Full Circle

Well. I guess it's time for something of a more direct update.

You can always rummage through this trash if you want some additional context.

Hmm. Where to begin? Ah yes. This year has sucked. Well, just about every year sucks for me. But this one particularly sucks.

Getting pneumonia in November and nearly suffocating on my own fluids was delightful all in itself, but as it turns out this would only be the first of my worries. One of my cats, Snowball, developed a perpetual eye ulcer in December that, to this day, has proven to be a nightmare straight from the pits of heck to deal with.

Sparing you the pages-long and depressing story of how I lose my life savings to vet bills and have had very little time to sleep in the last months, the cat is finally somewhat stable and will remain on drugs for the rest of her life. I'm not sure how or why cats and herpes seem like old lovers, but it's a real darn tootin shame every single one is destined for some kind of problem with it. Throw in old age and a truly shitty dust-filled house to live in, and problems are bound to occur. I just didn't think they would be this severe or troublesome to deal with.

The eye has a red spot on it currently, known as a Sequestrum, that is effectively going to take months to years to heal, if it ever heals at all. The eye surface has been sensitive enough that simply touching it causes it to just come right off. Between confusion, extensive lengths of time wasted, more confusion, and then finally proper drug treatment, the surface is mostly under control, but not before significant damage was done. The surface is still a problem, but not as much as it was. The cat wears a cone and has worn a cone for many months. Since it's a Persian, that means I have to take the cone off for every feeding. Of course, at times I was also rolling 6+ drugs a day at precisely timed intervals day and night. The drugs are down to 2 now, and the time has come to start weening her off the cone. Yes, the very process of simply taking off the cone has proven to be troublesome, as the cat's insatiable lust to clean itself leads to it rubbing its face and thus fucking up the eye again. It has happened probably a dozen times by now, often times without me being able to stop it.

Ah, the stress and depression. Good bed fellows of mine. From scares of leukemia to potentially paying over $500 a trip for a taxi to an eye specialist 45 minutes away, the road has been as long and winding as a black man's undulating shaft. Very much a reminder of my bloody past, a reminder I certainly don't need, as I am constantly stuck in memory of those days regardless.

The death of Retribution was certainly a blow as well. Spending all that time working on my last effort to redeem all of my wasted years of life, and then being unable to actually do anything with it plainly due to incompetence, was a conclusion I saw coming from miles away. After all, I haven't done anything meaningful since 2003.

These and other circumstances lead me full circle to fucking around with Starcraft 2, one of the most hostile environments for custom content I can think of. From the suffocating limits of battle.net 0.2 to the totally undocumented functionality of the game, to the performance problems and pipeline mishaps, this game has always been the black sheep when it comes to considerations for a project. Alas, HKS has been pestering me for some time, claiming he would be willing to do work for such a project. With the memory of such promises leading Retribution, and many other projects, on a suicidal path, you may understand my reluctance.

It is not that I am unfamiliar with Starcraft 2, it's that I hate it. You might remember my blogs of yore, detailing such ventures.

First and foremost, Starcraft 2 hasn't had a way to put assets into the game in a feasible manner for years. Only with the 3ds max exporter that Blizzard clung onto with a death grip until virtually all artists had lost interest could you actually hope to get good results. The existing third-party "tool" is, at best, a proof of concept. To this day, that ancient plugin remains the only importer, which happily fucks up everything it touches because it was built in an era of necessity and the creator abandoned it very quickly. Deja vu, neh? Yes, this is actually how Brood War was in a lot of ways. Outdated, incomplete tools that are used out of necessity because no one existed to make ones that weren't a trial by prickly dicks to deal with.

Nonetheless, like any non-artist, I was willing to try my luck regardless. I've dealt with worse. Or so I thought.

My original concepts of rebuilding Armageddon Onslaught were delusional. It's impossible inside Starcraft 2. At least, as it was in Brood War. So I decided I would stick to single-player projects, and avoid the whole cockblock of b.ent to begin with. Of course, I road the bandwagon of TPS projects and envisioned a reconstruction of my conceptual Warcraft 3 campaigns in which you played as a hero in an RTS-controlled combat world. Something like dota, but not scripted, and with actual base construction and shit going on.

