Click here to comment!
1...This sounds great, how soon can we get it into SDS to see if
2...How does this align with Doug's call for better handling of
Architectural Snag Linking Versioning, Eric Armstrong
Linking, Versioning Architectural Snags, Eric Armstrong
Architecture Snag Linking Versioning, 001018 Eric Armstrong
Architecture Problems KM Linking Versioning, Eric Armstrong, 001018
Versioning Architectural Snag KM, Eric Armstrong
Error Recovery Processing Occurs in AI Organizing Email Automatically
1008 - ..
1009 - Summary/Objective
101001 - Follow up ref SDS 29 0000, ref SDS 28 0000.
101003 - Paul explains types of error recovery that may be impacting OHS
101004 - architecture Eric Armstrong is developing. SDS design uses computer
101005 - aided thinking that avoids AI automated organization of subjects and
101006 - documents, because existence and human thought are too complex for
101007 - reliable processing of organic subject structures. Paul's explanation
101008 - points to ideas for possible open source add-ons that could enhance
101009 - SDS, ref SDS 0 BO4J Paul's proactive use of SDS records on the
101010 - Internet illustrates the opportunity for Communication Metrics to
101011 - improve productivity of knowledge work. ref SDS 0 5P2B Jack Park
101012 - submits support for open source development, ref SDS 0 ADK2, which
101013 - offers more evidence that delivering intelligence on the internet
101014 - provides context to grow useful knowledge about critical issues.
101015 - ref SDS 0 ADK1 This shows potential to grow a new work practice of
101016 - Knowledge Management, where SDS and Com Metrics might contribute.
101017 - ref SDS 0 5P2B
101023 - ..
1013 - Progress
101401 - Received ref DRT 1 0001 from Paul Fernhout commenting on a question in
101402 - the record on 001023 about error recovery, ref SDS 29 HZ5G, where he
101403 - says in a letter...
101405 - Another way to think of this is as "classes of validation" or
101406 - "classes of error recovery". So, some errors one expects the
101407 - system to handle routinely and other errors one expects a human
101408 - to have to detect or deal with (ranging from a minor automated
101409 - error recovery confirmation to a day long debugging session).
101410 - ref DRP 8 XG5J
101411 - ..
101412 - I set an action item...
101414 - Need example of errors that require "recovery" capability by the
101415 - system. What is the "system" doing that can make an error?
101416 - ref SDS 29 715O
101418 - Paul explains...
101420 - ...email system automatically categorizes based on content and word
101421 - frequency. (IBM's doing some cutting edge work on this by the way.)
101422 - Every once in a while you might review some of the categorizations
101423 - and correct them. For example, the system might put all your email
101424 - from this mailing list under UnrevII but you might want some but not
101425 - all of it to also be categorized under "Knowledge Management". The
101426 - "error" here is a miscategorization based on your wishes, which you
101427 - may not have made clear to the system, or which the system
101428 - algorithms may not be quite able to handle 100%. ref DRT 1 716I
101430 - Organizing the record in the manner Paul describes is a core
101431 - task in KM, called variously subjects, ontology, category and so
101432 - on.
101434 - Requirements for organizing the record change over time, so it is
101435 - risky to have a computer do this unattended. Paul recognized
101436 - the complexity issue in a discussion with Jack Park on 001025.
101437 - ref SDS 30 RH5H
101439 - Paul seems to cover this point later in his letter.
101440 - ref DRT 1 BI4I
101442 - Paul's letter on determining "meaning" in another record
101443 - today, further recognizes the difficulty of depending on
101444 - technology to adequately organize the record. ref SDS 32 TH4J
101446 - Additionally, SDS intends for the knowledge worker, e.g., an
101447 - engineer, executive, doctor, lawyer, carpenter, baker, or
101448 - scientist, secretary and in some cases a Communication Manager to
101449 - review documents and decide what is involved. The exercise of
101450 - assigning subjects is a big part of understanding correlations,
101451 - implications and nuance, which is the objective of KM. So, we
101452 - really don't want an automated system to deny users the
101453 - opportunity to understand information from email or other means.
101454 - ..
