CONTACTS 

SUBJECTS
Personnel, Training, Management Skills
Engineering Life Cycle
Usability Engineering
Mediators, Future Information Systems,

0606 -    ..
0607 - Summary/Objective
0608 -
060801 - Follow up ref SDS 19 0000.
060802 -
060803 - Finally completed analysing this article.
060804 -
060805 -
060806 -
060807 -
060808 -
060809 -
0609 -

SUBJECTS
Personnel, Training, Computers
Software Design, Data Bases, Information Systems
Mediators, Decision Support
Automated Management
Information Management (IT)
POIMS definition, DBMS Criteria
Knowledge & Ideas, Distinguish from Information
Artificial Intelligence
Librarian Organizes Record and Retrieves
Agent Mediator, Stanford

2413 -
241301 -  ..
241302 - Artificial Intelligence Research at Stanford Proposes Mediators
241303 -
241304 - The Acknowledgements indicate that this article is a follow up report
241305 - on a research grant Stanford has, and incorporates work at DEC and
241306 - research on HIV. ref OF 7 0R5L
241307 -
241308 - Widerhold recognizes that increasing information, ref OF 7 S19F,
241309 - reduces effectiveness of decision making, i.e., managment. ref OF 7
241310 - 0001   He envisions hardware will be developed, ref OF 7 AO4O, capable
241311 - of supporting "artificial intelligence" (ref OF 7 GX4G, using
241312 - "automated integration," ref OF 7 0234, or "mediator" software, to
241313 - manage information from different databases and programs. ref OF 7
241314 - 0519
241315 -
241316 - A manual implementation of mediators is at ref OF 7 YW9H, and
241317 - biblography on this topic is provided. ref OF 7 QQPP
241318 -
241319 -
241320 -  ..
2414 -
2415 -
2416 - General Analysis
2417 -
241701 - The author identifies some of the needs for better tools to create,
241702 - preserve, retrieve and apply knowledge.  Implementation is weak,
241703 - because there is not an adequate definition of "knowledge" to guide
241704 - construction of the tools.  The author uses the term "mediator" (see
241705 - definition, ref OF 7 1213) to identify whatever it takes to meet
241706 - undefined needs, in order to avoid association with efforts in
241707 - artificial intelligence.  A parody of "mediator" might be:
241708 -
241709 -      Integrated molecular melarky superimposes suboptimal hueristics
241710 -      to distribute confusion efficiently in a dynamic, multi-variant
241711 -      environment.
241712 -
241713 -        [On 000221 Tom Gruber defines "agents" that use an ontology
241714 -        vocabulary, that sound like mediators. ref SDS 25 2622
241715 -
241716 -  ..
241717 - Mediator - Automating Librarian Role Aided by Experts
241718 -
241719 - Mediator is defined in the appendix. ref OF 7 1213
241720 -
241721 - The authors seem to expect that mediator technology can strengthen the
241722 - librarian role so it can be used to support daily work similar to
241723 - accounting. ref OF 7 YW9H  Mediators somehow help proactive expert
241724 - "librarians" to organize information according to needs, i.e.,
241725 - context, find and present it when needed, similar to article in HBR on
241726 - IT work at IBM and Hallmark, that was reviewed 940510. ref SDS 14 WJ4H
241727 -
241728 - Below, knowledge specialists are called out, similar to Com Manager
241729 - role that implements SDS. ref SDS 0 9503
241730 -
241731 -     [On 960626 librarians are becoming knowledge navigators.
241732 -     ref SDS 22 R59H
241733 -
241734 -     [On 970418 Communication Manager using SDS accomplishes this, per
241735 -     US Army Corps of Engineers' report, ref SDS 24 958L
241736 -
241737 -     [On 000623 Jack Park planned to create an engine to organize the
241738 -     record. ref SDS 26 2915
241739 -
241740 -     [On 001025 Doug Engelbart proposed in the OHS/DKR Launch Plan that
241741 -     Knowledge Integrators be used for applying technology that aids
241742 -     information management. ref SDS 27 8094
241743 - ..
