PicLan - Networking for Pick

(C) Copyright 1990-1998 Modular Software Corporation. All rights reserved.

Version - February 28, 1998

   Goto:  Table of Contents   Index
Chapter:  1 2 3 4 5 6 7 8 9 10 11 12 13 14

5: Installation

Installing PicLan involves a number of steps:

5.1: Planning

A PicLan installation, because it involves multiple computer systems, typically involves more planning than would normally be required for a single centralized system. As such, an understanding of the overall goals and network configuration is helpful in ensuring a successful network installation.

PicLan may be installed into either a stand-alone PicLan Ethernet network installation or may be installed in conjunction with Novell NetWare or other IPX protocol supporting network. The presence of NetWare and other network components affects how you install and configure both the PicLan hardware and software.

The previous section on network configuration should be used to design the configuration of your network installation. Pay particular attention to limitations as to placement of Pick host systems and requirements for network hardware for Pick host and DOS workstation systems.

5.2: Installing the Client Software onto a Windows Workstation

The next step should be to install the PicLan client software onto a Windows workstation. The main reason that you want to install the client software now is to get access to the PicLan technical notes that are stored in Windows help format on the client disks. In order to install the client software onto a Windows workstation, you should insert the number one client installation diskette into a diskette drive and run the SETUP.EXE program on that diskette. SETUP.EXE must be run from within Windows. After the SETUP program completes, a PicLan program group is created which includes an ICON for the on-line documentation that is included with PicLan.

5.3: Installing Pick

Some versions of PicLan require that you install Pick in a manner that correctly configures PicLan ports.

PicLan does not require any special installation parameters.
PicLan requires that AP "Extended Memory" be enabled.
PicLan does not require any special installation parameters.
PicLan requires that PicLan ports be configured with the UpBoard configuration program and that the UpBoard system be booted with the PicLan PL-UPBRD.EXE driver program.
PicLan requires allocated R91 "dynamic" ports that must be configured when R91 is initially loaded.
Pick/64+ 2.2
PicLan does not require any special installation parameters.
Pick/64+ 2.3
PicLan requires that PicLan ports be pre-allocated during the Pick/64+ virgin install or by performing a 'C' (Configuration) restore from floppy diskette.
PicLan requires that PicLan ports be pre-allocated during the Sequoia PRO virgin install or by performing a 'C' (Configuration) restore from floppy diskette.
PicLan requires that PicLan ports be pre-allocated during the PC/OS virgin install.
Mentor PRO
PicLan requires that PicLan ports be pre-allocated during the Mentor PRO virgin install.

5.4: Installing Hardware

PicLan network interface hardware consists of PicLan Ethernet network interface adapter(s) and associated cabling. The PicLan network interface hardware should be installed before you attempt to load the PicLan software.

5.4.1: Pick Host System Network Adapters

R83 AP/N Pick/64+ PC/OS Seq/PRO
Each Pick host system that is to communicate on the PicLan network must have a PicLan Ethernet network interface adapter installed. These adapters are included with the PicLan package from Modular Software or from your dealers.
AP/Pro Men/PRO
Each Pick host system that is to communicate on the PicLan network must have a compatible Ethernet network adapter installed. These adapters are not supplied as a part of the PicLan package unless you dealer has chosen to do so. It is your responsibility to choose a network adapter that is compatible with PicLan. Refer to the previous section for a discussion on types of network adapters that work with the PicLan software.
The UPBOARD version of PicLan can use either PicLan Ethernet adapters, or other network adapters when used in conjunction with DOS driver software. If you are building a NetWare or NDIS based network, it is recommended that you use DOS drivers because this will give you more flexibility in configuring your PicLan Pick host system.
The R91 version of PicLan utilizes General Automation ArcNET network interface adapters. Consult the GA documentation for information on installing GA ArcNET adapters.

5.4.2: Choosing Adapter Configuration

Installing the PicLan Ethernet network interface adapter requires that you configure several hardware addresses on the adapter. With some types of network adapters, these settings are configured with jumpers. With other types of adapters, these setting are configured using a software setup program (usually a DOS program). Refer to the adapter's manual for more information on how to setup the adapters configuration.

NE2000 and WD/CPX Style Adapters - I/O Addresses

NE2000 style adapters use an I/O address range and an IRQ setting. Both of these values must be setup on the adapter and in software. Legal I/O address values are:

This list only describes conflicts with very common PC peripherals. If your system has SCSI controllers, multi-port serial controllers, sound cards, or other adapters, you should consider their I/O address range usage as well. NE2000 adapters use all 32 addresses (hex 20) in each range, so a board that uses addresses from 2B0-2B7 will conflict with an NE2000 card set at I/O address 2A0. Many brands of NE2000 cards are only addressable to I/O 300, 320, 340, and 360.

If your network adapter is addressable to I/O 380, this is the preferred address to use. With the exception of some sound cards, I/O 380 appears to not conflict with any other peripherals. If you adapter is not configurable to I/O address 380, you should try I/O address 300 and 320 next, paying attention to conflicts with Pick SCT tape controllers.

