Print

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

[Tormy]

It's been a long time since I've posted on this forum, but I just want to comment on this post. First of all, I am an old DF fan, I've been following the DF development since many years. DF is simply awesome, even in it's current state. Bad interface, no graphics [Toady, you still haven't separated txt/gfx - we've talked about this years ago! ], but who cares? Once you have some experience with the gameplay mechanics & the interface, you will have a lot of fun for a very long time. Gameplay >>> everything else in my book.

That being said, I can understand that the OP is frustrated.  The new version was released like 15+ months after 40d. I didn't followed the development in the last year, I was waiting for the new version. When it finally came out, I thought that the Army Arc will be fully functional finally [creating armies/attacking and conquering other sites etc.] after 15+ months of development. Needless to say, that I was very surprised that it's far from being complete. [I sent in a donation anyways ]However maybe it's understandable....afterall Toady is the only coder, and the game is free.....and after I've inspected the 2009/2010 devlog, I realized, that he worked A LOT on the game in these 15+ months.

So what is the point of my post? I am pretty sure that the OP wasn't trolling. I am not trolling [maybe some old DF lads still remember me] at all, and I can understand his points even. Who wouldn't want to play with the next major version in 5 months, instead of 15 months?  However, we all know, that Toady won't accept any help in coding [which is understandable, not to mention, that this is a known fact since many years], so it's pointless to make these posts. [I've seen many similar posts here, because I was visiting the forums occasionally.]

[SmileyMan]

This is something Toady has addressed pretty closely, in one of the DF Talks:

I leave enough comments that I can come back in six months and know what I've done. I don't use perfect variable names but I use pretty good variable names. I remember reading some forum somewhere where people were like 'well he probably just names his variables 'a', 'b', 'c', 'd' and 'foo fah foo flam blum blum blum' like he's an idiot or something' but you give things meaningful names and you need fewer comments, right? I can pretty much read through my code, it reads pretty straightforward to me, because things have sensible names.

I'd go with the Toad on this.  With meaningful variable names and function names, you should only need comments to explain non-code stuff (e.g. a mathematical algorithm, and why you've used it)

A long-winded function name might be a pain right now, but the compiled code won't care, and your future self will love you forever.

[spooq]

Are you all completely convinced that variable names determine the quality of code?

[SmileyMan]

Are you all completely convinced that variable names determine the quality of code?

Define "quality."

Do descriptive variable names contribute to more maintainable code?  Yes.

[spooq]

Bad variable names can be fixed with a search-and-replace.

Bad algorithms and APIs can be extremely difficult to fix.

It just surprises me that everyone mentions variable names instantly in these sort of discussions.

[SmileyMan]

It just surprises me that everyone mentions variable names instantly in these sort of discussions.

Perhaps it may be true then?  The Internetistas are allowed to be right once in a while.

The advantage that descriptive variable names have over comments, is that they are self-maintaining.  The compiler doesn't give a monkeys if the 20 lines of comments above the pathfinding routine aren't changed to reflect the new pathfinding algorithm.  But if you don't change DijkstraNodeVisitor to AStarNodeVisitor everywhere, then it likely won't compile.

It's not just about the "search and replace."  If you make a habit of avoiding comments through the liberal use of descriptive variable names, then you are required to go through your code with the diligence that we all pretend to our boss we have, but deep down know we don't.

[Hyndis]

Toady has done so much in such a short time, particularly considering that he is only a single person and is creating a game that utterly dwarfs one of those big budget AAA games in terms of complexity and gameplay.

Take any game from Electronic Arts or Ubisoft. Can one of those games keep your attention for years? They're all pretty and full of flashing, shiny things. Some of them even have pretty good gameplay. But none of them have much in the way of replay value. Even a city builder game like the Anno series or Sim City will eventually get old. You can only do so much. The sandbox is tiny.


DF allows you to do nearly anything you want. Do you want to make a gigantic dwarf statue out of gold who craps magma all over the landscape at the pill of a lever? You can!

Want to make a working mechanical computer inside the game? You can!


The complexity is utterly mind boggling. I don't even know how you'd go about building a computer within a game within a computer. I'm just going to at this point declare it to be witchcraft and look on in awe. 

[spooq]

Compilers care as little about variable names as they do about comments. If I don't change DijkstraNodeVisitor at all, or change it to DatabaseRowCount, everything still works exactly as before. Therefore, they are unreliable as documentation.

