]> GREENSTONE DIGITAL LIBRARY Ian H. Witten and Stefan Boddie Department of Computer Science
University of Waikato, New Zealand Back to manual index Back to top index Greenstone is a suite of software for building and distributing digital library collections. It provides a new way of organizing information and publishing it on the Internet or on CD-ROM. Greenstone is produced by the New Zealand Digital Library Project at the University of Waikato, and developed and distributed in cooperation with UNESCO and the Human Info NGO. It is open-source software, available from http://greenstone.org under the terms of the Gnu General Public License. We want to ensure that this software works well for you. Please report any problems to greenstone@cs.waikato.ac.nz Greenstone gsdl-2.50 March 2004

This document explains how to install Greenstone so that you can run it on your own computer. It also describes how to obtain associated software that is freely available—the Apache Webserver and Perl. We have striven to make the installation procedure as simple as it possibly can be. The software runs on different platforms, and in different configurations. Consequently there are many issues that affect (or might affect) the installation procedure. Section mentions some questions that you will need to consider before installing Greenstone. Section details the installation procedure for all the different versions; you need only read the part that relates to your operating system. Section describes the demonstration digital library collections that are included in the distribution. Section explains how to set up common webservers, Apache and Microsoft PWS/IIS, to work with Greenstone. Section describes various Greenstone configuration options, and Section shows how to make a personalized home page for your digital library installation. Finally, an Appendix lists pieces of associated software and how to obtain them.

The complete set of Greenstone documents include five volumes: Greenstone Digital Library Installer's Guide (this document) Greenstone Digital Library User's Guide Greenstone Digital Library Developer's Guide Greenstone Digital Library: From Paper to Collection Greenstone Digital Library: Using the Organizer

Copyright © 2002 2003 2004 2005 2006 2007 by the New Zealand Digital Library Project at the University of Waikato, New Zealand. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the section entitled “GNU Free Documentation License.”

The Greenstone software is a collaborative effort between many people. Rodger McNab and Stefan Boddie are the principal architects and implementors. Contributions have been made by David Bainbridge, George Buchanan, Hong Chen, Michael Dewsnip, Katherine Don, Elke Duncker, Carl Gutwin, Geoff Holmes, Dana McKay, John McPherson, Craig Nevill-Manning, Dynal Patel, Gordon Paynter, Bernhard Pfahringer, Todd Reed, Bill Rogers, John Thompson, and Stuart Yeates. Other members of the New Zealand Digital Library project provided advice and inspiration in the design of the system: Mark Apperley, Sally Jo Cunningham, Matt Jones, Steve Jones, Te Taka Keegan, Michel Loots, Malika Mahoui, Gary Marsden, Dave Nichols and Lloyd Smith. We would also like to acknowledge all those who have contributed to the GNU-licensed packages included in this distribution: MG, GDBM, PDFTOHTML, PERL, WGET, WVWARE and XLHTML.

The Greenstone software runs on different platforms, and in different configurations, as summarized in Figure .

There are many issues that affect (or might affect) the installation procedure. Before reading on, you should consider these questions: Are you using Windows or Unix? If Windows, are you using Windows 3.1/3.11 or a more recent version? Although you can view collections on 3.1/3.11 machines, and serve other computers on the same network, you cannot build new collections. The full Greenstone software runs on 95/98/Me, and NT/2000. If Unix, are you using Linux or another version of Unix? For Linux, a binary version of the complete system is provided which is easy to install. For other types of Unix you will have to install the source code and compile it. This may require you to install some additional software on your machine. If Windows NT/2000 or Unix, can you log in as the system “administrator” or “root”? This may be required to configure a webserver appropriately for Greenstone. Do you want the source code? If you are using Windows or Linux, you can just install binaries. But you may want the source code as well—it's in the Greenstone distribution. Do you want to build new digital library collections? If so, you need to have Perl, which is freely available for both Windows and Unix. Is your computer running a webserver? The Greenstone software comes with a Windows webserver. However, if you are already running a Web server, you may want to stay with it. For Unix, you need to run a webserver. Do you know how to reconfigure your webserver? If you don't use the Greenstone webserver, you will have to reconfigure your existing one slightly to recognize the Greenstone software. Versions of Greenstone are available for both Windows and Unix, as binaries and in source code form. The Greenstone user interface uses a Web browser: Netscape Navigator or Internet Explorer (version 4.0 or greater in both cases) are both suitable. In case you don't already have a Web browser, Windows versions of Netscape are provided on the CD-ROM.

