KOSH [Kommunity Orientated Software Hardware]

Weekly Summary

Week Commencing: 13th February 1999

Number: 008

Mailing List: kosh-general

In the mailing list this week, the following items were discussed. Please do
not email the scribe regarding any of these topics, it is not his job to answer
these questions but merely to report  the topics of conversation. If you have
any queries about this summary, please email ben@kosh.net, stating the Summary
Number, and Mailing List Name, and he will try to answer your queries.

a)

Subject Object Ocean

Summary of debate: An object ocean made up of many object seas plus network
                   objects which belong in no particular sea, but are
                   accessible from all was suggested.

                   An application server could send out precompiled binary
                   across just such a network. This may not be a "physical
                   server" but instead a function of the operation of the sea.
                   If a machine is isolated from the sea and then becomes a
                   part of the sea then all the common resources of the sea
                   should become available to it.


b)

Subject: Cloning

Summary of debate: A variation of migration is where a task is duplicated on
                   another machine but the original remains in place.


c)

Subject: More book text

Summary of debate: The Macmillan Personal Bookshelf has a "side service",
                   Betabooks which has current text of books-in-progress
                   available online.
                   The URL is http://www.mcp.com/betabooks


d)

Subject: Character sets (continued)

Summary of debate: Alternative character sets such as Japanese or Chinese
                   should not be treated separately, but instead we should use
                   UTF-8 or Unicode which could support such character sets.


e)

Subject: Slim binaries and anti-piracy

Summary of debate: Fixing slim binaries will also solve some serious problems
                   in anti piracy, it will enable a portable keyfile system to
                   be used across various platforms.


f)

Subject: UTF-8 or Unicode instead of ASCII?

Summary of debate: Would it be possible to use UTF-8 or Unicode instead of
                   where conventional systems use ASCII? The two main problems
                   with ASCII are that it is a vaguely defined standard and
                   only has 256 characters which is not enough. Unfortunately
                   most other systems use ASCII and programs would need
                   translators.

                   Alternatively make a new character code standard,
                   "KOSHKode". Each letter would be 32 bits. The low 16 bits
                   are standard Unicode but the upper 16 bits could be used for
                   font size and font selection. This could help make a
                   standard for wordprocessing. However only "KOSH-aware"
                   software would be able to read the new standard. The general
                   consensus seems to be that such a new standard would produce
                   a number of problems including compatibility, and over-large
                   files etc.

                   See http://unicode.org/ for further information on Unicode.

                   http://www.amazon.com/exec/obidos/ASIN/0201483459/joelnewkirk
                   (shameless plug for Joel's agent thing there) is a URL with
                   a book called "The Unicode Standard, Version 2.0",
                   ISBN 0201483459, cost $63.

                   See: http://www.cs.monash.edu.au/~jono/ISIS/Solo-bio.html
                   which is a URL detailing Ray Solomonoff and his invention of
                   Algorithmic Probability - and was involved with setting up
                   ASCII originally.

                   Scribe's note: Please see Paula Lieberman's (paal@gis.net)
                   post dated 16/02/99 "Re: Character Codes" which is 19k's
                   worth of excellent text relating to this thread and includes
                   numerous URL's like the one above - but which would double
                   the size of this summary if included here:P


g)

Subject: Distribution and precompilation

Summary of debate: Code that is generated and distributed could be either
                   precompiled on the host machine or compiled on-the-fly
                   depending on the CPU. Perhaps a basic form of this could be
                   built into KOSHv1 to see what the response to it is. This
                   would be tied into the ability of the software to be
                   migrated from one machine to another.

                   Such abilities would eventually allow a home network with
                   transparent resource sharing (including distributed
                   processing) and free and transparent migration across
                   machines.


h)

Subject: Uninterruptable Power Supply

Summary of debate: How about a 10 second UPS on KOSH machines - which should be
                   enough for at least a partial shutdown (including finishing
                   disk activity)?


i)

Subject: Software disk ejecting

Summary of debate: Suggested that software disk ejecting (a-la-Mac) could be
                   implemented to reduce the risk of disk corruption.
                   Alternatively make it so the manual disk eject button is
                   disabled until disk activity is finished. Both these ideas
                   could cause problems if the system crashed and locked out
                   the floppy drive (although a "paperclip" eject like Macs
                   could also be used).


j)

Subject: Delayed messaging

Summary of debate: Could include a delayed message send feature in KOSH so that
                   when a file/email/whatever is sent from one KOSH machine to
                   another on a LAN/WAN and the recipient machine is turned
                   off, it receives the "thing" the next time it is turned on.


k)

Subject: KOSH timekeeping

Summary of debate: Instead of an internal timeclock for KOSH machines on the
                   internet we could implement an online-timekeeping device
                   instead to maintain correct system time.


l)

Subject: Proposed Working Groups

Summary of debate: Joel Newkirk has proposed the following working groups:

                   Marketing - the what and how of KOSH production and sales.

                   Object Distribution - slim binary/ANDF, security, licencing,
                   anti-piracy, etc.

                   Nomenclature - document and as needed devise naming
                   conventions used within KOSH and to monitor usage of such
                   terms, etc.

                   Informational/Educational Resources - track expertise of
                   KOSHans, identify deficits in the community as a whole,
                   maintain resource references, etc.

                   Host OS Resources - subdivided into teams dealing
                   specifically with particular OSses to move towards being
                   able to host KOSH on such systems, etc.

                   Host Hardware Resources - would determine the level of
                   hardware support available within and outside of KOSH and
                   make attempts to increase this.

                   Industry Standards - Identify and assess various industry
                   standards available and their impact on KOSH.

                   Migration - research and propose a system of modems allowing
                   objects to be transferred from one KOSHer physical system to
                   another.

                   Object Model - (perhaps already covered in KOSH) to try to
                   develop an organic modem of /what/ an object needs to
                   encompass. Also emphasis on container objects for data
                   import from other platforms.

                   Scribe's note: Full details of the above that were proposed
                   can be found in an email by Joel Newkirk dated 21/02/99,
                   entitled "WorkingGroups..."