NE2000 and WD/CPX Style Adapters - IRQ Usage

NE2000 style network adapters also use an IRQ setting. Most adapters can be configured to the following IRQ settings:

For most systems IRQ 10 is the best starting place. Some systems like AP/Pro and Mentor PRO do not work with IRQ 11 because this conflicts with SCSI disk controllers (this conflict occurs even on systems that do not use SCSI disks!). Multi-port serial adapters also use IRQs, so you must avoid conflicts with these. If you decide to use IRQ 15 and have a motherboard with an on-board dual-channel IDE controller, you must disable the second IDE controller in order to use IRQ 15 (some motherboards do not allow you to disable the second IDE controller). If you decide to use IRQ 3 or IRQ 4, you must not have a COM2 or COM1 controller present in your system even if you have not configured Pick to use the device. Some motherboards have on-board COM1 and COM2 ports which you must disable either with jumpers or in CMOS setup in order to use IRQ3 or IRQ4. On some motherboards, even if you disable a COM1, COM2, or secondary IDE controller in CMOS, the IRQ is still unavailable.

WD/CPX Style Adapters - Shared Memory Address

WD/CPX style adapters also use a shared memory range to transfer data to and from the network adapter across the ISA bus. NE2000 adapters do not use a shared memory range. The shared memory address is setup completely within the PicLan software and you do not need to set any jumpers or run configuration programs to setup the card shared memory range. 8-bit WD8003 adapters use an 8K range. 16-bit WD8013 adapters use a 16K range. Compex adapters (both 8 and 16-bit) use an 8K range. You can set the range to any value provided that other adapters are not also using the same shared memory range and provided that the system motherboard allows the specified address range to be mapped across the ISA bus.

A good default address for the shared memory address is DC00. This is an address at the top of the range typically mapped across the ISA bus and allows for a 16K window. You should not need to change this address unless you have another adapter installed that conflicts.

Some motherboards are not compatible with shared memory cards. If you have one of these systems, you will need to install an NE2000 style adapter or jumper a Compex multi-personality adapter in NE2000 mode.

PCI Adapters - CMOS Setup

PCI adapters are "theoretically" auto configuring. The auto configuration is performed by the system CMOS and typically maps the adapter at I/O address 6000 and "snoops" the ISA bus for an available interrupt, usually in the sequence 2,10,11,15.

Some sites have reported problems using PicLan with PCI adapters where the system supports "Plug and Play" BIOS operations. If you have trouble, you may wish to disable plug and play options in the system CMOS setup. Also, if you are running some types of Pick systems, notable AP/Pro or Mentor PRO, and the system happens to map the IRQ to 11, you may have problems. You can sometimes work around these by telling the CMOS to not try to map the IRQ to 11.

PCI Adapters - Interface Port

The PicLan configuration program will list all possible interface specifications for each chip/board type. If the network adapter you are using is listed, you can select directly from the list. If you are using a network adapter that is not specifically listed, you may need to try several settings to see which one (if any) works with your particular adapter.

Most twisted-pair interfaces allow you to select between full and half duplex. Most networks are half duplex and setting the adapter to full duplex will reduce your network performance by causing a increase in the number of network collisions. Only networks using switching hubs exclusively are full duplex.

Common settings for PCI interface port settings in PicLan are:

Avoiding Conflicts With Other Installed Adapters

The purpose in allowing PicLan network adapters to be user configurable is to allow the user to install the network card in a variety of systems and avoid potential conflicts. The following known conflicts should be considered:

Conflicts with VGA Adapters

Many VGA adapters allow their ROM BIOS to be operated in 16-bit or 8-bit mode. The manufacturers state that 16-bit mode is desirable in order to achieve best performance. Conflicts can arise when the VGA adapter is an 8-bit adapter (or is plugged into an 8-bit slot) and installing PicLan network interface adapters into such systems nearly always requires that the VGA adapter be set to 8-bit BIOS mode exclusively. This is true even for those adapters that sense addressing conflicts. Note that this is a fairly rare conflict, since most VGA adapters are 16-bit boards.

Conflicts with Async Adapters

Pick async adapters use both I/O addressing space and IRQs. Be sure to configure PicLan networking addressing and IRQ setting to avoid any installed async cards. Most async cards use I/O addresses between 0100 and 02FF and either IRQ 3 or IRQ 4.

Conflicts with Pick Async Driver Code

Some Pick releases (R83) auto-sense the presence of Async adapters. For this reason, do not install the PicLan network adapter with an I/O port address between 0100 and 02FF as this is the most common range of addresses for async adapters. If you install the PicLan network adapter within this I/O address range, unusual Pick behavior including port and system hangs may be encountered.

Conflicts with Tape Controllers

The standard setting for R83 and AP Pick tape controllers is IRQ 5, I/O base 0320. Some tape controllers run IRQ 5, I/O base 0300 or 0308. Be sure to configure the PicLan network adapter card to IRQ and I/O addresses that do not overlap those of the tape controller. Similar conflicts occur with Pick/64+ and UPBOARD tape and disk controllers.

