It's good that FD specialized on video games

I, for one, certainly haven't set any kind of deadline for them regarding this update. So they could happily keep testing if for one more week or two (or how much time it would take them to spot hidden bugs like the module integrity or power priority one, let alone the FPS drops and the 10Cr bug).

Thing is, projects in business always have deadlines. Investors will always expect to see progress being made.

I'm not happy with the state of some of these bugs either, but this seems to be the way the games industry works. I don't think Frontier could be accused for being worse than their competitors and expecting them to be better.... well I think that goes beyond optimism.
 
The non-existence or "non-possibility" of bug-free code is a myth, an urban legend planted by the "let's go cheap & agile" Industry to get away with lowering quality standards.
Yeah, it's certainly doable. It's just not easy or cheap. I'm writing some code at work which basically has to be bug free first time, and it's taken several times longer than the normal "a few minor bugs that we can fix later are acceptable" standard would have done - and not just in my time to write it, but in a lot of other people's time to repeatedly review the design and implementation to make sure I'm getting it right.

And in the relevance to Frontier, the slower pace has caught other people out who expected it to be finished and working rather sooner than it will be - game is dead, maintenance mode, if they don't push out a new release every other week - but in this particular case I can argue it absolutely needs to be bug-free before it goes ... whereas actually while it's really obvious and annoying, and I'm not sure how they missed it, something like the restock bug is also basically harmless in terms of gameplay impact.

And believe me, it's not that complicated to catch sight of something like the power priority or the 10 Cr bug.
Much as it might sound like it, I'm not actually defending Frontier's current release procedure or the priority they give to testing and polish (and I'm currently building up a list of potential examples for a major conceptual problem with the new bug reporting system too - but I need more evidence that it's a real problem and not just insufficient data here). Every release seems to come with a few really obvious new bugs which should have shown up very quickly on a basic end-to-end test. But there's a big gap between what Frontier currently do and "bug free", and neither of those are where players actually want them to be - fewer bugs, but we're not going to either pay for the programmers to have the time to do bug free, or wait the extra several months between releases it would need to get there.

In this release, the 10 Cr bug, yes. Some of the weird (un)docking behaviour, yes. The frame rate issues some people are having, yes. The power priorities one not necessarily, because unless you expected that power priorities once set, tested and confirmed working would spontaneously and randomly unset themselves later, you might not think to include in an end-to-end test and point unit tests would never spot it because all the individual bits still work fine. It's a serious bug, but potentially tough to spot with conventional test approaches without spending a lot of time on it.

(As I've mentioned elsewhere, I think any significant release should have a 1 week beta just to check for issues that are really obvious when hundreds of people play the game for tens of hours each on diverse hardware, but might get overlooked because the trigger conditions are slightly too obscure for formal QA)
 
What this looks like to me is a lack of automated regression testing, which is near-universal best practice when developing large and complex systems (or for that matter small and simple ones). And yes I have done (and do) develop those for a living. If that is not possible - presumably because they have never built a test harness for the client - then the only way forward is a regular human-driven Beta test cycle, as many have suggested. I have no doubt Frontier would be overwhelmed with volunteers for that.
 
In this release, the 10 Cr bug, yes. Some of the weird (un)docking behaviour, yes. The frame rate issues some people are having, yes. The power priorities one not necessarily, because unless you expected that power priorities once set, tested and confirmed working would spontaneously and randomly unset themselves later, you might not think to include in an end-to-end test and point unit tests would never spot it because all the individual bits still work fine. It's a serious bug, but potentially tough to spot with conventional test approaches without spending a lot of time on it.

Idk, the power priority bug took me exactly five minutes to discover.
But okay, deadlines.
 
Last edited:
Idk, the power priority bug but took me exactly five minutes to discover.
But okay, deadlines.
Yeah, but you're actually playing it. Pretty sure you didn't find it in a code review pass.

This isn't defense of FD, but redirection to the real issue. This is more a huge flag for FD to institute a test server than us bothering to take cheapshots at their doods, which just starts a teacup fight going "they suck" vs "ur mean", with liberal doses of people bragging about their coding chops and such.

This one bug shares a history with a whole family of weird, easy-to-spot-by-playing bugs, so I'd wager there's a huge hole in their internal testing abilities of some sort. Having a public test environment would cover a lot of areas that don't seem to currently have oversight/direction/aliens ate their pizza/whatever. The failreasons don't matter as much as getting a better test method in the first place.

Hrm, just Devil's advocated my own idea tho, given FD's "air of mystery" release methods, I wonder how hot the hotbuild can be while also leaving out their fabulous seekrits. They may decline to have a test server to keep their seekrits really seekrit.
 
