Network Computingwww.networkcomputing.com

Reviewed by Walter Alan Zintz

To help celebrate this first appearance of UnixWorld Magazine on the World Wide Web, I'm starting this month by reviewing books about information applications on the Internet. The first book--immediately following--covers setting up and running information servers and the second book looks at going beyond simple hypermedia in making information more accessible on the net.

Managing Internet Information Services

By Cricket Liu, Jerry Peek, Russ Jones, Bryan Buus, and Adrian Nye; with Greg George, Neophytos Iacovou, Jeff LaCoursiere, Paul Lindner, and Craig Strickland
630 pages. Includes 19 pages of contents; code; diagrams; contacts; bibliography; 34 pages of index
ISBN 1-56592-062-7, paperbound, $29.95

Technical Level: computing--professional to wizard, subject--newbie to experienced
Information: concepts--passable, practice--good
Readability: textbook--passable, reference--good
Summary: The technical aspects of many Internet information server packages, explained in varying depth.

( Skip to the next review. )

Publisher: O'Reilly & Ass ociates
To order directly from the publisher:

• Web site: http://gnn.com/ora/
• E-mail: order@ora.com
• Toll-free: (800) 889-8969
• Fax: (707) 829-0104

Gopher, World Wide Web, FTP, WAIS and Majordomo servers all get multiple chapters in this book. A lot of smaller information-server tools are covered at suitable length, too. Who would want a book on setting up, using and maintaining all these packages when no one site is likely to put up more than a handful of them? Especially when no book a mortal could lift off the shelf could go very deeply into all these complex packages?

Anyone who finds a package's documentation impenetrable, for starters. For most of these applications there are no other books in print that deal with their installation and administration. Information-server consultants and wannabees may want it as an orientation, a warm-up for the piles of manuals. Analysts shou ld find this book deep enough to be very helpful in deciding which package is technically best for a particular site. And the book's hidden benefit, a brief course in managing information (as distinguished from managing the hardware and software that handle the information), might make the book worthwhile to people who primarily will deal with a server's information content.

When this many authors have a grip on the pencil, you can expect a book that varies in many qualities from one part to another. There are noticeable variations in technical depth here, but that's about the worst of the unevenness. The upside of very plural authorship is here, also--two good chapters on legal factors, which I didn't expect in a book this technical.

My one serious disappointment with this book is the major packages it covers. Frankly, the Internet community is stampeding toward World Wide Web as the information server of choice, and I don't know of anyone planning to set up a Gopher or WAIS server these d ays. The 10 chapters and 3 appendixes devoted to these waning packages might better have been used to cover Listproc, a mailing-list handler just as important as Majordomo.

In the book's favor, it contains a lot of useful information, and that information is much more accessible here than in the official manuals in most cases. Nor will there be rival books any time soon; I haven't read advance notices or even net rumors of any other books on most of these packages.

Text excerpt:

Gopher 2.x assigns a number of attributes to each file and directory. The default attributes for all items are set in the gopherd.conf file. You can set the name of the organization by changing the Org token, the physical location of the server by setting the Loc and Geog tokens, and the default language by changing the Language token. The gopherd.conf file also contains the server administrator's name and email address, set by the Admin and the AdminEmail tokens, respectively. Here are some sample settings for a fictitious server called ``The Wildlife Organization.''

Admin: Gopher Park Ranger
AdminEmail: <gopher@wildlife.ora.com>
Org: The Wildlife Organization
Site: New York Wildlife Preserve
Loc: 100 Elephant Avenue, Buffalo NY, 55405
Geog: 44 58 48 N 93 15 49 W
Language: En_US

Main Table of Contents:

Internet Service Concepts

Introduction to Information Services

Finger-, Inetd-, and Telnet-based Services

Setting Up an FTP Archive

The WU Archive FTP Daemon

Maintaining an FTP Archive

Creating an Internet Database Server with WAIS

Creating WAIS Sources with waisindex

Gopher Introduction

Gopher: Compiling the Server

Gopher: Managing the Server

Gopher: Preparing Info rmation

Gopher: Linking Services Together

Gopher: Incorporating Databases

Gopher: Veronica and Jughead

Gopher+ Forms and Other New Features

Introduction to the World Wide Web

Setting Up a Web Server

Authoring for the Web

Web: Gateways and Forms

Web: Access Control and Security

Introduction to Email Services

Simple Mailing Lists

Automating Mailing Lists with Majordomo

The Majordomo List Owner and Moderator

Ftpmail

Firewalls and Information Services

xinetd

Legal Issues

Protecting Intellectual Property

Appendixes:

gopherd Options

Gopher: Client Compilation Options

Gopher Tools and Gateways

Web: More HTML Tags

Web: httpd.conf Directives

Web: srm.conf Directives

Web: access.conf Directives

Web: For More Information

Challenges in Indexing Electronic Text and Images

Edited by Raya Fidel, Trudi Bellardo Hahn, Edie M. Rasmussen, and Philip J. Smith
306 pages. Includes 2 pages of contents; tables; bibliographies; 8 pages of index
ISBN 0-938734-76-8, hardbound, $39.50

Technical Level: computing--experienced to wizard, this subject--newbie to professional
Information: concepts--good, practice--passable
Readability: textbook--good, reference--passable
Summary: Facing some major gaps in current hypermedia practice

Publisher: Information Today
To order directly from the publisher:

• Phone: (609) 654-6266
• Fax: (609) 654-4309

Hypermedia, boolean text searches and the like are marvelous improvements in information access--as important as space between words in text, which was unknown in most medieval writings. But we're not even close to the ultimate in information accessibility with those.

