Publications
|
10 Principles of Knowledge Base Publishing
We use the phrase "Knowledge Base Publishing" to describe
the integration of print, Web, and database publishing models in a business
context. Integration involves both editorial and technical issues. While
traditional publishing is linear and segmented, knowledge base publishing
is web-like, interconnected, and integrated into the user's work flow.
It's a balance between managing knowledge from the bottom up, where the
user designs his own information tool kit, and from the top down, where
enterprise culture, technologies, and policies encourage knowledge sharing
across organizational boundaries.
Until the appearance of the personal computer in the
1980's, knowledge base publishing was only a dream. In the mainframe and
even the mini era, specially trained intermediaries were needed to make
computers do the business person's bidding. But with microcomputers, the
layman could create applications all by himself using spreadsheets, word
processors, desktop publishing programs, and databases. The Internet added
global communications, document sharing, and a new kind of publishing
(the World Wide Web) to his tool kit.
The microcomputer-plus-Internet revolution outpaced
the ability of organization structures, business models, and professional
education to assimilate all the changes. It's relatively easy to use new
technology to automate a manual process — e.g. turn a collection
of paper documents into a searchable electronic archive. But these automation
projects never occur in isolation. Software vendors come and go, new business
models are introduced, the business environment changes, and users are
implementing new desktop features. Productivity increases in one area
can be partially neutralized by changes elsewhere.
The goal of Knowledge Base Publishing is to create
a new system that delivers productivity increases greater
than the sum of individual knowledge technology projects. It means
minimizing rework costs every time technologies and vendors change. It's
also about minimizing knowledge worker frustration and time wasted due
to incompatible applications, outdated intellectual asset policies, and
failure to integrate and adapt practices of the information disciplines.
But what does this mean in practical terms? Through
our own experience as well as what we've learned from our members
and colleagues, we've identified 10 essential Knowledge Base Publishing
principles:
1. Focus on user productivity. There are many
ways to measure intellectual capital (see "Measuring
intellectual assets"), but with Knowledge Base Publishing, we're
concerned with its performance on the front lines — time saved,
risks minimized, opportunities discovered, assets leveraged. The impact
can't always be measured in quantitative terms until after the fact —
sometimes years afterward.
2. End user development. There's a great deal
of variability among knowledge workers — not just in the same industry
or function but also in individual preferences. Some people are visual
thinkers; others do better with words. Some are neatniks while others
work best in what seems like a mess. To achieve the productivity we're
after, it's necessary to give employees the freedom (within limits) to
design their own information environments. This requires a different kind
of IT support and easier to use, mix-and-match tools (see "Get
ready for end user development").
3. Hybrid information management structure.
Allowing each business unit the freedom to manage its own knowledge works
because it facilitates exchange among small groups of workers with common
interests, but it provides just a fraction of the potential benefits of
exchanging knowledge on a company-wide scale (see "Do
we need internal knowledge markets?"). Knowledge Base Publishing
requires a hybrid structure where some things (e.g. security, basic IT
infrastructure, publishing standards and policies, and global thesauri)
are managed at the enterprise level which everything else devolves to
the business unit or individual (see "Upstream
knowledge management").
4. Knowledge facilitators. Knowledge exchange
requires human facilitators, just like markets in goods and services do.
They're needed at the enterprise level to create content quality standards,
craft company-wide organization schemes (taxonomies), and administer author
incentives. They're needed at the business unit level to help authors
package and promote their knowledge products, help knowledge workers streamline
information gathering, archiving, and reuse, and create domain-specific
organization schemes. Some of them need to be "at large" boundary
spanners who encourage subject matter experts to create new knowledge
objects and translate concepts from one discipline to others (see "What
is Knowledge Base Editing?").
5. Blending print, Web, and database formats.
Although more content and business processes are moving to the Web, print
will continue to be valued for its convenience, portability, and prestige.
Databases will remain a mainstay for business applications and financial
analysis. Knowledge Base Publishers need to be able to blend all three
formats for maximum utility and adapt the quality control practices of
print and database publishing to the Web. For example, many textbooks
are now being created by mixing and matching book segments and journal
articles from a database that includes copyright permission and payment
data.
6. Integrating the information disciplines.
Over the years, librarians, journalists, book editors, technical writers,
and database administrators have developed tools, techniques, and standards
that define their craft. For example, journal editors have developed the
concept of "peer review" to weed out crackpot or poorly written
articles. Trade and professional publishers add tables of contents, A
- Z indexes, and cataloging in publication data to their books to make
them easier for readers to use and easier for librarians to classify.
These best practices need to be divorced from their physical format (i.e.
book, article, user manual, newspaper) and adapted to the blended format
of Knowledge Base Publishing.
7. Author responsibility. The current practice
of using search engines to serve up thousands of poorly written, out of
date, and irrelevant documents is a big drag on employee productivity.
Enterprise content managers have discovered to their dismay that neither
search engines nor taxonomies yield good results unless authors create
high quality content. By this we mean well-constructed documents relatively
free of jargon with accurate metadata, table of contents, definitions
and links, and (for long documents) an A - Z index. Knowledge facilitators
need to help authors understand how good writing can enhance their reputation,
lay the foundation for a promotion, make them more effective managers,
and increase the productivity of their colleagues.
8. Embedded and integrated information. In
the old days, you'd go to the library to do research. Today, you don't
have to leave your spreadsheet or word processor. A "research panel"
on the side of the current document lets you look up a company profile,
consult a thesaurus, translate a passage, or do a Web search. In many
commercial information services, you can search for articles on a topic,
mark the ones you're interested in, and download their metadata (i.e.
author, title, publication, abstract, etc.) into your own private electronic
library. We'll see more of this kind of streamlined workflow in the future.
9. Metadata repositories. As more corporate
applications and external services become automated, it's necessary to
identify, coordinate, and normalize the metadata they use. For example,
Amazon.com might use "Release Date" and "Manufacturer"
to describe a book, while the corporate library's automated card catalog
might use "Publication Date" and "Publisher." To integrate
book data from one system into another (instead of having to reenter it
from scratch), some mechanism is necessary to relate Amazon's terminology
to the library's terminology. A metadata repository does that.
10. Thesauri. Metadata repositories often
include thesauri — data structures that show relationships in the
same way that Roget's thesaurus lists synonyms. In an A - Z index, these
relationships get displayed as "see also" references or as a
topic hierarchy. Thesauri make it possible to see that "Myocardial
Infarction" is a synonym for "heart attack," that "AARP"
refers to "American Association of Retired Persons," and that
"polymorphism" is a feature of computer software.
|