Conflicts with Bus Mouse Adapters

Even though a Pick host system cannot actually use a mouse, many system have bus mouse adapters installed. PC bus mouse adapters use an IRQ value of 2, 3, 4, or 5. IRQ 5 is the most common value in most installations. If you have a bus mouse adapter, be sure to select a configuration that does not overlap with the bus mouse IRQ usage.

Conflicts with LPT1 Adapters

Some printer adapters use IRQ 7 for LPT1 support. If you configure PicLan to utilize IRQ 7, be sure that your LPT1 printer adapter does not use IRQ 7. Most Pick implementation do not activate the IRQ functions of parallel printer adapters, so an adapter that supports IRQ 7 does not preclude use of IRQ 7 with network adapters.

PC/OS Men/Pro
PC/OS and Mentor PRO uses IRQ7 for LPT1: printing functions, so this IRQ is usually not available with these systems.
PC/OS Men/Pro AP/Pro
PC/OS, Mentor PRO, and AP/Pro all use special processing to support SCSI disk controllers on IRQ 11. You should not configure the PicLan network adapter to use IRQ 11 on these systems.
Final Notes on Hardware Settings

The default settings of IRQ 10, IO 380, and RAM DC00 (for WD/CPX cards) and IRQ 10 and IO 300 (for NE2000 cards) were carefully chosen to minimize conflicts. While some sites may need to change these settings, you should be conservative: only change those settings which you must change, since changing any particular setting can easily introduce a new conflict.

5.4.3: DOS Workstation Hardware Installation

The process of installing PicLan hardware into DOS workstations is governed primarily by whether PicLan is being installed into an existing Novell NetWare site.

Network Workstations

Workstations running Novell NetWare or other supported DOS or Windows networking software already have all of the network hardware required to run PicLan. No additional hardware installation is required.

Stand-alone PicLan Networks

In stand-alone PicLan networks, the DOS workstations must have a PicLan Ethernet network interface adapter installed. This adapter can be an NE2000, WD/CPX, or DEC 21x4x PCI style adapter.

Choosing Adapter Configuration

Installing a PicLan network interface adapter requires that you configure the adapter in the same manner as installing the adapter into the Pick host system. You also must be conscious of the same potential addressing conflicts in the DOS workstation as in the Pick host system. In addition, because of the wide variety of DOS peripheral hardware and device-driver type software, you should also consider the following elements:

Using Software like QEMM or 386^MAX

The PicLan network adapters use a shared memory area for communication between the software driver and the hardware. Since the memory area doesn't get mapped into the DOS address space until the driver software is actually executed, some conflicts can occur with DOS memory management software. If you are using software like QEMM or 386^MAX, or you are using DOS 5 be sure to configure the memory management software or DOS to exclude the network shared memory area from the memory it manages. Consult your memory management software's documentation for more details.

DOS Services Gateway Installation

The PicLan DOS Services gateway is simply another PC DOS workstation connected on the network. It requires the same hardware as any other DOS workstation. The DOS Services Gateway may run on a NetWare non-dedicated router but performance issues may be relevant.

5.5: Installing Software

5.5.1: Pick Host System Installation

The Pick host software installation is identical for all systems (with the possible exception of the type of media and how to access the floppy or tape drive. For each system you:

    SET-FLOPPY (HA)                               most systems
    SET-SCT                                             GA systems
    T-SELECT 0                                      PC/OS or Mentor PRO systems
    SET-VTAPE a:\pic-lan.tap               UPBOARD systems
    T-LOAD MD (O)
This PicLan INSTALL proc will execute. Answer the install proc's questions as appropriate. The install proc performs the following steps:

After the PIC-LAN account is created, you need to take the following actions to actually activate the PicLan driver.

R83, R91, and PC/OS

AP/Native, Pick/64+, and Sequoia PRO

AP/Pro and Mentor PRO


5.5.2: The PicLan Account Maintenance Menu

When you log to the PicLan account, the system will display a menu of PicLan functions that are performed in maintaining the PicLan software. Each of these functions is also available as a separate PicLan TCL command:

1. Display or printing release notes.

This select is mostly obsolete. PicLan release notes are now located on the DOS/Windows client diskette in Windows help file format. When you install the PicLan client files, an icon for the release notes will be created in the PicLan group.

2. Upgrade SYSTEM and ERRMSG files.

This select re-executes the initial install function. If you restore the PIC-LAN account, you may need to re-execute this function to setup SYSTEM and ERRMSG items that an account-restore does not include.

3. Configure PicLan Software.

This function is used to run the PL-CONFIGURE program. PL-CONFIGURE is used to enter PicLan license, hardware, and software settings.

4. Install PicLan into the system COLDSTART function.

This selection will display help text about how to add the PL-LOAD command to the system COLDSTART procedure. Some systems add the PL-LOAD command automatically while other require that you do this manually.

5. Setup another account with PicLan verbs.

If you wish to execute PicLan function from another account, this selection will copy the necessary verbs and q-pointers to the other accounts master dictionary.

6. PicLan status display.

This selection executes a PL-STAT command. You can only use this selection after the PicLan PL-LOAD program has completed successfully.