101455 - A system that shows choices correlating word patterns with
101456 - the organic subject structure could help. Paul, Jack and others
101457 - have discussed from time-to-time topic maps, for example, that
101458 - might turn out to be useful. This needs to be investigated.
101460 - On 000602 Jack Park cited Ontologos as a program that has made
101461 - progress automating subject management. ref SDS 12 2496
101462 - Research seemed to later establish that Ontologos has not
101463 - actually accomplished this capability. ref SDS 14 0001
101465 - SDS addresses likelihood of errors organizing the record by
101466 - providing a variety of different ways to get information when
101467 - needed. Documents are a small, though important, aspect of
101468 - knowledge management incorporated into SDS records. There may be
101469 - 100+ subjects associated with a record, plus other discreet
101470 - pointers that provide a rich environment to make document
101471 - management reliable in SDS.
101473 - Never-the-less it would be very useful to develop some code that
101474 - could do the 4 or 5 routine tasks required to get a letter
101475 - received via email set up in an SDS record ready to review.
101477 - This is a good opportunity for an open source task, per
101478 - planning on 000503. ref SDS 10 6903, and earlier on 000423.
101479 - ref SDS 8 8000
101481 - Eric is focusing on email as the core of his CDS system (see
101482 - for example on 000505. ref SDS 11 4392) This work may yield
101483 - useful tools to save time handling documents, which would make
101484 - knowledge work more productive.
101486 - ..
101487 - Paul continues...
101489 - Another example might be that you realize there are two Doug
101490 - Engelbart's sending you email but they are categorized together --
101491 - you go in to the system and review all the correspondence and split
101492 - them apart (perhaps using tools that help you differentiate based on
101493 - email address) and then build a way to always keep them separate. Or
101494 - likewise, you get email both from an "Rod Welch" and "Welchco" and
101495 - you decide you want them categorized together. ref DRT 1 KO7I
101497 - Sounds like Paul is describing setting parameters for an
101498 - organizing program, along the lines Jack described on 001025,
101499 - ref SDS 30 LR6N, and previously on 000623. ref SDS 16 4752 This
101500 - idea of an automatic ontology generator seems to have initially
101501 - posed by Jack's letter on 000602 commending Ontologos as having
101502 - developed capability that could improve SDS. ref SDS 12 2496
101504 - SDS requires constant attention to making these distinctions, and
101505 - so tools that expedite the work could help, per above.
101506 - ref SDS 0 BO4J
101508 - Paul says probably the best example of error recovery requirements
101509 - is that you run some badly written piece of code (or code with
101510 - assumptions or prerequisites your system does not adhere to) and it
101511 - adds bad or incomplete data to your database. You then add some good
101512 - data you don't want to loose, so rollback isn't easy. You might then
101513 - need to go in by hand and find and fix the bad data based on an
101514 - understanding of the code and what it did wrong. That wouldn't be
101515 - easy to automate. ref DRT 1 VR9G
101517 - Off hand, this might conceivably arise, but generally it seems to
101518 - conflict with the definition of knowledge as human experience,
101519 - which is the objective of SDS.
101521 - Like human memory, knowledge grows daily. Like tending a garden,
101522 - the SDS record is under constant scrutiny using a process of
101523 - continual learning. Of course the program isn't learning
101524 - anything; it merely aids learning and memory of people, which is
101525 - a big part of human reasoning, as reported on 900319. ref SDS 1
101526 - 0702
101529 - Paul cites potential for hardware failures...
101531 - Finally -- your database suffers physical (disk) corruption but you
101532 - want to save what you can. A person might go in and see what is
101533 - salvageable and direct various tools to recover parts of the data.
101534 - ref DRT 1 RS9O
101536 - SDS is an environment for thinking and doing daily work; the
101537 - proverbial little engine that can produce knowledge. It contains
101538 - all of the information and connections people recognize as
101539 - important enough to require focused attention. as Doug envisions
101540 - for the OHS reported on 001025. ref SDS 30 7350 Integrity of the
101541 - system is therefore critical. Back up is done daily, and the
101542 - record is regularly placed on different machines, so if a
101543 - hardware problem should occur, loss of the record is minimized
101544 - through redundancy and backup. Rather than hire someone to
101545 - recover data from a hard disk in the event of a hardware failure,
101546 - a new hard drive is installed and the record is transferred or
101547 - restored.
