RE: [PROJECT] Project core

["Broadfoot, Kieran J" <Kieran.Broadfoot@xxxxxxx>]

Is there anyone in particular I need to ask to join the UKHA sourceforge
group?  My id is broadk

Anyways back to the important things.

John,
I agree that having a meet is not a bad thing but I would prefer to
continue
this work online purely because it allows more voices to be heard.  I
wouldnt want to see decisions made at an early stage by a smaller group of
people.  Id rather see the internet used for what it does best which is
produce some excellent code built by groups of people across the globe.

Back to the XML thing.  I was actually thinking that the only device that
has to understand XML in the sense of being able to encode/decode it is the
management console which to be honest is probably going to be PC based to
start with.  Given we would byte encode the XML each device only needs to
be
able to talk binary in effect.  I would like to see us design and implement
a control protocol which would allow the management software to determine
how to use a device once its arrived on the network.  In other words the
device can actually inform the console how to operate it.  The advantages
of
auto discovery are numerous especially for systems used by the less
technical amongst us (me included!)

Here at work I use XML a lot to define a data structure which I can then
use.  In some cases I dont actually use XML for message passing or anything
complex like that but rather just as a tool to help me think about data
structures.  Its a good definition language.

So to formalise my thoughts a little more:

Management console:

Java based - ensures both wintel and linux crowds are both kept happy.
manages dhcp leases and network topology (i.e we need NAT and private IPs
etc)

Handles all message passing between devices hence all devices only need to
communicate with the console (although we should build it so you could have
a hot spare console in case we can run this software on a much smaller
hardware platform (TINI etc).

My thinking is that a device would power up and request an IP address.
Every ten/fifteen seconds the console sends a discovery packet to the
network.  Any new device can respond to this with details regarding its
type/operation etc.  The console keeps track of this new device and can
then
compare this to the user defined functionality of said devices.  If matches
the console then starts to operate the device as described by the user.  If
the device has not been configured for use its very easy for the console to
determine the functions of the device and present this in a graphical form
to the user.

And yes I am waffling.. its still early.  Anyone have any thoughts on this?

Thanks
Kieran


-----Original Message-----
From: Dr John Tankard [mailto:john@xxxxxxx]
Sent: Monday, May 21, 2001 8:36 AM
To: ukha_d@xxxxxxx
Subject: Re: [ukha_d] [PROJECT] Project core

Yep I agree, talk is cheap producing a list of ideas is the easy part. I
wonder if Mark H's idea of a face to face meet would be the next step a
lot of the groundwork could be done face to face. People with development
kit could bring along the portable bits so others who want to do code can
see the various options. I know getting a date fixed up might be a
problem but it is worth a try.

Back to the underlying compatability of devices, I am worried about the
XML route, mainly because I dont know much about it I might be in the
minority.

If various projects spring up some will be PIC some TINI perhaps some
rabbit or PC based, peoples skills might be low level or high level, I
can see that we are going to have to produce code for lots of platforms.
Are the prople with XML skils going to be happy coding for platforms they
have no intention of using ?

Another approach might be to settle on a bace platform, get the
underlying coms part working and then  split into groups working on
actual projects. We also come to the cost to support the various
platforms are we prepaird to share equipment, it seams daft if we all
rush out and buy ICEPIC's when perhaps people with them could work
removing difficult bugs that can only be found useing ICE.  Also If se
settle on a platform we cousl design standard bace boards and get them
made, of a team member could produce them, its takes much less time if
one member etches 25 cards rather than 25 members each etch different
cards.

Just my 2p

John

John





Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/




Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/