7. Load PicLan ABS

This selection load the PicLan ABS code on OA platforms (AP/Native, Pick/64+, and Sequoia PRO). You must execute this selection before the PL-LOAD program will run. You need to re-execute this selection if the system ABS file becomes corrupted and you have had to re-load system ABS.

5.5.3: Configuring the Host Software

The PicLan software is configured by executing the PL-CONFIGURE verb. This program will prompt you to enter all of the required PicLan configuration information. The PL-CONFIGURE program begins by displaying a menu of 3 areas of configuration information that you need to enter in order for the PicLan software to execute.. These three areas are License, Hardware, and Software.

Entering License Information

The first step is to enter the PicLan license information. AP/Pro and Mentor PRO do not have license information (the license is obtained directly from Pick or GA). In order to complete this section, you must have received a PicLan License Configuration from Modular Software Corporation. Simply enter the information from the License Configuration directly.

Entering Hardware Configuration Information

Next, you will need to enter the hardware configuration for the PicLan adapter that you have installed. This information includes the driver type, parameters for the driver such as I/O address and IRQ, IPX addressing information, and network retry time values. The fields prompted for are:

NE2000 Drivers

WD/CPX Drivers

PCI Drivers

ArcNET Drivers (R91 only)

IPX Drivers (UpBoard only)

Once hardware information is entered, the PL-CONFIGURE program will ask:

If you are running a single segment IPX network, you can answer no, and PicLan will auto-configure it's IPX addressing. If your network is multi-segment and you are have IPX routers, answer yes and PicLan will prompt for:

Supply the information that matches the IPX router environment you are using. If you are using NetWare file servers as IPX routers, you can execute a CONFIG command on the server console to display all loaded LAN drivers, their frame types, and their associated IPX network numbers. You should enter the frame type and IPX network number into the Pick host that matches the frame type and network number bound to the IPX driver for the segment that the Pick system is connected to.

When you must enter a network address and frame type

If you are configuring a PicLan DOS Services Gateway and are going to run the DSG on a Windows 95 or Windows NT workstation, then you must configure the Pick host to use IPX routers even if you do not have routers present. This is because the Win95/NT systems have trouble advertising IPX services to a network that does not have a network address and frame type. In this case, you can use frame type 802.2 and network number 1. You will also need to configure the Win95/NT workstations for the same frame type and network number.

If you are configuring any type of remote access software to allow users to dial-in to your network and access your Pick host, then you must configure PicLan and your workstations and remote access devices for IPX routers. Remember that the remote access server is a router.

Network Timeout Fields

PicLan includes three fields that you can use to tune network error retries. Most users do not need to change the values of these field. Some of the fields are completely theoretical in nature and changing their values has not been needed at "any" sites.

Retry time

This field controls the amount of time that the Pick host system will wait for an acknowledgment packet before sending a retry. The default is 500ms (one half second). This value is a good compromise for most local and responsive wide-area networks. If you have a local area network with extremely high collision rates, setting this value to a lower number (like 200ms) may improve terminal echo "feel" If you have a wide-area network with long round trip delay times, then setting this number to a higher value (like 2000ms) may make connections more reliable and reduce WAN traffic. If you change this value, you should also change the Timeout=... value in the PL-CFG.INI file.

Auto XOFF time

This field contains the amount of time before PicLan will automatically "flow control" an outbound terminal channel that has no activity. This parameter is specifically for workstations that are using on-demand routers (typically ISDN). There are no known cases where this parameters ever needs to be set on a Pick host.

Watchdog time

This field contains the number of seconds between PicLan watchdog packets. Watchdog packets are packets sent from the Pick host system to make sure that clients are "still alive". The default is one watchdog packet every 5 seconds globally (not per connections). There are no known reasons why you should need to change this value.

Entering Software Configuration Information

The Software Configuration section controls how the PicLan software presents services to clients on the network.

Choosing a Host System Name

PicLan Pick host systems, because they operate in a network environment, must have a unique name by which they may be referred to. When you install the PicLan software, you will need to specify a host name for each Pick host system that you install. The following guidelines apply to naming Pick host systems.

In choosing a host name for your Pick system, try to use a name that makes sense to the users. Remember that the host system's name is used often in accessing the Pick system via the network and simple, easy to remember names are best.

In NetWare networks, the PicLan host system names are separate from other network names such as file-server names. You are technically allowed to use the same name for a PicLan server and for a NetWare server without conflict but this may prove confusing to some users.

Setting Up Terminal Port Assignments

In order to access Pick terminal session over the network, you must configure Pick ports that are to be used by the network. In configuring ports, keep the following guidelines in mind:

The PicLan port configuration information can be modified at any time, but the changes do not take effect until the system is next booted.

