KOSH [Kommunity Orientated Software Hardware]

Weekly Summary

Week Commencing: 27th March 1999

Number: 014

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: Cymraeg, Gaelic, Cornish and other local languages

Summary of debate: It would be inclusive and a positive step for the Kommunity
                   of Koshans to include support for languages that are
                   normally missed out in mainstream systems. This could be
                   done via the KOSH localisation system.

                   Suggested that we contact relevant local organisations
                   sometime soon about this.

                   There was (and still is??) an Amiga Translator's
                   organisation. Perhaps we could contact them to see if they
                   would do KOSH work too? - Does anyone have the contact
                   details? Suggested contact is Gary Peake who was apparently
                   in contact with them sometime earlier on this mailing list.


b)

Subject: Colour clash and dithering

Summary of debate: It was noted that certain colours on certain coloured
                   backgrounds (eg: red on blue, black on red) can be
                   difficult to view by a large number of people. Further,
                   dithering and patterns such as checks with text on top can
                   also present difficulties. Therefore in KOSH documentation
                   we should avoid such designs.

                   It was also suggested that where possible we so not use
                   dithering in KOSH as a whole, particularly as we are looking
                   at supporting new fast 24bit+ graphics cards (admittedly AGA
                   256 colour dithering will be needed). All of this obviously
                   only applies to full colour KOSH systems.


c)

Subject: KOSH documentation formats

Summary of debate: It was asked what sort of documentation KOSHans would like.

                   A printed manual or manuals would be liked as online
                   documentation is often difficult to use at the same time as
                   the software it refers to - both being on the same system.

                   As well as this, online (HTML?) documentation that
                   configures itself to represent the settings that the user
                   has setup (as mentioned in a previous summary).


d)

Subject: Functional languages

Summary of debate: It was suggested on the subject of what computer languages
                   to use with KOSH that a functional language that includes
                   complex computations could be used.

                   However functional languages are not very straightforward as
                   they often have bizarre syntax.


e)

Subject: Application generators

Summary of debate: Traditionally application generators are always tailored for
                   a specific type of application. However, as the UI in KOSH
                   will have personalities the applications for it will not
                   have to be constant. We could have a KOSH-Script builder
                   with another layer on top that changes with the personality.

                   An applications generator that produced KOSH scripts could
                   come in different versions, each specialized for one type of
                   work.


f)

Subject: More COMAL

Summary of debate: See:
                   http://www.macharsoft.demon.co.uk/unicomal/aboutcml.html
                   Macharsoft produces unicomal, a current COMAL implementation
                   which they claim incorporates OO concepts. NB: the whole
                   site is dedicated to COMAL.


g)

Subject: Structured summaries

Summary of debate: Marcus Peterson has gone through the summaries and
                   restructured them according to categories instead of summary
                   dates.

                   Having both date and subject-based summary formats available
                   will make locating particular information easier.


h)

Subject: Wider world survey

Summary of debate: Greg Webb, greg@gpwebb.freeserve.co.uk would like help with
                   conducting a survey of the wider world beyond us Koshans to
                   see what the computer world at large would like in something
                   like KOSH. If you can help (particularly but not exclusively
                   if you have experience with statistical analysis) then
                   please email Greg directly.


i)

Subject: KOSH programming and semicolons

Summary of debate: It was pointed out that C brings up an error when it is
                   missing a semicolon but then pinpoints the exact location
                   that the semicolon should have been placed at. Therefore it
                   was implied that KOSH should be able to automatically insert
                   such characters.


j)

Subject: One KOSH programming language for all / A KOSHan beginners language

Summary of debate: It was suggested that one global KOSH programming language
                   could be designed (possibly based on existing languages)
                   that was suitable for beginners to programming all the way
                   up to experts in OOP and structured programming.

                   An opinion was expressed against BASIC for something like
                   this in that it can make structured programming difficult.
                   However it was noted by another scribe-e that this may only
                   apply to some versions of BASIC.

                   Such a language could ship with KOSH and be used to
                   encourage people to program. However this may only teach them
                   how to program in the cosy well thought out world of KOSH
                   which could limit their ability to program with other
                   systems.

                   Was suggested that the critical features and design
                   decisions that must be made for a beginners language are:
                   1) Garbage collection - this may be too complex for a
                   beginner. The language must support garbage collection.
                   2) No traditional C-style pointers thereby eliminating a
                   possible memory leak, and making implementing garbage
                   collection easier.
                   3) Run-time array bounds checking or unbounded arrays to
                   protect against inaccurate writes to memory.

                   It was mentioned that eliminating pointers eliminates a lot
                   of power and convenience.

                   It was noted that Java almost fits these criteria except
                   that it doesn't actually support run-time bounds checking
                   but uses a theorem prover instead.

                   Suggested goals for a programming language for beginners
                   are that the system should be protected from accidental
                   damage, the language should be immediately effective with
                   something visible after only a short amount of work and the
                   language should help teach better habits, possibly by
                   requiring the syntax to be commented although this latter
                   suggestion may make the language wordy.

                   A language for KOSH should aim to reduce the gap between
                   having ideas and being able to implement them in a computer
                   language.


k)

Subject: ISO Modula-2 compiler

Summary of debate: Suggested that in the fullness of time we could look at
                   created a ISO Modula-2 compiler for KOSH.


l)

Subject: Installing KOSH

Summary of debate: Proposed that the KOSH v1 installation (on CD?) should
                   include a setup routine to install a base version of Linux
                   as well, with options available for a number of platforms.
                   This would allow those who own platforms on which a hosted
                   version of KOSH has yet to be written to use it, via the
                   Linux hosted version.

                   The complete install could be available as an FTP directory
                   located somewhere.


m)

Subject: Terminology WG

Summary of debate: Could anyone interested in joining the above named Working
                   Group for KOSH please email Ruward F. Leenstra at
                   uart@xs4all.nl


n)

Subject: Loading and coping with older versions of objects

Summary of debate: The following questions were posed: what happens if you have
                   have software which has a number of objects that contain
                   data and when the next version is released, these objects
                   have changed? How should this be handled?  How would older
                   versions of objects be loaded?

                   Old versions should be loadable. It should be possible to
                   just plug the old objects into the new handles or use an
                   adaptor for this.


o)

Subject: Sharing computer experiences

Summary of debate: A way to share experiences of the various computer systems
                   that we all use (including Mac, BeOS, Amiga, Linux, Unix,
                   BSD, NeXTStep, Windows, OS/2 and many others) could be
                   useful. This could take the form of a large section on the
                   web site available (to members only it was suggested) for
                   downloading in parts that would contain emulators, trial
                   versions and other relevant software to allow people to try
                   other systems.

                   The same area could also contain languages like COMAL and
                   Delphi that people could try out.

                   Jason Radford the administrator for the server that the KOSH
                   web site is on agreed that some of the 4gig of drive space
                   available for KOSH could be used for this purpose. He
                   suggested that if enough interest and ideas came forward
                   that ftp.kosh.net could be created.


p)

Subject: Case-insensitive KOSH

Summary of debate: Suggested that throughout KOSH an Amiga-style system where
                   case is stored but operations are case-insensitive is used.


q)

Subject: Standardized diagram construction package

Summary of debate: It was suggested that we try to find a standardized
                   package to construct diagrams for KOSH. A set of brushes
                   that could be used would be useful.


r)

Subject: Executive Scheduler

Summary of debate: KOSH could implement a scheduler similar to the Amiga's
                   scheduler.

                   Further suggested that this should be a separate object. The
                   basic system could ship with a scheduler that does the basic
                   round-robin scheduling.