241744 - Since it appears from the acknowledgements that considerable
241745 - collaboration occurred, this article may reflect some of the best
241746 - thinking in 1991 and 1992.  If so, it indicates the importance for
241747 - theorists and developers to gain daily experience trying to use the
241748 - ideas they promolugate, in order to "understand" objectives and
241749 - implications. If Stanford, or anyone else, has implemented even 10% of
241750 - the capability proposed in the article, there should be a rich body of
241751 - knowledge to support work on "mediators," yet it remains merely a
241752 - concept.
241753 -
241754 -
241755 -  ..
241756 - Contacting Author
241757 -
241758 - Author can be contacted at the Department of Computer Science.
241759 - Stanford University, Stanford, CA 94305-2140. His e-mail address is
241760 - wiederhold@cs.stanford.edu.
241761 -
241762 -
241763 -  ..
241764 - Terminology
241765 -
241766 - This shows the liturature seems to relate "artificial intelligence"
241767 - with "planning," "decision support," and "automated integration."
241768 - "Fusion" seems to be used in the way POIMS uses "linking" to create
241769 - knowledge. ref OF 7 G152
241770 -
241771 -     SDS meets this challenge by using common human "hooks" of time,
241772 -     people, objectives and documents, where the objectives are organic
241773 -     and can therefore be viewed at different levels.
241774 -
241775 -     For some reason, there does not seem to be a recognition in the
241776 -     liturature or in products, of the importance of integrating time
241777 -     and information, as defined in POIMS, ref OF 2 line 190, also at
241778 -     ref OF 3 line 88.
241779 -
241780 -          ..
241781 -         Communicate from Past to Present
241782 -
241783 -         Author recognizes this "remembering" function at ref OF 7 line
241784 -         201, but does not present it as a component of integrating
241785 -         time and information. ref OF 7 QL4G
241786 -
241787 -     Artificial Intelligence Replaced by Mediators
241788 -
241789 -     The author attributes work from artificial intelligence as the
241790 -     basis or inspiriation for his ideas on "mediators."  Since
241791 -     "mediators" do not exist, it appears to be merely a substitute for
241792 -     "artificial intelligence" because the latter has fallen out of
241793 -     favour. ref OF 7 G842
241794 -
241795 -        [On 000221 Tom Gruber defines "agents" that use an ontology
241796 -        vocabulary, that sound like mediators. ref SDS 25 2622
241797 -
241798 -  ..
241799 - GUI Does Not Solve Information Overload
241800 -
241801 - Author says that without a new processing paradigm information will be
241802 - hidden in ever larger volumes of detail, scrollable on ever larger
241803 - screens, in ever smaller fonts. ref OF 7 S19F
241804 -
241805 -     SDS provides the "new paradigm" by integrating information, time
241806 -     and objectives, with the management cycle through the synergy of
241807 -     integrated scheduling and reporting.  POIMS "controlled visibili-
241808 -     ty" lets people step into the level of detail needed for a partic-
241809 -     ular matter, see review of Covey at ref SDS 7 line 334.
241810 -
241811 -     [On 950710 Landauer's book "The Trouble with Computers" relates
241812 -     research showing GUI does not ensure usability. ref SDS 20 IP6I
241813 -
241814 -     [On 960629 the Internet began moving people toward correlating
241815 -     computer disk directories with a hierarchial subject structure.
241816 -     ref SDS 23 5555
241817 -
241818 -  ..
241819 - User Friendly Not Useful for Better Communication and Management
241820 -
241821 - Authors do not see a need for a user-friendly interface.  What is
241822 - needed here is a machine-and-communication-friendly interface.
241823 - Application programs executing on the users' workstations can provide
241824 - the type of display and manipulation functions appropriate to their
241825 - users.  Omitting the criterion of user friendliness avoids the
241826 - dichotomy that has led to inadequacies in the Structured Query
241827 - Language (SQL), which tries to be user friendly, while its predominant
241828 - use is for programmed access.7 Standards needed here can only be
241829 - defined after experience has been obtained in sharing these resources
241830 - to support the high-level functions needed for decision making.
241831 - ref OF 7 2T6H
241832 -     ..
241833 -     SDS design improves communication to enable using good
241834 -     management practices consistently.  The user interface makes
241835 -     capturing data and information, and adding connections that add
241836 -     intelligence to management for creating knowledge efficient, i.e.,
241837 -     fast, easy and fun.