If you are a Unix user, please skip ahead to Section . For Windows users, if you want just a simple, straightforward installation, go through the following “simple installation” procedure. The Greenstone system occupies about 40 Mb of disk space. For Windows Vista user, go to the for more instructions on getting Greenstone running on Vista. If you choose anything other than the default setup, you will have to decide whether you want to install the binary code or the source code. If in doubt, choose the binary code. The installation procedure is the same for both. The following sections tell you more about the options you will be presented with. When you've finished installation you should skip ahead to Section . To install the Windows version from the CD-ROM, insert the disk into the drive (e.g. into D:). If the installation procedure does not start automatically after about 20 seconds, click on the Start menu, select Run and type D:\setup.exe, where “ D ” is the letter that identifies your CD-ROM drive. For Windows 3.1, select Run from the “File manager” and type D:\Windows\win3.1\setup.exe. For the simplest installation, just accept the default at each point by clicking the Next button. That's all you need to do! Greenstone is installed in the directory C:\Program Files\gsdl. Once installation is complete, to start your Greenstone system click on the Start button, open the Program menu, and select Greenstone Digital Library. This brings up a dialogue box: just click Enter Library. This automatically starts your Internet browser and loads the Greenstone Digital Library home page, which should look something like the example in Figure . You enter the Greenstone Demo collection by clicking on its icon.

There are two separate Windows binary programs on the CD-ROM: the Local Library and the Web Library. The default installation described above selects the Local Library version. We strongly recommend that you use this version. The Web Library, which is much harder to set up, is only necessary if you already run a web server and want to use it for Greenstone. Despite its modest name, the Local Library offers a complete, self-contained, web-serving capability. Local Library. This enables any Windows computer to serve pre-built Greenstone collections. The Greenstone Demo collection will automatically be installed; you can also install the other collections on the CD-ROM (Section ). The Local Library software is the same as that used on CD-ROMs produced by the Greenstone system. The Local Library is intended for use on standalone computers or computers that do not already have webserver software. It contains a small built-in webserver so that other computers on the same network can also access the library. (However, the webserver has limited configurability.) The Local Library software automatically determines whether your computer has network software installed or is connected to a network. It operates correctly under any combinations of these conditions. However, there are two possible problems that may be encountered. Greenstone may cause an unwanted telephone dialup operation; fail to run because network software is installed, but installed incorrectly. A restricted version of the Local Library is supplied which is intended for use in these situations. The restricted version only works with Netscape (not Internet Explorer). When you invoke the Local Library version of Greenstone, the dialogue box contains a button that allows you to use the restricted version instead. Unless the above problems arise, you should always use the standard version. Web Library. This enables any computer with an existing webserver to serve pre-built Greenstone collections. As with the Local Library above, the Greenstone Demo collection will automatically be installed. You can also install the other collections on the CD-ROM (see Section ). The Web Library differs from the Local Library because it is intended for computers that already have webserver software. To run the Web Library, you also need Webserver software. One possibility is Apache (see Appendix ). The Collector. This component, which is included in both the Local Library and the Web Library, allows you to build collections containing material of your choice. (You will not be able to use the Collector on a Windows 3.1/3.11 machine.) An advantage of the Local Library version of Greenstone is that it runs “out of the box” and does not require any special configuration. For the Web Library version, however, you will have to make some adjustments to your webserver setup. If you already have a webserver, some small changes have to be made to its configuration to make your Greenstone installation operate. The install script explains what these are for the Apache webserver—see Section for instructions for configuring the PWS and IIS webservers. You may need help from a system administrator to reconfigure an existing webserver—they should be able to understand the instructions printed by the install script. If you do not already have a webserver, you will have to install one. (See the Appendix for information on the Apache webserver.) Then you will have to configure it appropriately. Section gives a detailed account of the parts of a webserver installation that affect Greenstone, and how they need to be altered. It comes down to including half a dozen or so lines in a configuration file. The Greenstone source code occupies 50 Mb of disk space, but to compile it you will need about 90 Mb. To compile the source on Windows you need The Microsoft Visual C++ compiler. (We are currently sorting out some minor problems in compiling Greenstone with various Windows ports of GNU GCC.) (You do not need GDBM, the Gnu database manager, because it is included in the Greenstone source distribution.) It is unlikely that you will be able to compile Greenstone on a Windows 3.1/3.11 machine. In the event that you recompile Greenstone and wish to use the recompiled version to create CD-ROMs, you should note that code produced by recent versions of the Visual C++ compiler does not run under Windows 3.1/3.11, although there is no problem with later Windows systems (95, 98, Me, NT, 2000). If you want your CD-ROMs to operate on early Windows machines, you will need a different version of the compiler. Moreover, Greenstone uses STL, the C++ standard template library, and although these compilers sometimes come with STL, the provided version does not always work properly. Hence to recompile Greenstone in such a way that it produces CD-ROMs that work on early versions of Windows, you need The Microsoft Visual C++ compiler, Version 4.0 or 4.2. An external version of STL, the C++ standard template library. STL is packaged with Greenstone for use with these compiler versions. Note that the Windows installation procedure does not attempt to compile Greenstone for you if you choose to install the source code. For platform- and compiler-specific instructions on compiling Greenstone, see the Install.txt document which is placed in the top-level Greenstone directory (C:\Program Files\gsdl by default) during the installation procedure. If you are using Windows Vista, you need to run the Greenstone installer as an administrator. If you have problems running GLI, it could be because access is denied. Here is a solution thanks to Amanda Hahn. Running Greenstone in XP compatibility mode does not work. Instead, you MUST change the permissions for the entire greenstone folder otherwise you will keep being told "Access is denied." You must also be an administrator to change this property. Below are the steps to change the permissions: Right click on the entire Greenstone program folder (normally under Program Files). Click Properties. Then Click on the Security tab. Under Group or User Names, click on Users, then click Edit. Click on Users again. Click the checkbox in the window beneath that says Full control in the Allow column. Here you MUST hit Apply. If you just hit OK the properties won't change. Then hit OK.

