KOSH [Kommunity Orientated Software Hardware] Weekly Summary Week Commencing: 02 January 1999 Number: 002 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 summaries@kosh.convergence.org, stating the Summary Number, and Mailing List Name, and he will try to answer your queries. a) Subject: Applications for KOSH Summary of debate: It will be hard to get people to move from what they are used to (Win9x, MacOS, AmigaOS etc) to a new platform/environment as they will need a good reason to do so. Killer applications can help to shift people. Even with KOSH specialist applications (eg: Photoshop type app) we may need specialist programmers, therefore there is only a point to which an OS like the KOSH-OS can simplify your work. Hopefully with KOSH pogrammers we can forget about the OS when we have it and concentrate on producing the applications. As KOSH will have the ability to run above existing OSs people may become interested in becoming KOSH programmers as they will not have to buy lots of new hardware or software. KOSH could be caught in a circle: a decent system but no apps-> not much interest from those wanting killer apps -> killer apps not developed because they cannot be sold-> and so on... Therefore KOSH needs to provide standard apps preinstalled, Therefore it will start as a niche system and hopefully grow from this. b) Subject: KOSH making money Summary of debate: No one should expect KOSH to make money for a couple of years. But it may well be attractive for developers by then. c) Subject: Hosting KOSH Summary of debate: Using Linux as a base for the moment would allow people to run KOSH on most current systems. By hosting KOSH on Linux, support for many processors can occur - x86, Alpha, PPC, 68K etc. It appears that KOSH may well be at the mercy of the host system if the host crashes; but when there are independent KOSH machines these will be OK. Running KOSH over windows may annoy as many people as are intrigued by it due to the reaction that "KOSH crashed my machine", ignoring the part that windows may have played in the crash... By showing that KOSH on its own is stable we may get some people to -consider- a move away from windows. However, running KOSH on win32 may be a no-starter as the 2 concepts are completely different. d) Subject: Demonstration versions of KOSH Summary of debate: A layer version as a free demo could lead to a native demo (no underlying host OS); then people could download a larger selection of objects to become a full system. KOSH must also be able to interact with other OSs transparently to their users over a network. A KOSH demo should start and end like an application, and clear out of memory when you exit. Initially as an "application" KOSH could use DirectX or some other method to access the existing device drivers, later on we can write our own. Maybe KOSH could stop multitasking and take control of the whole computer when necessary like with the Amiga, and could use such a feature with its host OS when an application needs to be realtime. The minimum installation for KOSH should work on 80x25 text terminals as it costs nothing to the project to support this and adds to potential uses for KOSH. A version of KOSH enabling the minimisation of windows to allow stability and productivity during use would be useful. e) Subject: A GUI for KOSH Summary of debate: If the original KOSH GUI is well made using components, then altering components and/or adding to them to create the GUI that YOU want will be easier. f) Subject: Scripts for KOSH Summary of debate: A graphical scripting language is suggested. Further, the graphical scripting should be refinable via a text editor, and a text-written script should be refinable in a graphical editor. But with KOSH both such physical implementations would be plugins upon a logical "user" interface. Graphical scripting would also allow pan and zoom to be used to look around the script. g) Subject: Macro recording in KOSH Summary of debate: KOSH should implement some form of macro recording - not just mouse movements and clicks, but which buttons were clicked in which windows etc. h) Subject: Standards for developers Summary of debate: As KOSH is a kommunity it may be very much dependant on its developers. We need some standards which can be used by developers (the FDE WG is on this) and a developer support plan defined by the developers. i) Subject: Living with KOSH Summary of debate: The KOSH "mentality" should eventually enable people to work from home living how they want to, developing objects and selling them at object "fairs", ie: computers should enable the enjoyment of life, not hinder it. j) Subject: Learning with KOSH Summary of debate: People do need to experiment with KOSH so they will know what an object sea is, multitasking and how to manipulate UIs to make your computer do what you want it to. k) Subject: Shutting down KOSH (continued) Summary of Debate: Shutdown can just be a quick check to finish up any disk activity and put the system into a safe state ready for power down. KOSH should fully support power-management features available on ATX boxes. Auto power on when sensing a modem or LAN activity would be handy. Plus power save option should be configurable. There should bean option to auto-save all important things every mins to reduce losses from a crash should be a config option. Perhaps KOSH could have a low power mode for backups or compiling large programs over night etc. that would allow the machine to use as much power as it needed for the tasks. Alternatively after an over-night task has been completed, the system auto powers down. But a low power mode surely gets that way by shutting down resources and therefore may not work when apps are running. Additionally, once powered down it takes time to power up - what is an acceptable time? ->1min, 1s? 1ms? x? The specs for ATX are at