Britain's rail ticket-booking systems go TITSUP* Train companies across the UK are unable to take online ticket bookings due to what has been described as a "national issue" with the country's reservation service. The outage – which appears to have affected all UK train operating companies (TOCs) today – means that users can't book tickets on websites or on mobile apps. The …
There's a feature that I quite like in Veeam (bear with me!)
When you're copying your VM's over from one place to another, or backing them up, or whatever, it tells you where the bottleneck is.
Source, Processing, Network, Target, etc. You know what's holding up the line.
Cloudflare do something similar, with a checklist of "Well, we're up, but the site you're trying to access is down" icons. It's about time, we all follow such lead and have the systems automatically lay blame.
"Please try again later" isn't acceptable. "Sorry, the national ticket-booking service is down", or "Sorry, this station isn't able to connect to the Internet to check that information", or "Sorry, information about Virgin trains isn't currently available from their systems."
Short, easy, subtle, but at least lay the blame where it needs to be laid rather than "Error". And it's literally the case of changing the output of a few exception catchers / error paths into something friendly.
However, Veeam caught me out the other day through exactly this problem. "Could not process the VM". Apparently that means "You need to take a Full before you can take an Incremental, pillock" but fails to tell you that and instead gives you something that's both heart-wrenchingly worrying, and almost impossible to narrow down without working blind.
Sensible error handling is so last century. It is much better to not do any error checking and to throw exceptions for relatively expected events and to then process these in the usually unhelpful "an error may have happened" kind of response.
Muppets. While there are near religious flame wars about error handling vs exception handling, they both have their place. Error handling is for the expected failures, exception handling is for the unexpected ones - use both to their advantage in other words.
>> Do trains stop for any other reason??
They do if the signalman at Wokingham sets the points wrong and an electric SWR train that's meant to go to Waterloo instead goes onto the Guildford line, which has no third rail.
The trains behind then get held for 3.5 hours until they can figure a way of getting the unpowered train back where it belongs.
>> That's a strangely-specific example, RichardT. What made you think of that?
I was stuck in one of the trains behind it.
It was on a Sunday afternoon. We reversed back to Winnersh station and waited. The train crew strongly advised us to catch one of the next trains back to Reading. However, it also happened to be the last day of the Reading festival and they were already putting out loads of barriers at Reading station when we left earlier, so I didn't fancy getting mixed up in that.
After a few hours the guard managed to get some taxis to come from Maidenhead but, just as they arrived, the line was cleared and we could proceed. A lot of the people on my train were going to Gatwick and probably missed their flights.
That's true. I got this one last week courtesy of WPF:
'Cannot attach type "TextBoxEnterKeyUpdateBehaviour" to type "TextEdit". Instances of type "TextBoxEnterKeyUpdateBehaviour" can only be attached to objects of type "TextEdit"'
Of course it being WPF meant that when I eventually tracked it down it was only tangentially related to the XAML line where the error was being reported.
Meh. It's been impossible to buy a season ticket from Chiltern Railways for the last two weeks due to configuration and deployment issues. Because the incompetent web managers can't grasp that in-house staff accessing the same system may, in fact, be accessing a different system (internal vs external DNS) they have been carefully denying the problem. Only little gems like "pick a station from the list" followed by "station not recognised" and wonders of "modern development" like that and different station lists appearing depending on whether or not the user is in-house or not.
This is on top of the standard issues with idiot web developers attempting to replicate standard browser functionality using JavaScript.
Since the only reason any of these contracts get award is the ridiculous cost of pitch and the liability cover, as opposed to any sort of competence other than shovelling money and entertainment. Surely it’s time to claim a refund. That goes for the hardware failure as well, be it the pointy wirey things or the borrowed worn out machinery.
For the record, selling off the property, for less than the annual rent, to your soon to be partners, is a fucking national scandal. #hydrogengoesbang
Okay so I'm paraphrasing a bit, but my local stations are all trying to phase out cash transactions as quickly as they can and force everyone to use electronic payment systems.
Most of my travel is on one network but my 'home' station is on another; as a result, I can't use any sort of discounts or their equivalent to the Oyster card because ONE station is part of "the Opposition"s network unless I buy at my DESTINATION station. I've tried paying on the train but apparently the remote terminals can't cope with anything except tickets at full price, so I have to hope the ticket office is open in the evening and that the queue isn't so long I miss the train home and have to wait nearly an hour for the next...
Add to that the fact that my rail ticket is approaching the cost of the petrol to drive in for the week (it started off at about 20% of the cost!) and the rail company is looking increasingly likely to lose my business...
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
