Volume 1 Number 17+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ Table of Contents
a. Knowledge Management is Now Legitimateb. What's going on at Smith Weaver Smith
2. On The Design of a Tool for Knowledge Work
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ KnowBits
a. Knowledge Management is Now Legitimate An interesting piece of literature came to me by mail the other day. It was an announcement about a conference. The conference is
entitled, "Transforming Knowledge-Intensive Processes: Improving Business Performance through Knowledge Management." Conference leader: Dr. Michael Hammer. And with that Knowledge Management is now legitimized. In case you are
interested you can find out more at www.hammerandco.com. First there was process – no people – just process. Then Dr. Hammer made an amazing discovery – you have to consider the people too. And now, low and behold, process has
something to do with knowledge. I think the fact that Dr. Hammer is now talking about knowledge management lends credibility to the knowledge management field. All of us who are practitioners should find it much simpler now to
sell our knowledge management efforts to management !! b. What's going on at Smith Weaver Smith Last week's newsletter that included the list of categories for the taxonomy prompted quite
a discussion about the content of the technology and the nature of ontology and taxonomy. The discussion is still going on at
www.smithweaversmith.com.
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ knowldgWORKS News Volume 1 Number 17 On The Design of a Tool for Knowledge Work At the risk of violating the cardinal rule of knowledge
work--"It's the people, Stupid"--it is still necessary occasionally to consider appropriate tools that simplify, facilitate, or enhance knowledge work efforts. Just remember that the tools are a secondary consideration
to the problem or task at hand and should not become the primary consideration. Writing these newsletters has prompted me to think about how their value might be increased. Consider this: each and every week you receive a
newsletter. If you are anything like me and you find the newsletter useful, you might decide to save it in a file so you could retrieve it when needed. However, since each and every newsletter is archived on the web, you could,
instead, return to the web site and examine newsletters anytime you wished. Without adequate web page indexing, however, in order to search for a specific topic, passage, or piece of information, you would have to re-read one
or more newsletters until you found the item you needed. Therefore, indexing is a tool that may be worth further consideration. Indexing Lycos, Yahoo, and Altavista are programs that index web pages. Their spiders crawl
around on the web indexing pages according to categories. Interestingly enough, the major failing of these programs is that they tend to miss lots of stuff and they are not readily controlled by their user. Their index is created
behind the scenes. Nevertheless the process of finding pages on the web is greatly facilitated by these programs. Local Indexing Local indexing duplicates the process used by Lycos and the other search engines with two
differences. First, the set of documents can be constrained and second, the terms for indexing can be precisely controlled. One interesting program that does this is dtSearch (
www.dtsearch.com). One version of this program that I have used on my PC can search through many different types of file
formats for words and phrases. An impressive feature of this program is that it is capable of scanning different formats including zipped versions of the formats. dtSearch now has a version for the web and I would recommend looking
at this software if you require this type of search/index functionality for a web.When I think of indexing, I think of a list of terms that are important to some domain. Attached to each term is one or more pointers. In the
traditional case, these pointers are page numbers. In a more webified version, the pointers can be links to the text locations of terms. Cross Linking Suppose we have multiple documents (like an archive of newsletters) and we
want to create a global index of all newsletters for a certain set of terms. Unlike the traditional index that points to terms in a single document, a cross linked index points to terms across multiple documents. A pointer to a
term in a document now becomes a pointer a particular document and a location within the document. A cross link is defined as follows: A link between two or more documents containing the same term. A cross linked index
facilitates examining documents containing similar terms. Therefore, by having an index like this it becomes possible to view a concept across many different views. Let's think about how a cross-linked program should operate. We
want the capability of answering the question, "Show me all places in the specified set of web documents where the specified word or phrase is used. Do this by creating a cross link index of the specified set of
documents." Our vision for this tool is that it will process a set of documents and create a new set of documents from these. The new set of documents will have inserted in them one or more anchors positioned at the specified
word or phrase. Links placed in the cross linked index will point to these anchors. In order to achieve this behavior we will first need to specify the URLs of one or more web pages. We would also like to be able to specify the
index terms and allow the capability to automatically generate a list of potential index terms. Once the list is generated, users should be able to remove any words or phrases they do not wish to include in the indexing process.
Since the list of auto-generated terms might be potentially large it should be possible to delete terms that are automatically generated. Another feature we would like to have is the ability to specify the number of words in
auto-generated terms. Terms of 2 words, 3 words, 4 words, 5 words, and 6 words can be generated. It would also be nice to allow patterns like regular expressions to be specified. As I mentioned earlier, the output of this program
would be a set of new pages - with anchors added where terms are found - and a hyperlinked index page. The hyperlinked index page would consist of a list of terms and links to where the terms were found. We can envision an output
something like that shown below. |
