ROM drive in every computer is usually not necessary. If you are building a network from scratch, you can save some money by purchasing only one or two computers with CD-ROMs, and sharing them so that everyone can use them.

You may want to share business-oriented CD-ROMs with all of the users on your network. For example, nationwide telephone directories, searchable dictionaries, and technical support reference CDs all can be valuable resources for network users. Before purchasing a product for this purpose though, make sure that multiple users can legally access the disk at the same time. A product intended only for a single user may have some sort of copy protection that prevents multiple users from executing the program. Some programs may not run at all over a network, even for a single user.

Although their use in a network environment is more problematic, you can also share removable disk drives like floppies and optical drives. This can be convenient when you need quick access to a file stored on one of these devices, but the problem with using them for regular network storage is that you must constantly make sure that the correct disk or cartridge is in the drive. In addition, for the person using the machine hosting the share, a constant stream of requests to change the disk can get old very fast.

Assuming that you have properly installed a network interface card and the software components required for network communications (as discussed in Chapter 6, Installing Network Hardware), you have only to perform two minor configuration tasks before you can begin to create shares.

The Access Control page of the Networking Control Panel (shown in Figure 7-3) is where you specify what form of security you will use on your shares.

The two available options are as follows:

Share-level Access Control

User-level Access Control

Because it requires a server to function as the source of the user and group names, you cannot select user-level access control on a network that consists of Windows 98 machines only. Since share-level access control is the default, you do not have to make any changes here; the setting is only mentioned in order to introduce you to the two forms of share security available on Windows 98 systems.

User-level access control is inherently more secure than share-level, because you can assign access permissions to specific users and exercise much more precise control over the access granted to them. Share-level access control provides only three options: no access, read-only access, and full access. User-level enables you to customize user access to the share with the individual permissions shown in the Change Access Rights dialog box shown in Figure 7-4.

If file system security is a major concern on your network, that is, if you have files and directories for which you need to assign more complex access permissions than read-only or full access, then perhaps you should consider adding a Windows NT server to your network in order to implement user-level access control on your Windows 98 shares.

After you install File and Printer Sharing for Microsoft Networks, you activate the service by clicking the File and Print Sharing button on the Networking Control Panel's Configuration page and filling the checkboxes to allow network users access to the computer's files and printers. Clearing the checkboxes is also a good way to disable all of the shares on a given machine in one step.

To create a share out of a drive or directory, you must access the Sharing page in the Properties dialog box for that drive or directory, as shown in Figure 7-5. You can do this is several different ways, as follows:

• Highlight the drive or directory in the My Computer window and select Sharing from the File menu.
• Click the secondary mouse button on the drive or directory in My Computer or Windows 98 Explorer and select Sharing from the context menu.
• Open the Properties dialog box for a drive or file by highlighting it in My Computer or Windows 98 Explorer, select Properties from the File menu or the object's context menu, and click on the Sharing tab.

To share the drive or directory, click the Shared As radio button and specify a name for the share in the Share Name field. You can also include a descriptive text string in the Comment field that will appear next to the share name in Windows 98 Explorer. The names that you choose for your shares are an important part of creating an easily navigable network. By default, Windows 98 uses the drive letter or the first 12 letters of the directory name as the share name. In some cases, this may be acceptable, but it is a better idea to supply a more descriptive name of the directory’s contents.

If for example, all of your users store their document files in the \My Documents directory, this could result in every machine having a share with the same name, causing confusion among your users. Even worse would be every machine having a share called C because you have shared the root of the drive without changing the share name. Remember that in a business environment, share names are not just something that users see on the computer monitor, they are a part of the everyday communication between fellow workers. Picture one user describing the location of a particular document to another user on the phone or in an e-mail and ask yourself if you can make this communication process easier by assigning more logical names to the file system shares.

In the Access Type box, you specify what degree of access to the share your users should have. The available options are as follows:

Read-Only

Full

Depends on Password

In the Passwords box you supply a access password for the option you chose in the Access Type box, after which Windows 98 prompts you to confirm the password by retyping it. In the case of Depends on Password, you supply a separate password for each access type.

When you click the OK button, Windows 98 creates the share, causing the drive or directory icon in My Computer and Windows 98 Explorer to appear with an outstretched hand, denoting that the object has been shared.

When you use Windows 98 on a peer-to-peer network, the security measures available for protecting your shares are limited, but there are steps that you can take to make it more difficult for unauthorized users on the local network to access shared directories. Unauthorized access from the Internet is an other problem entirely that is discussed in Chapter 3, Planning Your Network. The fundamental security problem with share-level access control is that each share requires its own individual password. In cases where you configure a share for both read-only and full access, two passwords are required. Creating, maintaining, and using passwords for many different shares can be difficult for administrators and users alike.

