Welcome to the deep dive. We're here to help you get the essential knowledge from sources.
Fast yep, straight to the good stuff.
Today, we're plunging into a really interesting space where software engineering meets games. But you know, not just for fun. We're talking serious applications.
Exactly, think education, training, maybe tackling real world problems. It's about more than high scores.
So if you're listening, you're probably looking for a let's say, quick but solid grasp of this, trying to get those aha moments without wading through tons of jargon.
That's the idea and our main guide here is the book Software Engineering for Games in Serious Contests Theories, Methods, tools and Experiences. It's edited by Kendrick Cooper and Antonio Book your Own.
And what's really impressive, I think is the sheer range of voices in this book. It's not just one or two perspectives.
No, not at all. We're talking forty five authors from nine different countries.
Wow. So that gives you a really broad view right on how game design ideas and solid software practices are well coming together in some quite important ways.
Yeah, a really rich picture.
So our mission today is basically to dig into the core concepts, understand the challenges, you know, in using software engineering to build serious games that actually work and keep people hooked.
Okay, sounds good. Where should we start maybe with just how complex these things are to.
Build good idea? Yeah, because it's not just coding, is it. It's inherently interdisciplinary.
Absolutely not just coding. The book really lays it out. You've got inputs from well, the arts, behavioral sciences, business, education, engineering itself, humanities, even physical sciences.
And math, and then domain specific stuff like healthcare too.
It's a real mix, and it's worth thinking about what each brings. Like the arts give you the visuals, the sound, the feel right two D, three D assets, music, even performance informs, animation, and the.
Humanities that brings in things like ethics, game addiction maybe or storytelling from literature, you know, crafting the narrative of the setting.
The behavioral sciences are huge. Anthropology helps with understanding social norms, communication theory for getting the message across, sociology for multiplayer.
Stuff, and psychology for engagement motivation, how rewards work the whole gameplay experience.
Really exactly. It underlines that building a good serious game rests on this incredibly diverse base of knowledge, which.
You know, must make collaboration tricky sometimes compared to more traditional software projects. You need different ways to talk to each other.
Yeah, communication is key and connecting it back. This complexity shows why having strong software engineering practices underneath it all is so vital. You're juggling so much.
Okay, so complex landscape check. But the influence of games goes beyond just making serious apps look like games, doesn't it.
Oh definitely. The book talks about game full engineering and gamification, but applied within software engineering itself.
Right. I saw that table table one point one. It lists areas like requirements engineering, architecture testing, even project management and QA getting gamified.
It's kind of surprising, isn't it, seeing these, you know, seemingly fun techniques popping up in the core processes of building software.
Yeah, really is.
And there's this idea of directly transferring strategies. Chapter Sun by Vanisawanik and her colleagues digs into this. They look at it from three angles. Yeah, player agency giving players meaningful choices. Then there's serious game modification or modding, and also the role of AI emotion modeling to design better games.
Interesting. So the stuff that makes regular games fun and Sticky isn't just for entertainment anymore. It's being used strategically, even behind the scenes in software.
Creation, which naturally leads to the question, how do we know if these serious games are actually well working? How do you measure the user experience effectively?
Ah good transition that brings us to chapter two by soteririst kre Genus. He points out that game ux depends on things.
Like flow, that feeling of being totally absorbed.
Right, an immersion, maybe frustration or tension with the right amount of psychological absorption, and even the social setting.
And there's this breakdown by Roto into three phases of experience, what you expect beforehand, what happens during play, and the overall feeling after and he argues.
The during interaction part is the most crucial for making improvements.
Right exactly, And then Lalaman adds more detail to that during phase. He says it's influenced by the person playing, their mood, motivation.
Needs, okay, the human factor.
Yep, then the system itself, it's complexity, purpose, usability, and finally the context of the environment, whether it feels meaningful voluntary.
So loads of factors how do they suggest evaluating all this, especially if you want quick and thorough insights.
Well, the book covers several methods. There's expert evaluation, basically getting designers or ux pros to give feedback. It's usually quick, cost effective.
Class mentions two ways for that, right, experts as evaluators or evaluators supervising experts.
's then you've got focus groups, more qualitative, subjective, good for exploring perceptions, figuring out what questions to ask, finding unexpected.
Issues, kind of exploratory.
Yeah, and surveys. Of course, you can gather objective stuff like demographics, playtime, and subjective things like attitudes, emotions.
Pretty standard but useful, especially combined with other data like game analytics.
Maybe absolutely. Surveys are easy to deploy and can be powerful when triangulated. But then it gets really interesting. They talk about physiological.
Measurements ah, okay, like biometrics.
Sort of electrodermal activity or EDA that measures sweat gland activity basically indicating arousal or mental effort, hum.
And cardiovascular heart rate YEP, heart rate variability and so on. Gives insight into overall physical and emotional response, and even electro biography EMG which measures muscle tension.
Muscle tension, what does that tell you?
It can indicate things like anxiety or even cognitive load or strain.
Wow, So really getting under the hood of the player's experience.
Definitely, And it's important to distincish between this objective data from machines and subjective data what the player reports.
Right, there's always a trade off there in reliability versus richness.
