Again, handling in GT is classified as "simcade". I'm not going to discuss that with you because there are people more knowledgeable than me who have already given their perspective.

About the other (more objective) points:

• damage is important, because professional drivers do have crashes;
• TOCA games sold much less than GT (hence, less money to invest) and had 50-80 real licensed tracks, so no excuses for GT;
• money is fictional in all games that have money; TOCA, PCARS and most PC simulators do not have money; career mode is scheduled-based and you progress by being hired by racing teams.

Finally, just for fun, here's a video comparison between GT5 and Grid: https://www.youtube.com/watch?v=PvAb2Vg949k

Enjoy!

PS: I'm not a GT hater. I played it a lot in the PS2 era. I'm just frustrated about many of its features. But I know it still has its place.

A little bit far-fetched. ;)

"simcade" perhaps compared to iRace or something on that level, but certainly not against Project Cars.

Damage is important, but i seriously doubt any simulator really simulates even a farcry of the impact of the damage in the handling compared to real world situation, so it's kinda pointless complain about it. And unless we are talking about minor damage, no one would keep driving the car after damage in the real world.

GT have a lot of real world tracks and fictional.

How many cars does the other games have and allow you to exchange? If you preffer to be stuck on a carreer determing your car for several races, your choice.

At least it's very specific compared too 'adventure'. Your definition of adventure was this: "I would define adventure as a series of events that happen in a context far from our normal daily lives, but which don't involve much interaction with other people (otherwise, it would be crime/war/horror). "

 adventure as a series of events that happen

I think you got it right there. This is how most people would define it, but it's way to broad to be really useful.

Our 'normal' daily lives
Well everyone has their own 'normal' daily life. Based on this alone you'll get different interpretations. That's not what you want when organizing video games in a unified way. That's why I believe that definition is not desirable. 

Here's proof of different interpretations:

http://store.steampowered.com/tag/en/Adventure/#p=0&tab=NewReleases

Did you know that 4.201 out of 10.162 games on steam are tagged adventure.

Not what's said on the links I've given you.

What's said there is that PCARS's objective is to simulate. Whether they can do it as accurately as iRacing or rFactor is a question of the amount of resources they had (it's kind of an indy game) and the number of things they wanted to include in the game (they had other aspects to develop like dynamic weather, a good number of championships and tracks, etc.).

GT is different. It's the most expensive racing game of all time (GT5 alone has cost 80M€), so resources were not a problem, even considering the amount of content they put in the game. Dynamic weather, damages and proper sound were clearly missing, as their focus was always the handling physics.

When you have games like F1 96 where your car was affected differently depending on what part was affected (front wind, rear wind, etc.), it's not pointless. Even if we're far from simulating 100% of the impact, having a reasonable approximation is not the same as nothing.

If it's minor damage, they continue. If not, they're going to the pit stop. In simulation games, that's what happens.

TOCA RD1 had 50 real tracks, while GT3 had only 2. Only then they started to increase. My point is: simulation was never their priority.

But that's simulation! What GT does is compromising simulation in favour of more fun/entertainment. I'm not saying it's a bad thing. It's what they believe is best for their game. It's their philosophy. It's "simcade"!

I think you are confounding car handling simulation with "life simulation", "carreer simulation", "real track simulation", "sound simulation". Those aren't even needed for the first to be good.

Project Cars can be said to have a better simulation on carreer or sound? Perhaps, but their despiction of the real tracks isn't that accurated for my taste, the car handling is bad and even their damage isn't very well done.

And having the car handling different depending of the damage, but being nowhere near the real life effect is pointless, that would put closer to arcade than to sim, because it's just a feature (or if Carmaggedon is more of a sim because cars get destroyed?).

If you want to have a real simulation on GT for damage, after you get any major bump just hit "start" and use the option forfeit race, because in real life that is what would happen, you would end up losing.

Simulation was always a priority, maybe not on what you think simulation is, but on the handling of the car and number of different cars.

For you Flight Simulator to be a propper simulator would need board service, cabin communication, passengers complaining of stuffs and whatnot or how to properly fly an aircraft? You are really confunding DRIVING SIMULATOR (that is what is described on the box even) with a lot of other things that don't really are necessary for the simulation.

Right, they have their definition, which doesn't mean that I cannot have mine.

In my definition, for instance GTA5 and Mirror's Edge would not be classified as Adventure, but as Crime. So, I wouldn't have the problem Steam has with nearly 50% of their games being tagged Adventure.