One of the decisions you must make as the administrator of the network is who will be responsible for creating file system shares and assigning passwords to them: administrators or users. This decision should be based on the level of security required for your documents and the computing skills of your users. Large networks avoid share-level security precisely because it is impractical for administrators to create and maintain the shares on many different machines, but on a network of less than twelve systems these tasks are not overwhelming.

Passwords are the key to protecting shares from unauthorized access, and most computer users are less than vigilant when it comes to selecting appropriate passwords. For this reason alone, it is a good idea for the administrator to create the shares on each computer and assign the passwords. This method also enables the administrator to exercise control over which directories are shared and how many shares are created on each computer.

If you do not require security for your shares (that is, if all users are to be granted full access), you can create simple, obvious passwords or use no passwords at all. However, when you select passwords for shares that require controlled access, you should take pains to select passwords that will not easily be penetrated by potential intruders. Using people's names, birthdays, and other easily discovered information for passwords obviously provides little security.

Passwords can be up to 8 characters, contain letters, number, and symbols, and are case sensitive. A good password would be one that consists of mixed numbers and letters, letters of various cases, and symbols interposed with letters or numbers.

Using these techniques, you can create complex passwords that would be extremely difficult to guess, but the trick is to also create passwords that are easy to remember. When users are faced with long, complicated passwords, they are more likely to write them down than try to remember them, which is always something to be avoided. You should take pains to come up with passwords that use some kind of logical pattern, rather than a string of random characters. For example, a password that consists of a word with numbers between the letters, like s1p2i3n4e, is more likely to be remembered than a meaningless sequence like s42gi8lkb5. Another good method is to use the initials of an easily-remembered song title. It isn't likely that anyone would guess that a cryptic password like splhcb68 stands for Sgt. Pepper's Lonely Hearts Club Band, which was released in 1968..

The use of varying case for the letters in a password is a good way of making it less penetrable, but there is no convenient point of reference that enables users to easily remember which letters use which case, unless you vary the case using a pattern. Using the previous example, you can alter the case of every other letter to complicate the password further, as in S1p2I3n4E.

Creating secure passwords does not place as much of a burden on the user as it may seem at first. Windows 98 is capable of caching the share passwords after users supply them for the first time. When users fill the Save This Password in Your Password List checkbox while connecting to shares, the system saves the passwords in an encrypted file in the \Windows directory on the local drive. The password files are named using the logon name of each system user, with a .pwl extension.

When connecting to the same shares during subsequent sessions, there is no need to supply the passwords again. The system accesses them from the password file and connects to the share immediately.

Since the passwords are stored in the .pwl file in encrypted form, there is no way to view them. However, Windows 98 includes a utility called the Password List Editor that enables you to view a list of the resources for which passwords have been stored in the file. The program enables you to delete passwords to specific shares from the .pwl file, but you still cannot view the passwords.

The Password List Editor is located in the \Tools\Reskit\Netadmin\Pwledit directory on the Windows 98 CD-ROM. The program consists of an executable and an information file, called Pwledit.exe and Pwledit.inf respectively. When you run the program, it loads the password file for the user that is currently logged on to the system. You cannot use the editor to open another user's file.

In a situations where users leaves the organization or changes jobs and no longer requires access to the same shares, you can easily remove all of their cached passwords from a system by deleting their password file. The system will then require a correct password before connecting them to a share.

It is possible to create shares within other shares, enabling you to modify the type of access granted to your network users for specific directories. When you attempt to create a share out of a directory that is already contained in another share, the Sharing page of the directory's Properties dialog box specifies the name of the parent share.

For example, you can create a read-only share out of the \My Documents directory on a particular system, enabling users to browse through and open the files stored there. You can then create full access shares out of specific directories beneath \My Documents, so that users can modify or delete some of the document files.

Be aware, however, that even though the directories forming the two shares may be nested, they appear as individual shares to network users (as shown in Figure 7-6), with their own individual access rights. This means that if you reverse the preceding procedure by creating a full access share out of \My Documents and a read-only share out of a subdirectory, the nested, read-only share does nothing to prevent users from gaining full access through the \My Documents share.

Printer sharing is the single most compelling reason for networking computers. Windows 98's printer sharing features make it easy to create a printer share and access it from a network workstation, but there are a number of administrative issues that you will want to address in order to share your printers efficiently and keep them running smoothly. Before you actually begin setting up the printer and creating shares, you should ask yourself the following questions:

Where will the printer be located?

