Devnote Tuesday: We’re back!

[MaxPeck]

Identify a core problem.  There are several to choose from.  A lot of the reports on the bug tracker are symptoms or variants on a few core issues.  Pick one of those core issues and work THAT PROBLEM until it's fixed, before moving on to the next problem, or trying to tackle concurrent issues.  You have the team working on 10 things at once, in different directions, and then trying to consolidate the changes and hoping for the best.  Pick a problem.  Work the problem as a team, and then move on to the next problem.  Repeat until you run out of problems.  Don't even discuss shiny new features (like comm relays) until the current release is feature complete and stable.  That's how every design team I've ever worked with has operated.  Do one thing, do it well, do the next thing.

[Alshain]

4 minutes ago, JJE64 said:

Identify a core problem.  There are several to choose from.  A lot of the reports on the bug tracker are symptoms or variants on a few core issues.  Pick one of those core issues and work THAT PROBLEM until it's fixed, before moving on to the next problem, or trying to tackle concurrent issues.  You have the team working on 10 things at once, in different directions, and then trying to consolidate the changes and hoping for the best.  Pick a problem.  Work the problem as a team, and then move on to the next problem.  Repeat until you run out of problems.  Don't even discuss shiny new features (like comm relays) until the current release is feature complete and stable.  That's how every design team I've ever worked with has operated.  Do one thing, do it well, do the next thing.

Well, that's not how team development works.  At most, two developers on a problem, more than that and it becomes too many cooks in the kitchen.  It actually makes matters worse.

[smjjames]

10 minutes ago, Alshain said:

Except it doesn't matter.  1.1.3 won't fix the outstanding issues because 1.2 clearly says it will.  Sure it might fix some of them, but I have a feeling that without the Unity update it won't even fix all of the most egregious ones.   We are pretty confident the wheel issues and landing leg issues are Unity, which means 1.1.3 will be as useless as 1.1.2.  Thus, the bugs will not be fixed before new features are added.

I don't suppose they could take a stab at it for 1.1.3 as far as the landing legs go? Or at least do a baindaid solution to make it not so explodey and find a real solution with the Unity patch. Though it's probably not an optimal solution.

Edited May 25, 2016 by smjjames

[MaxPeck]

6 minutes ago, Alshain said:

Well, that's not how team development works.  At most, two developers on a problem, more than that and it becomes too many cooks in the kitchen.  It actually makes matters worse.

The thing is, these problems have multiple legs.  There are what? 1100 open items on the bug tracker?  I don't think there are 1100 actual problems, I think there are 1100 symptoms of maybe one or two dozen core issues.  So start by consolidating those items into core issues, and then tackle those.  So, say 1.1.3 will address CTD issues exclusively.  Gather all of the tracker items that relate to CTD, freezes and memory overflows.  Tackle those.  Then make 1.1.4 specifically address the mess that is the surface colliders.  Tackle those.  then make 1.1.5 specifically address landing gear and landing legs.  Address those.  Memory leaks?  1.1.6.  And so on.

And if the answer to any of those problems is "we need to push the next version of U5" then do it.  Don't wait for an arbitrary 1.2 if the solution to the problem in front of you is to upgrade the engine.  Just do it.  If doing so fixes other upstream problems, BONUS!  But don't waste time on bandaids today because we have to wait for 1.2 for the next engine upgrade.  And don't have 10 people working in 10 directions and hope those projects will mesh and then rush to release.  Because then you get fixes that conflict, you get users in a perpetual state of beta testing, and you get a lot of unhappy customers.  Consolidate your problems, address them systematically and don't move on or add new features until the one in front of you is complete.

There.  KSP fixed in 500 words or less.  Work your magic, game developers.

[Cornholio]

4 minutes ago, JJE64 said:

The thing is, these problems have multiple legs.  There are what? 1100 open items on the bug tracker?  I don't think there are 1100 actual problems, I think there are 1100 symptoms of maybe one or two dozen core issues.  So start by consolidating those items into core issues, and then tackle those.  So, say 1.1.3 will address CTD issues exclusively.  Gather all of the tracker items that relate to CTD, freezes and memory overflows.  Tackle those.  Then make 1.1.4 specifically address the mess that is the surface colliders.  Tackle those.  then make 1.1.5 specifically address landing gear and landing legs.  Address those.  Memory leaks?  1.1.6.  And so on.