You usually cannot re-assign serial ports that are controlled by an intelligent serial board device driver.
Terminal port assignments for the UPBOARD are established by the UPBOARD configurator program.
PicLan network ports for R91 must be set to system "dynamic" ports that are configured when the operating system was initially loaded. You cannot overload ports that have serial hardware associated with them.
PicLan port assignments for AP/Native system should typically be set up to occupy Pick ports starting from the highest port number. For example, on AP/Native version 5.2.4, you would typically assign ports 33-64 as network ports. Also, the maximum number of ports you can assign is 32 and it is often appropriate to allocate all 32 ports even if your Pick license or PicLan license is much smaller.
PicLan port assignments for AP/Protected Mode should be set up to occupy Pick ports that do not have any installed serial adapters. You can assign ports starting at 256 and counting backwards. The PicLan driver allocates memory for AP/Pro dynamically, so do not assign more ports than you actually will be using if you have a system with a very small amount of memory.
Pick/64+ 2.2
The PicLan installation program will automatically configure PicLan for any unallocated Pick ports at install time. You can subsequently add additional ports that overlay Pick/64+ serial adapters.
Pick/64+ 2.3 Seq/PRO
PicLan utilizes Pick ports that are configured when Pick/64+ 2.3 is initially installed. You can also change this configuration information by performing a 'C' (Configuration) restore from the Pick/64+ 2.3 System 1 diskette. These ports are configured starting at port 1.
PicLan utilizes Pick ports that do not have serial adapters associated with them. You should initially install PC/OS with enough PIBs allocated to accommodate both serial port and network users. Changing the number of PIBs requires that you completely re-load PC/OS.
Setting Up PicLan Server Process Port Assignment

In order to use PicLan file transfer and printing functions, you must assign a single, dedicated Pick port to act as a "server-process" for the PicLan software. The PicLan server process is a Pick process that executes continuously and processes requests from a number of sources involving network file transfer and printer functions. The PicLan server process sleeps most of the time and only activates when it has work to do. Where the server-process is assigned varies from system to system. On some systems, the server-process executes on a real port. On other systems, the server-process executes on a phantom.

The server-process port on an R83 system should be set to an otherwise unused port number without a network assignment and without a serial controller present. In general, you should use the highest numbered user port on the system as the server-process port.
The server-process port on an AP/Native system runs on an AP phantom port.
The PicLan server-process should be configured to execute on a port that has been assigned as a network port by the UPBOARD configurator program. In general, you should use the highest numbered PicLan network port as assigned by the UPBOARD configurator. If you assign the server-process in this manner, that port will no longer be available for network connections. If you assign the server-process port to another port on the system, both server-process and overall system performance will be significantly degraded.
Pick/64+ Seq/PRO
The SERVER-PROCESS port for Pick/64+ is automatically assigned to the highest phantom port number.
The SERVER-PROCESS port for R91 must be configured to a system "dynamic" port. In general, you should assign the server-process port just above the PicLan network ports.
The SERVER-PROCESS port for PC/OS and Mentor Pro should be assigned to the highest available port. This ports operates as a PC/OS background port and will count as a Pick login within your licensed user count.
Moving the Pick Host

If the Pick host is being relocated to another segment on a NetWare network, you must use the PL-CONFIGURE verb to change the hardware assignment for Network Number. The Pick host Network Number must always match the NetWare network number for the Ethernet segment the Pick host is on.

Changing Network Adapter Cards

On systems with PicLan license authorization codes, PicLan uses the node address of the installed network adapter as a hardware key. If you re-configure your Pick system's hardware, you must be sure to install the PicLan network interface adapter that matches your PicLan license configuration.

The UPBOARD PicLan license is keyed to the UPBOARD co-processor hardware board. You may freely change network adapters without re-entering your configuration license information. If you need to change UPBOARD co-processor hardware, contact Modular Software for a new software authorization code.
The R91 PicLan license is keyed to the GA system serial number. You may freely change network adapter node id settings without re-entering your configuration license information. If you need to change licensed system serial numbers contact Modular Software for a new software authorization code.

5.5.4: Workstation Installation

Installing on DOS Workstations

PicLan is installed on a DOS workstation by running the INSTALL.EXE program located on the first PicLan client diskette. This program will create a C:\PICLAN directory and load all of the PicLan driver and utility files into that directory. You can also optionally load NDIS and Packet-Driver support files into the C:\PICLAN\NDIS and C:\PICLAN\PACKET directories.

Installing on Windows Workstations

PicLan is installed on a Windows workstation by running the SETUP.EXE program located on the first PicLan client diskette. This program will create a C:\PICLAN directory and load all of the PicLan driver and utility files into that directory. The SETUP program will also create a PicLan group and copy some support DLL files to the Windows SYSTEM directories. After the SETUP program completes, you should execute the PL_SETUP program (available as an Icon in the PicLan group) to setup your network driver configuration.

Installing from Downloaded Files

If you have received the PicLan client in the form of a electronic download instead of on diskette, you must create DOS installation diskettes first. The PKZIP archive that you downloaded includes a README file and a utility program that will create 1.44Meg floppy disks from the files that you downloaded. This program must be run under DOS, Windows 3.1, or Windows 95 (it is not compatible with Windows NT). After you create the installation diskettes, you can run INSTALL.EXE (for DOS) or SETUP.EXE (for Windows) from the first diskette.