This section is written for Unix users. (Windows users should skip ahead to Section .) You need to choose whether to install the binary code or the source code. The binary code occupies about 50 Mb of disk space; the source code requires about 160 Mb to compile. The binary code requires an Intel x86-based Linux distribution which includes ELF binary support. Distributions that meet these requirements include: RedHat 5.1 SuSE Linux 6.1 Debian 2.1 Slackware 4.0 More recent versions of these distributions should also work. You will need a webserver: we recommend Apache. We also strongly recommend you to install your webserver before installing Greenstone—this will make it much easier to answer the questions that are asked during the Greenstone installation procedure. If you want to build new digital library collections, you will also need Perl if this is not already on your system. To check, open a terminal window, type perl —v, and see if a message appears specifying, amongst other things, the version number. For most versions of Linux, Perl is installed by default. The Appendix gives information on how to obtain Apache and Perl. The source code is the same for Unix as for Windows. It has been compiled and tested on Linux, Solaris, and Macintosh OS/X; it should be a fairly routine matter to port it to other flavors of Unix. To compile the Greenstone source code on Unix, you need GCC, the Gnu C++ compiler. GDBM, the Gnu database manager. To run the Greenstone software, you also need a Web server and Perl, as described above under Unix binaries. To install the Unix version from the CD-ROM, insert the disk into the drive, and type If you already have a webserver, some small changes will have to be made to its configuration to make your Greenstone installation operate. The install script explains what these are. You will probably need help from your system administrator to reconfigure the webserver—he or she should be able to understand the instructions output by the install script. For your convenience, the output of the install script is written to a file called INSTALL_RECORD in the directory into which you installed Greenstone. If you do not already have a webserver, you will have to install one. The Appendix gives information on Apache. Then you will have to configure it appropriately. Section gives a detailed account of the parts of an Apache webserver installation that affect Greenstone, and how they need to be altered. It comes down to including half a dozen or so lines in a configuration file. You do not need to be the Unix “root” user to go through the installation procedure above. When it comes to configuring an existing Apache server, however, you may need “root” privileges—it all depends on how Apache is set up. If you install Apache yourself, you can do it as a user without “root” privileges. If you need to work your way around an uncooperative system administrator, you can always install a second Apache webserver on your computer—even if one exists already.

