[Date Prev] [Date Next] [Thread Prev] [Thread Next] Indexes: Main | Date | Thread | Author

[ba-ohs-talk] 3-layer Architecture Purpose of DKR?


Jack,    (01)

Thanks for helpful clarifications of the NexistWiki architecture. 
There seems to be another issue popping up.  A few years ago, Doug
asked that OHS/DKR do a work up on the meaning of knowledge to guide
development of OHS/DKR, see for example on 000307....    (02)

http://www.welchco.com/sd/08/00101/02/00/03/07/073530.HTM#4820    (03)

This reflected prior discussion during the Colloquium Doug conducted
at Stanford, as shown on 000120....    (04)

http://www.welchco.com/sd/08/00101/02/00/01/20/080146.HTM#5063    (05)

As you recall, there were difficulties understanding the meaning of
knowledge.  I took a shot and mentioned POIMS that talks a little
about knowledge...    (06)

http://www.welchco.com/03/00050/01/09/01/02/00030.HTM#0367    (07)

...in relation to an architecture for human thought and cognition.
This was not well received, as you noted.      (08)

On 000405 Doug tried a simpler idea of suggesting people put
addressability in correspondence....    (09)

http://www.welchco.com/sd/08/00101/02/00/04/05/223813.HTM#L661502    (010)

....to create a connected record, which over time would build a DKR
and give people experience to discover the meaning of knowledge.  This
led to efforts with weblogs, Zwiki and Wiki, eventually leading to
NexistWiki, but we still don't have a lot of success getting a
connected record with addressability.    (011)

Where addressability has been furnished demonstrating the power of
connections, this has been met with significant resistance because
people feel drawn to open every link, rather than merely those
relevant to immediate concerns.  In short, there are now calls for
ways to provide intermediate formats that shield people from the
actual connections, because as Eric noted on 000125, the spectacle of
knowledge "boggles" the mind....    (012)

http://www.welchco.com/sd/08/00101/02/00/01/25/071445.HTM#3867    (013)

Now it seems that your 3-layer architecture might solve this by
providing a shield between the actual record that has the connections
needed to triangulate accurate understanding, and a work product that
is less demanding with just the links needed for the moment.    (014)

In any event, the important point is there is considerable cultural
resistance to working directly with a DKR.  People don't like seeing
the links that connect cause and effect to convey knowledge.    (015)

Recently, evidence seems to be accumulating that not only is a
connected record psychologically debilitating, so, too, is the process
of creating links that provide context for converting information into
knowledge.  Again, recall that Doug provided Purple Numbers in the OHS
Launch Plan, and later Eugene began adding addressable anchors in
OHS/DKR email, yet it is not clear that anyone has ever in two (2)
years used one of those links, including Doug and Eugene, as shown in
the record on 020812...    (016)

http://www.welchco.com/sd/08/00101/02/02/08/12/091456.HTM#TV9K    (017)

For a variety of reasons, we can discount Doug in this discussion. 
But, as shown on 020812 even our most stalwart advocates and teachers
of KM cannot muster the strength to invest 10 seconds it takes to add
a link or two to original sources.  On 020530 Eric offered that he
thought that if people had tools like NexistWiki to create a connected
record, people would use it.  However, there is no evidence in the
record to support that thesis, except for your letter today that shows
a breakthrough in resistance to KM by adding a link.    (018)

Being facetious, one might ask if this was a mistake, a moment of
weakness perhaps, where someone let their guard down for a moment and
let in the light of knowledge.    (019)

More seriously, is this something you plan to build into NexistWiki
3-layer architecture?  How can we get KM advocates to invest just 10
seconds to do KM?  There are objections to using a connected record
and objections to using links in a connected record that save time and
money.  What then is the objective of a DKR/    (020)

Why create an OHS to create a DKR to solve world problems, if we are
not going to use the DKR?  It's a dilemma?    (021)

Rod    (022)