As of the "normal daily life", of course it's subjective. Most things are. For instance, what's Horror for me may not be for you. What we need to do is to look for objective criteria that help us making it as objective as possible, like: "events that happen to less than 10% of people" or "less than once a year" or something similar. My goal with this Taxonomy is not to go into such detail. I just want to have a basic framework that proves to be useful.

My focus is not even the categories, but the dimensions. And I admit Adventure may not be a good category. However, without it I wouldn't know how to classify (in the Theme dimension) games like Crash Bandicoot, Sonic, Tarzan, etc.

Ofcourse you can have your own personal definitions. This works great if you want a basic framework that proves useful to yourself. 
Just keep in mind that your definitions might not be so useful for most gamers. It gets complicated really quickly. I would like it to be universally appealing as a basic framework for Voice commands.

My basic requirement is "If it doesn't make sense with voice commands it's probably not good enough" +  "dimension/category SHOULD be one word otherwise it's not clearly defined. 

Wiki

• taxonomy - a formal list of concepts, denoted by controlled words or phrases, arranged from abstract to specific, related by subtype-supertype relations or by superset-subset relations.

These three lines could be helpful

What you play is about the game. (Game Classification) <-This thread
How you play is about hardware. (Hardware compatibility)
Why you play is about experience. (MDA framework)

A Interaction/ Controlls 

1 Physical Input (Digital, Analog, Motion, Neural)
I feel like your examples should be categories. Lightguns have triggers and buttons, they inherit all thise three values. Controllers are also digital, analog, and motion (PS3 sixaxis). Maybe this should be in hardware compatibility?

2 input time / space
This one is named after continuous time/space. While it should be a supertype of continuous/discreet. Maybe call it interaction type?

3 interaction time
This one feels the same as 2, it only has a different perspective.  Real-time or turn-based (synchronous and asynchronos gameplay) that's what you mean right?

 B Environment
4 Space dimimensions
I suggested Viewpoint since it's more descriptive and you use the 'space' word a lot which makes it confusing. There is some interestion stuff in the categories. It's not about spacial dimensions but camera perspective. Look at your own notes.

5 Space Connection
Open World and Closed Levels could refer to loading times as immersion breaking events while playing a game. 
This is more about level design it woud fit Content better them environment. This way environment can be changed to viewpoint.

C Structure/Rules
6 Single player, multiplayer co-op competative doesn't gamemode fit better than players? It also deserves to be a dimension 'C' and not a sub-dimension '6'. It's already a best practice in the industry (give more credit, since it's pretty useful)

7 Behavior is about AI? Or is it about story i'm not sure. Would be nice if we could compare smartness of Game-AI. Unreal vs Quake Bots or CoD vs Battlefield.

8 Ah is Scripted story progression? Because then you can drop non-scripted and add scripted right here. Progression is meaningful enough to be it's own dimension?

Again, great hints! Thank you for the time you've put into this. Let me now go step by step.

I agree with you. It should make sense with voice commands. But that's not the only, and probably not the most important, requirement. In some cases, we may have to compromise it in favour of a more accurate, better organized and/or more comprehensive classification.

Usecases for end users
filtering your own library to quickly find what you want to play.
new way to organize a storefront to quickly find what you want to buy.

Usecases for industry professionals
Exploring possibilities ideation

Both groups have different backgrounds so one basic framework that's understandable for both would be impossible.
It seems like you are leaning more to wards a taxomony for industry proffesionals, while I tend to think more about end-users.

The question we were discussing was: between Adventure and Mithology which was a better term for a game like Tomb Raider. And honestly, I believe the former is. Another question was: does the Steam classification make sense to people? Do gamers see GTA as an adventure game? Again, I don't think so.

Ancient Mythology is more specific than adventure. When you filter games based on Adventure you'll get a long list of games. Filtering based on Ancient Mythology would result in a shortlist of results. If you like the theme of TombRaider, then I would recommend Uncharted or Indiana Jones. Crash Bandicoot should not be a recommendation at all. Adventure is not wrong, it's just not enough.

Steam classification makes sense to gamers, since it's user generated by gamers. It's not perfect because a tag could be a category or a sub-sub category. I would like to see color-coded tags based on 5 - 7 dimensions.
GTA is hard to classify because it mixes many genres. But the theme of the game is pretty clear, which is crime/mafia/gangster theme. 