If you are using the Local Library, simply run the Greenstone program from the Start menu. This automatically opens a dialog box that starts your Internet browser and loads the Greenstone Digital Library home page. The Greenstone Demo collection should be accessible from this page. The dialog box contais a File menu item that allows you to change the default browser used by Greenstone. It doesn't matter whether you use Netscape or Internet Explorer, except that if you are running on Windows 2000, we recommend that you use Internet Explorer. If you are using the Web Library, once you have installed the software and configured the webserver, use this URL to enter your Greenstone system: http://localhost/gsdl/cgi-bin/library The Greenstone Demo collection should be accessible from this page. A link to the Collector is provided on the digital library home page. A link to the Administration pages is provided on the digital library home page. The “administrator” user is called admin, with a password that you specified during the installation process. The administrator is authorized to add new users, and to build collections.

The Greenstone Librarian Interface (GLI) is a tool to assist you with building digital libraries using Greenstone. It gives you access to Greenstone's collection-building functionality from an easy-to-use “point and click” interface. GLI is installed automatically with all distributions of Greenstone. It is placed in the subdirectory gli of the top-level Greenstone directory (C:\Program Files\gsdl\gli by default). Note that it runs in conjunction with Greenstone and will not work properly unless it is placed in a subdirectory of your Greenstone installation. If you have downloaded one of the Greenstone distributions, this will be the case. To use the GLI, your computer needs to have the Java Runtime Environment. If it doesn't, the installer will offer to install a version that is included on the CD-ROM. On Unix, you will also need to ensure that Perl is installed (for Windows, Perl is already included in the Greenstone software). Please report any problems you have running or using the Librarian Interface to greenstone@cs.waikato.ac.nz. To run GLI under Windows, browse to the gli folder in your Greenstone installation (e.g. using Windows Explorer), and double-click on the file called gli.bat. This file checks that Greenstone, the Java Runtime Environment, and Perl are all installed, and starts the Greenstone Librarian Interface. To run GLI under Unix, change to the gli directory in your Greenstone installation, then run the gli.sh script. This script checks that Greenstone, the Java Runtime Environment, and Perl are all installed and on your search path, and starts the Greenstone Librarian Interface. The Greenstone Librarian Interface has extensive on-line help facilities. You get help by clicking the Help button at the top right of the screen. This opens up the text to a section that relates to what you are doing—which of the GLI panels you are on. You can click around the help text to learn what you need to know. Use it. If you have downloaded the Greenstone source distribution, you will have the Java source code of the Librarian Interface. To compile it, your computer needs to have a Java Development Kit. The Appendix gives information on how to obtain this. To compile the source code, run the makegli.bat(Windows) or makegli.sh(Unix) files. Once compiled, you can run GLI as described above.

To test Greenstone, point your Web browser at the Greenstone home page and explore the Demo collection and any other collections that you have installed. Don't worry—you can't break anything. Click liberally: most images that appear on the screen are clickable. If you hold the mouse stationary over an image, most browsers will soon pop up a message that tells you what will happen if you click. Experiment! Choose common words like “the” and “and” to search for—that should evoke some responses, and nothing will break.For more information, see the Greenstone Digital Library User's Guide.

To learn more about the innards of your Greenstone installation, consult the Greenstone Digital Library Developer's Guide. It includes (for example) details of the directory structure that has been created, and information about how to configure your Greenstone site. Several demonstration Greenstone collections are included on the CD-ROM. If you have Web access, many others can be downloaded, in either pre-built or unbuilt form, from the New Zealand Digital Library Project website (nzdl.org). The Greenstone Demo collection is a small subset of the Humanity Development Library (HDL), a polished collection. It illustrates that relatively rich browsing capabilities can be provided (so long as suitable metadata is available). It is included automatically when the software is installed. Greenstone also comes with some well-documented example collections whose “about” page describes how they are constructed. They demonstrate various capabilities of Greenstone. The install dialogue will ask you whether you want to include them in your Greenstone installation; the approximate amount of disk space needed for each collection is shown below. In this section we describe how to set up your webserver to work with Greenstone. Note that all this is unnecessary when using the Windows Local Library, because this software works “out of the box” and does not require a webserver. We discuss both the Apache webserver, which is freely available for both Windows and Unix (see the Appendix for details) and Microsoft's Personal Web Server (PWS) and Internet Information Services (IIS) webserver. PWS is the standard Microsoft server for Windows 95/98; IIS is the standard webserver for Windows 2000 and the forthcoming Win dows XP ; Windows NT can use either. The Apache description applies equally to the Windows Web Library and Unix versions (though we use Windows-style terminology and pathnames); the PWS/IIS section applies only to the Windows Web Library. Once you have installed your webserver, the next step is to install Greenstone. We will assume that during the install procedure you have taken the default action for each stage by clicking on the Next button. The result is that the directory C:\Program Files\gsdl is created and the Web Library binary is stored there, along with some supporting files. All webservers use the special URL “localhost” to denote the computer that the webserver is running on. Thus when you install a webserver, you can get at your html documents by typing the URL http://localhost into a browser. If your computer has a domain name set up, this is used instead of localhost to identify your computer from remote sites. Thus on the New Zealand Digital Library's computer, http://nzdl.org and http://localhost are equivalent. If you type http://nzdl.org on your computer you will get the New Zealand Digital Library webserver, whereas if you type http://localhost you will get your own computer's webserver.