And if the answer to any of those problems is "we need to push the next version of U5" then do it.  Don't wait for an arbitrary 1.2 if the solution to the problem in front of you is to upgrade the engine.  Just do it.  If doing so fixes other upstream problems, BONUS!  But don't waste time on bandaids today because we have to wait for 1.2 for the next engine upgrade.  And don't have 10 people working in 10 directions and hope those projects will mesh and then rush to release.  Because then you get fixes that conflict, you get users in a perpetual state of beta testing, and you get a lot of unhappy customers.  Consolidate your problems, address them systematically and don't move on or add new features until the one in front of you is complete.

There.  KSP fixed in 500 words or less.  Work your magic, game developers.

This guy has it figured out. 

[Frybert]

3 hours ago, KasperVld said:

I scratched my arm trying to replace the headlight bulb on my motorbike

The word that has me wondering is 'trying'. Did you manage it? Or did the scratch prove to be too much?

[Tourist]

6 minutes ago, Alshain said:

Except it doesn't matter.  1.1.3 won't fix the outstanding issues because 1.2 clearly says it will.  Sure it might fix some of them, but I have a feeling that without the Unity update it won't even fix all of the most egregious ones.   We are pretty confident the wheel issues and landing leg issues are Unity, which means 1.1.3 will be as useless as 1.1.2.  Thus, the bugs will not be fixed before new features are added.

I suppose it depends on what issues are bothering people the most. From reading the devnotes Squad seem to  suggest that 1.1.3 is aimed largely at stability, ie, crashing... which actually seems pretty important and what I would generally regard as being a vital issue to get right before adding new features. Many of the commenters seemed to have ignored this paragraph entirely and gone to Squad is not fixing anything before adding stuff.... and a thing which annoys me more than exploding landing gear is selective reading.  

But I take your point that they are not dealing with all of the outstanding issues before new features are added. 

Ultimately I'd not have responded if these comments had said, "hey, they say they will deal with the issues, but I don't believe them" rather than just ignoring the fact they said it at all. 

 

8 minutes ago, Cornholio said:

I don't think it's much about the semantics of the paragraph.  It's more that the Dev Notes message wasn't "WE'RE BACK, GUYS! Don't worry, we're here to save you, and quickly!" but instead was a bit more like, "Business as usual!"

Its not semantics that is important, rather what was actually said.... there is always a risk in interpreting text by using words or impressions that a not directly raised by the text... the risk being that the "impression" or interpretation you get is just confirmation bias. 

 

[Hodari]

1 hour ago, Alshain said:

 

I have it on good authority that someone ate ice cream too fast and got a headache.  It was touch and go for a while but he did survive.  Naturally they didn't want to mention it out of respect for the victim and his family.

I hear he is on the rocky road to recovery though.

[Alshain]

36 minutes ago, JJE64 said:

The thing is, these problems have multiple legs.  There are what? 1100 open items on the bug tracker?  I don't think there are 1100 actual problems, I think there are 1100 symptoms of maybe one or two dozen core issues.  So start by consolidating those items into core issues, and then tackle those.  So, say 1.1.3 will address CTD issues exclusively.  Gather all of the tracker items that relate to CTD, freezes and memory overflows.  Tackle those.  Then make 1.1.4 specifically address the mess that is the surface colliders.  Tackle those.  then make 1.1.5 specifically address landing gear and landing legs.  Address those.  Memory leaks?  1.1.6.  And so on.

And if the answer to any of those problems is "we need to push the next version of U5" then do it.  Don't wait for an arbitrary 1.2 if the solution to the problem in front of you is to upgrade the engine.  Just do it.  If doing so fixes other upstream problems, BONUS!  But don't waste time on bandaids today because we have to wait for 1.2 for the next engine upgrade.  And don't have 10 people working in 10 directions and hope those projects will mesh and then rush to release.  Because then you get fixes that conflict, you get users in a perpetual state of beta testing, and you get a lot of unhappy customers.  Consolidate your problems, address them systematically and don't move on or add new features until the one in front of you is complete.

There.  KSP fixed in 500 words or less.  Work your magic, game developers.