Making a Hard-Disk Based Install Directory

If you are running an underlying DOS or Windows network, you may wish to create a shared install directory that eliminates the need for diskettes at each workstation. Simply create a directory on your network that is accessible for all users. Copy the contents of all of the PicLan client diskettes into that temporary directories named ...\disk1, ..\disk2,...\disk3, and ...\disk4 and then run INSTALL.EXE or SETUP.EXE from the first directory.

Installing the Utility Programs on a LAN

When installing on a NetWare LAN, the PicLan utility files may be copied to a shared search drive on a network file server. In this case, do not place the PL-CFG.INI file in the same directory. (In fact, do not place any copy of PL-CFG.INI on any global network search drive.) Rather, each user should have a copy of PL-CFG.INI in a private directory that is on their DOS PATH - this will allow each user to have their own configuration, instead of all of the users sharing the same configuration. This type of installation is not recommended for systems running Windows as a number of shared DLLs are required and these DLLs are configured differently for different types of Windows installations

5.5.5: DOS Workstation Configuration

Before PicLan can be used from a workstation, you need to configure PicLan and the workstation driver software for PicLan functions. This configuration varies depending on the type of network and the type of drivers in use:

Configuring with NetWare

PicLan does not require any special configuration when used with NetWare IPX drivers. PicLan will operate with either newer ODI style drivers, or with older IPX.COM style drivers and is also compatible with most versions of NetWare shells including NETX and VLMs.

Configuring with NetWare Lite and Personal NetWare

PicLan does not require any special configuration when used with NetWare Lite IPX drivers.

Configuring with other IPX Drivers

PicLan does not require any special configuration when used with third-party IPX drivers such as Spry Concurrent.

Configuring a Stand-alone PicLan Network

If PicLan is to directly drive an installed Ethernet adapter, you must configure the DOS workstation with:

Modular Software does not test or endorse particular brands of adapters and it is your responsibility to obtain adapters that are compatible with PicLan.

The following types of adapters are known to NOT work with PicLan in stand-alone mode:

After you have installed a compatible network adapter, you must update the PL-CFG.INI DOS configuration file with the configuration information for the adapter you have installed. The PL-CFG.INI file is a DOS text file that can be edited with any DOS text editor such as EDLIN or the DOS 5 EDIT program. You will need to update the following fields:

NE2000 Style Adapters

With NE2000 style adapters, you must set the EnetIRQ and EnetIO settings to match the adapter. You should set the EnetRAM setting to 0.

WD8003/8013 Style Adapters

With WD style adapters, you must set the EnetIRQ, EnetIO, and EnetRAM settings.

PCI Adapters

With 21x4x style adapters, you must set the EnetMedia field. These settings use the same numbers as the PicLan host uses.

Configuring with Packet Drivers

PicLan supports the Clarkson University Packet Driver specification with both Ethernet and ArcNET networks. In order to use PicLan with a packet driver, you must first install the packet driver TSR using the instruction that came with it and then load the PicLan PL-DEV driver TSR. PL-DEV will automatically locate the installed packet driver and use it.

PicLan supports multiple frame types with packet drivers, although the packet driver specification itself often does not. For this reason, you should always use Ethernet-II frames with packet driver installations and you need to set the EnetFrame field in the PL-CFG.INI file to:

If you wish to verify that PicLan is using the correct driver, include the /i (information) option when you load PL-DEV as in:

    PL-DEV /i
The PL-DEV program will then display the driver that is actually in use. DRV_EPKT refers to the Ethernet Packet driver and DRV_APKT refers to the ArcNET Packet driver.

Configuring with NDIS 2.0 Ethernet Drivers

A number of networks use NDIS version 2.0 network drivers on DOS workstations including:

PicLan supports the NDIS 2.0 real-mode network driver specification by using the DIS_PKT.DOS NDIS to packet-driver converter program. DIS_PKT.DOS allows NDIS Ethernet drivers to appear to DOS software as both an NDIS and a packet driver Ethernet driver concurrently. In order to use the DIS_PKT.DOS program, you need to:

    drivername = PKTDRV
    bindings = SMCMAC
    intvec = 0x61
In this example, a SMC Ethernet adapter is in use and SMCMAC refers to the MAC driver that you are using. You installation will probably use a different MAC driver name, so you should substitute this name for the 'bindings' field in the PKTDRV section.

The "intvec" setting in this example specifies a software interrupt vector that the packet driver uses. If the specified vector is in use, you should specify another value between 0x60 and 0x7F.

The DIS_PKT.DOS program only supports Ethernet-II format frames, so you need to specify:

in the PicLan PL-CFG.INI file.

If you have configured everything correctly, you should see the message

    The command completed successfully
when the NET START command executes. You should also see a copyright notice for the DIS_PKT.DOS driver as it loads. If you do not see these messages, some piece of information is missing or has been entered incorrectly in the CONFIG.SYS or PROTOCOL.INI file.

After you have configured the CONFIG.SYS and PROTOCOL.INI files, you can load the PL-DEV PicLan driver TSR. If you load PL-DEV with the /i option, it should display the type of network driver you are using. NDIS installations should be using DRV_EPKT bound to the MAC/DIS converter.