Try changing a published API and see what happens. Try adding threading to code that was written without threading in mind and see what happens. These things cannot be done by clicking "replace all".

I don't have the option to pretend to be diligent about my code working or not. On the other hand, noone is interested in my variable names either.

[einstein9073]

... Despite Einstein's trolling attempt...

Yeeah, NO U.

...it has nothing to do with messiah worship.  ...  For those of us who have enjoyed DF for YEARS, having it (and Toady) insulted can ONLY result in outrage. 

You don't see the conflict there? I'll confess to hyperbole, sure.
But "The Community" overreacts really strongly to any criticism - real or perceived - of the man.

The problem with that is the assumption that the codebase remains static long enough for simultaneous development to occur without stepping on toes.  Even using source repositories and merge tools, multiple people working on the same file can be disastrous.  Secondly, you are assuming that someone brought in would have the knowledge necessary to know where in the code the bug would be.  Toady knows his code inside and out.  The ramp-up time for supplemental developers would be enormous.

If Toady One doesn't want to do administrative stuff and Just Code, perhaps the solution is to add administrative staff, then. Let someone else manage the merging and delegation and just let the man himself design and code.

Maybe Version One plays in his head, and is powering his fey mood, demanding that he bring it into being or perish.

I like that explanation better.

Nah, if he were really in a fey mood he'd either immediately fix the bugs or kill himself trying. And he'd *never* release incomplete versions. He'd just toil in his workshop for however many years and then we'd have fully-featured DF out of nowhere.

In that case, there would be an even chance between DF being awesome and complete, and being an obsidian earring that menaces with spikes of silk.

[Deon]

Would you build a fortress with one dwarf?

Definitely yes. I want adventurer skills so badly!

[HatfieldCW]

DF is a complex topic.  I'm a guy who likes video games, and I love DF, but I don't think DF is a very good video game.  Let's face it:  This is a public alpha of a game that will probably be in development for decades.  It's being developed by two people and coded by one.  A project this ambitious will take thousands and thousands of man-hours to be realized, and with just Toady doing the heavy lifting, that translates to a very long dev time.  It lacks polish, it lack graphics, the interface is famously punishing and just about everyone who takes the game seriously modifies it and runs one or more helper utilities to make it work better for the player.  When I introduce a friend to DF, I make sure they have Dwarf Therapist and a few changes to the .init file before they try their first embark.

I like that, to some extent.  The progress of the game, the community's relationship to it and and big ideas that are being folded into DF all make me very happy.  I sometimes think that the development of DF is more interesting and fun than the game itself (depending on the version).  I think it's entirely possible that this will never be finished.  Either Toady will lose interest and do something else, or the community will be unable to support him and he'll go get a "real" job, or he'll just work on it until they nail his coffin shut.  A boxed version on store shelves is pretty much unimaginable to me.

In years to come, though, DF will have a big impact on game design, I think.  It's taught us a lot about how much complexity players are willing to tolerate, and shown some of the rewards that can only be earned through incredible attention to detail and massive scope.

[Rumrusher]

Would you build a fortress with one dwarf?

Definitely yes. I want adventurer skills so badly!

the modify question for me would be would you want to build a fortress outside of a city with an adventurer.
for just one dwarf in fort mode well didn't darius updated his utility so that you and edit the embark race?
so one could change the embark file with 6(or mode if you want more wood) wagons and one dwarf and end up with 1 dwarf and 24 pieces of wood for building.

[Rotten]

Compilers care as little about variable names as they do about comments. If I don't change DijkstraNodeVisitor at all, or change it to DatabaseRowCount, everything still works exactly as before. Therefore, they are unreliable as documentation.

lolwut? The point of variable names is for the coder to understand it, not the compiler...
This may seem like an accusation, which it isn't, but do you actually program long programs? Because in my experience, after my code is even over a few hundred lines I can't remember what "x" does as opposed to "xt" or "XLTds" or"i" or "j" or why j=6*(sqrt(x1*x-i)), especially when you're writing cellular automata map generation or something similar that uses a lot of math functions that don't make any sense by themselves.