Perhaps kind of ideally you want both. The book also talks about formative versus summative evaluation. Okay, formative is during development, right to guide design like think clouds or interviews.
Exactly, it's more dialogue based. Summative is at the end assessing overall performance against goals, often using those psychophysiological measures or self assessments.
Got it now, shifting gears a bit? How about assessing what someone's learning in the game, but without stopping the game for a test.
Ah, Yes, that's the idea behind stealth assessment.
From chapter three. Sealth assessment sounds intriguing. It is the.
Goal is to measure those harder to pin down twenty first century skills, creativity systems, thinking, persistence, things traditional tests often miss.
And crucially, without breaking that flow that engagement we talked about precisely.
It provides feedback without interrupting the experience.
And they link this to something called CAGE, content agnostic game engineering.
Right. CAGE tries to solve a common problem with older serious games. They were too tied to specific educational content, made them hard to reuse sometimes less engaging.
So CAGE aims to make the core game mechanics and the assessment reusable across different subjects.
That's the goal. Build the assessment into the framework in a content agnostic way so you can potentially reuse it, maybe speed up development for new topics.
How do they actually do that stealth assessment within CAGE?
They mentioned methods like mouse tracking, analyzing cursor movements, and emotion tracking software things like Visa, just DK or aft dex which look at facial.
Expressions facial expressions yea to detect boredom or frustration.
Yeah potentially, though with mouse tracking there's a performance consideration. Track too much and it might slow the.
Game down, a classic trade off.
All this data stealth or not feeds into educational data mining EDM looking for patterns.
Okay, and they also mentioned indogenous versus exogenous games.
Ah, Right, Indogenous means the game mechanics and the learning content are tightly linked, intrinsically connected. Exogenous is where they're separate, Like playing a fun game than getting a quiz.
Old chocolate covered broccoli problem kind of.
Yeah. Endogenous games generally lead to better learning and more enjoyment. It feels more integrated.
Makes sense. And this all feeds into a.
Student model exactly. It's like a profile that accumulates data, emotions, skills, knowledge based on how the player interacts. It helps understand behavior, give feedback, and this is key. Dynamically adapt the.
Game, adapt the game on the fly, Yeah.
To try and create an optimal learning zone, keep it challenging but not frustrating. The book mentions validation work on cage showing benefits for reuse and speed, and that adaptation helps with boredom, though maybe less so for learners who are already knew the material well. They even use dynamic basion networks for the adaptation part.
Complex stuff. Yeah, Okay, let's move from the player experience and assessment to the actual bones of the game. The software architecture Chapter four right.
Chapter four tackles architecture, especially for mobile serious.
Games, mobile ads, constraints, doesn't it memory processing.
Power definitely, So how you structure the game the components, how they talk to each other is critical for handling data efficiently and distributing resource as well. Design patterns can really help here.
So thinking about non functional requirements from the start, things like complexity, coupling, performance, reusability absolutely crucial.
Interestingly, research by Mizutani mentioned in the chapter suggests that while data driven design is common, things like test driven development and applying standard design patterns are less frequent in serious games than you might expect.
HU wonder why that is.
Could be various reasons, maybe team composition, project pressures. The chapter uses an example Android game to illustrate points, even mentioning how UI color choices might impact usability like blue.
For calm, subtle things matter. How do they suggest evaluating the architecture itself?
They look at metrics complexity like cyclomatic complexity CCM and coupling between objects CBO. Basically how complicated and interconnected the code is?
Okay, standard software metrics yeah, and.
They use statistical correlation Spearman correlation to see how these relate and to derive a sort of quality factor for code paths. And the main point is evaluating the design early saves time and money later by catching potential problems like defects and helping prioritize testing efforts.
Makes sense, find problems before they get baked in. Did they find patterns helped?
Yes, they found that implementing patterns like wrapper and MBC tended to result in code sequences with better quality scores based on those complexity and coupling metrics.
Okay, so good architecture matters. Now let's look at a very specific, quite high stakes application Chapter five and this en Trust platform.
Ah. Yes, en Trust. This is a serious game platform specifically designed for assessing clinical decision making in medical residents.
Right, judging their readiness for entrustment, basically being trusted to perform tasks independently.
Exactly. Medical education is shifting more towards competency based assessment. It's not just did you pass the test? But can you actually do this safely and effectively?
And traditional observation has limits, right, depends on the case that they happen to see faculty time potential bias.
Precisely, Virtual patient simulations like interruest offer objectivity, reproducibility. You can create standardized scenarios measure performance consistently potentially reduced bias.
What does en trust actually do? What are its features?
It's pretty comprehensive. There's a case library for creating scenarios, specifying how interventions affect vials, scoring actions, a virtual patient generation tool for customization age, sex, BMI, even physical abnormalities.
Really detailed customization for diversity.
Yeah, you can set starting vital signs and choose algorithms for how they change realistically, like simulating shock or sepsis. Specify physical exam findings, what orders are available labs, meds procedures and their outcomes and scoring effects.
It's very granular, sounds powerful. Did they test it? Was there a pilot study?