101549 - The likelihood of system failures is one reason why schemes for
101550 - using a server on the web as the primary storage medium seem
101551 - unworkable, except for delivering mail, as reviewed on 960304.
101552 - ref SDS 2 9007
How Soon Can I Get This in an SDS Record
130301 - ..
130302 - Reviewing the Record and Following Up Core Objective of KM
130304 - Paul mentions in a post script...
130306 - Welchco system provides an interesting example of knowledge
130307 - management & integration issues from using multiple systems. You
130308 - have questions but I don't know of them if I only look at the email
130309 - list. So either I need to check several systems myself, or I need a
130310 - system sophisticated enough (Apple's Knowledge Navigator prototype?)
130311 - that it does if for me and presents it in a common format.
130312 - ref DRT 1 00W3
130314 - [On 001116 Paul comments SDS again. ref SDS 37 H23M
130316 - It is gratifying to see proactive efforts by Paul and others
130317 - reviewing the record for pending issues. (See for example Jack
130318 - Park's letter below. ref SDS 0 ADK1) If participation like this
130319 - grows, then SDS and Com Metrics can succeed. As well, if Doug's
130320 - open source engineers take the initiative, as Paul has done in
130321 - hunting up an action item in SDS, then OHS will be a great
130322 - success.
130324 - Paul's work addressing action items without express notice shows
130325 - how SDS is applied to improve daily work by amplifying
130326 - "intelligence" across tradational boundaries of space and time.
130327 - We need to get people designing cars, writing laws, and directing
130328 - projects, engaged in using the record, as Paul demonstrates
130329 - today. This is how world problems are solved and avoided, and how
130330 - time and money are saved, i.e., using KM in the work place, which
130331 - was considered on 000920. ref DIP 2 006P
130333 - Complexity of life and daily work, which Paul and Jack have been
130334 - discussing, for example on 000125, ref SDS 30 02TX, drive demand
130335 - for KM technology to maintain alignment. The only issue is how to
130336 - give the opportunity visibility in a market place where bandwidth
130337 - is zero because people are chasing solutions down blind alleys,
130338 - i.e., looking for love in all the wrong places.
130339 - ..
130340 - So, we potentially we have a customer profile here.
130342 - What led Paul to proactively review the record? If we can bottle
130343 - that we will have a customer? In the mean time, its great to see
130344 - enthusiasm showing people care about improving knowledge work,
130345 - discussed in the letter on 001004. ref DIP 3 ON5G For those who
130346 - care enough, the single, biggest step a CEO can take to improve
130347 - earnings is to improve understanding, shared meaning, creativity,
130348 - and collaboration, and, also, reduce mistakes, by saying at the
130349 - end of every meeting...
130352 - This sounds great, how soon can we get it into SDS to see if
130353 - everything is aligned?
130356 - The merits of Paul's immediate query go the purpose of SDS, see
130357 - for example the typical day scenario on how SDS is used.
130358 - ref OF 2 0001
130360 - Every question doesn't necessarily need an immediate answer.
130362 - SDS enables users to identify questions and action items, and
130363 - work issues in an orderly way, as the need arises. Issues are
130364 - never lost, and are always available when the time is right to
130365 - take up pending matters.
130367 - For example, On 000929 Doug planned to have a meeting on action
130368 - items that have been building up on his OHS project with SRI,
130369 - ref SDS 23 6M8I, per the record on 000922. ref SDS 20 0001 He
130370 - initially indicated this would be a good idea, but has not
130371 - followed up, possibly because SRI wants to take a crack at
130372 - creating an SDS program themselves, per telecon with Pat Lincoln
130373 - on 000829, ref SDS 18 1056, and the meeting on 001017.
130374 - ref SDS 26 9622
130376 - In any case, people don't need to feel a duty to hunt for issues
130377 - from an SDS user. SDS enables users to develop questions and
130378 - answers as needed. Paul is correct this is a big part of KM, and
130379 - SDS handles it pretty well. However, to be effective it needs to
130380 - be used, and sometimes leadership enhances results, as set out in
130381 - NWO... ref OF 3 5653
Management Cannot Make Assignments
Contributors Use Links in Launch Plan to Comment on Plan
Links to Sources Omitted, Listen to Doug
SDS Record Use Shows KM Can Succeed
Listening to Doug Alignment Doug's Writings Work of Colloquium
Linking Aids Alignment Understanding Avoid Meaning Drift
Feedback Shows Attractiveness of SDS Record on Web
Context Management Positions Information in Patterns of Chronologies
391101 - ..