How will the printer be connected to the network?

Who will be responsible for printer maintenance?

Who will use the printer?

In most cases, network printers are left unprotected, enabling anyone to use them. However, there may be cases in which you want to limit access to a particular printer. For example, higher-end color printers can cost anywhere from fifty cents to a dollar or more per page to use. Multiple inks or toners and special papers are expensive, and color printing can use them up very quickly. In addition, some people are liable to see the printer as a toy and use it for non-business purposes.

You can protect a shared printer against unauthorized access through the network, just as you can protect a file system share. By assigning a password to the share, users must be authenticated before they are able to install the printer.Sharing Windows 98 Printers

In order to share a printer that is connected to a Windows 98 workstation, you must first install the printer on the local machine in the usual manner, using the Add Printer wizard in the Printer Control Panel. Once you do this, and there is an icon representing the printer in the Printers window, you can click the secondary mouse button on the icon and select Sharing from the context menu, just as you would when creating a file system share. On the Sharing page of the printer's Properties dialog box (shown in Figure 7-7), you specify a name for the share and, optionally, a description and password.

Once the system creates the share, the icon appears with the outstretched hand that designates it as a shared printer. Users can then view the share in the Network Neighborhood and Windows 98 Explorer utilities and use it to install support for the printer.

Setting up a third-party print server on your network is a bit more complicated than using Win98 for the print server. Because the device is connected to the network and not directly to the computer, Windows 98 has no built-in means of communicating with it. The software included with the device enables the system to detect the print server on the network and address it as it would a printer.

The HP line of JetDirect print servers supports all of the major network operating systems, including Windows 98. When you use them with a Windows NT or Novell NetWare server, they store the queued print jobs on the server drive as they are waiting to be processed by the printer. On a peer-to-peer network, however, the print server operates in direct mode, meaning that each workstation queues its own print jobs and feeds the data to the printer as fast as it can process it.

In essence, Windows 98 uses the print server just as it would a local printer. The only difference is the port that the computer uses to transmit data to the printer. Local printers use parallel or serial ports to transmit print data. The JetAdmin software supplied with the HP print servers creates a new port on the workstation that addresses the print job data to the print server's network address.

The JetDirect print servers support both the TCP/IP and IPX/SPX protocols (as well as specialized protocols used by other operating systems). When you install the print server on the network, it makes its presence known to the computers by responding to the broadcast messages generated by the JetAdmin software.

Installing an HP print server is as simple as plugging the device into a power source and plugging in the cables connecting to the network and the printer. When you press the test button on the device, it prints out a configuration page that lists the device's network address and other identifying characteristics.

When you install and run the JetAdmin software on a Windows 98 system, it detects the print server (as shown in Figure 7-8) and configures it for operation on your network. On a TCP/IP network, you must supply an IP address for the print server, that all of your computers will use to communicate with the device.

In the same way that you assigned IP addresses to your computers in Chapter 6, Installing Network Hardware, you must choose a unique address for the print server when the software prompts you for it. The software then configures the print server by writing its address to a ROM chip in the device. When you install the JetAdmin software on additional workstations, they will detect the device, including its IP address.

Once the print server is configured, the software creates a new port on the workstation using the IP address you supplied, or in the case of an IPX/SPX network, the device's hardware address. From this point on, the printer installation is totally conventional. You select the printer's manufacturer and model from the standard Add Printer wizard interface, and the system installs the appropriate driver. The print server only provides a means for the computer to communicate with the printer over the network. The format of the print job does not change at all, so you can use the same drivers with an external print server that you would with a locally-connected printer.

Once you have installed the JetAdmin software on a single machine and configured the print server, you can proceed to connect your other workstations in one of two ways. You can install the same software on the other computers, or you can share the printer on the first computer in the usual manner.

If you choose to install JetAdmin on the other machines, you will not have to supply an IP address again, since the print server has already been configured. You will, however, have to identify the manufacturer and model of the printer on each machine so that the system can install the appropriate driver. If you decide to share the printer you installed on the first machine, you create a standard printer share in the usual manner.

If you share the printer, the computer hosting the share also functions as print server, just as if you were sharing a locally-attached printer. Print jobs generated by other workstations will be queued on that system until they can be transmitted to the printer, subjecting that machine to the same performance degradation described earlier. However, the advantage of this method is that point and print is in effect, enabling other systems to install support for the printer without selecting a driver.

Installing the JetAdmin software on each machine is a task that is more likely to fall to the administrator, particularly if you have more than one JetDirect print server on the network. If this is the case, the software will detect all of the print server devices, and you must select the appropriate printer driver for each one. Since each computer sends its jobs directly to the print server, no single machine is overburdened by other users' print data.

