[SystemSafety] The Mailing List(s)

[Peter Bernard Ladkin]

Kevin, Paul,

yes, I only sent my comment on the Australian conference to the Bielefeld List. That is the only 
list which I am currently sure works as intended by the list maintainers.

Tim communicated the results of the poll to myself, Tom Anderson, Brian Jepson who maintains the 
SCSC on-line presence, and Zoe Squires of the IET.

Apparently he got about a quarter of the membership voting and the results were

> 100 in favour of the SCSC
> 88 in favour of (including those already transferred to) Bielefeld
> 4 for Professional Body (IET/BCS)
> 1 who wishes no further involvement in any list

He suggested that SCSC and Bielefeld could "consult" on how a "joint option" could be pursued. We 
already did that two months ago, of course. He didn't say anything about York providing some sort of 
continuing service.

To my mind, a "choice" between SCSC and Bielefeld is relevant only in terms of whom you trust to run 
a SW service. The technical details as discussed can obscure the actual maintainer. I think there is 
a feasible plan on the table which will get slowly but surely executed. See the next paragraph but 
one, and following.

The option of IET/BCS versus SCSC/Bielefeld was a viable choice.  It would be great if the IET 
professional society would again decide to support a community which it decided a couple of years 
ago to dump. It is indeed reconsidering, thanks in no small part to the efforts of the new Convenor 
of the IEC Maintenance Team for 61508 Part 3, Audrey Canning. As also the IET Information Technology 
Policy Panel, which includes three primarily safety people including the chair. And Carl Sandom, 
supported by Margaret Fanagan and Zoe Squires, who puts on the IET System Safety Conference each 
year. And John McDermid is active at the BCS. But I also think the professional societies could well 
ask themselves why 188 people to 4 don't trust them to run a mailing list.

So, the planning. We are thinking of registering a new domain name for use by the list and its 
archives. Bielefeld is in process of installing and maintaining archiving SW. We have such archiving 
SW, and we have a version which looks good. But it requires a few hours of time to install it 
properly on an open server according to the Bielefeld TechFak standards and at the moment they are 
up to their ears in work, as am I.

When we are sure how it all runs and what the level of effort is, we will likely register and switch 
to the new domain name and link the old list address to the new name. At that point it will not 
matter technically where the list and its archives are hosted: it is just an entry in a table at the 
nameservice provider to which the list administrator, whether at Bielefeld or SCSC Newcastle or York 
or Timbuctoo, has direct access.

The Bielefeld list includes quite a few people who were not on the York list, and indeed are not 
involved with the British safety scene in any way.

Apropos the IET System Safety Conference, I hope the videos and slides will soon be available on 
iet-tv. In contrast to the SCSC, it may well be that access will be restricted. I would find that a 
great shame. In particular because potential viewers will miss my appalling flute playing at the 
beginning as IET technicians were fixing their IT (which apparently goes to sleep when displaying 
PDF slides. Or maybe just my slides. Which may or may not be a sign of implicit good taste on its 
part) and also the splendid coda by the magnificent Mr. Lorne MacDougall in full regalia who 
completed my talk with a rendering of Gordon Duncan's splendid piece Pressed for Time. The IET could 
put that bit on YouTube.

My slides will be up on my WWW site soon.

I shall also have a longish piece about on-demand versus continuous system functions on the blog. I 
had intended it for this morning, but was waylaid. It turns out that some eminent system engineers, 
including Dave Parnas, Martyn Thomas and John McDermid, think the distinction is spurious. Whereas 
almost all mechanical and electrical engineers I have met, as well as myself, think it fundamental. 
How can that be? Well, the digital logic with which one implements either type of function is 
formally very similar. You need to poll some kind of a sensor for either, just at different rates, 
and react differentially to the sensor values that are returned. But not all systems follow digital 
logic. Indeed, most systems are not digital. Most systems are straight cause-and-effect. When the 
sun shines on your photovoltaic panels, you get electricity coming out. There is no sensor telling 
the system that the sun is shining. When a mouse runs past the sleeping rattlesnake, the rattlesnake 
just reacts and the mouse is gone. It's not as if the rattlesnake is continuously waking up to see 
if a mouse is running past. I think it inadvisable to conflate a system and its functions with the 
means of its implementation.

PBL

-- 
Prof. Peter Bernard Ladkin, Faculty of Technology, University of Bielefeld, 33594 Bielefeld, Germany
Tel+msg +49 (0)521 880 7319  www.rvs.uni-bielefeld.de