5.5.6: Windows Workstation Configuration

If you are using PicLan from a Windows workstation, you should always run the Windows version of the PicLan install program (SETUP.EXE) and then run the PL_SETUP.EXE program to correctly configure the PicLan drivers. The PL_SETUP program will also give you a list of driver configuration steps that you need to take to configure the software.

Configuring with Windows 3.1

PicLan runs with Windows 3.1 as a DOS hosted network. You should configure the DOS network drivers (stand-alone, Packet, NDIS, or IPX) prior to starting Windows and then load the PL-DEV program before you start Windows.

Configuring with Windows for Workgroups version 3.1

PicLan is compatible with Windows for Workgroups version 3.1 as either an NDIS 2.0 DOS client or by using the MSIPX program to create an IPX driver.

Configuring with Windows for Workgroups version 3.11

PicLan is compatible with Windows for Workgroups (WFW) version 3.11 when WFW is configured to use real-mode NDIS-2 Ethernet drivers. PicLan will not work with WFW 3.11 protected mode (NDIS-3), protected/real-mode (NDIS-3/2), or ODI Ethernet drivers.

If you network is using Novell IPX drivers (IPX.COM or IPXODI), you should not follow these installation instructions. Follow the instructions for use with IPX drivers instead.

In order to configure your workstation for PicLan and WFW support, you need to follow the following procedure:

    [network driver]
    A copyright message for your network adapter driver.
    A copyright message for the MAC/DIS converter (DIS_PKT.DOS)
    The message `The command completed successfully'.
Configuring with Windows 95

PicLan runs with Windows 95 either as a Windows 3.1 client with real-mode IPX drivers (Novell IPX.COM or IPXODI.COM) or with 32-bit drivers using the NDIS IPXSPX stack that is built into Windows 95 or included with the Novell Client-32. The PicLan Windows 95 client allows you to choose between the following client driver types:

Real Mode IPX

You can install Windows 95 as a Windows 3.1 client with real-mode IPX drivers (usually older NetWare IPX.COM or IPXODI drivers). You then load the PicLan PL-DEV TSR program in AUTOEXEC.BAT. This installation method is not recommended for most installations and prevents 32-bit PicLan client programs from functioning.

32-bit IPX support with WSOCK32.DLL

This is the default configuration for a Windows 95 workstation. By using IPX calls that are available in the Microsoft supplies WSOCK32.DLL (this is the 32-bit Winsock support DLL), PicLan can use IPX services directly from multi-threaded 32-bit applications. You should use this configuration unless you are using Novell's Client-32 requestor.

32-bit IPX support with WS2_32.DLL

WS2_32.DLL is a new Microsoft supplied interface that supports the new Winsock 2 interface specification. This DLL is not available on Windows 95, but should be available on Windows 98. Support for this interface is included in PicLan primarily for Windows NT clients, but should also work with Win98.

32-bit IPX support with Novell's Client-32

The NWSIPX32.DLL can be used as an IPX interface for clients that are running Client-32 for Windows 95. If you are running Client-32, you should use this option, although the WSOCK32.DLL option also works in most cases. If you are running Client-32 with the Novell IPX/IP or IP/IP gateway installed, then you must use the NWSIPX32.DLL option.

Configuring with Windows NT

Windows NT is configured nearly identically to Windows 95 with the exception of not having real-mode IPX stacks available.

Additional Configuration of the PicLan Utility Programs

In order to use the PicLan utility programs, you must first configure these programs for operation on the network. In general, this configuration process involves customizing the PL-CFG.INI file.

The PL-CFG.INI file is a normal DOS text file which contains the configuration information for the PicLan DOS Terminal Emulator, Pick to DOS copy functions, etc. It can be changed by using any text editor to edit it. DOS EDIT or Windows NOTEPAD are preferred editors for editing this file (If a word processor is used to edit PL-CFG.INI, be sure to save the file as text only, without any formatting information.)

If you wish to use a file with a name other than PL-CFG.INI, you can use the environment variable "PL-CFG.INI" to specify the name. For example, the DOS command

    set pl-cfg.ini=c:\etc\my_plcfg.ini
will cause the file MY_PLCFG.INI, located in C:\ETC, to be used. (When setting the "PL-CFG.INI" environment variable, be careful not to include any spaces on the command line on either side of the equal symbol.) Consult your DOS documentation for information on environment variables, and the SET command.

The PL-CFG.INI file is searched for using the following algorithm: If the "PL-CFG.INI" environment variable is defined, then its value is the name of the file; otherwise, "pl-cfg.ini" is the name. If the name of the file includes a path specification, then the file must be located in that directory - no searching will be performed. If the name does not include a path specification, it is searched for first in the current directory, then in the directory that PL-TERM.EXE (or other PicLan utility program) was loaded from, and finally in all of the directories on the DOS PATH.

See the comments in the PL-CFG.INI file for instructions on how to use it to configure the Emulator. Before editing PL-CFG.INI, you should save a backup - it is possible to cause the Emulator to be non-executable if incorrect configuration information is supplied.

