Neil Nelson wrote:

> 

> Dear Jure,

> 

> You mentioned yesterday that the Console has messages for

> disconnecting and I was looking through the messages and

> trying to see what I could do with them along this line.

> 

> The Console sends the following message to the icsa when

> 'Remove node' is used.

> 

> --- Entry by CFP Default on 07-09-2002 03:07:54

> DisconnectName Win_Susan

[Snip!]

> And so it would seem we would want to have an ability to

> use DisconnectName without losing the icon, otherwise we

> could not reconnect.

[Snip!]

> And so finally, we seem to have four somewhat separate

> goals:

> 

> (1) Remove a user from the Console-icsa files so that

> connections are not subsequently made to that user.

> 

> (2) Synchronize the list of users between the Console

> and the icsa.

> 

> (3) Disconnect from a user with likely an ability to

> reconnect during the same session.

> 

> (4) Disconnect from all connected users when the

> Console and icsa are closed.

> 

> The last seems to be substantially working.  Goal

> (2) seems like an important near-term goal and that

> would seem to be accomplished by using the

> 

> --- Entry by ICSA on 07-09-2002 03:34:22

> Shell: ID,Win_Robert

> 

> messages provided to the Console when the icsa starts.

> The icsa is able to do (3) but the Console seems to

> tying that action with (1).  And then there is no icsa

> command available to remove a user name from the

> names.dat file though it would be easy to provide one.

> 

> This seems to be an area you are concerned with and I

> have provided the above to get myself coordinated with

> what you have in mind.



Yes, (2) is exactly what I had in mind. 



The Console does not care what users exist, only what are online and of

course display only those, since the others are unreachable anyway and

would only fool the user and other programs in thinking they can send

them messages. Therefore, when a node goes offline, it transmits a

remove message and the record is automaticaly removed from the

databanks. And when a node goes online, it automaticaly sends some

synchronization messages, to which all Consoles respond by automaticaly

adding the node's records to their databanks and of course respond with

the appropriate info (actualy ANY message from a specific node will

cause it to be added). 



Therefore, if the local node was on all the time, it would record all

the logons and logoffs and therefore keep a perfect record of the status

of all nodes.



Of course here is the most obvious problem that it is a general issue

that the node is not online all the time, otherwise the entire procedure

would be pointless. Therefore the ICSA should help the local Console get

into sync with the actual online/offline status. For just as obvious

reasons, this would only need to happen once, on startup, the rest would

be tracked by the Console itself.



This is why we agreed upon a vocabulary. This very same vocabulary is

being currently used in my new attempts, where we have connection

interfaces other than the ICSA possible. [My last idea was that the NLU

we are working on, would best work if it considered it's user(s) as

another (set of) node(s) and itself as another connection interface, so

I have re-tested the vocabulary a bit, it works great!]



The correct message format is:



	--- Entry by ICSA on 07-09-2002 03:34:22

	SET SOURCE = 0.0.0.0

	DO Shell: ID,Win_Robert



...which was the problem in your attempts.



And I expect this is all there is to it from the ICSA's side. I have

just finnished up an interface module for the Console, written in VB,

you might want to know what all is in it, because all of that is

considered required to set up a connection over a connection program

like ICSA.



C'ya!



--

Cellphone: 0038640809676 (SMS enabled)



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

reply.

[AC/HFA(AS) -- no suprize]



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