Normally that would be the responsibility of the customer service team to determine similar issues and get them consolidated into a bug tracker system so the developers can dive in.  I told you I've worked on many development teams, that is my job, I am customer service for my company.  I do a little development as needed as well, but mostly CS.  However, Squads Customer Service team is it's customers, as evidenced by the public bug tracker.  You shouldn't be inputting bugs that way, there should be a middle man between you and the developer.  So you have this conglomeration of 1000 bug reports where (best guess) 40% of them are duplicates?  The lack of real customer service team to parse these out is why that happens.

Now really I am exaggerating a bit, Squad does have some customer service, but it simply isn't set up to be efficient at what it does, and the fact that the developers have to interpret reports is a big time waster.  Customer service has to be able to gather relevant information and squeeze out the irrelevant or they might as well not be there.

Again, this goes back to Squad being a young development team, as I said earlier, it is their first project.  It does take time to set these systems up, and often it results in some "culture shock" which can be difficult to overcome.  For example, I once had a developer who absolutely would not stop calling customers, eventually his boss game him instructions in no uncertain terms that he was not to be in direct contact with the customer.  It just wastes his time, customers can be unclear and wishy washy and can take hours just to get a good report sometimes, (often there is no bug and the customer doesn't know what he/she is doing) his talents were better served coding.

Edited May 25, 2016 by Alshain

[swjr-swis]

4 hours ago, KasperVld said:

I scratched my arm trying to replace the headlight bulb on my motorbike

Motorbikes: even more dangerous than rockets!

[linuxgurugamer]

Will 1.3 require any significant changes to mods?  

 

[MaxPeck]

48 minutes ago, Alshain said:

Normally that would be the responsibility of the customer service team to determine similar issues and get them consolidated into a bug tracker system so the developers can dive in.  I told you I've worked on many development teams, that is my job, I am customer service for my company.  I do a little development as needed as well, but mostly CS.  However, Squads Customer Service team is it's customers, as evidenced by the public bug tracker.  You shouldn't be inputting bugs that way, there should be a middle man between you and the developer.  So you have this conglomeration of 1000 bug reports where (best guess) 40% of them are duplicates?  The lack of real customer service team to parse these out is why that happens.

I think this right here also hits the nail on the head.  Squad encourages any- and everybody to report bugs via the bug tracker, and so you get hundreds of duplicate bugs described in hundreds of different ways under hundred of various conditions.  They (or their proxies in the moderators) squelch "complaints" either by shaming or by directly silencing them, and then encourage people to contribute to submit a ticket for bugs they encounter, which has turned the bug tracker into a hot mess.  As far as I can see, and I can't say this for sure because I can't see how many chefs are in the kitchen, but it appears there is little to no project management going on, and a lot of people working on projects of their own volition, or by committee consensus. I saw an embryonic attempt at project management on the bug tracker site, but much to my shock, few, if any of the items on the progress items showed any attempt at progress - sometimes for years.  That tells me that there is no unified effort, and that issues are being addressed in a haphazard way, rather than in a systematic way.  I almost get the feeling that all the devs are equal and there is no head developer or project manager setting direction, which is why releases feel so ... disorganized. 

If it were me, and it's not, but from the safety of my keyboard I'm free to dream... I would freeze KSP as it is.  I would close adding new bugs to the bug tracker.  I would spend a week or two going through the bugs, analyzing and classifying them, develop core issues to address, and then create a roadmap to addressing them.  I would put someone in charge, form action teams and start systematically addressing the bugs according to the roadmap.  I would not start the next action item until the previous one was resolved, and then once all of the items on the roadmap was addressed and the program was stable, well, now what you've got is a production-ready product.  Release time.  NOW you start dreaming about console ports and comm relays, until the next design freeze.  And then you repeat the process for each iteration of KSP.  In the end, it's a whole lot less work, because you don't have to shotgun hot fixes.  Those should be the exception, not the rule.

Now I know you said you've got dev experience.  Many people on this board have the same, there are a lot of software devs on here, and a lot of us with project management experience.  What we're effectively doing right now is - and pay attention Squad - we're consulting for your company for free.  You have experts from various disciplines here in the community who are wiling to help out with suggestions, and not everyone who criticizes or complains is doing so out of malice.  I can't code in c#.  I don't have any experience with Unity.  Some people do, and they can be of great help in those areas.  But I do have experience managing large projects and working on complex design teams, and this is where I see their biggest breakdown being.  Their vision for the game is awesome, and their coding and software development skills are far beyond what I could ever hope to achieve, but from what I've seen, their organizational setup isn't ideal, and their approach to design and implementation is where they fall down.