Yes, and the pilot study showed promising results. It was able to discriminate between surgical residents at different training levels PGI levels when assessing their management of inguinal hernias.
So higher PGI level generally meant better score on entrust.
That's what they found a positive correlation. It also seemed to capture performance on critical decision points in complex cases like strangulated hernias.
So it suggests feasibility and some initial validity for using this kind of platform for objective assessment, potentially informing those crucial entrustment decisions exactly.
It's a really interesting application.
Building on that idea of assessment and adaptation. Chapter six dies into using machine learning mL in serious games.
Right, this is a growing trend using mL not just for assessment like we touched on with Stealth assessment, but also to automatically adapt the game.
Itself, personalizing the experience basically to reduce frustration keep players engaged.
That's the aim. The chapter lays out a kind of general four phase methodology if you want it to take an existing non adaptive game and add mL to it.
Okay, what are the steps?
First, identify a game that could actually benefit from adaptation. Second, model the gameplay tasks and how they relate to the learner's ability or the competencies you're targeting.
Makes sense, Know what you want to adapt and why.
Third is the technical part, Actually build the mL models and integrate them into the game's code. And fourth, critically evaluate if the adaptive version is actually better than the original does it improve learning or engagement?
That evaluation part sounds crucial, But are their challenges in doing this?
Oh?
Yes?
The chapter highlights five common ones first, selecting the right data to feel the mL model garbage in.
Garbage out right, always it Else.
Choosing which game elements to adapt what makes the most impact. Then there's the.
Cold star problem, like a recommendation engine with no.
History exactly, how do you personalize when you have no data yet? Maybe use existing data or generate synthetic data to start Tricky. Another is figuring out how often to adapt between sessions levels during tasks. There are trade offs in computational cost and impact on.
The player, right, Adapting too much could.
Be jarring and maybe the biggest challenge proving that adaptation actually improves learning, not just engagement. Sometimes more fun doesn't equal more effective education a.
Really important distinction to make. So evaluate both the engagement and the learning value.
Definitely you need evidence for both.
Okay, so we've covered complexity, evaluation, adaptation. Where does the field go from here? What are the big future directions or challenges?
Yeah?
Chapter seven and thirteen touch.
On this, right they do. Chapter seven revisits that idea of t ansfering strategies from entertainment games. They give examples like games tackling serious themes effectively hell Blade and Psychosis as mentioned.
Or those simulator's truck driving PC building that are surprisingly engaging but also teach real world skills sort of exactly.
They also talk about emergent possibilities, how game design can lead to unexpected player experiences, and modding again as a way to modify and transfer serious games. Emphasizing contexts and.
AI emotion modeling comes up again too.
Yes, the idea that understanding player emotion can lead to better design. They introduce this concept of personality vectors for NPC's non player characters.
Personality vectors like giving AI characters more nuanced behavior.
Kind of They give examples like patrolling guards or poker players behaving more realistically. This could be huge for intelligent tutors or health scenarios, making interactions feel more genuine.
That sounds potentially very powerful. What about the grand challenges laid out in chapter thirteen?
Chapter thirteen summarizes several big ones, like the need for software engineering methods, specifically for serious games. We need tailored approaches.
Not just borrowing from general software engineering right.
Also establishing common terminology, a shared ontology, so everyone speaking the same language. Improving reusability and maintainability always a challenge in software, but maybe more so here.
Quality assurance and testing too, I imagine have unique aspects.
Definitely, and developing better ways to monitor and adapt gameplay safely avoiding unwanted side effects, enabling truly player centric design. They stress needing uniform data, logues and iterative design.
Consistent data is key for that monitoring and adaptation.
Absolutely, and finally, a big opportunity creating effective serious games for teaching software engineering itself, using gamification and AI for personalized learning paths for developers.
Turning the tools back on ourselves essentially.
Pretty much So. Yeah, lots of challenges, but also lots of opportunities for research and advancement.
It really sounds like a dynamic field with a lot still to figure out.
No question, these challenges are where the next breakthroughs will likely happen.
Okay, let's try and wrap this up. We've covered a lot of ground in this deep dive. We started with that inherent complexity pulling from so many disciplines.
Right then, how gamification is sneaking into software development itself and the crucial need for solid user experience evaluation using various methods, even physiological ones.
We looked at stealth assessment measuring learning without interrupting flow and dynamic adaptation using machine learning, plus the architectural nuts and bolts, especially for mobile.
Don't forget the specific applications like the Entrust platform for assessing medical decision making showing real world impact true.
And finally, we touched on those future directions transferring ideas from entertainment, AI personality, and the big challenges ahead like specialized methodologies and better adaptation.
It's clear that blending rigorous engineering with engaging game design gives us some incredibly powerful tools for education, training, maybe much more so.
The final thought for you, the listener, think about those principles of engagement and motivation from games. How could they apply elsewhere in your life? Maybe in learning something new, or even making mundane tasks feel a bit more well game like and rewarding.
What game mechanics could make tricky things more approachable? Definitely some interesting food for thought there.
Absolutely If this sparked your interest, do check out the book Software Engineering for Games and Serious Contexts and the related research for a much deeper look. Thanks for joining us on the deep dive.