[HN Gopher] Y2K20 Parking Meter Software Glitch Causes Citywide ...
       ___________________________________________________________________
        
       Y2K20 Parking Meter Software Glitch Causes Citywide Snafu
        
       Author : lelf
       Score  : 41 points
       Date   : 2020-01-04 11:46 UTC (11 hours ago)
        
 (HTM) web link (gothamist.com)
 (TXT) w3m dump (gothamist.com)
        
       | iforgotpassword wrote:
       | It should be sufficient if we hot-patch the end date to 2030 -
       | we'll surely have updated or replaced those systems until then.
       | 
       | I'm curious what part of the system even required specifically
       | coding an end date, and not just sanity checks relative to the
       | current date.
        
         | officialchicken wrote:
         | > I'm curious what part of the system even required
         | specifically coding an end date, and not just sanity checks
         | relative to the current date.
         | 
         | It's working as designed... liability and support was shifted
         | to the owners (cities) prior to the event.
        
         | jiveturkey wrote:
         | my unsubstantiated guess is that this is a licensing scheme.
         | 
         | reminds me of the parking garage in hoboken that wouldn't pay
         | for a new software license. omg was that over a decade ago? i'm
         | getting old.
         | 
         | https://www.govtech.com/magazines/gt/Robot-Garage-Hijacks-Ca...
        
         | nottorp wrote:
         | I'm sure that's what they also thought in 1981 when storing
         | only 2 digits of the year and we all know how that turned out.
        
       | coin wrote:
       | > Y2K20
       | 
       | It's 2020. Why use a cryptic abbreviation with _more_ characters?
       | If they want to use the SI "k" prefix shouldn't it be Y2.02k?
        
         | M4v3R wrote:
         | It's probably just a play on the original "Y2K" name for the
         | problems in year 2000.
        
           | glenngillen wrote:
           | Might also be actually related to the same problem: https://t
           | witter.com/jef_poskanzer/status/1213258774938415104...
        
         | analog31 wrote:
         | Perhaps it's a reference to a standard for labeling electronic
         | components. For instance a resistor with 3300 Ohms will be
         | labeled "3k3" on a diagram. It eliminates the decimal point,
         | which might be hard to read or even fall off if it's printed on
         | a tiny component.
        
           | a1369209993 wrote:
           | That would be "2K02". It's actually a (bad) abbreviation of
           | "two[2] thousand[K] twenty[20]".
        
       | mickotron wrote:
       | We had to do an emergency patch to the config of a well-known
       | data analytics platform, because their regex to identify/extract
       | 2 digit years did not cater for anything beyond 2019... We dodged
       | that bullet, but shows a lack of foresight by said company. Lucky
       | we don't have much data that has 2 digit year dates.
        
       | megaframe wrote:
       | This is so funny I just had a similar bug yesterday and named it
       | our y2k20 bug, just a path name change that someone forgot to map
       | correctly
        
       | ChuckMcM wrote:
       | One of a number of errors that are showing up. My favorite are
       | the websites that regex the '20' out of the year to get just the
       | decade, "Whoops!"
       | 
       | In many ways I'm glad we didn't bother to fix the bugs this year
       | like we did for the 1999 - 2000 transition. Way too many people
       | think that Y2K was just some big 'scare tactic' and nothing ever
       | happened. It is sad that operations is always like that, you work
       | your arse off things operate smoothly and the unknowing just say
       | "I don't see what the big deal was all about."
        
         | Unsimplified wrote:
         | Reliability work is unfortunately under-observed and under-
         | valued... until its absence triggers an incident.
        
       | mxuribe wrote:
       | I have no experience with parking meters, but surprising that
       | software updates need to be done at the site/hardware (as opposed
       | to, say, online or "over the air" or updates).
        
         | mannykannot wrote:
         | If we can judge the vendor's competence from the problem
         | reported here, then if they had made them updatable wirelessly,
         | they would likely be easily hacked.
        
         | Jamwinner wrote:
         | Easier to secure and administer I would guess. More and more
         | orgs are realizing that IOT admin is a bigger burden than a
         | liberator to those who are responsible for upkeep.
        
       | alias_neo wrote:
       | Who wants to bet that is not a numeric bug, but a certificate
       | expiry?
        
       ___________________________________________________________________
       (page generated 2020-01-04 23:01 UTC)