Jump to content

Update 1.2 has entered Experimental testing!

UomoCapra

By UomoCapra
in 2016

 Share

Recommended Posts

Giskard

Giskard

Posted
Posted
1 hour ago, severedsolo said:

If you set out to break the game, and the game breaks. That's not unintended or unexpected beahviour.

That is true. But what I was trying to say was that sometimes these extreme use cases can reveal bugs that are also affecting the normal gameplay. There is no clear border between intendend and not intended use of a software. Sometimes it might help to push it to the limit to make it all more stable under more normal circumstances as well.

Link to comment
Share on other sites
steve_v

steve_v

Posted
Posted (edited)
1 hour ago, severedsolo said:

If you set out to break the game, and the game breaks. That's not unintended or unexpected beahviour.

Er, that's what QA testing is all about. To pull random internet quote I saw recently:

QA Engineer walks into a bar. Orders a beer. Orders 0 beers. Orders 999999999 beers. Orders a lizard. Orders -1 beers. Orders a sfdeljknesv.
— Bill Sempf (@sempf)

If the bartender says: "We don't serve sfdeljknesv here", all is well. If the bartender vanishes and the player falls through the planet, not so much.

Adding "Do not order sfdeljknesv at the bar." as a footnote in the game manual is not a good solution.

WRT KSP: If you do something wacky in game, and the physics engine delivers a realistic result (within game definition of 'realistic') then that's cool. If it causes the universe to implode, violates the laws of motion / thermodynamics,  accelerates parts to impossible velocities, or causes any critical physics parameters to become NaN, then it's a bug.

What Danny2426 does could well be called edge case testing.

If KSPs physics engine did not have bugs, we would not (and we should not, IMO) have kraken drives.

As physics simulations go, KSP is a bit wacky - that's cool, if it's by design. But the universe going kablooey because the player does something unanticipated is a QA fail.

Edited by steve_v
Addendum.
Link to comment
Share on other sites
Corw

Corw

Posted
Posted
5 minutes ago, steve_v said:

Er, that's what QA testing is all about. To pull random internet quote I saw recently:

QA Engineer walks into a bar. Orders a beer. Orders 0 beers. Orders 999999999 beers. Orders a lizard. Orders -1 beers. Orders a sfdeljknesv.
— Bill Sempf (@sempf)

If the bartender says: "We don't serve sfdeljknesv here", all is well. If the bartender vanishes and the player falls through the planet, not so much.

Adding "Do not order sfdeljknesv at the bar." as a footnote in the game manual is not a good solution.

WRT KSP: If you do something wacky in game, and the physics engine delivers a realistic result (within game definition of 'realistic') then that's cool. If it causes the universe to implode, violates the laws of motion / thermodynamics,  accelerates parts to impossible velocities, or causes any critical physics parameters to become NaN, then it's a bug.

If KSPs physics engine did not have bugs, we would not (and we should not, IMO) have kraken drives.
 

As physics simulations go, KSP is a bit wacky - that's cool, if it's by design. But the universe going kablooey because the player does something unanticipated is a QA fail.

The devil is in the details. QA takes time and effort, so you would prefer to optimize to check for things that people would more likely do instead of covering think not even Whackjob would do. To complicate things even more, it is not just KSP gameplay that you are testing, but also bugs in underlying Unity engine that is out of KSP devs reach and in a lot of times they had to work around Unity bugs.

Link to comment
Share on other sites
steve_v

steve_v

Posted
Posted (edited)
8 minutes ago, Corw said:

The devil is in the details. QA takes time and effort, so you would prefer to optimize to check for things that people would more likely do instead of covering think not even Whackjob would do.

In most games, sure. Don't bother testing that part of the map because you can't get to it without cheating, etc.
But KSP is essentially a sandbox, there are no real guidelines as to how one should play it, so to a certain extent there's no such thing as "things that people would more likely do". To wit: making hinges out of thermometers or bearings out of landing-gear. Who would have thought?
And when you're still adding features, leaving edge-case bugs unfixed will likely bite you in the donkey later on...

