Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Andir]

There are a couple lines of though in development.  One of them is pushing on with features and leaving optimization for later.  It's usually sound, but I do believe there is room for some preemptive optimization.

[Naros]

For me, what turns me off of DF currently is bugs. Namely the population limit bug.

If I set the pop limit to 30, I do not want immigrants to take me up to 55 pop, activating nobles and all that on top of giving me more dwarves than I like.
I can kill 'em off, sure, but the nobles and related features are there to stay.

[Jerg]

An idea I have about migrant wave control that doesn't affect the lore of DF (or rather, the system of how DF dwarven society works), is that pop cap won't work instantaneously as per usual, BUT you can decide to turn DOWN a newly-arrived migrant wave if you so choose.

Even better, have it so that you can decide whether to let a whole wave in, to let a select few of the wave in, or reject the entire wave. But this would be very complex and consequential systems will need to be implemented alongside it to make it sensible (e.g. those let in might become sad that their companions weren't, those rejected could go rogue / go back to origin civilization and defame the player's fortress, etc).

[jarodw]

Setting aside the two elephants in the room to say what other "bad" things there are about the game is kind of moot...  Ascii, and controls are the two top things that scare people off... but if like myself and many others here, they got past the moon sized mountain of the ascii and controlling the game, the rest is intellectual...  the depth and detail... I think the game is just TOO complex for most...  very very good thing to me and a lot of others, but the majority of gamers have been preconditioned as of late into angry bird simpletons...

Honestly keep it the same...  attracting more people to this game would be a good thing, but to change it in favor of those less willing to learn the ropes now would break it for me!   a good in game tutorial would help,  maybe a small quest system pointing a potential player in the right direction of making a successful fortress, but none of that dumbing down thing I BEG YOU!  also should you put such a tutorial in...  make it have an ON/OFF switch.

people like myself, Captnduck, Nagidal146, TheZemalf, and other youtubers try to spread the good that is DF...   I know I have hooked a few players for you and Tarn with my videos... captnduck hooked me...  I say just give it time...  If you build it, they will come

[KFK]

There are a couple lines of though in development.  One of them is pushing on with features and leaving optimization for later.  It's usually sound, but I do believe there is room for some preemptive optimization.

I'm generally a "no code will be optimized before its time" kind of person, but there comes a point when you have to show some discipline and clean up what's there, even if there's a legitimate concern you'll have to rewrite it later anyway. It can make pushing forward easier at times when you have a tidier base to build from. I can't say for sure if DF has reached that point, but my gut says we're actually WAY passed it. I'd wager that DF2010 was the time to really take a hard look at what's already there.

The so called presentation arc is way more important than some people think. There are long term benefits to both the users AND development that are being overlooked.

the majority of gamers have been preconditioned as of late into angry bird simpletons...

Thank god for Jim Sterling

http://www.escapistmagazine.com/videos/view/jimquisition/4081-Angry-Birds-Is-Not-Sh-t
http://www.escapistmagazine.com/videos/view/jimquisition/5047-Hardcore-Hypocrisy

[zilpin]

Yes, like Knuth says, premature optimization is the root of all evil.

Key word being 'premature'.

Much of DF speed problem does not come from lack of optimization, but from overly entwined structures and naïve implementations.
That's not a matter of optimization, that's just a matter of good professional practice.  But everyone fudges things for release, and players always push him to release more & bigger & sooner.

As for UI, I still think Toady is better off giving easier ways for external programs to be UI.  StoneSense has come so far, and DF becomes unplayable without Therapist, but making a true UI is not feasible without Toady wanting it to happen.

We don't need an API, linked DLL.  Plenty of games have a simple client-server design even for single player.  He already separated interface commands & rendering from core logic, so all he need expose is just what's in his UI now.  This would reap the greatest benefits for UI, and save Toady the most time.  UI programming in C/C++ is really boring and tedious for a mathematician who groks three dimensional fractals (e.g. DF's mapgen).
Don't send Einstein to dig a trench in the mud; we'll gladly do it for him, if he only lent a shovel, point, and say 'over there'.


But I'd always prefer bugfixes before anything.  You often find that bugs hurt performance, and clever optimization has a habit of introducing hard-to-find bugs, so you're always better debugging first.

[[NO_THOUGHT]]

Lack of tileset support: I like ASCII as much as the next roguelike, I even voted for you in the annual roguelike contest, but when I want to play with a tileset, I want a tileset. The staircases and bins should not look alike, the text should not look like dingbats, ect.

Designation is not very intuitive: 'd' 'enter' 'downarrow' 'rightarrow' 'enter', down, down, right, down... repeat forever... receive simple circle. PLEASE let us have some real designation tools. I know that dwarf fortress is tile based but behind those tiles is a mathematical simulation of legendary proportions! You're telling me the only way I can effect that simulation is through rectangles? Circles, outline shapes, polygons, lines, thickness, flows, ect. The dwarf fortress community is a legion of artists given an infinite canvas, and to paint they get a smooth wooden ball the size of their fist. Not surprisingly, the amazing simulation of Dwarf Fortress is enough to keep anyone hooked long enough to make that situation turn into something beautiful (read: Raynard Fractal rooms *drool*), but the things we could do with those tools! It's probably me being a spoil CAD brat, but spoiled I remain. Pardon how rant like that turned out.

