THE WELCH COMPANY
440 Davis Court #1602
San Francisco, CA 94111-2496
415 781 5700
rodwelch@pacbell.net


S U M M A R Y


DIARY: April 28, 2000 03:15 PM Friday; Rod Welch

Received notes on meeting yesterday for OHS/DKR project at SRI.

1...Summary/Objective
2...ThinkTank Experience Suggests Caution About Adopting WBI Transcoding
3...WBI Useful, Not Core Component of DKR, Too Slow, Not Stable Record
4...Email Enhancement Start with Hackable Program, Modular Architecture


..............
Click here to comment!

CONTACTS 

SUBJECTS
OHS/DKR Meeting at SRI 000427
Guide to Running NICs, 000427

0604 -
0604 -    ..
0605 - Summary/Objective
0606 -
060601 - Follow up ref SDS 2 0000, ref SDS 1 0000.
060602 -
060603 - Received ref DRT 1 0001 from Eric Armstrong setting out results of the
060604 - meeting yesterday at SRI. ref SDS 9 0001
060605 -
060606 - Received 3rd letter, ref DRT 3 0001, from Eric on Doug's presentation
060607 - of WBI Transcoding and "intelligence" gathering at meeting yesterday.
060608 - ref SDS 9 3995
060609 - ..
060610 - Received 4th letter from Eric, ref DRT 4 0001, submitting more
060611 - information on "intelligence" and transcoding presentation, yesterday,
060612 - and providing additional observations.
060613 -
060614 - Eric advises that he is unclear about scope for a "Guide to Running to
060615 - a NIC", and indicates guidance is needed on how to contribute on this
060616 - new initiative set by Doug during the progress meeting at SRI
060617 - yesterday. ref SDS 9 2867  Eric says "It's pretty clear to me that I
060618 - don't know the first thing about it." ref DRT 4 3000
060619 -
060620 -
060621 -
060622 -
060623 -
060624 -
060625 -
0607 -

SUBJECTS
WBI Transcoding Vector
WBI 3 Tier Server Architecture, 000427
WBI Not Suited for DKR, 000427

1605 -
160601 -  ..
160602 - ThinkTank Experience Suggests Caution About Adopting WBI Transcoding
160603 - WBI Useful, Not Core Component of DKR, Too Slow, Not Stable Record
160604 -
160605 - Follow up ref SDS 9 4674.
160606 -
160607 - Eric submites in his 4th letter today the opinion that WBI is
160608 - fundamentally the wrong approach for the DKR.  Its a great tool to
160609 - investigate, and undoubtedly has uses -- even within this project --
160610 - but not as the fundamental component, ref DRT 4 6975, because...
160611 -
160612 -     •  Server-based manipulation of the user's view too slow
160613 -
160614 -        Having to make a round-trip to the server every time you want
160615 -        to change your view is, in my opinion, simply untenable. I find
160616 -        that kind of system unacceptable on a high-speed local network.
160617 -        I hate to think of what it would be like over the Web,
160618 -        especially across a modem. ref DRT 4 7488
160619 -
160620 -     •  Degradation of Information
160621 -
160622 -        Links would degrade over time as pages are moved or modified.
160623 -        That's the way the Web is today, true. But the problem would be
160624 -        exacerbated by an order of magnitude -- the average number of
160625 -        paragraphs per document. Even worse, since modifying a page
160626 -        would change the tags, any links I created might still be
160627 -        "valid", but pointing to the wrong location.  The result would
160628 -        be gibberish as text I linked to or included inline from
160629 -        another document would be replaced by some other text in that
160630 -        document.
160631 -
160632 - The result of these two observations is that the system, as presently
160633 - envisioned, will be both slow and ineffective. I saw ThinkTank
160634 - singlehandedly destroy the outliner market by building a truly awful
160635 - interface and marketing the hell out of the product. They got a
160636 - tremendous amount of mind share and investigation by early adopters --
160637 - all of whom quickly concluded that the concept was useless. The fault
160638 - was the interface, rather than the concept, but by the time we could
160639 - prove that, the damage had been done and the market had all but
160640 - disappeared.
160641 -
160642 - While an architecture built around WBI might be something we can learn
160643 - from, I can only hope that it does not get promoted too widely, lest
160644 - it have the same unfortunate impact on public perception that
160645 - ThinkTank did.
160646 -
160647 -      [On 001025 multiple views incorporated into OHS launch plan.
160648 -      ref SDS 13 D6H8
160649 -
160650 -
160651 -
160652 -
160653 -
1607 -

SUBJECTS
Email Improved by Webmail, 000427
ArsDigita Email System, 000427
Email Enhandement Tools, Hackable, 000428
Email Enhancement Hackable Program, 000428
Email XML Seed Team Use What is Available, 000302
Hackable Email XML Seed Team Use What is Available, 000428
Small Scale Low Risk Easier to Sell
Hack Nucleus, 000208
Email Hackable WebLog Organizing and Thought Tracking, 000428
Zope Email System, 000427
SlashDot Email System, 000427
Modular Architecture Facilitates Hacking

2814 -
281501 -  ..
281502 - Email Enhancement Start with Hackable Program, Modular Architecture
281503 -
281504 - Follow up ref SDS 9 5985, ref SDS 1 5073.
281505 -
281506 - Received 2nd letter from Eric on understandings from meeting yesterday
281507 - to evaluate a range of starter technologies in relation to SlashDot,
281508 - for improving email, ref DRT 2 0001, which is reported on 000427.
281509 - ref SDS 9 4836
281511 -  ..
281512 - Eric comments that, all things being equal, the team should choose the
281513 - system that is most "hackable" -- that is, the one with the most
281514 - modular architecture and the most reasonable language for hacking in.
281515 - ref DRT 2 1683
281516 -
281517 -    On 000426 Jack Park explains architecture guides product design for
281518 -    accomplish project requirements set by marketing, engineering and
281519 -    vision. ref SDS 7 0304  SDS provides proof of concept. ref SDS 7
281520 -    3315
281522 -  ..
281523 - Eric initially rules out ArsDigita (Tcl) and SlashDot (Perl +
281524 - installation process). ref DRT 2 5103
281526 -  ..
281527 - That leaves PHP/SlashDot and Zope. ref DRT 2 6237
281528 -
281529 -     [On 000502 Eugene Kim commented that none of the email enhancement
281530 -     programs reviewed by the project team improve upon conventional
281531 -     email. ref SDS 10 0001
281532 -
281533 -     [On 000503 Eric expects Zope, SlashDot can be improved by hacking.
281534 -     ref SDS 11 3465
281536 -  ..
281537 - The questions he sees are...
281538 -
281539 -     •  What kinds of evaluation mechanisms do they offer?
281540 -
281541 -     •  How good a language is PHP for our purposes, vs. Python?
281542 -
281543 -         [On 000524 Paul Fernhout submits to Eric information on
281544 -         Pointril Data Repository system implemented in Python.
281545 -         ref SDS 12 0584
281546 -
281547 -     •  Since PHP/SlashDot uses the Apache Server and Zope is a full
281548 -        server in its own right, how important is it for us to hack the
281549 -        server? Can we be fully non- denominational with respect to
281550 -        servers, or will we need to make important hacks to the server
281551 -        to implement the functionality we envision. (The former argues
281552 -        for PHP/SlashDot. The latter for Zope.)
281553 -
281554 -
281555 -
281556 -
281557 -
281558 -
281559 -
281560 -
281561 -
281562 -
281563 -
281564 -
2816 -