1. Home
  2. Electronics Design
  3. OT: An expensive software error, year 2010 bug: 240 Million Euro!

OT: An expensive software error, year 2010 bug: 240 Million Euro!

The French manufacturer Gemalto confesses it made an error in the software of the eurocard money cards. This caused an error on 1 Jan 2010, resulting in the cards being rejected by some terminals. and 30 million people could no longer draw money or pay with the cards. A temporary solution is in progress, downgrading the terminals so the cards are no longer addressed via the secure "EMV system", but with the old insecure "electronic cash ecc" system, or the magnet stripe based "electronic cash Spur 2" system.

At 8 Euro per card, this makes for 240 million Euro damage if all cards are replaced. It is not clear who will pay for it. Shares of Gemalto dropped 4% today. Source in German:

formatting link

Reply to
Jan Panteltje
Loading thread data ...

On a sunny day (Wed, 06 Jan 2010 15:13:28 GMT) it happened Jan Panteltje wrote in :

some terminals.

are no longer

cash ecc" system,

replaced.

formatting link

Was not John Larkin in France recently?

Reply to
Obe

some terminals.

are no longer

cash ecc" system,

replaced.

formatting link

I'll have to post pics of my new private jet, helicopters, and yachts.

John

Reply to
John Larkin

some terminals.

Is this anyway different from the 1990 year problem ?

One of my customers used the YWWD format (Y=last digit of year, WW=week number and D=1..7 day of week) to schedule work in a large mechanical manufacturing plant.

The system had been originally created in the early 1980s on a mainframe and ported to a minicomputer and fortunately the problem was spotted in the late 1989.

Fixed that problem with some significant work, but I announced in 1989 that the year 2000 problem is going to be a much larger mess with other 1990/2000 issues.

Reply to
Paul Keinanen

some terminals.

I've adopted time-stamping my designs with YYYY_DD_MM_hhmm.

Makes it a whole lot easier to track revisions which, in a typical chip design, may number as much as 100, when you have a customer with wandering thoughts ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

On a sunny day (Wed, 06 Jan 2010 18:35:44 +0200) it happened Paul Keinanen wrote in :

some terminals.

I am waiting for 2038 when Unix time_t overflows... , from libc.info:

- Data Type: time_t This is the data type used to represent simple time. Sometimes, it also represents an elapsed time. When interpreted as a calendar time value, it represents the number of seconds elapsed since 00:00:00 on January 1, 1970, Coordinated Universal Time. (This calendar time is sometimes referred to as the "epoch".) POSIX requires that this count not include leap seconds, but on some systems this count includes leap seconds if you set `TZ' to certain values (*note TZ Variable::).

It was originally a *signed* 32 bit integer.

Nice here:

formatting link

The minimum representable time is 1901-12-13, and the maximum representable time is 2038-01-19. At 03:14:07 UTC 2038-01-19 this representation overflows.

See also Year_2038_problem:

formatting link

I think time_t is now 64 bits, but I am sure a zillion 32 bit applications are still around..

Perhaps when Unix was invented nobody anticipated it would stay around so long, amazing feat in itself.

Reply to
Jan Panteltje

Jan Panteltje Inscribed thus:

I don't think it matters ! In 28 years time computers could just about have become implants.

--
Best Regards:
                     Baron.
Reply to
Baron

On a sunny day (Wed, 06 Jan 2010 20:15:30 +0000) it happened Baron wrote in :

Then it would matter even more, your heart could stop at 2038.

Reply to
Jan Panteltje

I would be surprised if some industrial or infrastructure systems delivered today, would _not_ be used in 2038.

Looking backwards 28 years (1982), I would not be surprised, if some of my systems are still in use. Some systems will have a quite long life time.

Regarding long life systems, some hydroelectric generators had their century overhaul just a few years ago :-).

Anyway, I plan to be retired well before 2038 and in any case, I would be dead by 2100 or 2106, when the next problems occur.

Reply to
Paul Keinanen

Lemme see, will I make it to 98 (2038) ?:-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Jan Panteltje Inscribed thus:

True ! If I'm still around by then. ;-)

--
Best Regards:
                     Baron.
Reply to
Baron
[...]

Last summer I overhauled a 5 HP 'Century' repulsion/induction organ blower motor made in 1919.

The brushgear was adjustable to take up the slack each time the (disc) commutator was skimmed - but until this year it had never needed adjustment. It wouldn't have needed adjustment this time if the motor hadn't been thrashed to death by an incorrect replacement spring which forced it to run on its starting circuit until the brushes wore away and arced continuously, badly pitting the commutator.

The next major overhaul will be due in 2099 if it is not looked after properly. Otherwise we could safely delay it a bit longer.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

I'm curious about why you chose a mixed endian YYYY DD MM instead of the big endian YYYY MM DD, as suggested by ISO 8601.

formatting link

Regards, Allan

Reply to
Allan Herriman

I like the old DEC convention

06-Jan-2010

except they used all caps.

John

Reply to
John Larkin

some terminals.

Yes.

It's one thing to write software in the 70s with the expectation that it will be obsolete before the end of the century, and quite another to write software that will break a few years hence.

Sylvia.

Reply to
Sylvia Else

some terminals.

are no longer

cash ecc" system,

replaced.

formatting link

This same glitch appeared in other places than Europe, and seemingly for other card types. Media reports - for what they are worth - stated that cards (either MC or Visa, I forget) ex the National Australia Bank had widespread issues in Oz with the terminals, not the cards. Report indicated the *terminals* rejected cards as the date was deemed to be 2016 rather than 2010, so the cards were deemed expired.

Initially I wondered whether this was a hex coding issue.

Reply to
who where

The date representations should be such that they can be easily sorted in chronological order. The ISO convention YYYYMMDD works well. I have adopted a serial number scheme that used a two-digit year followed by a numerical sequence, such as 890001 and 990099, but at Y2K I opted to use a hexadecimal form of the year so that the first s/n in 2000 was A00001. Now it changes to B0001. I don't think I'll be around to deal with what happens after Z90001. That will occur in the year 2259. I suppose one could continue on the ASCII chart to [00001. Not my problem.

I like the Delphi TDateTimeField representation which is a real number with the integer portion as the number of days from 31Dec1899 to the date represented, and the fractional portion is the fraction of a 24 hour day. It works for just about any date and time from dtBigBang to dtEndOfDays.

Paul

Reply to
Paul E. Schoen

some terminals.

Why do you use the day in front of the month? It makes no sense.

--
Greed is the root of all eBay.
Reply to
Michael A. Terrell

by some terminals.

I mis tryped :-( ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

On a sunny day (Thu, 07 Jan 2010 07:57:03 +0800) it happened who where wrote in :

some terminals.

are no longer

cash ecc" system,

replaced.

formatting link

I looked at my card, and it has written 'Gemalto UK' on the back.... So that is not France, and it worked after Jan 1 2010 in the shop using the card's mag stripes, and also in the small crypto calculator I have that uses the card's chip for online banking. I did read today that the un-official fix seems to be to put some scotch tape over the chip, that makes the reader think it should use the mag stripes only. It also seems to clog up some cash terminals... LOL Now they talk about modifying the terminals so the card are reprogrammed when you insert them. No if somebody could log that reprogramming sequence, then they could program their own cards soon. Oh well, there is no 100% security.

Reply to
Jan Panteltje

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.