No manual automation: As it stands, no fort can run for any period longer than a season without a spanner getting thrown into the works. Now unpredictable chaos is the charm of dwarf fortress, but there is a whole ton of predictable chaos that makes it hard for me to play long. Workshop queues being <10 places long and no numerical orders being the big one. You try and you try to make enough beds for everyone, but when a 20 migrant caravan comes you ask for 10, hope you will remember to make 10 more but then only remember to do so when the next caravan arrives. Then you're behind, the migrants get angry, throw tantrums, everyone loses. I could designate repeated beds, only to find 300 beds made, my wood stock destroyed, no trees left, and a migrant cap of 150.

Those are the three biggest things that make me stop playing forts after a while, but I always start again after a while because new forts don't need automation, designation planning is interesting (because I plan a lot), and the dingbats don't bug me as much. One day I'll try to get past all that to actually reach nobles, but that's for another day.

[zwei]

No manual automation: As it stands, no fort can run for any period longer than a season without a spanner getting thrown into the works. Now unpredictable chaos is the charm of dwarf fortress, but there is a whole ton of predictable chaos that makes it hard for me to play long. Workshop queues being <10 places long and no numerical orders being the big one. You try and you try to make enough beds for everyone, but when a 20 migrant caravan comes you ask for 10, hope you will remember to make 10 more but then only remember to do so when the next caravan arrives. Then you're behind, the migrants get angry, throw tantrums, everyone loses. I could designate repeated beds, only to find 300 beds made, my wood stock destroyed, no trees left, and a migrant cap of 150.

Try using manager (u-m) for larger batches of production.

It is actually a lot more comfortable than manually queuing items in workshops.

Problem is still there. With older fort, you get lots of industries going on which could theoretically easily be automated: for example, if Skull item appears in stockpile, i would like automated workshop order for making totem. If there is ore in ore stockpile, i would like smelt job queued.

One can set such things on repeat, but if raw materials run out, you have to set up production again ... and remember to check later to reqeue.

Not that some things work like this already (cloth production, tanning... ).

[Phmcw]

The more I know about it, the more I think that DF in not really multithreadable. So yeah, hope for faster processors.

[Starne]

The biggest stumbling blocks for me when it comes to me are:

 The UI. It's almost impenetrable. I've played DF off-and-on for about two years now, and I still barely understand some parts of it.

To lesser extent: Lack of Automation. Eg, 'If booze amount is less than X: Create y brewing jobs ' 'Notify if amount of stored food is less than x'

One of the biggest things in the automation department would be for the game to auto-task dwarves who've been idle for an extended period. Lot's of chopping or mining jobs piled up? That soaper who's been standing around the meeting hall for three months should try to find an axe or pick and go help out. Another thing here would be for dwarves to atleast try to defend themselves using whatever is at hand, instead of just mindlessly fleeing until they're killed, die of dehydration, or escape(could see unarmed civilian dwarves throwing rocks or using other improvised weapons)

Bit of a tangent off that: Have things like dwarves finding something to do instead of standing around, or defending themselves in a pinch be tied to their personality. A lazy dwarf might never bother to find something to do, a cowardly dwarf would attempt to hide or flee, rather than try to defend itself.

Of course, not done properly, implementing this sort of thing would just make the UI even more complex.

[Loud Whispers]

Nothing really turns me off DF.

Except when I run out of ideas.

But then I go on bay12...

And the !FUN! begins anew.

That would make a nice poem.

[zwei]

The more I know about it, the more I think that DF in not really multithreadable. So yeah, hope for faster processors.

Well, it is matter of outlook. You could get nice returns with "easy naive threading" that does not require synchronization or data locking. I am talking about lookups.

Each "tick" of gameplay several things happen that can happen concurently: Looking up suitable items for hauling/use and looking up patch for dwarf. Stuff that does not modify data which is thus safe for threading.

Game can spawn threads with this and wait for them to finish before continuting to next tick.

That is wishfull-thinking theory, of course.

[blaize9]

One of the things that slowed me down in the beginning was being confused by [v] and [k], and I think the fact that they're so easily confused yet are both very necessary is one of the biggest interface problems.

I agree with this, they could just combine them into 1 letter.

[Loud Whispers]

One of the things that slowed me down in the beginning was being confused by [v] and [k], and I think the fact that they're so easily confused yet are both very necessary is one of the biggest interface problems.

I agree with this, they could just combine them into 1 letter.

NONONONONONONONONONONONONOOOOOO

TERRIBLE IDEA D:

[Telgin]

I constantly screw up q, v, k and sometimes the t command.  They do need to stay separate though, since otherwise it takes away from controller screw and moves it into interface screw, hehe.