This, too, was of course impossible. The game's performance is just not good enough. Of course, there's all the bugs and such, but really at the heart of my problems was the lack of custom content support and performance.

Familiar?

That was 2010, and was the last major attempt at a project until 2012, with Apex.

Apex was a much more traditional B&D campaign concept. Something like the Blizzard campaigns, but actually challenging and with gameplay. My overarching goal was something with AI that could defend itself, something that literally doesn't exist in campaign functions. Another key aspect of Apex A's original concept was custom Tactical AI that could give computer units "intelligence", at least enough so they could use spells and micro their units. The first iteration of this was the Raven by JademusSreg, something demonstrated to an extent in the MGEC video covering the project.

However, getting the computer AI to actually work was just not happening. The defense solutions Sreg was able to come up with were just not satisfactory. Despite his mastery over the game and its functions, the mystery of the computer AI remained firmly hidden under Bisu's swarm of corsairs.

Then 1.5 came along and deleted 80% of my work. I probably stated something along the lines of "fuk dat" in a more lengthy and enraged manner. Needless to say, despite the considerable effort placed into research, the project was destined for failure.

Fast forward to near-present day. JademusSreg responds all of once to an inquiry and then decides to just ignore me. HKS, with zero experience in programming, scripting, or whatever, decides to tackle the computer AI conundrum that has held up B&D campaign concepts since I first attempted them. In a week, he has a fully functional example of everything we want.

A deathball that reflects the scale of opponent you could face in a Tutorial map.

Truth be told, the whole idea of using the Campaign AI and then tacking functionality onto it was the root of the issues. Through months of intense testing and research, I came to understand Starcraft 2, particularly Galaxy and its hardcoded AI-related functions, as being very unexplored and, in many cases, broken. Starcraft 2 actually leaks AI time, at a rate of 33% every 10 minutes, making everything reliant on AI time slower and slower as the game goes on. Not to mention that, from what I understand, Galaxy as a language is incredibly limited.

Despite this, the new AI framework as presented by HKS is leaps and bounds of anything ever conjured in my previous conceptuals. The AI expands, defends, builds structures as its expansions semi-reliably (you have no idea how hard this is to do without manually triggering it, and to be honest, fuck that), and attacks with enormous 300 supply armies.

Of course, it has its problems. The AI can't path very well, and armies tend to be strung out in this slow-moving congo line. Their persistence in attacking needs to be far more aggressive, and their defense needs to happen faster. Playing with numbers here and there has lead to increased responsiveness, but without knowing Galaxy, and without any documentation at all, progressing past here is slow. Of course, if I ask my "friends" in Blizzard about a Galaxy API, I am met with total silence. I firmly believe that the melee computer AI in Starcraft 2 is the work of a single intern and no one else in the company actually knows how it works. If they did, they wouldn't use the stripped-down editor triggers for their campaign.

I've found traces of posts about sc2 melee AI projects back in WoL, but no actual documentation, no information, and certainly no useful code. My inquiries are only met with hopelessness. Where every question has only been met with stout confusion and "no one knows how it works" as an excuse, it seems this AI may very well be the only of its kind to work as well as it does in a campaign environment.

The biggest, most severe blockade of a potential Starcraft 2 project is essentially gone.

What of the Art tools, though? Do they solve my issues with the previous option?

In a word, yes. The art tools have a lot of bugs, but they are mostly pipeline-related, and I've been able to get past them to reasonable success.

My attempts to get a physics-enabled dress were far less than successful, so this was one of many things I was forced to abandon. While the one side of the dress is rigged and functional, so much as introducing any other kind of physics object to the scene, even if it's nowhere near those ones, causes everything to enter SSJ and flip the fuck out. I tried to bake physics animations into it, but be damned if I can figure out how to get those to work properly, either...

I can get both Tera and Lineage 2 models into the game with relative ease. I can make normal maps, speculars, whatever for Lineage 2 and WoW models without any real problems. Mindtex, a normal map generator that showed up on steam, has consistently produced superior results compared to nDo for me.