391102 - Listening to Doug
391103 - Feedback Key Metric of Understanding, Communication and Intelligence
391104 - Context Requires Links to Original Sources to Enhance Understanding
391106 - Follow up ref SDS 30 ADK1.
391108 - Received ref DRT 2 0001 from Jack Park commenting on analysis of open
391109 - source development, ref SDS 30 ADK1, provided through a link in a
391110 - letter to the team, ref DIP 5 0001, requested by Doug Engelbart in his
391111 - letter asking for feedback on the OHS Launch Plan, ref DRP 9 0001, he
391112 - submitted on 001025. ref SDS 30 G3W8
391114 - Jack initially says pointers in SDS records that provide access to
391115 - original sources are helpful. ref DRT 2 ET6G
391117 - Jack's willingness to invest time providing feedback supports the
391118 - concept of "intelligence" in Communication Metrics theory set out
391119 - in POIMS, ref OF 1 3774 and ref OF 1 2200, and also in the NWO...
391120 - ref OF 3 0936
391122 - Growing application of technology to aid human memory and
391123 - reasoning, i.e., augment intelligence, takes time, but Jack's
391124 - participation shows it may eventually take root to yield a rich
391125 - harvest of knowledge and ideas that can lift a team, a company,
391126 - community, indeed civilization, per analysis above of Paul
391127 - Fernhout's proactive inquiries. ref SDS 0 5P2B
391128 - ..
391129 - Jack's letter extracts and comments on portions of the record
391130 - for 001025, ref DRT 2 ET6G; he provides no links that facilitate
391131 - reviewing context. This conflicts with an objective of KM to maintain
391132 - context that aids understanding, as pointed out by Gil Regev in his
391133 - letter on 001019. ref SDS 28 TF5H Earlier on 000926 Henry van Eykan,
391134 - cited advantages of context. ref SDS 22 375K On 000824 Eric Armstrong
391135 - wanted email in context. ref SDS 17 G92D Jack supported these calls
391136 - for context in a letter to Paul Fernhout, which Paul cited on 001025.
391137 - ref SDS 30 I58G
391138 - ..
391139 - The same day, 000125, Doug Engelbart requested that the team
391140 - comment on the OHS Launch Plan by linking to specific provisions.
391141 - ref SDS 30 00VU Over the past 9 months Doug set an objective for OHS
391142 - to provide addressability, which means in part adding anchors to
391143 - facilitate linking into original sources. Doug's letter on 001025
391144 - applies and tests the usefulness of this capability, as set out in the
391145 - Launch Plan. ref SDS 30 D3W8 Analysis on 001025 indicates that
391146 - linking into original sources helps maintain context, ref SDS 30 00VU,
391147 - which meets objectives cited by Jack, Gil and Eric, but it is resisted
391148 - for a lot of good reasons. ref SDS 30 XF4N
391149 - ..
391150 - On 000424 Adam Cheyer encouraged the DKR team to listen to Doug,
391151 - and follow his lead; Jack concurred. ref SDS 9 1590 Since Doug has
391152 - expressly called for links, and since anchors are on every para and
391153 - heading in SDS records and documents, understanding is aided and the
391154 - work is expedited, saving time and money, by providing links into the
391155 - context of the record, as called out by Doug.
391157 - If there is not enough time using conventional methods, consideration
391158 - might be given to using SDS, although that is a big step that entails
391159 - implementing KM.
Open Source Development for Unspecified Compelling Reasons
Open Source Termites Contributors Develop Own Designs Yields Stronger
Study Needed Demonstrate Open Source Best Strategy, 000601
80% Solution Typical Open Source Result
100% Solution Planning Needed for Open Source, 001017
SRI Not Convinced Open Source Best Strategy
Engelbart Convinced Open Source Best Strategy
Feel Good Management Empowerment Do What We Want
Open Source Do Something Even If Wrong
Product Design, Control Lost
Management Cannot Make Assignments
Empowerment Creativity Termite Design Open Source Rejects Management
Open Source Effective Management Method Based on Unsupported Referenc
541601 - ..