http://www.teleport.com/~atx/ - it is not pc-only and could be used for KOSH. l) Subject: Squeaking with KOSH Summary of debate: Squeak at http://squeak.cs.uiuc.edu/ which is a SmallTalk based pure object system is given a plug as an example of one way that a KOSH environment could work. m) Subject: Reducing development costs Summary of debate: It has been suggested that perhaps we should not go for custom hardware for KOSH - the Wintel-world people do all the R&D with the costs and this can be passed on to others (like KOSH). As the hardware gets better and better we won't therefore fall behind. n) Subject: Virtual workspace Summary of Debate: The BT Virtual office/desk where a user on their chair becomes part of their working environment (using VR etc) could be one implementation of KOSH. URL: www.labs.bt.com may be for this. o) Subject: Virtual networking screens Summary of debate: A beta program called "Anywhere" allows one TCP networked computer to display a remote computers screen on your local screen - perhaps this could be ported to KOSH eventually. http://www.orl.co.uk/vnc for details. p) Subject: QNX distributed processing Summary of debate: It has been suggested that the QNX approach to distributed processing and networking could be used in KOSH. q) Subject: Expert Systems (continued) Summary of Debate: Suggested not to go down the path of AI and Expert systems at the moment as some past examples of LISP machines have disappeared, along with the companies running them. r) Subject: God could be KOSHan Summary of Debate: More support for the idea of some image of God to represent KOSH (although using a computer wearing sunglasses as an icon for a deity seems curious). s) Subject: Merchandising (continued) Summary of Debate: Perhaps Eric Schwartz could be asked to do an animation for KOSH - seems to be the consensus is that we would love to get him involved. t) Subject: "Alternative users" Summary of Debate: KOSH could be targeted at -amongst others- "New Age People" who may not like huge corporations like M$, to quote: "New AGE people are a virgin market for computer corporations." u) Subject: KOSH on the www (continued) Summary of debate: A KOSH "Pic 'n mix" where you can buy the exact objects that you want online for a small amount of money, therefore allowing you to buy EXACTLY what you want would be useful. Further to this could be "KOR" - KOSH Object Repository offering a central site for obtaining objects that all KOSH programmers could have access to (with money split between them and KOSH). v) Subject: Summaries Summary of debate: A general-summary ML for those without time to read all the emails in the main ML's would be useful. w) Subject: Thanks Summary of debate: Fleecy thanked Ben and Ben thanked Fleecy. x) Subject: Reference Counting (continued) Summary of debate: A possible alternative implementation for KOSH that does not rely on Freezing/Unfreezing objects (which could apparently create problems for programmers who may find themselves tied to 1 set of H/W rather than being H/W-independent which is partly the KOSH way) is to use something like the forth style to handle storage. y) Subject: Optimisation Summary of debate: KOSH should be very highly optimised unlike -certain- OSs + apps. Therefore ppl who have limited resources can get into it. z) Subject: Thoughts Summary of debate: Quote: "The object sea metaphor ensure(s) that there will be plenty to explore for users who are fascinated by computers themselves." However, KOSH must be aimed BOTH at such people and also those who do not want to (for whatever reason) tinker with the system. A stable OS that can be used "out of the box" and that after some experience with it allows unlimited configuration (without this getting in the way of being productive on the computer) seems to be the way to go. a1) Subject: Developing the KOSH concept Summary of debate: A multi-prong approach to KOSH may be the way to go: 1 "team" could handle HW/host abstraction, one could handle each prospective environment and the latter handling 'native KOSH', with a timeline 1-2 years beyond the primary target of the others.** People should be free to swim in the object sea and dive deeper into its workings when they feel ready, but KOSH should enable this in such a way that they do not drown simply because they overestimated their ability to swim. b1) Subject: Networks Summary of debate: There could be a tremendous opportunity with LANS that KOSH should not miss. If KOSH can successfully incorporate transparent processing, office LANs can be a target. KOSH could then utilise the power of all machines on a network reducing the need to upgrade the entire office to new cpu speeds+memory. This idea can be used in the home where ppl often have more than one machine due to the wintel "need" to upgrade. c1) Subject: Web-casting exploration Summary of debate: http://tech-head.com/amiga.htm - Anyone interested in web-casting exploration should read this (appears suggestion is everyone should read this). d1) Subject: Digital ARTS WG Summary of debate: Previously suggested was a WG exploring Digital Art/cinema/multimedia aspects (technical content), also incl MIDI. More support for this shown, and Greenboy and Dave Haynie have a "midi shopping list" together that should prove useful. e1) Subject: Employee Ownership of KOSH Summary of debate: Jeff Gates, an expert on ESOP (Employee Ownership) has been contacted about the KOSH project and he is interested in helping us with some advice. As he is busy we need to create a suitable document for him to have a look at. A WG is needed for this - volunteers please? f1) Subject: Which platforms do KOSHans use? Summary of debate: A request to find out the numbers of people using various platforms who are involved with KOSH has been posted. Stats on this would allow us to see what platforms have not got informed of KOSH.