The Apache webserver is usually installed in C:\Program Files\Apache Group\Apache and is configured so that the cgi-bin directory is in the subdirectory \cgi-bin and the document root is the subdirectory \htdocs. It is reconfigured by editing the configuration file C:\Program Files\Apache Group\Apache\conf\httpd.conf. This is a text file: it's quite easy to read it to see how things are set up. Depending on how your computer's networking software is set up, you may have to add this line to Apache's httpd.conf configuration file: ServerName localhost If this line is not included, the system attempts to find your server's name. However, there are bugs in some versions of Windows that cause this to fail. In this case, Apache will exit immediately when you start it up. It does display an error message, but it is immediately erased and you probably can't read it. Cgi-bin is a directory from which the webserver treats documents as executable programs. Apache's ScriptAlias directive is used to create a cgi-bin directory. Note that this directive can make any directory a cgi executable directory—it doesn't have to be called “cgi-bin”! Conversely, a directory called “cgi-bin” isn't special unless ScriptAlias has been applied to it. When installed, Apache has a cgi-bin directory of C:\Program Files\Apache Group\Apache\cgi-bin. This means that if presented with the URL http://localhost/cgi-bin/hello, the webserver will attempt to execute a file called hello from within the above directory. There is one Greenstone program, which is called “library.exe”, that needs to be executed by the webserver; it in turn reads a file called the Greenstone site configuration file, or “gsdlsite.cfg”, which needs to be located in the same directory. The best way of arranging this is to use Apache's ScriptAlias directive to create a new cgi-bin directory. Here's the excerpt from Apache's httpd.conf configuration file that adds C:\Program Files\gsdl\cgi-bin as an additional cgi-bin directory: ScriptAlias /gsdl/cgi-bin/ "C:/Program Files/gsdl/cgi-bin/" <Directory C:/Program Files/gsdl/cgi-bin> Options None AllowOverride None </Directory> (It's a curious fact that Apache configuration files use forward slashes in place of standard Windows backslashes.) This means that any URLs of the form http://localhost/gsdl/cgi-bin... will be sought in the directory C:\Program Files\gsdl\cgi-bin, and executed by the web server. For example, if presented with the URL http://localhost/gsdl/cgi-bin/hello, the web server will attempt to retrieve the file C:\Program Files\gsdl\cgi-bin\hello and execute it. However, the URL http://localhost/cgi-bin/hello looks in Apache's regular cgi-bin directory for the file C:\Program Files\Apache Group\Apache\ cgi-bin\hello and executes it, just as it did before. The document root directory is the root of your webserver's directory structure. When installed, Apache has a document root of C:\Program Files\Apache Group\Apache\htdocs. This means that if presented with the URL http://localhost/hello.html, the webserver will attempt to retrieve a file called hello.html from within the above directory. Several files within Greenstone need to be read by the webserver. The simplest way to arrange this is to use the Alias directive, which is just like ScriptAlias except that it applies to ordinary web pages, not cgi scripts. Insert these lines into your Apache configuration file, after the ScriptAlias directive, to add C:\Program Files\gsdl as an additional place to look for documents. Alias /gsdl/ "C:/Program Files/gsdl/" <Directory C:/Program Files/gsdl> Options Indexes MultiViews FollowSymLinks AllowOverride None Order allow,deny Allow from all </Directory> This means that any URLs that match the first argument of Alias (gsdl) are sought as files in the place corresponding to the second argument. In other words, URLs of the form http://localhost/gsdl/... will be sought as files in the directory C:\Program Files\gsdl. For example, if presented with the URL http://localhost/gsdl/hello.html, the webserver will attempt to retrieve the file C:\Program Files\gsdl\hello.html. However, the URL http://localhost/hello.html looks in the regular htdocs directory for the file C:\Program Files\Apache Group\Apache\htdocs\hello.html, just as it did before. Be sure to add the Alias directive after the ScriptAlias directive. Instructing Apache to alias /gsdl before /gsdl/cgi-bin would match the URL /gsdl/cgi-bin/library against the Alias directive rather than the ScriptAlias, and it would be interpreted as a request for a document rather than the result of executing a program. The outcome would be to “display” the binary program file as a page in the Web browser, instead of executing it. You should be aware that if the web library version of Greenstone is set up as instructed above, anyone will be allowed to download any file in the gsdl directory structure. This includes the index files and source documents of any collections you make, the user database, usage logs, etc. If you are concerned about this, you can easily tighten up your webserver configuration to improve security. For the Apache webserver, put these lines into the configuration file instead of those given in the previous subsection: Alias /gsdl/ "C:/Program Files/gsdl/" <Directory "C:/Program Files/gsdl"> Order allow,deny Deny from all <FilesMatch "\.(gif|jpe?g|png|css|mov|mpeg|ps|pdf|doc|rtf|jar|class)$"> Order allow,deny Allow from all </FilesMatch> </Directory> This means that only files whose extensions match the regular expression in the FilesMatch line may be downloaded.