All of the particle-related assets from Retribution, except for the particle files themselves, can be translated 1:1 to Starcraft 2. Starcraft 2's particles are much more cull-happy though, so getting them to appear ingame reliably is proving to be troublesome. Nonetheless, I have managed to produce particles comparable to those in Retribution with only a day's worth of effort.

So, what now?

Although I placed much effort into the UDK, it seems that I will never be able to do anything with it. I'm not a programmer, and I'm not an artist. In order to have a chance with a UDK project, I'd need to become both. That won't happen. I'm too old, and I can't learn such complex things as organic modeling, texturing, or C/C++. Just understanding HKS' hacks to the melee code in Starcraft 2 gives me a headache, and that shit is void of all the normal things you see in programming, like arrays.

I am stuck in very much the same situation I was stuck in during my Brood War years. Making graphics is incredibly difficult, and I can't do advanced stuff. Hell, even the data editor in sc2 is going to be a multi-year challenge, given how it takes 5-20 seconds to input individual values and nothing is organized. I figured if I can conquer the hardcode labyrinth that is brood war, I can conquer this.

Perhaps, though, it is too late. I am almost 27 years old, and I feel the weight of time heavily on my shoulders. I've been modding since 1999, and very little has moved forward since I first started. Physically, and mentally, I am exhausted. My environment is volatile and threatens to fall apart at the earliest opportunity, and what then? Only death awaits me.

The concoction of mental stability, neutrality, and focus grows ever more wobbly, like a particularly fat man in a golf cart. I never could focus for shit, and I find it very hard to focus on learning something like Starcraft 2. HKS is pulling weight for the first time since he last put major effort into a Brood War project, which is great, but I still don't know if I can do anything with that work.

For the time being I am starting with terrain. Simple brush work, then doodads. I'm a firm believer that doodad spam is unnecessary. If I need 50 doodads to make something, I am better off ripping a model from something else and using 1 doodad. Performance is a great concern. Sc2 performs terribly. The pathing collision, the AI, mineral shaders, lights, everything is fighting for that single core, and fight they do. Mass doodads is something I must avoid, lest I drench the mountain in even more dwarf vomit. We'll never get any migrants if I can't keep the terraining minimalistic, since every map will be 256x256.

My mission design remains fairly much the same as it was in previous B&D considerations. Multi-base gameplay, multiple players, often times multiple teams. The diversity will come through the total conversion aspects, the AI being able to actually think for itself and not just arbitrarily spawn wave sof shit that a-move into it, and of course the terrain itself. I derive much of my design from what I learned watching Brood War professional play for countless years.

Were such an effort to actually get somewhere, it would of course be a private release. I'm not big on the whole public thing. I would allow viewers to opt-in, and that's as close as I care to get.

Alas, the whole idea of being able to even manage a private video demonstration of a functional mission seems like a distant, impossible dream. I have functional concepts, but actually getting a project far enough I can say, "Hey, look, everything here is working and demonstrable" seems like something I haven't been able to say since 2009. And, even then, it was a half-assed pile of shit.

Added difficulties lay in the fact that custom models rely much more heavily on textures and UV's in an isometric view. My modeling skills are far too limited and juvenile to produce something as simple as some new Terran vehicles, because they need textures I simply cannot give them. The idea of relying on someone else for such a vital task is a two-fold hopeless one - there is no one who can produce my standards of results in consistent and reliable timeframes, and I absolutely despise having to rely on anyone for anything.

Of course, I used to know people who could deliver. They became too good for me, I guess.

I could not tell you what the future holds. I could not tell you what will become of my experiments, or of the cat. I cannot tell myself. I refuse to allow myself to consider such things anymore. Thinking is for nerds. Success in any walk of life for me depends very much on a yoloswag420 kind of approach, I think. Taking anything seriously is where shit starts to go downhill. Maybe that's why Retribution's death was assured.

At the very least, I can waste time on this. Time is there to be wasted, after all.

For better or for worse, I must strike the earth, because that is what I was born to do. Bring on the tentacle demons.

Full Circle

Completed

Ongoing

Upcoming