**************    (023)

Jack Park wrote:
> 
> Hi Rod,
> Good questions! Thanks.
> 
> If you visit http://www.nexist.org/wiki/Doc2740Page you will see some
> sketches of Version 2 which is a different (functional) view of three layers:
>          user interface
>          middle tier
>          datastore
> 
> Now, within that architecture, there is room for the logical three layers I
> and others (like Lee Iverson) have referenced in the past.  That logical
> layout, from my point of view was
>          user interface
>          knowledge layer
>          raw information layer
> 
> User Interface is and always will be a kind of "lens" (Mary Keeler's word)
> into the information universe.  It's also the interface where user's create
> new information resources, which reside (always) in the raw information
> layer (which is the datastore).
> 
> Knowledge layer is where the action is.  It functionally resides in the
> datastore, but it is manipulated and maintained in the middle tier (which
> does not exist in NexistWiki Version 1 -- what you play with on the Web
> now; instead, part of its functionality is hard wired in the guts (not a
> middle tier) of Version 1).
> 
> Middle tier will be (as you can see at the link above) essentially a
> coordination system with a "society of software agents" each of which has
> some functionality related to building and maintaining Knowledge layer.
> 
> In relation to your discussion at
> http://www.welchco.com/sd/08/00101/02/02/07/30/191538.HTM#YR6F
> I remain convinced that SDS functionality is an important demonstration of
> the power that can and will eventually be captured and utilized in a full
> monty OHS implementation.  I also believe that SDS functionality can be
> built right into NexistWiki; perhaps you and Gary might imagine how that
> could be done.  As I have stated earlier, the "society of software agents"
> capability enables proprietary plug in widgets to be dropped into
> NexistWiki at will.  Of course, the licensing of NexistWiki intends to
> prevent proprietary additions to the NexistWiki suite from locking out the
> larger project's natural evolution in  the same or similar
> directions.  This has been one of Douglas Engelbart's concerns and it is
> one of mine as well.  In the spirit of the OHS vision, nothing should get
> in the way of the project evolving to suit the ever-changing needs of its
> users.
> 
> Version 1 of NexistWiki is, as I said, an engineering prototype. I just
> don't think I'm smart enough to see or understand all the potential use
> cases and scenarios for which it may be suited, so I put it up to see what
> might happen.  Right now, there are only a couple of dozen users.  I have
> no idea how the program will react when hundreds or thousands of users log
> on; it's on an ancient wintel box right now, awaiting installation on a new
> G4 PowerMac that I bought for the project.
> 
> The primary hypothesis being explored with NexistWiki is this:
>          A seamless integration of evolvable story telling structures with
> orthogonal discussion/argumentation structures will facilitate palpable
> improvement in the ability to articulate, discuss, and evolve solutions to
> complex, urgent problems.
> 
> Truth be told, I got forced to bring the site online now because it is
> referenced on the back cover of the book _XML Topic Maps_ and there's
> nothing worse than the famous 404 HTTP error message when owners of the new
> book go looking for its home page on the Web.  I suppose a site with a
> terribly unintuitive user interface is almost as bad, but the site exists
> and it is evolving as fast as I can move it along.
> 
> Cheers,
> Jack
> 
> At 11:25 AM 8/13/2002 -0700, you wrote:
> >Jack,
> >
> >Great news on advancing toward the OHS/DKR with Nexist Wiki, as
> >related in your letter today.  Does this apply the three (3) layer
> >architecture you an Mary Keeler discussed at SRI on 000518....
> >
> >http://www.welchco.com/sd/08/00101/02/00/05/18/160027.HTM#8439
> >
> >This came up recently on 020730 in a letter from Gary Johnson, and I
> >asked at that time if Gary's ideas for SDS implement or in some way
> >address your planning for a 3-layer architecture...
> >
> >http://www.welchco.com/sd/08/00101/02/02/07/30/191538.HTM#YR6F
> >
> >Of course you have something entirely different in mind, and so this
> >gives rise to asking the master architect.  For example, what
> >scenarios, or use cases, are envisioned for NexistWiki.  If is is
> >readily integrated in a way that adds value to existing methods, this
> >is a big boost toward the OHS/DKR objective.
> >
> >Rod
> ><snip>
> > > ---------------------------------------------------------------------------
> > > XML Topic Maps: Creating and Using Topic Maps for the Web.
> > > Addison-Wesley, ISBN 0-201-74960-2.    (024)