About dimension/category, these are different concepts. Dimension is the characteristic you are looking at when classifying (e.g. interaction time). Category is the value that you assign a given game (e.g. turn-based). In a biological classification, the dimensions are the ranks (kingdom, phylum, class, etc.) and the categories are the instances those ranks can assume (animalia, chordata, mammalia, etc.). But this one is a hierarchical taxonomy (which is what is described in your citation). Taxonomies do not need to be of that fashion ( https://en.wikipedia.org/wiki/Taxonomy_(general) ). Mine is not. But do you think it should? Why?

It shouldn't the wiki link was helpful for me with superset and subset.

That's an interesting perspective. I kind of mixed the three in the Taxonomy. Dimension 1 (Physical input) is more related to hardware and Dimension 10 (Purpose) is more about why you play. Separating these three lines is something to be considered.

Hopefully you can get a clearer picture by using these Master-dimensions, because it's a little mixed up. Those pillars are atleast easy to grasp. You could also say 1. Software 2. Hardware 3. User. 

But I still have questions about this. For instance, the physical input (and output - WiiU has changed a bit this one too) is not just about hardware, but the game (or at least the operating system where the game runs on) needs also to be designed accordingly.

Games are about interaction, that makes it hard to draw a line between hardware and software. INput could be hardware Output could be software. Wii U controller / VR HMD are in both.

About the MDA framework, I think mechanics and dynamics is already included in the Taxonomy. Aesthetics is a higher level dimension, which should result from the building blocks of the lower level dimensions of this Taxonomy.

 I said MDA framework, while I meant just Aestetics. This topic is about psychology, and very subjective when applied to individual games. I don't feel like exploring (aestetics/user) any further. (too complicated) 

1) That would be a way of looking at it. However, there would be a lot of categories and they would result from combinations of the current ones. Moreover, lightguns and dualshock 3 would be separate categories, although having the same characteristics (digital, analog and motion). I like to break things in their building blocks, so prefer to keep it this way. Or another possibility would be to turn these categories (digital, analog, etc.) into dimensions.  The categories would all be binary though. In the way it is right now, I understand the issue you're pointing. Still, we could overcome it by considering that some games can be hybrids (digital + analog) or that the categories are nested (analog includes digital, motion includes digital and analog, etc.).

You don't need to know the innerworkings of a car to drive a car. I feel like it's not really useful for gamers, It might very wel be for industry professionals. I want to know whether I can play Duckhunt with a Lightgun Not that it's (Digital + Motion). I believe that trigger was just digital (not sure).

Are thumbssticks considered analog, because ever modern would become a hybrid.

2) The problem with "interaction type" is that it could represent different things. I know my organization is not completely clear. But the idea was to distinguish first between supertypes, where discrete would correspond to discrete space and time, and continuous could be further divided.

Maybe interaction intensity?

3) Do you have another suggestion for the name of the dimension? About the categories, it's exactly that! Those names are probably more scientific, but the ones I used were more the terminology used by gamers.

Maybe interaction timing

4) You're right! Probably camera perspective or viewpoint are better names.

5) I agree it's about level design, like viewpoint is also about level design. And it's not just about loading times. Imagine that in Super Mario 64, the loading times were reduced to 0. That wouldn't make it open world, because each level would still be accessed by a single portal. On the other hand, GTA Vice City has loading times, but it is still open world, because there's a continuous interface between both islands.

6) I don't know. Gamemode is also too generic. For instance, in a racing game, gamemode can be arcade, career, time attack, etc. But I agree this dimension deserves a lot of credit.

Gamemode is a bit confusing, because it describes stuff like career, deathmatch, story mode etc. Should be something different. That fits singleplayer, multiplayer. Co-op. How about playmode? How you play with others if at all.

7) AI is a very important aspect and maybe it deserves a separate dimension. But this one wasn't just about AI. It's about everything that makes a game more or less scripted, so it includes AI, randomness (e.g. casino roulette) and how sequential events depend on each other (less dependency = more scripted).

Procedural Generation is where AI pretty much dictates everthing about the game. It's a trend, should be includded somewhere.

8) No, a scripted story game would correspond to mission-based (I don't know if it's the best name), while non-linear, collaborative story (RPG) games would correspond to skill-based. This is not the same as (although it may be related to) the scripted/non-scripted behaviour of the previous dimension. That one is about what happens inside the mission/match. This one is about how missions/matches are structured in the game.

You gave a lot of inputs, and I still have to think about some of them. Meanwhile, if you have more ideas or comments on these questions, please let me know! Thanks!