Rick, Do you have any idea how common PDP-11 systems are in the electrical industry? I noticed that the N.H. survey listed their RDMS as non-compliant and being PDP-11 based. The PDP-11 OS is non-compliant and I don't believe it can be made compliant short of substantially re-writing the operating system. If this is a common system in the plants thats not a good sign.

-- Anonymous, November 26, 1998

Answers

PDP-11's and their younger siblings the DEC VAX's were the "original" embedded system core. They are very prevalent. I would imagine that any moderately sophisticated automation system that was installed between the early seventies and the mid eighties would incorporate at least one.

--AJ

-- Anonymous, November 27, 1998

PDP-11's are pretty common in the electric industry. Without significant modifications, PDP-11's and the RSX operating system will not support Year 2000 transition. As far as I'm aware, there's only one company supporting Y2k upgrades for PDP-11's:

http://www.mentec.com/Y2K/Y2K_MAP.htm

There are a few Digital folks monitoring this forum; hopefully, one of them will pipe in with a direct answer to your question.

-- Anonymous, November 27, 1998

This is a very common system used for many different applications. Per below URL, about 600,000 of them are (or have been) used worldwide. I wouldn't be surprised if 400K are still running (cheap, reliable, fast). Anyway, its operating system, RSX, is non-compliant. A single company, Mentec, still supports RSX and does offer operating system updates. I wonder if a single small company can support 400K users? Anyway, their methodology is clumsy at best. Below is how they represent post y2k dates.

See http://www.mentec.com/Y2K/Y2K_MAP.htm

The two bytes will be encoded as follows: Low byte-Low order decimal digit of year, in ASCII. High byte-Quotient of years since 1900 divided by 10, plus 608. This represents the ASCII high digit of the year from 1900 - 1999. In 2000, however, this will result in ":" being stored for the high digit of the year. For years 2010 through 2019, ";" will be stored, and so forth. This change should result in another 200 years or more of file representations Table 1-1 Storage of Dates in RSX Systems after 1990

Calendar Year Year Field Representation 1900-1999) 00 - 99 ( 2000-2009) :0 - :9 ( 2010-2019) ;0 - ;9 ( 2020-2029) <0 - <9 ( 2030-2039) =0 - =9 ( 2040-2049) >0 - >9 ( 2050-2059) ?0 - ?9

Sounds great Huh? Except, how many application packages can understand this date representation? Answer - none, unless you wrote it that way very recently. The PDP-11 often uses archaic application packages sold by vendors that are now so much cyber history dust. In many of its real-time applications, I consider the PDP-11 part of the embedded system of systems. As a company, unless you are willing to stand in a long, long line awaiting operating system updates from Mentec AND you are willing to rewrite your application code, your PDP-11, and whatever application it is controlling, is toast. Why should you non-codeheads care? This is a firm example in the embedded/real-time control area that you can point at and say THIS WILL FAIL. You should be asking your neighborhood refinery, power plant, pipeline control station, water purification plant, etc. do you have a PDP-11?

-- Anonymous, November 28, 1998

We have a PDP-11 based building energy management system that the Y2K issue provided the necessary "incentive" to fund a replacement project. Annual maintenance is getting expensive on these systems also.

VAX/VMS systems need to move up to VMS 7.1 or higher.

Jim

-- Anonymous, November 30, 1998

Not just PDP-11's. SEL(Systems Engineering Labs), later Gould then Encore Series 32 and Concept 32 computers are fairly common in scientific, industrial, military, system simulation and power production and control. And as far back as 1987 I found out the hard way that the RTC's are incapable of making the century transition. I found out as we were testing a military flight simulator which used a brace of Concept 32 machines. Our software was specified to use a four digit date and I did the old "set the clock forward" test to see. All of the cpu's froze the instant the clock rolled over. There was no recovery procedure which worked, they were locked-up and that was it. It was my first introduction to hardware having a Y2K problem. It took installing new RTC boards to get back up and running.

At the same time we were testing our simulator, there was a nuclear power plant simulator undergoing test, and they used those same model computers. Based on my experiance, I'd be willing to bet that ninety percent of nuclear power plant simulators use these machines. Probably many of the fossil fired plant simulators do too. Through exposure to other users I've also become aware that many remote data monitoring and control systems for nuclear facilities had this hardware installed.

If the PDP-11 problem is limited to an OS problem, then it's at least a workable fix. If problems get into a fatal hardware flaw which causes a computer to lock-up in a non-recoverable halt, then things are really in trouble.

Check six! WW

-- Anonymous, December 06, 1998

While at work today, something I learned a few years back occurred to me. Several different manufacturer's late-eighties vintage graphics display systems (my experience is Sanders Graphics) are based on PDP- 11 emulations, to the point of running the PDP-11 instruction set. As the PDP-11's OS is problematic, so too may be many graphics display systems currently installed in operations and control systems.

What good is a working instrumentation and control system if the displays are dead?

WW

-- Anonymous, December 07, 1998