241838 -
241839 -        [On 001130 Jack Park reports SDS has the user interface that
241840 -        converts the structure of knowledge into a useful tool for
241841 -        people. ref SDS 28 H17O
241842 -
241843 -        [On 010916 SDS cited for enabling amazing memory using systems
241844 -        that obviously work; desire for better user interface.
241845 -        ref SDS 29 O14M
241846 -
241847 -
241848 -  ..
241849 - Specialists Support Decisions by Segmenting Information into Context
241850 - Context Segments Information into Useful Knowledge for Decision Support
241851 - Design of Knowledge Management Solves Information Overload
241852 - Information Overload Solved by Segmentation and Knowledge Specialists
241853 -
241854 - Author recognizes people experience "information overload" because of
241855 - the difference between information and data, ref OF 7 S19F, and
241856 - ref OF 7 0695
241857 -
241858 -     This is supported by Drucker at ref SDS 9 line 335, and cited in
241859 -     POIMS at ref OF 2 line 14, and at ref OF 3 line 80.
241860 - ..
241861 - Authors say that making decisions requires applying a variety of
241862 - knowledge to information about the state of the world. ref OF 7 PQPS
241863 -
241864 -     This proposition aligns with Jeremy Campbell's analysis of human
241865 -     cognition reported in the record on 900303 that says people use
241866 -     "stories," ref SDS 3 3016, to apply "Baker Street" reasoning.
241867 -     ref SDS 3 5443
241868 -
241869 - Managing this variety requires specialists.  To manage the volume of
241870 - data, we segment our databases.  In these partitions, partial results
241871 - are produced, abstracted, and filtered. The problem of making
241872 - decisions is now reduced to the issue of choosing and evaluating the
241873 - significance of the pieces of information derived in those partitions
241874 - and fusing the important portions. ref OF 7 4F3J
241875 -     ..
241876 -     This reflects the SDS design of segmenting the record into
241877 -     context based on organic structure and using a flexible structure,
241878 -     and adding organization, alignment, analysis, summary connected to
241879 -     detail and feedback, i.e., intelligence.
241880 -
241881 -        [On 011102 OHS/DKR team trying to design KM technology that
241882 -        manages unstructured information. ref SDS 30 0001
241883 -
241884 -     Specialists reflect Com Manager role, which the authors see as
241885 -     "librarians," per above. ref SDS 0 8J6J
241886 -
241887 -
241888 -  ..
241889 - Knowledge from Processing Information and Data through Experience
241890 -
241891 - Authors do not describe a process that generates "knowledge," but
241892 - present a flow chart of knowledge and data feedback loops that show a
241893 - process relating knowledge from acquiring experience by taking action
241894 - that generate data and information. ref OF 7 NX6F
241895 -
241896 - The diagram seems incorrect in failing to link action with experience,
241897 - but showing a process of human cogntion is an important step.
241898 -
241899 -
241900 -  ..
241901 - Fusion of Information by Mediators Creates Knowledge
241902 -
241903 - Wiederhold identifies "knowledge" as abstract classes (ref OF 7 4054
241904 - -- similar to the "objectives" idea in POIMS), which seems to result
241905 - from "fusion" by "mediators".  The article deals extensively with
241906 - "knowledge," e.g., ref OF 7 06S8 and, ref OF 7 474E, yet there is no
241907 - mention that connecting cause and effect is critical to knowledge.
241908 -
241909 -            Author mentions "linking" tangentially, in discusison of
241910 -            SOD technology, (see below). ref OF 7 ES6K
241911 -
241912 -     This is important because the mind seems to "understand" or "know"
241913 -     by assembling pieces that must appear to the mind to fit together
241914 -     and serve an objective.  So without linkage, there can be no
241915 -     knowledge.  Indeed, arguing that "knowledge" can only exist in the
241916 -     human mind, is not an unattractive notion.
241917 -
241918 -      ..
241919 -
241920 -     Data and Knowledge are distinguished pragmatically, rather than
241921 -     resorting to common definitions. ref OF 7 I045
241922 -
241923 -     There does not seem to be a clear distinction between data and
241924 -     information, i.e., these terms are used interchangeably.