Edited by steve_v
typo
Link to comment
Share on other sites
Corw

Corw

Posted
Posted (edited)
22 minutes ago, steve_v said:

In most games, sure. Don't bother testing that part of the map because you can't get to it without cheating, etc.
But KSP is essentially a sandbox, there are no real guidelines as to how one should play it, so to a certain extent there's no such thing as "things that people would more likely do". To wit: making hinges out of thermometers or bearings out of landing-gear. Who would have thought?
And when you're still adding features, leaving edge-case bugs unfixed will likely bite you in the donkey later on...

Hint: most people wouldn't and it is not part of core gameplay, more in line of "look what a wacky thing I made". So it would be low on testing priority. If it works, great, if someone finds a bug there, it will go low in bug fix priority because it is a fringe use case that is not affecting any reasonable part of the user base. At this time the developers time (and cost) comes at premium.

 

The issue may be important to you, but 99.9% of the user base don't care.

Edited by Corw
Link to comment
Share on other sites
Curveball Anders

Curveball Anders

Posted
Posted
2 hours ago, steve_v said:

What Danny2426 does could well be called edge case testing.

Nah, I don't fully agree.

To take an example from flight sims (both games and commercial ones).

They focus on getting the simulation as good as possible within the approved flight envelope of the aircraft in question and also try to make sure that the edge cases are handled in a realistic way (as in when you are about to or just has left the envelope).

They often don't bother much about realistic handling way outside the envelope since you're never expected to be there and if you are, you've either ripped your aircraft to bits or are about to crash.

So I see a difference between edge cases and wildly outside any reasonable envelope.

Even the most expensive flight sim used for training commercial pilots will not be very realistic if you try to recover a fully loaded 757 from a flat-spin, just because it shouldn't ever happen, and if it happens there's nothing the pilot can do about it.

Link to comment
Share on other sites
Mako

Mako

Posted
Posted
5 hours ago, steve_v said:

What Danny2426 does could well be called edge case testing.

WRT the entertaining videos he produces I think theres an argument to be made that what he does could well be called finding undocumented features.

I, too, would love it if Squad could find and fix all the bugs - both known and unknown - because someone, someday, might do something unpredictable that breaks the game, but I also want new features. As with most things, it seems we must take the good with the bad.

And in this case I think the good is really good, and the really bad is often predictable and avoidable.

Link to comment
Share on other sites
Green Baron

Green Baron

Posted
Posted

 

The prerelease looks pretty stable on linux after 1 hour. Minor graphics glitches and flaws (inline communotron doesn't communicate, new lateral one does) are to be expected. Next step: set up a proper comms network ... :-)

 

 

Link to comment
Share on other sites
KroShan

KroShan

Posted
Posted
7 hours ago, Red Iron Crown said:

They're up now (subject to change due to feedback and testing, of course):

 

Oh - thx. didn't find them 'till now.

It will definitly help to tell bug and feature apart...

 

Has anybody infos about the mac version - i bought one for my brother and it didn' work well at that time (i think .25) so he quit after the first orbit. Maybe time to come back?
 

Link to comment
Share on other sites
Workable Goblin

Workable Goblin

Posted
Posted
On September 14, 2016 at 4:22 PM, KroShan said:

Has anybody infos about the mac version - i bought one for my brother and it didn' work well at that time (i think .25) so he quit after the first orbit. Maybe time to come back?
 

I play exclusively on Macs, and it's worked fine for me since 1.0.2 or some such. Right now 1.2 is a little funky, there's a weird momentary graphical glitch when loading the VAB (and I assume the SPH, but I haven't tried it yet) and the menu and VAB screens lag pretty noticeably for me (which didn't happen with 1.1.3), but once you get into the actual game everything works fine. 1.1.3 definitely works fine on Macs, given that you have sufficient horsepower. Don't expect silky-smooth gameplay on a MacBook displaying on a 4K external monitor, but for recent iMacs, at least, you ought to be fine.

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...