(I was out while all of you replied, so I didn't have a chance to reply individually).

I would not make your first game a procedural generated game. If you lack any experience with game design, you will create a game that's bound to fail. Sandbox and Procedural generation will not survive on it's own. Sure it sells well enough, because players are trying to find something that looks different. But when people start noticing that there's not much meat to it, the popularity of it's game will just run face first into an indestructible wall.

Also, what good is a massive land, if there's nothing there to catch the player's eye? This is another issue I find a lot of indie games are not actually thinking about. They make these open world survival games, but these enviorments are unimaginative, left completely up to a random generator, or slopped together. but the enviorment tells no story, shows nothing about what this land was like in the past. Has nothing that's forboding about it. No giant skeletons of some massive creature. No corpse of a massive creature that got mangled by something. etc.

No Ruins. No civilization. The animals don't even behave like animals, they're just interested in running their foreheads into you repeated. The list goes on.

It's like no one sat here and thought about what made Fallout and Elder Scrolls stand out so much. The worlds were alive, and were carefully hand crafted. Every square inch told some story about the land. And there were such miniscule details that were put in, that it added to the greater picture. Skeletal hand sticking out of the muck holding a sword. a bunch of skeletons piled up against the door, many of which have shattered fist. Skeletons lying in bed and embracing each other. A child's skeleton with a pillow over it's head.

Massive ruins embedded into the side of a mountain, with both ancient and newer items inside it. Burnt out torches, and freshly lit brazier's.

Just... trust me on this. This project should not be the first one you do.

It's actually not going to be my first project. I actually plan on getting this developed either through a studio I joined, a studio I found, or simply outsourcing it to another studio (the game is actually part of a larger franchise that starts with a series of original English-language light novels that I am currently writing. If it does get famous, at any time, I could have an adaption made).

That being said, it's still a far-off project, though unlike every other time I actually have a plan to get this made.

Mind telling me the specific reason?

In term of budget estimation you are way too optimist, as says Tangletail.

• Budget: for a game with the graphical fidelity of FFXI on a high-definition emulator, a total size (from all the game worlds added up) about 43 square miles big (30 of said 43 square miles is procedurally generated), partial voice acting (mostly for main characters and special scenes), and a customization and character progression system leagues deeper than any TES title, is this a realistic estimate for how much the game will cost to develop?

For your expectaction count at least a medium sized team (so 20-30 people) multiply it by an average salary of 2500 euros (average salary of a programmer) to have a (really) rough estimation of the cost, by month, of your game. And this is only the salary of the people, without taking into account the tools and the marketing. So yes, $100,000 may be a little light.

• Day/Night cycle: I believe a common complaint for this idea is that some will feel as if this is like playing two games at once. Even though I plan to elaborate on this later, do any of you think this way currently?

This is a good idea, in my opinion. If you manage to make the feel of day and night really different and not just a loss of luminosity, it is an interesting feature.

• Dialogue system: simply put, Is in interesting or refreshing?

Interesting? Not much, how can you talk about thing not related to environment? Refreshing, not really. The dialog mechanic is not as important as the actual dialog quality (except if you have a good AI which can talk inteligently (use different dialog lines depending on what happened, his status, player's reputation, etc...)

• game world: created due to budget constraints, the game world is not what makes this game a sandbox RPG (that would be the mechanical system, to be discussed in a future post). that being said, does anyone like this hub-network idea?

If you have a well crafted hub, and the other area are some kind of dungeon, it can excuse the lower quality of the procedural area.

• was something to vague or just left out? point it out.

I would like to hear more about the combat system.

The dialogue system actually can include characters (for example: You hover your cursor over the immediate conversation partner and three button icons show up. Press X to flatter. Press B to get serious. Press Y to compliment).

Also, the combat system is going to be in a later thread. I split this one up to avoid information overload for both myself and my readers.

Well, in light of the comments regarding the budget, I do plan to increase it, though it is staying under 900,000$. Costs can be saved through using pre-made assets; given that Unity is going to be used to create Somnion, it would be easy to find these due to that engine's community and level of support.

Something to keep in mind with budgets, is that it's not just salary, but benefits and whatnot as well, the rule of thumb is to usually double the salary to take into account having to pay for healthcare, product licenses and whatnot. At least if it's your own studio, outsourcing you don't have to do that.

And I disagree with only the dialog quality being important, the mechanics used drastically change how the dialog gets implemented and written. The branching dialog for a game like Mass Effect with a custom party would be written very differently than a JRPG with no choices and a static party.

I agree with others that the scope is too large. Assets from the Unity Store only mildly help, you can't rely on it as you end up with inconsistent styles and poly counts, and even assets from the store will probably need tweaking to suit your needs.

The Dreamworld stuff is a nice idea, as you can re-use areas and art with minor tweaks to make it feel new.

And as someone else said, a large sandbox of sameness is not great, there is nothing wrong with making a small sandbox that is amazing and full of fun stuff to play in. (Procedurally generated or otherwise)

Don't forget: No Mans Sky was done by 15 people. 15 highly talented people most probably working for 100k$ per year EACH. They worked 3 years on it, at LEAST. Result might be impressive at first glance, might be a technical achievement even at second or third glance, but clearly is not "GTA V but to a bigger scale" that many people expected when they were fooled by the hype. Its - just - not - possible. Not even with procedural generation. Yes, you get rerolled planets you get to name and whatever. No, those planets are not filled with setpieces the way GTA V's world is.

Then I'll try to control the hype. I did know that I wouldn't get setpiece moments through using procedural generation as well. Plus, Somnion isn't going to advertise itself on scope.

Well, thanks for the marketing strategy! Is there anything else you would like to point out?

Something to keep in mind with budgets, is that it's not just salary, but benefits and whatnot as well, the rule of thumb is to usually double the salary to take into account having to pay for healthcare, product licenses and whatnot. At least if it's your own studio, outsourcing you don't have to do that.

And I disagree with only the dialog quality being important, the mechanics used drastically change how the dialog gets implemented and written. The branching dialog for a game like Mass Effect with a custom party would be written very differently than a JRPG with no choices and a static party.

I agree with others that the scope is too large. Assets from the Unity Store only mildly help, you can't rely on it as you end up with inconsistent styles and poly counts, and even assets from the store will probably need tweaking to suit your needs.

The Dreamworld stuff is a nice idea, as you can re-use areas and art with minor tweaks to make it feel new.

And as someone else said, a large sandbox of sameness is not great, there is nothing wrong with making a small sandbox that is amazing and full of fun stuff to play in. (Procedurally generated or otherwise)

I may have been vague on this, but the game does not use one giant 43 square-mile sandbox as it's game world; Somnion uses several small sandboxes that, when added up, equal 43 square miles of space (including indoor and underground areas).

On the asset front, we won't be using many visual assets. Coding and scripting and maybe the occasional stock sound effect are probably all we would need from my perspective.

The new scope is definately feasible over a longer period of time. Something akin to the scale of Psyconauts, it seems, with a bit less content.

Esmurdoc stated that Psychonauts took about 4.5 years to complete — though that without all the complications the real development time was closer to 2 years — with a team of 42 full-time developers and additional contractors, with a final budget of $11.5 million

So you figure with 1/2 the content, simpler mechanics, and creating the basic maps with Terragen or something simliar, it should be doable over the span of 6-7 years with fewer people.

If the goal (as it seems to be) is to produce novels to fund this, it's not an impossible goal, but I'd worry about releasing great novels ahead of planning for the game, as the book market can be pretty tough from what I understand.

The funding is not going to come directly from novels actually.

It actually shares the same universe as the novels I mentioned, though Somnion: Darkest Dreams is actually a side story to the Somnion light novel series (despite taking place at around the same time).

The procedural generation would be used to fill in the generic content. Specific story content is very much harder (so that big companies really cannot do it much) to do. Random simple encounters (preferably a wide range of combinations) can employ some procedural techniques - but again needs to be SIMPLE lest you get overwhelmed by the rule logic a complex generator needs.

Also 43 square miles is fine (~6x6 miles) but its the small settings and unique situations which make the game (too many big games have terrain that is "pretty desert" with little of interest in it after youve seen it once. The traversal issue is also there, as to how fast you move versus distances will eat player time, and after becoming routine then becomes tedious and boring.

So most of the effort needs to be the story/theme related details and situations the player must encounter and 'handle' using the tools/resources they muster. However many play hours you design the game for you need a sufficient number of 'problems' the player must solve to advance (even being 'sandbox' it still needs progression to give the player a fealing of making progress).

Those situations being intricate generally have to be largely hand-crafted to be more than the 'simple encounters' you might use as filler.

As to the $100000 budget, that might be fine for a text adventure, but anything with assets - even just lots of static pictures to go with dialogue interactions - eats man-hours like crazy, as can story trees/networks of shifting context (and the more intricate it gets, the more testing and fixing becomes required to make the whole story system consistent and cohesive.)

This topic is so huge in my mind, I keep wanting to comment on it, but I'm not sure what to say.

Q: do any of you think it is an interesting, more involved mechanic that allows a player to have a greater effect on the game world, or are there still those who think that playing Somnion is still playing two games at once?

It's all about execution here. How well do the two mechanics integrate. A great example of two differing mechanics integrating well is a game like Splatoon that splits time between platformer-esque swimming and third person shooting expertly. As long as the mechanics are complementary, that is, one helps the other and vice versa, then they're just two modes of the same game. If changing things in dreams changes things in real life, then what you do in real life should enable you access to more dreams. Or vice versa, either way.

Q: does anyone like this hub-network idea?

Sounds awesome. Many games have done well with that. You're walking down a well trodden path with lots of great examples to look up to.

Q: was something to vague or just left out? point it out.

How does being targeted at tweens affect the design of your game as opposed to targeting teenagers or adults?
Why does the game world need to be so large? You've shared your budget, I've yet to see one person handcraft 13 square miles of content. Why not squeeze all your content close together? Would that remove anything valuable from your game?

Well, you pretty much nailed the the dream-world/real-world dynamic on the head there. What you do in the real world does unlock new areas of the dream world, and you can change aspects of the real world (namely electronics or people's behavior for the most part) by performing actions in the dream world.

As for the target audience, a concept change in the novels has rendered the entire Somnion universe (Both LNs and the game) to be more teen-oriented. That being said, most of the main characters are tweens and early teens.

For the size of the game world...I really don't know if downsizing the world area would remove anything from the experience or not. It's a good decision to ponder. That being said, the 13 hardcoded square miles are just the real world plus the starting areas of the dream world zones combined. the dream world zones are completely seperate from one-another; the 43 square miles are all the zones plus the real world added up.