Neil Nelson wrote:

> Linux uses a UTC time that is the number of seconds since the

> beginning

> of 1970 at 0 degree longitude. Getting the local Windows time to be

> UTC time is done (after some investigation) in C as follows.



This reminds me of some of my attempts to make time (digits) more easy

to understand. I created a precentage measurer that took a day for the

measure and then one year for dates. A few days ago we were on 95% of

the year and that tells a lot more than 20th of December. =] There are 4

feilds: "11.2233444%" and they're coloured to give them names. However,

I soon realized that people other than maybe programers don't understand

the precentages and serial hexadecimal codes (for names) as I do.



> Harold has provided another version of the Flexhash and is working

> on an application to provide authentication and file control as

> in the case of marketed CDs and DVDs.  This would also apply to

> our system in that only authorized users would be able to use

> specific ICI components.



This would not be all that hard for me to make. In the current version I

have added an option in the Function Server to disable any specific

add-on with a simple click (but I *forgot* to write the code for it ;).

The same could be done on remote-control *now* and for specific users

with some slight upgrade of the code.



> That is we will have certain

> applications similar to what VeriSign does where we know who is

> using particular ICI components and that they can only do certain

> things. I obviously need to get my work done and out to both you and

> Harold



It's good to hear that the rest of you are finnishing your parts too. I

will be sure to provide more when I find the time for it (it's as simple

as ever to code for ICI now, but it still is slightly time consuming).



There is one thing I have to say now tho. ICI and it's add-on functions

are one layer on top of Internet, they automate some of what the

Internet offers. Now I'm thinking about layering something on top of

ICI, that would automate ICI procedures and function calling, which is

very likely to be required; currently you can get any specific piece of

information by calling an appropriate ICI function and then use it for

whatever you want, but likely a joe-type user will not be resourcefull

enough to see the combination of things he needs to do to get some more

advanced result (for example the number of people doing X now, will

require the number of people on this world and the statistics for X).



The new top layer would definetely be a program (a set of ICI function

calls), but the architecture of such a program would be in question

here. ICI is a paralell processing system, while the top layer is a

sequence of commands! I have ended up with two architectures: 1.

Worm/agent mixtrue and 2. Tree-shaped database crusher.



(1) Is a program that is executed on individual computers and is posibly

very efficient and surely very very imposible to trace and terminate by

force. It is an ordinary program, set of ICI function calls and variable

storing methods that contains a special module that can, on error (when

a function in the prog does not exist on the local computer) find an

appropriate host on the ICI network, transfer itself there with all the

variables and the program offset stored in the package, then very simply

run itself on the remote computer and order this computer to delete it's

instance here. When the program would preform it's currently given task,

it would find an appropriate host to wait on for more orders. 



Maybe it does not immediately seem so, but this I could start coding

such a prog now and possibly finnish it in about a week. If you're

interested making true in Artificial Life, this is the way to go.



(2) Is a parallel processing architecture and is likely to be very fast.

It's trick is that parallel sequences are written (depening on I/O data

dependencies) in a 2D tree and of course made of ICI function calls and

variable storing methods. Such a structure is executed by starting on

the ends of the tree's branches when function calls are sent to the ICI

network and when the results arive, they are automaticaly filled in.

Then when the required information has arrived, the next function call

down the branch can be made (and the just processed is "chrushed",

removed).



This is obviously a bit easier to manage, but not to me, since I have no

experience with tree structures and those included with VB are not easy

to use to me. There is also the question of the ease of coding up a 2D

program, tho schematics could help here. This is more related to ICI

than (1), but not that it even matters.



We will very likely be implenting both architectures, tho it is probably

obvious that both are pretty nice achievments; the idea of having

independend programs running up and down the Internet preforming given

tasks (1) or having a mass paralell processing machine on the home PC

(2), is perhaps slightly bizare, but we will have to see what cyberspace

is for sooner or later anyway. =]



--



Don't feel bad about asking/telling me anything, I will always gladly

reply.



Digging for info? Try AI Meta Search:

Http://WWW.AIMetaSearch.Com



MesonAI -- If nobody else wants to do it, why shouldn't we?(TM)

Http://WWW.MesonAI.Com