Information retrieval specialists have long realized this, and they frequently think about better ways to access stored information. Some of their better ideas regarding machine-readable information are in the 15 separately-written articles in this book.

There aren't any canned algorithms here. The book's content is divided, roughly equally, into concepts for improving information access and analysis of problem areas still in need of solutions. Both categories deal mostly with matters that most of us have never realized are bottlenecks. For example, how many programmers have given much thought to indexing or searching for pictorial images, by semantic content rather than physical characteristics? The first three articles here evaluate a lot of good ideas, and still make it clear that a lot more thought is needed.

To borrow a buzzword from the suits, this is a book for proactive organizations. Those who only plan to produce smoother hyperlinks and more complex boolean sea rchers won't need it. But anyone who expects to be on the front edge of the next wave in online information access should read this book at least twice.

Table of Contents

Indexing and Accessing Images

Introduction
User Types and Queries: Impact on Image Access Systems
Thinking Ambiguously: Organizing Source Materials for Historical Research
Analyzing Art Objects for an Image Database

Indexing of Hypermedia

Introduction
Designing Hypertexts: Start with an Index
Online Help Systems: A Multimedia Indexing Opportunity
Hypertext and Indexing
Information Structure Management: A Unified Framework for Indexing and Searching in Database, Expert, Information-Retrieval, and Hypermedia Systems

Computer Support Tools for Indexers

Introduction
Knowledge-Based Systems for Indexing
Computing Support for Indexing a t Petroleum Abstracts: Design and Benefits
Computerized Development and Use of the NASA Thesaurus
Machine-Aided Indexing from the Analysis of Natural Language Text
Computerized Tools to Support Document Analysts

Indexing and Retrieval from Full-Text

Introduction
Automatic Indexing
The Role of Linguistic Analysis in Full-Text Retrieval
Text Based Applications on the Connection Machine

( Return to the top of this review. )

Wednesday, 15 Feb 1995

Subject: Your review of MIIS found on Unix World Online

Dear Walter:

First, I just say how glad I am to find Unix World's technical experts back in print (even if the print is only on my screen). After watching both Unix World and Open Systems Today become ``suit and briefcase'' publications, those of us trying to manage the technologies need experts to read. Welcome back.

Now, about your scathing review of MIIS ( Managing Internet Information Services )...in most cases you are correct in your assessments. No matter its shortcomings, though, the book fills a huge GAP in knowledge. We need this type of book and despite the problems with multiple authors' styles, it is a very good book.

In your review, you comment that too much space is wasted on non-WWW services since you don't know of ``anyone planing a gopher or WAIS server these days''. Well, we are doing that at this very time and for a very good reason.

While the world is beating a path to the door of the Web, I still believe in the best tool for the job, and sometimes the best tool is NOT the Web. Our mission is electronic information delivery from NASA to teachers (typically K-12 classroom teachers). If you look at the network statistics, you will learn that only about 1-3% of the schools are connected to the Internet in a way that will support use of the Web. We have built a server to deliver information t o teachers in ANY way they can get to us, be it via the Web or (mostly) via modem. Some states have built networks for teachers to reach the Internet, but most of those networks are still based on dial-up, VTxx terminal-based services.

We sought to serve the most teachers possible with the most services. Our single posting data structure has the Web, FTP and gopher servers pointing to it. We will have WAIS pointing to the data soon to meet a government mandate for data indexing. Four servers, four protocols, but one happy user base (most of whom have no idea the other access methods exist).

We decided on offering a gopher service primarily because the gopher client (which we offer to the general public who call us) is SO fast and efficient with user time. We offer Lynx (a VTxx Web browser) to our modem-only teacher-users who also get email and Usenet news. Lynx offered us the ability to build our teacher services around Web pages that only Lynx can see. For those teachers with full inte rnet access, our HTTP daemon serves up some of the best graphical pages we can build. Users with file transfer needs (or those without browsers, etc) can get the same great NASA data via anonymous FTP. Our data services model was built on a BROAD service base...most data to more people.

The government is slowly coming to grips with the equity issues of publishing vast amounts of information ONLY via the Web, when only a small percentage of the citizenry have Web access...some government agencies are working on solutions (like our system) under projects called 'equitable access' to public information.

Thanks for your time...I just wanted you to understand that there are needs for all the services even in this Web-crazy time.

Walter replies:

Yes, it is equitable and desirable to make net information stas hes available to that vast majority of net users who have no access to Web sites. That's why I wished that the MIIS book had covered the other important mailing list handler for Unix systems; because e-mail is the net service that everyone has available.

Your organization has a rare situation where a non-Web server makes sense--you are only trying to reach a membership group whose preferred access methods are known, and you have the resources of NASA to cover the cost of all those access methods. A typical organization putting up a new server right now wants to talk to everyone who will listen, and has a very modest budget for the server project. A good example is my own Tea Lovers shop, which wants to reach everyone with an interest in exquisite tea and has only a three-figure budget for going on the Internet. It's a real scramble, and we're betting on e-mail for wide access and the World Wide Web for speed, interactivity and impact.

I'll retract my comment on server trends to this extent: I now know of one organization building new non-Web servers, versus new Web servers popping up daily. But net statistics as well as word of mouth now have me convinced that, by the end of this year, Web will be second only to e-mail in the number of net users prepared to access it.

BTW, was my review of the MIIS book honestly ``scathing''? I still believe that I was being fairly evenhanded, praising the book at least as strongly as I criticized it. That's the hazard of putting too much nice-guy talk in these reviews: when I do say something at all negative, readers will blow it out of proportion.