241925 -     ref OF 7 PQPS
241926 -
241927 -
241928 -      ..
241929 -     Knowledge & Information
241930 -
241931 -     Author says decisions require "knowledge" available through
241932 -     administrative and technical staff at institutions.  Some
241933 -     knowledge is encoded in data-processing programs and expert
241934 -     systems for automated processing, ref OF 7 9502.  Author says
241935 -     mediators contain the knowledge of experts, ref OF 7 7902.
241936 -
241937 -        This is the view of knowledge as technical skill, like
241938 -        mathamatics, and overlooks the role of the mind to create
241939 -        knowledge by linking chronologies of cause and effect, which,
241940 -        in the case of most executive decision makers, is their
241941 -        personal experience, per Campbell's work reviewed on 900303.
241942 -        ref SDS 3 3002
241943 -            ..
241944 -            Author mentions "linking" tangentially, in discusison
241945 -            of SOD technology, (see below) at ref OF 7 line 811.
241946 -
241947 -     Actually, the author does show a relationship between experience,
241948 -     education, knowledge and information in the graphic at ref OF 7
241949 -     line 233.
241950 -
241951 -           ..
241952 -          Manual Mediator
241953 -
241954 -          A sidebar defines a theoretical manual Mediator department of
241955 -          experts who get and organize information. ref OF 7 YW9H
241956 -
241957 -
241958 -     Knowledge Difficult to "Fuse" without Human Input
241959 -
241960 -     Author gives an example of integrating "knowledge" in libraries
241961 -     that show maps for a trip, with "data" from an airline, and notes
241962 -     this is difficult to process ("fuse" see next section) without
241963 -     human input.
241964 -
241965 -        This seems like a false problem.  It reflects the notion that
241966 -        somehow a computer can figure out what someone wants.
241967 -        ..
241968 -        It is the "human" that wants information and for whom it
241969 -        has meaning relative to objectives. What humans need help with
241970 -        is capturing, preserving, linking and retrieving, their human
241971 -        needs and desires, i.e. objectives, with their perception of
241972 -        the world, i.e. maps and airline availability.  Computers can
241973 -        do this.
241974 -
241975 -
241976 -  ..
241977 - Mediators for Decision Making
241978 - Fusion & Integration
241979 -
241980 - Author explains "decision making" as a process to integrate and fuse
241981 - information to create ranked alternatives, considering risk and long-
241982 - range objectives in combining the results, ref OF 7 line 112, using
241983 - techniques developed in artificial intelligence, ref OF 7 line 765.
241984 -
241985 -     However, AI efforts, according to Jeremy Campbell have not worked
241986 -     out very well, ref SDS 3 line 47.
241987 -
241988 -
241989 -     "Mediator" seems to be a bunch of capability or "services" the
241990 -     author would like to see implemented, ref OF 7 line 825, see also
241991 -     ref OF 7 line 910.  Rather than a unifying concept, "mediator"
241992 -     seems to be a metaphor for whatever it takes to create "fusion"
241993 -     which itself is unclear, resulting in "confusion."  The author
241994 -     applies mediator to mean rules promulgated by "experts" for
241995 -     processing data, see for example ref OF 7 line 678.
241996 -
241997 -           ..
241998 -          Manual Mediator
241999 -
242000 -          A sidebar defines a theoretical manual Mediator department of
242001 -          experts who get and organize information, ref OF 7 line 1059.
242002 -
242003 -          Automated Mediator Features
242004 -
242005 -          Another sidebar at ref OF 7 line 1100, says mediation covers
242006 -          a wide variety of functions that enhance stored data prior to
242007 -          their use in an application. Mediation makes an interface
242008 -          intelligent by dealing with representation and abstraction
242009 -          problems that you must face when trying to use today's data
242010 -          and knowledge resources.
242011 -          ..
242012 -          Mediators have an active role. They contain knowledge
242013 -          structures to drive transformations. Mediators may store
242014 -          intermediate results.
242015 -
242016 -
242017 -     The distinction between "fusion" and "integration" is unclear, but
242018 -     in any event the author recognizes the mind needs more than mere
242019 -     "data" in order to make a decision. ref OF 7 line 112.
242020 -
242021 -
242022 -         ..