Although neither PWS nor IIS is installed by default on current Windows systems, they can easily be installed using the “Add/Remove programs” control panel . If they are not already on your Windows distribution CD-ROM you will have to download them from the Microsoft web site (www.microsoft.com). The setup procedure for Greenstone is identical for both PWS and IIS. Invoke the Personal Web Manager and perform the following actions. Select Advanced to get the Advanced Options screen. Select Home and click AddFill out the fields as follows: Directory field: C:\Program Files\gsdl Aliasfield: gsdl Access permissions: Read Application permissions: None Click OK This makes Greenstone files accessible to the webserver. Back in Advanced Options, select gsdl and click AddFill out the fields as follows: Directory field: C:\Program Files\gsdl\cgi-bin Alias field: cgi-bin Access permissions: None Application permissions: Execute Click OK This allows the Greenstone program library.exe to be executed by the webserver. Go to the URL http://localhost/gsdl/cgi-bin/library.exe Note: you need to specify the .exe file extension with PWS and IIS.

For Greenstone to work properly, access permissions for certain files must be set up appropriately. Also, there is a configuration file associated with each Greenstone site. The install procedure creates a generic configuration file based on your installation choices; however its contents can be tailored to cope with different situations. This section explains both of these issues.

This section is irrelevant for Windows 95/98, because these systems don't identify the owners of files. On Windows NT, 2000 and Unix systems, cgi scripts don't run as normal users, because users can't be identified over the Web. Instead, they run as the user who started up the webserver program (on Windows systems), or as a special user (commonly called nobody on Unix systems). Because of this, all files and directories within C:\Program Files\gsdl need to be globally readable (or at least readable by the cgi-script user, perhaps “nobody”). To test whether file permissions are set up correctly, run the program library.exe from the command line. If the files are in the right places but the permissions are set incorrectly, it will run from the command line—that is, when you execute it—but not from a browser—that is, when the “nobody” user executes it. Another test is to log in as another user to see if the file permissions are specific to your original user account. To work through a Web browser, all the Greenstone directories must be globally readable. Also, the C:\Program Files\gsdl\etc directory and all its contents must be globally writable. This is the directory into which the library program writes the usage log, error and initialization logs, and various user databases. If you're reluctant to make this directory globally writable, you can set permissions so that just the files errout.txt, initout.txt, key.db, users.db, history.db and usage.txt are writable by the cgi user. If file permissions are not set up correctly for C:\Program Files\gsdl\etc, you may find that user authentication and search history do not work, and that no usage log (usage.txt) is generated.