Idk, the power priority bug took me exactly five minutes to discover.
Yeah, I found it pretty quick as well. And it's the only one of the new bugs which is causing me serious issues, because many of my ships rely on power priorities working and so I currently have to leave them in the hangar. But of course anyone who runs a ship which fits under 100% to start with might never notice it, even if they'd previously set priorities.

But to find it during formal testing, you'd need to run a test with an underpowered ship which involved setting power priorities before supercruise and then using them afterwards [1]. That's well outside the realm of unit testing or even feature testing, which is getting into the expensive side of things to catch a bug which might never happen. What you need to find that sort of bug is just someone playing the game in a variety of ways - not to a fixed test script, which might have the "run an underpowered ship" bit of the test a long way separate from the "do some supercruise travel" bits and therefore never spot it - but just flying about doing stuff. That sort of testing is actually probably cheaper and quicker to run as a public beta than as a formal process - considering the great variety of ways people can play the game and the size of the player base compared with the size of the dev team.

[1] Edit: and even then you might miss it since it doesn't always do it, and it doesn't always flatten all the priorities, so the ship might still behave correctly after the first couple of jumps, and by then you'd be assuming it worked and moving on to making sure that pulse lasers didn't spontaneously turn into beam lasers if you did a neutron jump.
 
Sure. Are you willing to pay the approximately 25x-50x extra cost for software developed to that sort of standard (so bought at full price, around £1000 to £2000 for Elite Dangerous, and the same again for Horizons? ... but you could probably pick it up in a sale for £250 or so each if you were patient?)

It sounds expensive, but it'd be worth it to have no bugs at all, right?
So...less than Star Citizen then? :sneaky:
 
Idk, the power priority bug took me exactly five minutes to discover.
But okay, deadlines.
Don't underestimate deadlines. This derives from the worship of the neolithic god Daedlinius. The tribe who worshipped this particular god later evolved into middle management everywhere. Daedlinius was a wrathful god who meted out horrific punishments to all who disobeyed the central tenet of the faith which was to not finsh an edict that was given by one of the high priests.

Nowadays common sense applies across most industries and I could say things like something is dangerous/illegal and the project will be delayed until it's fixed. However, in industries where legal action is unlikely, such as in the gaming industry, worship of Daedlinius is still going strong. This is why the concept of not adhering to completing a task that has been ordained is still considered to be blasphemous.

Ok, yes, my post is ridiculous, but don't doubt that worship of the final deadline date exists.
 
Last edited:
They were not (necessarily) supposed to find it in a code review pass.
They didn't find it. That's what we know for sure, same as a bunch of other weirdo bugs. As far as why, "aliens ate their pizza" is as valid a complete guess as any others in this thread.

Having a test server would have allowed you to find that bug with your speedy bug-finding skills, and actually have it addressed instead of us all hanging out in a thread pretending like we really know how FD does internal business.
 
I never claimed it wasn't. But you surely can quote my post which made you think so.

Yeah... At the point I've got to explain things like that, when your misunderstanding is very obvious in the post I responded to, then I'll bow out and leave you to your ramblings. I've got a planet to drive round...
 
Reminds me of the old joke about how we're lucky Kawasaki limit themselves to building ships.

Honda and Yamaha both make reliable motorbikes and their engines often get used to power light aircraft.
Kawasaki, OTOH, make some sexy bikes but they have a reputation for being, to be blunt, a bit shoddy and you probably wouldn't want a Kawasaki engine keeping your plane in the air.

Course, I ride a Ducati and, frankly, I wouldn't trust a Ducati engine to power a cement mixer. :rolleyes:

😄 Yeah, yeah...

I've sailed in a ship powered by Kawasaki and I've ridden Hondas, Yamahas and Kawasakis.

The ship broke down*, but their bikes were flawless! In fact, the engine of one of my GPX750Rs survived a head on collision that destroyed the car and crippled me. 😲 Last I heard it was inside a custom frame, still pulling like a train...

My other bikes weren't always reliable- I've had a shonky Honda and a worn out Yamaha.

I wanted a Duke at one point, but the showroom staff blatantly ignored me when I turned up in riding gear. When I was younger I'd have just demanded they attend to me, but I can't be bothered these days. I took the family to Florida instead. 😁 I still had enough left over for a Triumph when we got back. Result!


*Our own fault. While practicing what we'd do if the ship actually broke down, we turned the engines on and off, repeatedly. Maritime diesels have huge turbos, which don't take too kindly to that sort of abuse. Bad news for our engineers, but I got an unexpected run ashore in Miami out of it. It's an ill wind...
 
i was thinking people in this thread could use a bit of sense of humor, then i read this ....

The advisarial relationship between Boeing and the local unions has created a situation where positions in Boeing are being staffed by union employees who aren't fully qualified.

:ROFLMAO:
 
Back
Top Bottom