541602 - Open Source Development for Clear, Compelling Reasons
541603 - Phase-1: OHS Launch Project: HyperScope with Open Source Development
541605 - Follow up ref SDS 30 ADK1.
541607 - Jack characterizes analysis on 001025 as
541609 - ...collected thoughts of an individual carrying the baggage of a
541610 - personal bias towards open source. ref DRT 2 A29O
541612 - He may have intended to say....... "bias against open source."
541613 - ..
541614 - Jack further says... Nothing in the listed citations pays homage
541615 - to the fact that thousands of individuals all over the world came
541616 - together to build Linux, an operating system that, indeed, has the
541617 - likes of Microsoft behaving in ways that are terribly unprofessional,
541618 - much like politicians who are trying to win office. Nowhere do you
541619 - cite any of the really important projects ongoing world wide that are
541620 - improving the lot of computer scientists, not to mention moms on the
541621 - internet doing their shopping and so forth. ref DRT 2 I44H
541623 - These are generalities.
541625 - [On 001101 Jack submits article explaining challenges Linux
541626 - faces using open source. ref SDS 34 HW9F
541628 - [On 001106 VA Linux stock drops 43%. ref SDS 35 0001
541630 - [On 010111 slow progress on Hyperscope using open source.
541631 - ref SDS 39 0002
541633 - [On 010112 no progress translating Augment to Java using open
541634 - source management. ref SDS 40 0001
541636 - What open source projects are producing KM work product? On
541637 - 000324 Lee Iverson reported that SRI research shows KM projects
541638 - have failed. ref SDS 6 4877 On 000602 there was a claim ontology
541639 - has been developed. ref SDS 12 8932 This was suspect because
541640 - subjects are difficult, per POIMS. ref OF 1 0582 On 000606 a
541641 - source claimed field tests establish utility of ontology.
541642 - ref SDS 13 0001 Research on 000607 showed the claim was a
541643 - mistake. ref SDS 14 0001 Accordingly, unsupported claims do not
541644 - lend confidence to open source for KM development.
541646 - [On 001127 letter to team explains this point, that if KM can
541647 - be developed without SDS, or devining requirements without a
541648 - painful discovery process, it would have been done. ref SDS 38
541649 - CV3L
541651 - Linux is an operating system (OS) for computers. There is a lot
541652 - of knowledge about OS dating back to the 1940s. As a result,
541653 - people can go to school and learn about OS; they get experience
541654 - using computers interacting with OS. This experience yields ideas
541655 - for improvements. Folks can attend seminars; look things up on
541656 - the Internet; exhange ideas about technologies via email. These
541657 - circumstances enable people to develop incremental improvements by
541658 - doing what they want, under the open source synergy theory of
541659 - going in the wrong direction, set out on 001012, ref SDS 24 B4P2,
541660 - ..
541661 - Knowledge Management is different.
541663 - KM is an operating system for people and organizations. It's new;
541664 - and so hasn't been around for 50 years. There is no school where
541665 - engineers can learn about KM. Education for KM was proposed to
541666 - Stanford on 000929, but it is not a priority. ref SDS 21 RT6N
541667 - There are a lot of of seminars, books and magazines on KM, but
541668 - none explain how to produce KM; everyone shows up and asks what is
541669 - it, as reported on 991217. ref SDS 3 9030
541671 - As a result, On 000615 the DKR team reported 6 months of research
541672 - at SRI showed there is not enough knowledge to develop Knowledge
541673 - Management. ref SDS 15 6271,
541675 - [On 001031 Jack Park notes he does not have enough knowledge to
541676 - develop KM, but expects it can be obtained from the Internet.