242023 -        Integration of Process Not Calculations
242024 -
242025 -        The author seems to define "integration" to support decisions,
242026 -        as merging data, or calculations, ref OF 7 line 205. I think
242027 -        this overlooks the advantage of integrating the management
242028 -        cycle, time and information.
242029 -
242030 -         ..
242031 -        Fusion Assesses Quality & Risks
242032 -
242033 -        Author seems to indicate decision is a leap of faith, based on
242034 -        hueristics that is emulated by ruled based expert systems, ref
242035 -        OF 7 line 211.
242036 -
242037 - Author says "mediator" is a software module that exploits encoded
242038 - knowledge about certain sets or subsets of data to create information
242039 - for a higher layer of applications, ref OF 7 line 445.
242040 -
242041 - Mediators are distinct modules, distributed over a network (ref OF 7
242042 - line 612), that contain knowledge beyond the scope of a database, ref
242043 - OF 7 line 620.  Mediators implement concepts of abstraction and pro-
242044 - vide intelligent processing of data, often dealing with uncertainty,
242045 - ref OF 7 line 625. They respond to queries from applications or to
242046 - triggers placed in databases. They do not interact intelligently with
242047 - each other; a hierarchy is imposed for every specific task. ref OF 7
242048 - line 719.
242049 -
242050 - ..
242051 - Mediators perform functions that are different in scope from the
242052 - tasks performed on computer workstations, ref OF 7 line 637.
242053 -
242054 -     This seems confusing.  People perform work at "workstations." The
242055 -     article seems to be describing a tool that helps people perform
242056 -     work.  How do people use mediators, if not at the workstation?
242057 -
242058 -
242059 -      ..
242060 -     Converting Information into Knowledge
242061 -     through Communication Metrics
242062 -     -------------------------------------
242063 -     The author seems to view "communication" as distribution of
242064 -     information, but notes that this does not guarantee understanding
242065 -     by the receiver.  Mediation considers the meaning assigned to the
242066 -     bits stored, ref OF 7 line 573.
242067 -
242068 -         This conflicts with Drucker's idea of communication as
242069 -         transferring common "understanding," ref SDS 9 line 336.
242070 -
242071 -         It is not clear what the author means by "receiver."  Is this
242072 -         a person or a machine?
242073 -         ..
242074 -         What takes place, and how does it become useful to
242075 -         people, for a mediator to "consider the meaning assigned to
242076 -         bits of stored data"??
242077 -
242078 -     POIMS argues that "meaning" must be discovered, and exists in
242079 -     relation to human drives which vary from person to person and from
242080 -     time to time, though over a long period and for large groups of
242081 -     people, drives occur within a constant range.
242082 -
242083 -     POIMS argues that "communication" occurs only in the mind, and
242084 -     that the best way to find out what takes place in the mind, is to
242085 -     write down what was heard, how it was considered and related to
242086 -     other stuff, and what was decided and why. Otherwise even the
242087 -     decision maker will not "know" why a decision was made.
242088 -     ..
242089 -     POIMS then argues that having made that investment to
242090 -     discover meaning, it can be recycled many times for other stuff
242091 -     that has yet occurred.  It can enhanced with subsequent
242092 -     connections, and leveraged by automation to be applied instantly
242093 -     when needed.
242094 -
242095 -
242096 -      ..
242097 -     New Technology Needed to Implement Mediators
242098 -
242099 -     Author feels new technology is needed to accomplish his idea
242100 -     because accessing one mediator at a time does not allow for
242101 -     fusion, and seeing multiple results on distinct windows of one
242102 -     screen does not support automation of fusion, ref OF 7 line 442.
242103 -
242104 -         Still seems unclear what is ment by "fusion."
242105 -
242106 -
242107 -  ..
242108 - Knowledge Maintenance
242109 -
242110 - Author's explanation of maintaining "knowledge" (ref OF 7 line 641),
242111 - is reflected in the SDS environment.  Experience yields new perspec-
242112 - tives that require integration into the Subject Index, and linkages to
242113 - existing information.  As the data base gets bigger, this becomes a
242114 - larger task.  Two points need consideration:
242115 -
242116 -     1.  There is an advantage in taking more time to "maintain" the
242117 -         subject index, since it affords an opportunity to make sure,
242118 -         unlike the mind, that things are updated that need to be. This
242119 -         means the mind gives itself an opportunity to correct itself,
242120 -         rather than act on false impressions and impulse.