I hope, sincerely, that they can look at some of these posts and see them not as complaints from whining complainers, but as people contributing the best way they can.

Edited May 25, 2016 by JJE64

[AlamoVampire]

7 hours ago, SQUAD said:

<snip>

 Results indicate that at least one or two crashing issues have indeed been fixed already.

<snip>

I hope its two. Heres hoping 1.1.3 and 1.2.X fix all of the instability.

Edited May 25, 2016 by AlamoVampire

[EliasDanger]

Awesome! Can't wait to wait for all my mods to update to 1.2!

[Dispatcher]

As Squad releases patches and updates, the mods will have to work overtime to try to keep catching up.  

Welcome back Squad!

[BeeGeenie]

Another "no poem in the dev notes" haiku:

The Devs have returned

Code ain't gonna write itself

Glad to have you back.

Edited May 25, 2016 by BeeGeenie

[roosterr]

VERY Well deserved break ladies and gentleman.  What a ride you have been on over the last few years and thanks for your wonderful world changing creation.  Kerbal IS Awesome

[kBob]

Three weeks, that's nothing, I'm back after two years, good thing I'm not one of your developers.  Anyway nice to see the progress being made and continuing to be made, I see several things on my wish list are now incorporated, love it.  Now if I can just keep from using KSP in too long sessions that cause me bad pain all will be great and hopefully no more two year vacations...but I suspect it's even more addicting than before.

[Sirad]

Devnotes:

' we see that some of you are still having various crashing issues, amongst a few other things we want to resolve '

I expected this, anyway. The will to resolve does not solve any Bug.

a statement like 'Bug x and Bug y will be removed as primary shedule' would have been a statement. all else is just babble to soothen the customers. Sleep well then.

 

[Superfluous J]

20 minutes ago, Sirad said:

a statement like 'Bug x and Bug y will be removed as primary shedule' would have been a statement. all else is just babble to soothen the customers. Sleep well then.

It'd have also been a lie. You can't promise you'll fix a bug that - even after 3 weeks of thought - you don't know you are capable of fixing.

I don't know about Squad, but where I work falsely promising the customers undeliverable things is Marketing's Job I mean a huge no-no.

Edited May 25, 2016 by 5thHorseman

[Rath]

11 hours ago, KasperVld said:

I scratched my arm trying to replace the headlight bulb on my motorbike

Kasper, How could you do this to me?  You have a motorbike?  That, that means, your cooler than me...

[Redneck]

14 hours ago, SQUAD said:

 

Next on the list is planning for update 1.2: this update will most likely contain a minor update of the Unity game engine which should fix various outstanding issues, more updates to the wheels, the implementation of some already announced features which didn’t make the 1.1 patch and various further items we’re looking into.

So see there? It was Unitys fault! Like I said it shouldn't require a squad "Super-coder" person to fix the wheel problem. Glad to hear of the Unity fix. Im curious though of the timing of the Unity update. Is the update to Unity game engine a direct result/responce of the error findings in KSP?

[Majorjim!]

Welcome back, once more into the breach dear friends!

 Does this mean that the  1.2 update is the unity update that fixes wheels and legs?

[Padishar]

37 minutes ago, Redneck said:

So see there? It was Unitys fault! Like I said it shouldn't require a squad "Super-coder" person to fix the wheel problem. Glad to hear of the Unity fix. Im curious though of the timing of the Unity update. Is the update to Unity game engine a direct result/responce of the error findings in KSP?

No, KSP 1.1.2 is currently using Unity 5.2.4.  Unity 5.3 has been released for some time and 5.4 has recently gone into beta (I believe).  Apparently, the wheels were already much improved in 5.3 but other issues in that version made it impractical to use it for KSP so they are waiting for (presumably, the official release of) 5.4

1 hour ago, Rath said:

Kasper, How could you do this to me?  You have a motorbike?  That, that means, your cooler than me...

Are you sure that follows?

[Jovus]

Are you kidding me? That guy's much cooler than any of us, unless you're typing outside in a Siberian winter.