541677 - ref SDS 33 ADK2
541679 - [On 001123 letter to team makes this point again. ref SDS 38
541680 - E55F
541682 - Without enough knowledge from school, publications and colleagues,
541683 - and without experience using KM tools, it is harder to develop
541684 - core technology for Knowledge Management than it is to support the
541685 - Linux OS. This record leads to the view that open source is
541686 - conducive for developing enhancements to established technologies,
541687 - as proposed on 000503. ref SDS 10 6903
541689 - Creating KM technology requires doing KM 24 7 in order to learn
541690 - what to create, and what to improve. On 001025 Doug requested
541691 - that contributors use links to specific locations in the record,
541692 - which is the minimum level of effort for doing KM. ref SDS 30 00VU
541693 - This has not been done for 10 months, ref SDS 30 MO6M, and as of
541694 - yesterday, Doug reported nobody has responded to his request.
541695 - ref SDS 31 0001
541698 - Jack asks... What, in your judgement would constitute *compelling
541699 - evidence*? Does the fact that several dotcoms devoted to open source
541700 - got great press, not to mention a lot of wall street interest, do
541701 - that? If not, then what? ref DRT 2 A55K
541703 - Define knowledge to guide development of technology that improves
541704 - information management.
541706 - Progress on OHS.
541708 - Evidence that anyone, anywhere has produced KM, i.e., show me the
541709 - work product.
541711 - Progress on 100+ action items in the record on 000922, ref SDS 20
541712 - 0001, would show open source can get things done.
541714 - Progress reviewing Doug's writings. For 10 months, open source
541715 - contributors have offered references to technologies and web
541716 - sites, but nothing is aligned with Doug's explanation of KM, for
541717 - example on 000327. ref SDS 7 3971
541719 - Jack says today that for SDS records to make complete sense,
541720 - ultimately, Welch thoughts need to be blended with those of
541721 - others, not by email, but in a massively similar web space.
541722 - That way, we can all make sense of all of our biases.
541723 - ref DRT 2 HQ5G
541725 - How does this align with Doug's call for better handling of
541726 - daily working information? ref SDS 7 3971
541728 - How does it align with the definition of "knowledge."
541730 - SDS blends Welch thoughts with Jack's, Paul's, Doug's,
541731 - Eric's..., et al, Drucker, Landauer, Kissinger, Hofstadter,
541732 - Pinker, Grove, Havelock, Plato, Aristotle and many more,
541733 - making SDS a continual blend of thoughts and sources.
541734 - Therefore, need example of a better "blend" to see what
541735 - Jack has in mind. What, for example, is a "...massively
541736 - similar web space"?
541738 - Why not get started on KM with partial sense? If we wait
541739 - for "complete sense," we may never get there.
541740 - ..
541741 - The meeting at SRI on 001017 indicates work on OHS is
541742 - pending award of funds from sponsors. ref SDS 26 4877 Open Source
541743 - nominally thrives on volunteer work. Over 10 months nothing has
541744 - been volunteered. This doesn't mean people are lazy, indifferent,
541745 - or un-talented. Doug has assembled an outstanding team, and SRI
541746 - is planning a strong core development team. However, if there is
541747 - not enough knowledge, progress is difficult. On 001017 Eric
541748 - notified the team he will volunteer source code soon, which is a
541749 - constructive sign, ref SDS 25 0001, but that conflicts with the
541750 - meeting at SRI where Eugene indicated the architecture will be
541751 - developed by a new team. ref SDS 26 TJ3G
541753 - [On 001106 Eric advises code release is delayed due to design
541754 - problems. ref SDS 36 0001
541756 - On 001025 Doug Engelbart requested comments linked to the OHS
541757 - Launch Plan. No one has done this because they are busy doing
541758 - what they want, reported on 001012, ref SDS 24 PT5M, and they
541759 - don't want to do KM, for reasons in the letter to the team on
541760 - 000920, ref DIP 2 719F, reviewed again in the record on 001025,
541761 - ref SDS 30 XF4N Software engineers want to work on interesting
541762 - technology; but, KM is not interesting, because there isn't enough
541763 - knowledge.
541765 - This record suggests open source is a good development method for
541766 - enhancing KM, proposed on 000503. ref SDS 10 6903 A core
541767 - capability may be needed to guide such an effort, as Eric noted on
541768 - 000120, ref SDS 4 3002, and discussed with Doug on 000327.
541769 - ref SDS 7 9975 Creating the core capability requires doing things
541770 - people don't want to do, which conflicts with empowerment precepts
541771 - of open source.