242121 -         ..
242122 -     2.  The task can be moderated by focusing maintenance on
242123 -         matters of current interest, and providing means to accomplish
242124 -         main- tenance in less time in the future by making multiple
242125 -         links, and explaining concerns without doing the detailed work
242126 -         at the moment.
242127 -
242128 - A third point in SDS, is that a lot of maintenance takes place in the
242129 - course of using the system.  Like the human mind, every time we access
242130 - the database, we are checking it against current perception based on
242131 - experience subsequent to the previous access occassion.  We necessari-
242132 - ly bump into related stuff, that can as well be updated.
242133 -
242134 -      [See "Tending the Garden of Knowledge" at ref SDS 21 line 424.]
242135 -
242136 -
242137 -  ..
242138 - Implementing Mediators with Actors and Agents
242139 -
242140 - Author cites "objected-oriented" concepts, "Actors" and "Agents"
242141 - terminology, used by PIM software, as means to implement his vision
242142 - of "mediators." ref OF 7 line 697.
242143 -
242144 -         This entire section is unclear.
242145 -
242146 -
242147 -  ..
242148 -
242149 -    Artificial Intelligence
242150 -
242151 -    Techniques developed in artificial intelligence will be employed,
242152 -    per ref OF 7 line 765, also ref OF 7 line 891 and ref OF 7 line
242153 -    903.
242154 -
242155 -
242156 -     ..
242157 -    Semantic Domains of Discourse (SOD)       [why not SDD??]
242158 -
242159 -    This is an AI approach being considered by a group at Stanford
242160 -    called:  Knowledge-Based Management Systems (KBMS).  The author
242161 -    says this effort will result in a specific kind of mediator, ref
242162 -    OF 7 line 783.
242163 -
242164 -    It has a language, SODAL, but requires further work to implement
242165 -    the author's "mediator" objectives, ref OF 7 line 824.
242166 -
242167 -    Uses "keyword" technology and assesses quality, bias, ref OF 7
242168 -    line 838.
242169 -
242170 -        I understand "keywords" which is pretty low-tech, but how are
242171 -        quality and bias evaluated?
242172 -
242173 -
242174 -
242175 -  ..
242176 - Mediators Can Learn
242177 -
242178 - The "rules" can be updated automatically by gathering new statistics,
242179 - per ref OF 7 line 744 and further at ref OF 7 line 757 and ref OF 7
242180 - line 906.
242181 -
242182 -
242183 -
242184 -
2422 -

SUBJECTS
Computers not used for management
Scribe, Communication Manager until Tools
Communication Manager Makes Technology
Subject Indexing
Capture the Record

2908 -
290801 -  ..
290802 - Managers Do Not Use Computers Because Information Not Fast Easy
290803 -
290804 - Author explains computers are used by staff, but not managers because
290805 - information from multiple sources is too diverse for effective
290806 - integration in a computer. ref OF 7 KM8O
290807 -
290808 -    Cultural inertia from experience using other systems resists using
290809 -    SDS that provides integration needed to make technology effective
290810 -    for management, and reported by Morris on 890809, ref SDS 2 6047,
290811 -    and more recently an article in HBR reviewed on 940510, ref SDS 14
290812 -    3829, explaining managers use conversation and printed documents,
290813 -    rather than computers to get information for decisions.
290814 -    ref SDS 14 0034
290815 - ..
290816 - It appears the author is "selling out" to executives who resist
290817 - using computers.  He seems to be saying the benefits of automation are
290818 - available without hands on-use, like an ancient ruler relying on
290819 - scribes to read and write, instead of learning to take pen in hand.
290820 - This overlooks the idea of creating knowledge, actually building it
290821 - through the writing process, ref OF 2 line 502, and with SDS the
290822 - linking process explained at ref OF 2 line 511, and further at
290823 - ref SDS 5 line 263.
290824 -
290825 -
290826 -
290827 -
290828 -
290829 -
290830 -
290831 -
290832 -
290833 -
290834 -
290835 -
290836 -
290837 -
290838 -
290839 -