"Nines" Problem in Railway Monitoring

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

A reader just sent me this URL:

http://www.iee.org.uk/2000risk/Casebook/eg-27.htm

I hope this isn't old news here.

-- Lane Core Jr. (elcore@sgi.net), June 27, 1999

Answers

Lane,

Well, this looks like one of those "dreaded nines" that has been verified to be waiting to fail. I had not thought there were very many problems that would arise due to the nines situation. But here is one. In fact it looks like it can't be programed out of the device. Could you comment any further on why the date of 09/09/99 and the end of file code of 999 will match up in this case? Also, and this is pure speculation, I note that the DoD has put the military "on alert" for callout from Sept 1, 1999 thru Mar 31, 2000. I have not seen any reason stated for the Sept 1, date, which seems early. Now, I wonder if they think there are going to be some dreaded nines problems in their own dinosaur legacy codes on Sept 9th?

-- Gordon (gpconnolly@aol.com), June 27, 1999.


Hey Lane and all--

So many tidbits and soooo little time. I came across this yesterday, but fell off the chair into zzzz's. Can anyone take a look here, and give an enlightened response to the left column of numbers? Particularly those with no following matrix numbers associated with the column marker.

gopher://solar.sec.noaa.gov:70/00/forecasts/GEOA/0627GEOA

June 27th will work, but so then will all I looked at.

Thanks in advance

PS. note the producer of said matrix, columnular report or whatever.

-- zMichael (mikeymac@uswest.net), June 27, 1999.


Before computers conquered the world, every railroad had one or more people standing by the tracks in a location where train speed is slow, listening to the sound of the wheels as the train passed by. A flat wheel makes a very distinctive noise. Any flat wheel, or any wheel where the brake shoe was dragging on the wheel, got a splash of paint to indicate that it needed attention. Eventually any car with a marked wheel would be cut out and sent to the shop.

No reason this can't be done now. It means hiring more people, and it will slow things down a bit. Or maybe more than a bit.

-- Tom Carey (tomcarey@mindspring.com), June 28, 1999.


zMichael (mikeymac@uswest.net),

Go to http://solar.sec.noaa.gov/Data/Reports.html -- the page title is "Reports and Special Products". Look about two-thirds of the way down on the left for "GEOALERT" and click on that, which will take you to gopher://sec.noaa.gov:70/11/forecasts/GEOA -- titled "Gopher Menu".

At Gopher Menu, the first link listed -- "About this Menu" -- will take you to a description of the format of the GEOALERT messages such as the June 27th one you reference above. However, this description does not include the meanings of the numeric codes. Apparently those are "described in detail in the manual I.U.W.D.S. Synoptic Codes For Solar And Geophysical Data, written by SESC", which I will leave for you to find if you wish. ;-> >the left column of numbers? Particularly those with no following matrix numbers associated with the column marker.

If you're referring to the "99999" numbers, I don't think there's any connection between those and a 9/9/99 date. I don't see anything in the pattern that indicates that the "99999" is going to be compared to a date, and so it seems to me that there's no danger of a glitch on September 9.

There are numbers such as "90627" which do seem to represent a date, but notice that they are _only_ in a fixed position (2nd following number) relative to the five-letter code word that precedes them. The data that precedes the "99999" values almost surely does not represent dates, so there wouldn't be a mixup between "99999" and "9/9/99" because the receiver of the data would not be treating any of that part of the number sequence as a date.

June 27th will work, but so then will all I looked at.

Thanks in advance

PS. note the producer of said matrix, columnular report or whatever.

-- zMichael (mikeymac@uswest.net), June 27, 1999.

-- No Spam Please (nos_pam_please@hotmail.com), June 28, 1999.


My apologies for the careless editing. The last part of my preceding message should have looked like this:

Apparently those are "described in detail in the manual I.U.W.D.S. Synoptic Codes For Solar And Geophysical Data, written by SESC", which I will leave for you to find if you wish. ;->

>the left column of numbers? Particularly those with no following matrix numbers associated with the column marker.

If you're referring to the "99999" numbers, I don't think there's any connection between those and a 9/9/99 date. I don't see anything in the pattern that indicates that the "99999" is going to be compared to a date, and so it seems to me that there's no danger of a glitch on September 9.

There are numbers such as "90627" which do seem to represent a date, but notice that they are _only_ in a fixed position (2nd following number) relative to the five-letter code word that precedes them. The data that precedes the "99999" values almost surely does not represent dates, so there wouldn't be a mixup between "99999" and "9/9/99" because the receiver of the data would not be treating any of that part of the number sequence as a date.

-- No Spam Please (nos_pam_please@hotmail.com), June 28, 1999.



Tom,

Hence in England we used to and still do have working mens' clubs (cheap beer) one of which still going strong is "the Wheeltappers And Shunters Social Club" - mainly found oop north...

-- Andy (2000EOD@prodigy.net), June 28, 1999.


Come to think of it this would be a perfect first job for y2k pro... nah, he'd screw up - fuggeddabadit

-- Andy (2000EOD@prodigy.net), June 28, 1999.

Gordon,

I am assuming that either (1) three nines is a typo or (2) the problem will actually exhibit itself on September 1, when the month/year becomes 9/99.

Or, they don't know what they're talking about. That's always a possibility. :-(

-- Lane Core Jr. (elcore@sgi.net), June 28, 1999.


NO spam--

I had supposed that it may have been an actual visual sighting of the *end of file markers* 99999 often referred to these days. There seems to be an appropriate repetition that could signify such. I had not actually seen this used as an end-of-file marker before, but I think this is a real case of this phenom. Have another look.

-- Michael (mikeymac@uswest.net), June 28, 1999.


Michael,

>I had supposed that it may have been an actual visual sighting of the *end of file markers* 99999 often referred to these days.

Oh, I see what you mean. Yes. Since I'm used to seeing all-nines used as EOF markers before in such contexts, it didn't occur to me to remark on that aspect.

>I had not actually seen this used as an end-of-file marker before, but I think this is a real case of this phenom.

Yes, I agree. The "99999" numbers seem very likely to me to be EOF markers.

And, it also seems likely to me that there is no chance of confusion between the "99999" EOF marker and a calendar date in this context. The instances of "90627" indicate to me that the only calendar date representations as 5-digit numbers are in the format YMMDD. There would never be a match to a real calendar date with either MM = "99" or DD = "99". Furthermore, the calendar date number seems always to be in a fixed location in the file, and thus would not be scanned for an EOF marker as the later, variable-length data would be.

-- No Spam Please (nos_pam_please@hotmail.com), June 28, 1999.



Moderation questions? read the FAQ