This thread on Hacker News terrifies me

[Gambler]

On 9/1/2018 11:42 PM, Nick Sabalausky (Abscissa) wrote:
> On 09/01/2018 05:06 PM, Ola Fosheim Grøstad wrote:
>>
>> If you have a specific context (like banking) then you can develop a
>> software method that specifies how to build banking software, and
>> repeat it, assuming that the banks you develop the method for are similar
>>
>> Of course, banking has changed quite a lot over the past 15 years
>> (online + mobile). Software often operates in contexts that are
>> critically different and that change in somewhat unpredictable manners.
>>
> 
> Speaking of, that always really gets me:
> 
> The average ATM is 24/7. Sure, there may be some downtime, but what, how
> much? For the most part, these things were more or less reliable decades
> ago, from a time with *considerably* less of the "best practices" and
> accumulated experience, know-how, and tooling we have today. And over
> the years, they still don't seem to have screwed ATMs up too badly.
> 
> But contrast that to my bank's phone "app": This thing *is* rooted
> firmly in modern technology, modern experience, modern collective
> knowledge, modern hardware and...The servers it relies on *regularly* go
> down for several hours at a time during the night. That's been going on
> for the entire 2.5 years I've been using it.
> 
> And for about an hour the other day, despite using the latest update,
> most of the the buttons on the main page were *completely* unresponsive.
> Zero acknowledgement of presses whatsoever. But I could tell the app
> wasn't frozen: The custom-designed text entry boxes still handled focus
> events just fine.
> 
> Tech from 1970's: Still working fine. Tech from 2010's: Pfffbbttt!!!
> 
> Clearly something's gone horribly, horribly wrong with modern software
> development.

I wouldn't vouch for ATM reliability. You would be surprised what kinds
of garbage software they run. Think Windows XP for OS:

http://info.rippleshot.com/blog/windows-xp-still-running-95-percent-atms-world

But in general, I believe the statement about comparative reliability of
tech from 1970s is true. I'm perpetually impressed with is all the
mainframe software that often runs mission-critical operations in places
you would least expect.

Telecom systems are generally very reliable, although it feels that
started to change recently.