Knowledge Management Argue

I engaged my brain. The discussion about how we should organise this page belongs on a discussion about how we organise pages, so I've moved it again, to ArgueAgreeIdiomArgue. I hope that seems logical. I think this page is a good place to argue points about KnowledgeManagement, leaving that (KM) page as a place for a free-for-all. But enough meta-discussion ! EdwardWelbourne

 -- Eddy/1998/August/26.

There follows the state of KnowledgeManagement as at this date, in its original Argue (ie threaded or diary) mode:

[1 Feb, 1998; YeshaSivan]: Knowledge Management is a new front for management. It is the practice that deals with the many aspects of Knowledge. Wiki seems to be some kind of a cool system to manage knowledge. And since a friend said it is cool thing I'm testing it -- to see if it can serve in part in system that manage knowledge.

[July 20, 1998; YeshaSivan]: Ok, few months have passed since I have created this page. A friend, YonatSharon, has just told me, that few people have actually looked at this and have added some value. I see that this may go somewhere. So here are few additions to the concepts, all taken from the work we do here in Israel.

I'm not going to define the term KnowledgeManagement; this was done in other places; and I hope the BooksAboutKnowledgeManagement will tell you all about it.

A related term to KM is KnowledgeInfrastructure.

Actually when you start to think about how to do KnowledgeManagement (as a person or as an organization) you can follow three steps:

Plan how to do KM.
Implement the KM plan.
Evaluate your implementation (and therefore update the implmenetation or the plan as needed).

This is what I call the PieModel (in this case of knowledge management).

[July 15, 1998 PeterMurchland]: Yesha, I am interested to explore how Wiki could be used to advance the creation and sharing of knowledge. Perhaps we should construct some sort of experiment here ?

[July 18, 1998; DenhamGrey]: I'm keen to experiment here as well. There seem to be interesting aspects of Wiki for KnowledgeAnnealing, for CoAuthorship? of documents, for JointWebsiteDevelopment?, for studying EmerGence? and for compiling a DisTinctionary.

This is an interesting alternative conversational medium.

[July 20, 1998; DenhamGrey]: I have this feeling that there is a close link between patterns as a representation and process and knowledge. Here is a theory of knowledge presented in the form of a pattern KnowledgePattern.

Perhaps we can explore this further at this site where both knowledge management and pattern folks have collected?

[July 23, 1998; YeshaSivan] again, I edited the top of the page to crate a clear distinction between the ropic -- KM -- and the discussion about the format. Which led me to the follwing point:

Why do I say that knowledge is linked with people and time? Well, because i feel that knowledge without context is of lesser value. There is an old saying "it is not what you know, but who you know" there is a lot hidden in this simple sentence.

[July 25, 1998; RafaelTeixeira] I can't understand what you're triyng to mesure..

[July 25, 1998; DaveHarris] I've also been thinking about patterns and knowledge and computation. Patterns are powerful partly because they are generative grammars. Generative grammars are powerful. I'm groping towards some insight involving the ChomskyHierarchy, where simple languages are recognised by finite state machines, more complex ones by push-down automata, and the most complex need full Turing machines. There's some connection between the language we use for talking (and thinking), and the computational power of a Turing machine.

I suspect this is utterly trivial stuff that I knew 15 years ago and I am just rediscovering :-(

[July 25, 1998; DaveHarris] On "Who you know" - provenance of ideas is important. If I know who wrote a page, I can discount for known stupidity, or pay more attention to known experts. This is important on the WWW because anyone can put up a web site containing any kind of rotten information. Schoolchildren need to be taught how to discriminate between web sources, and provenance is one of the tools.

Part of WhyWikiWorks is that because anyone can edit anything, the pages reflect, more or less, the consensus of the community. Bad stuff gets deleted, or at least annotated with an explanation of why it is bad. This means the provenance of individual paragraphs is less important for signalling quality. Rather then, "I'm not going to believe anything unless it has Ward's name on it", it's kinda enough to know Ward read it and didn't delete it.

This strikes me as rather radical and unlike the rest of the WorldWideWeb.

[August 4, 1998; ChristopheVermeulen] On "enough to know Ward read it" But again, how are you going to know that Ward read the page you are referring to ? (and didn't delete it). Maybe the change is very recent, or re-posted regularly...

More generally, I wonder what happen when people would edit the same page at the same time. In other words, what about WikiSecurity?, WikiFileLocking? and the like ...

[August 4; DaveHarris] There are technical issues with this Wiki. They're solvable. It's possible to keep track of who read what and when, and who the author was, etc. Also to detect edit conflicts, to recover accidental or malicious deletions, and the like. You can't reason from the limitations of this Wiki to the limitations of all Wiki-like environments.

[4Aug AustinDavid] In the sense of "manging useful information" all the above is great, and proven. Has anyone actually listened to, or read, the real KM-speak? It's either TheNextGreatThing? (to save the world) or TheNextSnakeOil?. Patterns had this sort of effect on me, but they've always (mostly) been grounded in models, diagrams, and code. RealKnowledgeManagement (at least to the extent that I've been exposed) is more about buzzwords and doublespeak. Anyone? (no, I'm not bitter)

[7th August; DenhamGrey] I agree there is much hype in the air around KM. The comments in this community on the superiority of patterns is an interesting case I think of GroupThink. I have seen references and appeals to the ScientificMethod and to logic, but patterns like process, carry a certain type of 'blindness', they too are social artifacts.

Many who come to KM will depart with a sense of shallowness and may feel some snake oil but there is an interesting and lasting quality to working with knowledge that parallels the TQM movement. Those who wish to get something of lasting value from KM should look to the social aspects of collaboration, communion and critique, seek opportunity in new perspectives of work practice, learning, innovation and knowledge sharing.

An interesting connection between KM and patterns is the DataMining movement who seek to discover patterns in data streams. Data mining uses algorithms to serach for possible patterns and relies on people to interpret the significance and meaning of the 'discoveries'.

Here is an article on the KnowledgeManagement metaphor: http://www.gwu.edu/~adsci/metaphor.htm

[21Jan2004 GarryHamilton]

Wow. I find myself somewhat bewildered by the discussion.

In my own little world, KM is applied to storehouses of relevant data whose usefulness is directly a function of how it is interrelated, queried, and retrieved.

The business of DataMining is much more mechanical and more related to StatisticalAnalysis? from a DataWarehouse or similar construct.

The main purpose (again, in my world) of KM is to shorten the learning and discovery paths for development, debugging, comparative evaluation, and so on.

In this light, it is of primary importance that the knowledge be organized, indexed, and stored in ways optimized for human consumption.

A really good example of this is the program "HELP.EXE" from the DOS days. It was eventually turned into a general purpose help engine with a database compiler that would take a body of marked up text and turn it into a searchable HyperText arrangement.

There were HELP datbases for DOS BIOS functions, DOS commands, and so on. You could search for key words or partial words, and relevant terms on the current page were linked to pages that explained them, with unlimited backward traversal.

There are certainly more sophisticated ways of relating data, but this is one area where ease of use and essentially natural access aren't just "nice to have" but quite necessary.

A wiki might serve for such a repository, but note that this particular wiki doesn't seem to support general term searches.

Grow this discussion by its tail.