Hacker News Reader

Comment by bradford | Hacker News Reader

I've played DF a lot.

Biggest problem that I had is that it eventually becomes a challenge to manage the framerate. I realize there are ways to fix this that are intertwined with the gameplay, but I'd rather play the game instead of butchering kittens and other things in an effort to keep the game performant.

Second issue is with military organization. I never really became confident in my ability to get the squads doing what I wanted them to do (wearing the proper uniforms, training with a crossbow).

Despite this, I've had a lot of fun with the game. Building, farming, and managing a metal industry is a lot of fun. The barrier to entry is still pretty high, and I don't think the announced additions are going to change that.

(haven't played in the last year or two, so my criticism may be outdated)

View on HN

I once had a mature fortress laid waste by a black dragon. My dwarves got it killed, but there were only three surviving dwarves (who had been fully plate armored) out of 200, and after the battle they were so stricken with grief they could hardly do anything.

I've had invasions of orcs and goblins thrown back by fortresses. What a complex and !FUN! game.

No the criticisms are still valid. There's still severe frame rate issues under quite a few circumstances. The military thing is more subjective though, it's still a complicated system, but such is the way of DF.

As much as I can see why Tarn wants to keep the code closed source, I really wish he'd allow someone to help him with some optimisations.

I'd disagree that the military issues are subjective, as of 0.44.x dwarves still put on a random assortment of what you assign them, normally missing something important like a right gauntlet or a shield.

Multithread w h e n ???

Also, early embark setup fatigue is real. Setup the stockpiles. Plan the rooms. Have the plans outpace your current productivity and have it take way too long. Do this every time.

Once that's over, though, you get to enjoy sorting through droves of migrants, assigning each to the tasks they're most appropriate to or which need the most dwarves right now (the game actually approximates this automatically in that it generates migrants vaguely in reflection to the fortress's needs, but it's not like it assigns all burly, tough, slow-to-tire dwarves to your melee squads or that talented bonecrafter to more useful crafting jobs, so it's up to your manic OCD). Do this every season.

Somewhere in there, have some !FUN!.

I prefer to embark next to volcanos and have the dwarves tap the volcano to make defenses (think magma traps and moats) and power furnaces and forges.

Once, my dwarves built a box structure on top of their underground fortress entrance. Built a wall around that, and filled the in-between space with magma to create a sheath defense. The ingenuity of dwarves is marvelous!

Multithread never. You know in your heart of hearts that DF wasn't built with modularity in mind and that the code is a horrific mess of dependencies everywhere and making it multithreaded would require a near complete rewrite of the system.

Actually, one of DF's biggest CPU-sinks is AI pathfinding, which is a: embarrassingly parallel and b: highly speculatable[0], so the only hard part would be keeping map walkablity data in-sync or read-lockable (depending on whether you want a separate copy or not).

0: ie, you can fire off a pathfind from your current position, keep walking for several steps (many game ticks) before getting the result back, and just do some minor fixups to make it work.

> the only hard part would be keeping map walkablity data in-sync or read-lockable

I suspect that this is entirely the problem. For quite a while not I've gotten the impression that all of the game state in in a single data structure.

Also I'm pretty sure that there are update anomalies galore even now. It's almost certain to be a herculean task. I might be wrong, but the fact that Tarn has never even attempted it makes me suspect that I'm right ;-)

Should rewrite it in rust

Given what we can see on the surface, DF is probably a mess under the hood. I'd bet a competent C programmer could speed it up a LOT with old-fashioned single-threaded optimizations.

By the time you synchronized all that state would there be a performance payoff? My gut tells me the overhead would eat up a lot of the benefits.

Don't forget how dwarves drop their personal items all over your fortress, even when they have an assigned bedroom. Some ahole leaves their sock in your front door and the vile forces of darkness just waltz on through.

But all of the half-implemented features, years-old bugs, "screw you" approach to UX, and general jankiness are worth it when a fire imp immolates 10 people in your great hall only to have its skull caved in when it tangles with the wrong toddler.

> The barrier to entry is still pretty high, and I don't think the announced additions are going to change that.

This would be a problem for a new game, especially in Steam's early access program - but at this point DF has a huge "brand", and people know what they're getting into. I think it could even be argued that the difficult UI is part of the experience, as weird as that sounds.

I would buy a copy just to support development and the past fun I’ve had. But I’m likely in the minority.

I'll likely buy a copy and continue to download and play the "free" version.

Reply on news.ycombinator.com