5.5.7: DOS Services Gateway

There are three different versions of the DOS Services Gateway program:

If you are running a Windows based network and wish to interact with Windows print queues, then you will need to run the PL-DSGW.EXE program. This program will run under Windows 3.1, Windows 95, or Windows NT (it is likely most robust under Windows NT).

If you are looking for maximum performance out of a DSG for transferring large files, the DOS versions are faster than is the Windows version. PL-DSG also has lower DOS memory usage than does PL-TERM.

In order to configure a DSG, you must setup the User= and DosServicesGateway= fields in the PL-CFG.INI file.

5.6: Troubleshooting

There are a number of techniques available for troubleshooting PicLan installation problems. The techniques that you can use are quite dependent on the configuration of your PicLan installation.

5.6.1: Troubleshooting the Pick Host

Troubleshooting usually begins at the Pick host. First, you should verify that the PL-LOAD program completes without errors and that it display a "reasonable" network node address. An Ethernet node address is a 6 byte number expresses as 12 hexadecimal digits. The number should not be all zeros nor should if be all FFs or a repeating pattern. If the node address is not correct this is an indication that PicLan is not recognizing the network adapter. You should double check adapter I/O address settings (both in PicLan and on the card) as well as verifying that the adapter is compatible with PicLan and setup for the correct operating mode. You should also verify that the adapter does not conflict with other cards installed in the system. Often, it is helpful to de-install all non-essential adapters (tape controllers and multi-port serial controllers) when initially installing PicLan to make it easier to identify conflicts.

If the PL-LOAD command appears to have executed without errors, the next step is to execute a PL-STAT (R) command from TCL in the PIC-LAN account. This command should display the name of the Pick host, at TP field value of 07 (01 is legal as well if you have elected to not start a PicLan SERVER-PROCESS), and a reasonable node address. Below the server list display, you should see a number of driver statistics that can help you pinpoint problems with your installation:


If you system is not communicating at all, this field probably means little. If you are communicating and this field has a high number (more than 1% of the "Receives" field), then this may point to a network cabling problem or other systemic network troubles.

Xmit Timeouts:

Typically this number should always be zero. If this number is not zero, then it indicates either an IRQ problem with the adapter, or a cable-off situation. Cable-off can also mean that the adapter is configured for the wrong type of interface (eg. TP instead of COAX).


This count indicates the number of connections that were broken because of communications errors. The most common cause is a workstation being powered down without shutting down their PicLan connection first.


A reopen is a particular type of retry. It is tracked for historical reasons, but is probably not a concern with current PicLan releases.


This is the total number of packets that have attempted to have been transmitted.


This is the total number of IPX packets addresses to this system that have been received.

BCast Receives

This is the total number of IPX broadcast packets that this system has received


This is the total number of non IPX packets or IPX packets with invalid network addressing received by this system.

If your PL-STAT (R) display shows zero values for all types of receives, this indicates that the network adapter is not receiving any traffic. If the connected network is running any other type of network system (such as NetWare, Windows for Workgroups, or Windows NT), then the Pick system should see at least some traffic. This indicates that the adapter is not functioning. You should check IRQ settings as well as verify that the adapter is compatible with PicLan.

5.6.2: PicLan / NetWare Networks

If you network is built around server-based NetWare, you can use NetWare function to determine if the Pick host system is transmitting. First, with this type of network, you should always configure PicLan for use with IPX routers and specify a frame type and network number.

If NetWare is displaying error message once a minute about "router as node xxxxxxxxxxxx is claiming network number xxxxxxxx, should be network xxxxxxxx", this indicates that the network number and/or frame type on the Pick host do not match NetWare. You should change the Pick host so that it's frame type and network number match those of the NetWare segment that you are connected to.

If NetWare is not displaying error messages, you can execute a NetWare console command DISPLAY SERVERS. This command will display all known IPX servers with their names and number of router hops from the NetWare server. The name of the Pick system should be included in the list (this is a good reason not to give the Pick host the same name as another server on the network). If the name is not included, this indicates that the Pick host is not transmitting on the network or is transmitting on a frame type that NetWare is not listening to.

5.6.3: PicLan / Windows NT server

If you have Windows NT server with "routing and RAS" installed, you can use the RRAS administrator to display IPX servers by server type and name. PicLan servers use an IPX server type of 03FF (hex).

5.6.4: Other Troubleshooting Techniques

There are several other techniques that you can use to troubleshoot the network. If you are trying to determine that the Pick host and its network adapter are working, you should boot the Pick host system to DOS and try to connect to the DOS network using DOS drivers (if this is applicable). This can help show you whether the Pick system's hardware is operable in an environment that may be more familiar to your or your network technical support personnel.

You can also configure your network to include a DOS Services Gateway (even if you have not licensed the DSG option) for the purpose of using the DSG as an advertising server under DOS or Windows. The PL-STAT.EXE command (or the DISPLAY SERVERS command from the NetWare console) should see the DSG system even if there is no Pick host present.

5.6.5: The Most Common Types of Installation Problems

Some installation problems happen time and time again. Here are the most common problems: