Roland Perry wrote on 02 Jan 2016 at 11:32 ...
In message , at 11:12:16 on Sat, 2 Jan
2016, Richard J. remarked:
The free travel thing points even more to a Y2K style problem, if the
charging system wasn't active yesterday. In the Nottingham case it
wasn't spotted until the 2nd because no buses ran on the 1st.

ps I note some NatWest debit cards had outages on the 1st: another Y2K
problem perhaps.

For decades, computer systems have exhibited faults after a holiday
period, caused often by problems in restarting hardware or software
after a holiday outage or reverting to normal operation after
non-standard holiday operation, or caused by changes to the system that
were applied during the holiday. Why are you assuming that this
particular instance was in any way similar to Y2K?

Firstly, because a very similar incident *was* tracked down to that
cause,

You mean there was a very similar incident 16 years ago? But what is
special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

and secondly the other routine issues you mention ought to be
well understood and planned for during a holiday period.

In theory, yes, but in practice there is always a greater risk of a
problem following a period of non-standard operations. This is
especially so if you're doing a major update of fare tables across the
network, which might be "routine" in the sense that you do it every
year, but is still a more obvious likely cause than some previously
unknown fundamental problem connected with 1/1/2016.

Plus the fact they are having to talk to their suppliers to work out a
fix, rather than applying a clue-bat to the sysadmin-du-jour.

Quite normal if there's a major system outage. It doesn't tell you
anything about the cause.
--
Richard J.
(to email me, swap 'uk' and 'yon' in address)

In message , at 16:58:52 on Sat, 2 Jan
2016, Richard J. remarked:
Why are you assuming that this
particular instance was in any way similar to Y2K?

Firstly, because a very similar incident *was* tracked down to that
cause,

You mean there was a very similar incident 16 years ago?

The Nottingham City Transport smartcard issue in around 2006.

But what is special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

Or 1/1/2006 compared to 1/1/2000. There's something about the range of
years for which 1/1/xxxx is recognised, which transcends xxxx=2000.
--
Roland Perry

In article , (Roland Perry)
wrote:

In message , at 16:58:52 on Sat, 2
Jan 2016, Richard J. remarked:
Why are you assuming that this
particular instance was in any way similar to Y2K?

Firstly, because a very similar incident *was* tracked down to that
cause,

You mean there was a very similar incident 16 years ago?

The Nottingham City Transport smartcard issue in around 2006.

But what is special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

Or 1/1/2006 compared to 1/1/2000. There's something about the range
of years for which 1/1/xxxx is recognised, which transcends xxxx=2000.

That's not the millennium bug then. It was about years being held in 2
digits so not able to handle properly to rollover of the century.

--
Colin Rosenstiel

In message , at 18:47:45
on Sat, 2 Jan 2016, remarked:
Why are you assuming that this
particular instance was in any way similar to Y2K?

Firstly, because a very similar incident *was* tracked down to that
cause,

You mean there was a very similar incident 16 years ago?

The Nottingham City Transport smartcard issue in around 2006.

But what is special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

Or 1/1/2006 compared to 1/1/2000. There's something about the range
of years for which 1/1/xxxx is recognised, which transcends xxxx=2000.

That's not the millennium bug then. It was about years being held in 2
digits so not able to handle properly to rollover of the century.

It's the same class of problem, which is why I put the "Y2K" in quotes.
--
Roland Perry

In uk.transport.london message , Sat, 2
Jan 2016 16:58:52, Richard J. posted:


You mean there was a very similar incident 16 years ago? But what is
special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

It is the first year 20xx for which xx cannot be stored in four bits.

If memory space was limited in a system designed a decade ago, someone
might have chosen to use just four bits for the variable part of the
year, thinking "Well, I'll not be in this job in 2016, ha ha!".


--
(c) John Stockton, Surrey, UK. Turnpike v6.05 MIME.
Merlyn Web Site - FAQish topics, acronyms, & links.

Dr J R Stockton wrote:
In uk.transport.london message , Sat, 2
Jan 2016 16:58:52, Richard J. posted:


You mean there was a very similar incident 16 years ago? But what is
special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

It is the first year 20xx for which xx cannot be stored in four bits.

If memory space was limited in a system designed a decade ago, someone
might have chosen to use just four bits for the variable part of the
year, thinking "Well, I'll not be in this job in 2016, ha ha!".

Well, as we soon learned, the problem wasn't any sort of Y2K issue. But in
any case, Y2K type bugs date from software designed 40 or more years ago,
when every byte mattered; Oyster is far too recent for the designers to
have been trying to save microscopic amounts of memory.

In message
-sept
ember.org, at 23:52:54 on Mon, 4 Jan 2016, Recliner
remarked:
You mean there was a very similar incident 16 years ago? But what is
special about 1/1/2016 compared to 1/1/2015, 1/1/2014, etc?

It is the first year 20xx for which xx cannot be stored in four bits.

If memory space was limited in a system designed a decade ago, someone
might have chosen to use just four bits for the variable part of the
year, thinking "Well, I'll not be in this job in 2016, ha ha!".

Well, as we soon learned, the problem wasn't any sort of Y2K issue. But in
any case, Y2K type bugs date from software designed 40 or more years ago,

Nope, plenty of things which actually failed were much more recent than
that.

when every byte mattered; Oyster is far too recent for the designers to
have been trying to save microscopic amounts of memory.

That doesn't explain why the NCT Smartcards broke on 1/1/2006.
--
Roland Perry

In message
-sept
ember.org, at 23:52:54 on Mon, 4 Jan 2016, Recliner
remarked:

Y2K type bugs date from software designed 40 or more years ago

Here's another I came across by chance today:

http://www.theregister.co.uk/2012/03...day_confirmed/
--
Roland Perry

Roland Perry wrote:
In message
-sept
ember.org, at 23:52:54 on Mon, 4 Jan 2016, Recliner
remarked:

Y2K type bugs date from software designed 40 or more years ago

Here's another I came across by chance today:

http://www.theregister.co.uk/2012/03...day_confirmed/

Wonderful! I'm pretty sure that even my distinctly amateurish code does
date checks in a smarter way than that (just subtract a date from today and
see if the answer is greater than 365 -- the system clock at least should
know all about leap years).

In message
-sept
ember.org, Recliner wrote:
Oyster is far too recent for the designers to
have been trying to save microscopic amounts of memory.

Rubbish. In my job we're *still* worrying about every single byte, even
on brand new chips.

--
Clive D.W. Feather | Home:
Mobile: +44 7973 377646 | Web: http://www.davros.org
Please reply to the Reply-To address, which is: