I'm now back to Shelter engine. I was adding quests to the main document, but the code got cold, and I stopped comprehending it. That was a mistake. I'm now working on "dropping items on the ground" and proper implementation of doors.

Ground items are quite a pain in the ass, because I have to determine when you're supposed to be able to click on them, i.e. when the objects in front of it are transparent so you can see it... among other things.

Then I'll be putting in the first area (Hillside Penitentiary) and quests directly into the game.  Also, the quest document will be merged with Zorlag-quest-editor, because plain-language quest descriptions inevitably differ from their implementation, and I don't have the time to keep them in sync.

It was initially planned to not have intelligence-influenced dialogue, and now I'm back to that idea. Nobody REALLY wants to play dumb ass characters, but people still want to have unique dialogue options each time they play.

This is why instead of INT checks, a lot of dialogue will be dictated by Player's in-game accomplishments (and failures).

I'll see both of you next week.

The quest file grows slowly, as I constantly have to reconcile the more ambitious designs with current (and projected) capabilities of the engine, and take into account economy of resources - for instance, we cannot afford to create an entire map filled with NPCs just for one quest where you nuke it.

Keeping track of the quests (how they interlink) is something I hoped to relegate to Zorlag, the automated quest compiler, but it cannot do so for human-language written quests document. So, tracking quest chains past the "what brain recalls at once" threshold is rather nightmarishly confusing at the moment, but I'll find a solution somehow.

Also, our sprite system will have to become layered, like Diablo2. Otherwise, even by the most conservative estimates, the amount of graphics that we'll have to generate will be unsustainable.

Speaking of graphics, this is a concept art image of the Player that the talented Mr. Bough created last year.

Asides from slowly adding to the quest database, the only things that stand out this week are my attempts to fix and enhance Shelter's renderer. They both failed.

The first was an attempt to fix the game not restoring itself properly after ALT-TAB. Managed textures, reinitializing Direct3D, nothing worked. Conclusion: I will cheat and make Shelter run in a full-screen window. This way you can never completely ALT-TAB out of it, and the game doesn't have to restore itself.

The second was an attempt to integrate pixel shaders into the engine, for a grittier look, a la Left 4 Dead post-processing. The attempt failed miserably, despite me doing everything by the book. Wasted hours looking for a solution, found nothing. Oh, it ran,  but the end result mangled the image in unholy ways.

For instance, this is the result of applying a cookie cutter shader that was supposed to post-process the image into black&white:

So close, yet so far. Gorramit :(

In other news, we're limiting Shelter ambition and scope, saving a number of features for the sequel, should there ever be one. Driveable vehicles is one such feature. Dynamic (reactive) quests may be another casualty. 

Mind you, by the very nature of our quest design, the quests only activate when context appropriate. They don't activate based on "status of other quests", but based on the state which the other quests impacted onto the world.

For instance, when someone needs an item, they need it not because the quest is marked "incomplete", but its the other way around... the quest is marked incomplete because they don't have the item, and will remain so until they have it.

But, I also planned to have reversible quests, and framework for that has been both in and out. Quests where say, you put someone in prison, but then you freed them by doing another quest, which fails the first quest (it becomes red), but then if you put them in prison again, it will all get reversed again, etc.

THESE types of quests may be too far to shoot for right now. The scope of the game has to be maintained, as we're trying to make its release in a feasible future, sometime before a bright light and a series of low concussions render our world obsolete. We're sticking to Fallout scope for Shelter. 

Now, in the sequel, "Shelter II: The Awakening: Origins - Resurrection: Redemption - Wrath Of The Transformed", we'll have driveable vehicles, more weather effects, more ambitious quest system, and more, well, emergent gameplay.

Replayability IS mandatory in a game like this. The player has to wonder what would happen if they used a different character or approached situations in different order, or picked different sides.

Reconciling a linear story arc AND "choice&consequence" is difficult. It is a learning experience in a category of its own. A number of issues have to be overcome, including these:

1) Controlling the player's progression, but loosely, without "hovering over the shoulder" or creating invisible walls in the way other games do.
2) Not having the main arc overshadow the secondary quests and make them seem small in comparison.
3) Not making secondary quests be immediately recognized as secondary quests.
4) Linking secondary quests into the main plot arc without destroying the arc.
5) Allowing for a more-than-superficial breathing room within just the main arc itself.
6) Not making the game easily breakable/unfinisheable due to player doing unforeseen things.
7) Not making the game in a way where player simply avoids certain situations the next time they start over. This makes "major twist endings" a difficult affair, but perhaps your key party NPC's shocking but inevitable betrayal the kind of thing better left for another medium, such as film.

I'm happy to report that the issues in the above list have been, for the most part, resolved. Shelter's design is inching forward with steadiness of a Soviet bulldozer and cleverness of a Soviet bulldozer.

Wait-..

I don't know if you, loyal readers (both of you), caught on to the fact that Shelter used to be Earth-based and post-apocalyptic, until it stopped being so.

Eons ago, we decided to limit Fallout's influence to the gameplay mechanics, and not steal the actual setting itself. There were two reasons:

1) It is very hard to ape Fallout without *aping Fallout*. The game was a quintessence of everything cool you'd expect from a post-apocalyptic world. Hey guys, let's have this cool town full of junked cars... let's call it, I don't know...  SCRAP CITY. And how about a place with caravans, a HUB of sorts, if you will... 

2) Post-apocalyptic is the new black. Post-apocalyptic movies and games are popping out everywhere. There's even a post-apocalyptic TV show, "Life After People". Somehow, the subject is wearing thin.

So...after several (wild) iterations, we've arrived at the Osiris-5, Earth colony, Warp Gate storyline.

During this week Shelter's quest database has grown, and main plot arc is being fleshed out. As in Fallout, the main arc is a mostly linear sequence of events which can be approached in several ways. As of now, it is a combination of both expected and bizarre.

Thus far we've managed to avoid these types of quests:

* killing caveloads of rats and spiders
* bringing someone 15 bear skins to make shirts out of
* bringing a boy his peg leg

... and nobody treats you like their best friend without reason to do so.

Not all stereotypical quest types can be avoided, however, due to inherent nature of such games. People sending you somewhere, and such. All we can do is to make them meaningful, connected to the plot.

Fingers crossed.