The install procedure creates a generic Greenstone site configuration file based on your installation choices. For our installation this file is C:\Program Files\gsdl\cgi-bin\gsdlsite.cfg and its content is: # Site configuration file for Greenstone. # Lines begining with # are comments. # This file should be placed in the same directory as your library # executable file. it should be edited to suit your site. # points to the GSDLHOME directory gsdlhome “C:/Program Files/gsdl ” # this is the http address of GSDLHOME # if your webservers DocumentRoot is set to $GSDLHOME # then httpprefix can be commented out httpprefix /gsdl # this is the http address of the directory which # contains the images for the interface. httpimg /gsdl/images # should contain the http address of this cgi script. This # is not needed if the http server sets the environment variable # SCRIPT_NAME #gwcgi /cgi-bin/library # maxrequests is the most requests a fastcgi process # will serve before it exits. This can be set to a # low figure (like 1) while debugging and then set # to a high figure (like 10000) when everything is # working well. #maxrequests 10000 You can customise your installation by editing this file, although you will probably not need to do so. The gsdlhome line simply points to the C:\Program Files\gsdl directory. httpprefix is the web address of the directory that Greenstone is installed in. We explained earlier how to create an alias so that URLs of the form http://localhost/gsdl/... are sought in the C:\Program Files\gsdl directory. Putting a line httpprefix/gsdl into the gsdlsite configuration file establishes the same convention for the Greenstone software. httpimg is the web address of the C:\Program Files\gsdl\images directory, which contains all the gif images used in the interface. In any standard Greenstone installation this will always be httpprefix/images, and the line in the file above is left untouched. gwcgi is the web address of the library cgi program. This is not required by most webservers (including Apache), and should remain commented out. Don't uncomment it unless you're sure you need to, because that may introduce problems. maxrequests is only used by versions of Greenstone that are compiled with the “fast-cgi” option on. The standard binary distribution does not include this option because not all webservers are configured to support it. Fastcgi speeds up cgi executions by keeping the main executable in memory between invocations of the software, rather than loading it in from disk each time a web page is requested from the Greenstone software. The trade-off is the amount of memory used, which can grow the longer the program remains in memory. Once maxrequests pages have been generated, the cgi program quits, thereby freeing any accumulated memory. To respond to the next request for a Web page, the cgi program is read in from disk again, and a new cycle of page requests is begun. Most installations use the standard cgi protocol, which means that maxrequests can be safely ignored.

Probably the first thing you will want to do once your Greenstone installation is up and running is personalize the home page. The file that generates the Greenstone home page is called home.dm, and is located in the macros subdirectory of the directory into which you installed Greenstone. (The default for Windows systems is C:\Program Files\gsdl.) This is a plain text file that you will have to edit to create a new home page. Instead of editing it, we recommend creating a new file, say yourhome.dm. This will be like home.dm but will define “package home”—which is the bit that does the actual work—in a different way. When you make a different home page, there must be some way of linking in to the digital library pages so that you can search and browse the collections on your system. The solution that Greenstone adopts is to use “macros”. That's why the home-page file is called “.dm” and not “.html”—it's a “macro” file rather than a regular html file. But don't quail: the macro file basically contains just html, sprinkled with a few mystical incantantations which are explained below. The macro language is a powerful facility, and only a small part of it is described below—see the Greenstone Digital Library Developer's Guide for more information.

Figure shows an example of a new digital library home page. Each of the “Click here” links takes you to the appropriate Greenstone facility. This page was generated by the file called yourhome.dm shown in Figure .

package home _content_ { <h2>Your own Greenstone home page</h2> <ul> <table> <tr valign=top><td>Search page for the demo collection<br></td> <td><a href="_httpquery_&c=demo">Click here</a></td></tr> <tr><td>"About" page for the demo collection</td> <td><a href="_httppageabout_&c=demo">Click here</a></td></tr> <tr><td>Preferences page for the demo collection</td> <td><a href="_httppagepref_&c=demo">Click here</a></td></tr> <tr><td>Home page</td> <td><a href="_httppagehome_">Click here</a></td></tr> <tr><td>Help page</td> <td><a href="_httppagehelp_">Click here</a></td></tr> <tr><td>Administration page</td> <td><a href="_httppagestatus_">Click here</a></td></tr> <tr><td>The Collector</td> <td><a href="_httppagecollector_">Click here</a></td></tr> </table> </ul> } # if you hate the squirly green bar down the left-hand side of the # page, uncomment these lines: # _header_ { # }