Printing in a network environment raises a number of unique maintenance issues. While printers do not require constant vigilance, they do need administrative attention more frequently than most network components. Even a properly functioning printer needs to have its paper trays reloaded and its toner/ink/ribbon cartridge changed occasionally. In addition, there are also usually paper feed jams to deal with, changes to different print media such as paper of various sizes and envelopes, and the occasional error message generated by the printer itself.

In a network environment, users have a tendency to evade these tasks, just as the person who drinks the last cup of coffee never seems to make a new pot. This is because in many cases no one is assigned direct responsibility for the printer. When a network printer is jammed or out of paper, the print jobs pile up in a queue, filling temporary hard disk space. The printer may take some time to catch up when it is finally cleared.

For this reason, it is a good idea to delegate the tasks involved in maintaining the printer to one person. Usually, this is the person closest to the machine, but you may want to experiment with a rotating schedule, to avoid sticking the same person with this onerous task all of the time.

Depending on your printer's capabilities and your users' needs, one of the maintenance tasks that may be required is the changing of paper trays or supplies to support different sized pages or alternative media, such as forms and envelopes. Many of the printer drivers in use today can recognize the size of the paper currently loaded in the printer, and will not print a job that requires a different size. The average laser printer used in an office environment will have some form of display that informs the operator when a paper size change is necessary.

When users access a printer through a standard Windows 98 share, a queued job requiring a paper change will halt the processing of all subsequent jobs until the paper is changed and the paused job is completed. When users all access an external print server directly, the same is true; all other jobs remain queued on their individual machines until the paper is changed or the job times out.

If your users require frequent media changes, you might be better off purchasing separate printers to support the different paper sizes, or a printer with multiple trays that requires no manual media changes. It can be extremely distracting for a worker to constantly be receiving calls requesting a different paper size. The lost productivity resulting from these continual, repetitive maintenance tasks may end up costing you more than the price of a new printer.

When multiple users are sharing a printer, one of the most common problems is that of getting the printed pages to the appropriate user. Pages exiting the printer often get mixed up, and one person's print job may contain pages belonging to someone else. To minimize occurrences like this, it is a good idea to configure your printers to use separator pages. A separator page (sometimes called a banner page) is an additional page printed before every job that identifies the user generating the document and the name of the file being printed.

The Properties dialog box for every locally-installed printer contains a Separator Page selector on the General page. Windows 98 provides two separator page options: full and simple, where full uses graphics and simple consists only of ASCII text. You can also specify the name of another separator page file, which can be any Windows metafile (.wmf) file.

From the user's perspective, accessing shared resources requires a bit of initial preparation, but after connecting to a share for the first time, subsequent uses are all but transparent.

When a Windows 98 system boots, the user is prompted to provide a user name and password in order to log on to Microsoft Networking. When Windows 98 is configured to use share-level access control, the logon name plays no part in granting access to shares (as it does with user-level access). The logon name will, however, be used to identify the user's print jobs and, if user profiles are enabled, will determine which profile the system will load.

The password that a user specifies when they log on for the first time will be saved in the password file for that user. The system will then require that password for each subsequent logon.

Once connected to the network, users can access drive and directory shares in two basic ways. Using the Network Neighborhood and Windows 98 Explorer utilities, users can browse through the shares on all of the other computers to which they have passwords, accessing the files and directories directly using UNC (Universal Naming Convention) references. Users can also map drive letters to specific network shares, enabling them to access the shared files on another system through the same interface they would use for local drives. Both methods have their advantages and their disadvantages.

The Network Neighborhood (visible both in Windows 98 Explorer and in the Network Neighborhood application itself) displays icons representing all of the computers currently operating on your network. When you expand a computer icon, you can see all of the shares on each computer. The shares themselves are advertised to all of the systems on the network, but this does not mean that anyone can access them. When you select a share, your system attempts to connect to the other computer and prompts you for a password if the other system's user or administrator specified one when creating the share.

When you supply the correct password, the other system grants you access and displays the contents of the share, as shown in Figure 7-9. At this point, you can browse through the directory structure on the other computer and utilize the files according to the access (read-only or full) that you have been granted.

When you reference a file on a Win98 system's local drive, you use a path that is composed of a drive letter, possibly directory names, and a file name. When you browse a network share in this manner, however, there is no drive letter by default, so the system needs another method for uniquely identifying the specific elements found in a share. To do this, Windows systems use the UNC, or Universal Naming Convention.

A UNC name is a path that uses the network system's computer name and the name of a share in place of a drive letter. The name format calls for two backslashes preceding the name of the computer sharing the resource and single slashes between the share, directory, and file names. A typical UNC path name appears as follows:

where server represents the computer hosting the share, and share represents a share existing on that computer. Directories and files are the same as they would appear on a local drive.

The advantage of a UNC name is that it is absolute; it always refers to the same resource, even when you use it on different computers. Applications that are network-aware can often use UNC names when referencing files and can browse the network using the Network Neighborhood interface. The disadvantages of UNC names are that they are alien to computer users without networking experience, and that applications that are not network-aware may be unable to use them. In these situations, you can use drive letters instead of UNC names.

Another method for accessing network shares that is more familiar to most users is to map drive letters. Windows 98 can associate the drive letters not used by local resources with network shares in order to provide the user with a uniform interface to all of the available storage media. Using mapped drives is second nature to PC users. In fact, when you configure your networked systems to reconnect to the same shares during every session, users are often unaware that they are accessing network drives at all (a situation that can be more of a liability than a benefit).

The problem with drive mappings is that different computers may use the same drive letters to point to different shares. This can lead to confusion on the part of relatively unsophisticated users that tell other workers that a particular file is located "on the F: drive." One person's F: drive may not be another's.

To avoid this problem, it is a good idea to create permanent drive mappings on all of the network systems, so that some consistency can be assured. A small network should not have that many shares to deal with; you can map the same drive letters to your network's primary data storage shares on all of your systems and make them persistent connections, so that the computer re-establishes then whenever it boots. However, this doesn't prevent users from mapping their own drives (or disconnecting the ones you've mapped for them).

To map a network drive, you highlight a file system share in Network Neighborhood or Windows 98 Explorer and select Map Network Drive from the File menu or the context menu displayed by clicking the secondary mouse button on the object. This generates a Map Network Drive dialog box like that shown in Figure 7-10.

The system chooses the first available drive letter for the mapping, which you can modify with the Drive selector, and specifies the UNC name of the share in the Path field. If you want to create a persistent connection, fill the Reconnect at Logon checkbox before clicking the OK button. If a password is required to access the share, the system prompts you to supply it, after which the new drive letter is added to the My Computer display. You can then use the mapped drive for any purpose, just as you would a local drive.

You can also open a generic Map Network Drive dialog box from the Tools menu of Windows 98 Explorer or by clicking the appropriate button on the toolbar. This box is identical to the one shown in the illustration, except that you must supply a UNC name for the share in the Path field.

Once you configure your workstations to use network shares, you should also be familiar with the capabilities of your applications when it comes to using them. If you want your users to store all of their word processing documents in directories on a particular network drive, for example, you should configure their word processing application to use that drive by default. Nearly all applications enable you to specify a location for the storage of data files, and when you map drive letters to shares, the fact that the drive letters represent network resources is irrelevant. Again, pre-configuring you workstations like this does not prevent your users from changing the default storage directory or storing files in other locations, but it does give them a nudge in the right direction.

Configuring a Windows 98 system to access a printer share is even easier than installing a local printer, thanks to point and print. You can launch the Add Printer wizard in any one of three ways:

• Open the Printers Control Panel and double-click the Add Printer icon.
• Open the Network Neighborhood, click the secondary mouse button on a printer share, and select Install from the context menu.
• Open the Network Neighborhood and drag an icon representing a printer share to the Printers window.

The first method requires you to specify that you want to install a network printer and either browse to the printer share or supply its UNC name. The second and third methods proceed directly to the few questions you need to answer, asking you whether you want to redirect LPT1 output to the printer (so that you can print from MS-DOS-based programs), and whether you want to make that printer the default for the system.

Because of Windows 98's point and print feature, you do not have to specify the manufacturer and model of the printer in order to locate the correct driver. After prompting you for whether you want to print a test page, the wizard downloads the appropriate driver from the system hosting the share and completes the installation process. Once the printer is installed, you can use it just as if it was a printer connected to the local machine.

All of the print jobs sent to a particular Windows 98 printer share are queued on the machine hosting the share until the printer can process them. Any user of the share can access the print queue and view the jobs currently waiting to be printed. When you double-click the appropriate icon in the Printers Control Panel, a dialog box like that in Figure 7-11 appears. When you have a print job pending, an icon is also added to the tray, that you can double-click to display the same dialog box.

In this dialog box, you can see a list of the currently queued jobs, their status, size, time submitted, and the logon name of the user that submitted them. You can use the commands in the Document menu to pause or cancel the printing of any document you submitted that has not yet started printing. You cannot, however, change the printing order of the jobs in the queue.