You can use Figure as a template for creating your own specialized Greenstone home page. Basically, it defines a macro called content. Inside the curly braces is ordinary html. You could insert additional text, along with any html formatting commands, to put the content that you want to see on the page. The text is regular html; if you want you can include hyperlinks and use all the other facilities that html provides. To make your new home page link in with other digital library pages, you need to use an appropriate magic spell. In this macro language, magic spells are words flanked by underscores. You can see these in Figure . For example, _httppagehome_ takes you to the home page, _httppagehelp_ to the help page, and so on. In some cases you need to include a collection name. For example, _httpquery_&c=demo specifies the search page for the demo collection; for other collections you should replace demo by the appropriate collection name. The definition of the macro called _content_ is plain html. Any standard html code may be placed within a macro definition. However, the special characters '{', '}', '\', and '_' must be escaped with a backslash to prevent them from being processed by the macro language interpreter. Note that the _content_ macro definition does not contain any html header or footer. If you want to change the header or footer of your home page, you should define _header_ and/or _footer_ macros, adding them to the yourhome.dm file in the form _macroname_ { ... } For example, the squirly green bar down the left-hand side of Greenstone pages is defined in the _header_ macro, and making this macro null will remove it, as indicated at the end of Figure .

You have to tell Greenstone about the new home page yourhome.dm. The system reads in the macro files that are specified in the main configuration file main.cfg, so if you create a new one you must include it there. Name clashes are handled sensibly: the most recent definition takes precedence. Thus to make the Greenstone digital library software use the home page in Figure instead of the default, first put the yourhome.dm file in Figure into the macros directory. Then edit the main.cfg configuration file to replace home.dm with yourhome.dm in the list of macro files that are loaded at startup.

You may want to redirect a more convenient URL to your Greenstone cgi program. For example, on our system the URL http://nzdl.org (which is shorthand for http://nzdl.org/index.html) is redirected to http://nzdl.org/cgi-bin/library. The Apache webserver accomplishes this with the Redirect directive. Along with other directives, this goes into the C:\Program Files\Apache Group\Apache\conf\httpd.conf configuration file. To redirect the URL http://www.yourserver.com to http://www.yourserver.com/cgi-bin/library, put this line into httpd.conf: Redirect /index.html http://www.yourserver.com/cgi-bin/library Then you will reach your digital library system directly from the URL http://www.yourserver.com. Instead, if you wanted a URL like http://www.yourserver.com/greenstone to be redirected to http://www.yourserver.com/cgi-bin/library, include in the httpd.conf file Redirect /greenstone http://www.yourserver.com/cgi-bin/library If your computer doesn't have a domain name (like the “www.yourserver.com” above), just replace www.yourserver.com by localhost in the lines above. So long as the browser is running on the same machine as the webserver—which it surely is if your computer doesn't have a domain name—this has the same effect as the above redirections. Instead of putting redirect directives into the file httpd.conf, you can equally well put them into a file called .htaccess within your server's document root directory. In fact, doing so has two advantages. First, changes to .htaccess take effect immediately, whereas you have to restart the Apache webserver to see the effect of changes to httpd.conf. Second, on Unix systems you usually have to be logged in as the “root” user to edit httpd.conf, whereas you don't to edit .htaccess.

Here is how to obtain the software packages mentioned above.

To run any version of Greenstone apart from the Windows Local Library version, you need an external webserver. Many installations, particularly larger ones, will already have a webserver. If you are using Linux, Apache may be on your installation disk but may not have been selected during the installation procedure. The Apache Webserver from www.apache.org is free, and easy to install.

Greenstone uses the Perl language when building collections. For Windows, Perl is already included in the Greenstone software. Most Unix systems already have Perl installed, but if not, source code and binaries for a wide range of Unix platforms are freely available at www.perl.com. Perl version 5.0 or higher is needed.

The Unix version of Greenstone compiles under the Gnu C++ compiler, GCC. Greenstone makes extensive use of the C++ standard template library (we've found it to be broken on some older versions of GCC; please tell us if you have STL problems). Note that this version of Greenstone does not compile under GCC 3.0.

All versions of Greenstone use the Gnu Database Manager, GDBM. It is supplied with all Windows versions of Greenstone and installed automatically during the installation procedure. Linux systems already have GDBM, so we do not provide it for Linux. Most other Unix systems have it, but if necessary you can download it from www.gnu.org.

To use the Greenstone Librarian Interface, you need a suitable version of the Java Runtime Environment. If you don't already have this, a suitable version is included on the CD-ROM, or you can download the latest version from http://java.sun.com/j2se/downloads.html. Version 1.4.0 or higher is needed.

To compile the source code of the Greenstone Librarian Interface, you must first install a Java Development Kit. You can download the J2SE Software Development Kit from http://java.sun.com/j2se/downloads.html. Version 1.4.0 or higher is needed.