Spoiler: "Proof of above" (click to show/hide)

In fact, find out what this code does:

Code: [Select]

for(a=0; a<y; xt++)
for(bs=0; b<x; a++)
g[a][bs] = T;

for(a=0; a<y; a++)
g[a][0] = g[a][x-1] = T;
for(bs=0; bs<x; bs++)
g[0][bs] = g[y-1][bs] = T;
 }






Gave up? Read this:

   

Code: [Select]

for(yi=0; yi<size_y; yi++)
for(xi=0; xi<size_x; xi++)
grid2[yi][xi] = TILE_WALL;

for(yi=0; yi<size_y; yi++)
grid[yi][0] = grid[yi][size_x-1] = TILE_WALL;
for(xi=0; xi<size_x; xi++)
grid[0][xi] = grid[size_y-1][xi] = TILE_WALL;
 }
It's a segment of a cave generation script I shamelessly stole off RogueBasin because I couldn't be bothered getting out my flash drive, but nevermind that. The first snippet is obviously impossible to understand without memorizing the variable names. The second (which is the original), if you came across it in the code, while not perfect (yi? Since it's a counter variable, ycount probably would have been better), is obviously dealing with generating the map, because it's creating tile positions (TILE_WALL). If it had a few comments on the important lines it would be even better.

Of course, I don't follow my own advice. My current most important variables are variations on the name "oti", for 1, 2, input version (it's an AI and needs a few different versions of the position variables). I have no idea how I'll decode that in a month.

On topic, I personally fine with whatever the Great Toad does. It's his dream-project-monstrosity simulator-thing, and if he didn't dream it up and put in the thousands of hours to code it up from a blank, white screen, we wouldn't have a damn thing. Trying to force Toady in divulging portions of the code he may particularly like or be proud of (Oh, come on. The pathfinder is actually quite good, I would be proud of it, considering how lightweight it is for 300-odd entities (tasked items path to)) just so we get faster updates is frankly, incredibly selfish.

[spooq]

DF is a complex topic.  I'm a guy who likes video games, and I love DF, but I don't think DF is a very good video game.  Let's face it:  This is a public alpha of a game that will probably be in development for decades.  It's being developed by two people and coded by one.  A project this ambitious will take thousands and thousands of man-hours to be realized, and with just Toady doing the heavy lifting, that translates to a very long dev time.  It lacks polish, it lack graphics, the interface is famously punishing and just about everyone who takes the game seriously modifies it and runs one or more helper utilities to make it work better for the player.  When I introduce a friend to DF, I make sure they have Dwarf Therapist and a few changes to the .init file before they try their first embark.

I like that, to some extent.  The progress of the game, the community's relationship to it and and big ideas that are being folded into DF all make me very happy.  I sometimes think that the development of DF is more interesting and fun than the game itself (depending on the version).  I think it's entirely possible that this will never be finished.  Either Toady will lose interest and do something else, or the community will be unable to support him and he'll go get a "real" job, or he'll just work on it until they nail his coffin shut.  A boxed version on store shelves is pretty much unimaginable to me.

In years to come, though, DF will have a big impact on game design, I think.  It's taught us a lot about how much complexity players are willing to tolerate, and shown some of the rewards that can only be earned through incredible attention to detail and massive scope.

Hey, nice post. I was going to write the brave "it isn't a good game" post as well, with some of my reasons laid out, but you did a better job than I was going to.

I still really like the game of course. Hopefully people can see you can still like something while admitting it isn't very good by some measures.

[spooq]

Compilers care as little about variable names as they do about comments. If I don't change DijkstraNodeVisitor at all, or change it to DatabaseRowCount, everything still works exactly as before. Therefore, they are unreliable as documentation.

This may seem like an accusation, which it isn't, but do you actually program long programs?

Yes. Do you actually modify other peoples code, some of whom stopped working for that company five years ago? The header file I was sent for the system I'm working on at the moment has a 1991 copyright notice on it. Wait till you see the variable names they came up with. Those names are used across a large multi-national corporation. I doubt anyone actually knows how many systems rely on them. You gonna be the one to rename them?

Btw, I "fixed" your code by replacing the variable names. Took me one minute, including fixing the bugs you introduced. Would you mind spending the same time to split the workload across the four cores on my machine? Thanks.

Next time your FPS is low because of path-finding, tell me how good the code is again. Tell me it does the job it is meant to. Post that using the idle core on your computer to process the forum webpage.

Personally, I don't even care about the FPS. My forts fall and/or I get bored with the current crop of bugs before I get to the unplayable point. Then I go play Team Fortress 2 while occasionally looking at my newsfeeds to see if there is a DF update. I'll always come back and see what is happening with DF because it is interesting on several levels to me.

I'm not forcing Toady to do anything. I'm just posting observations on a forum. My suggestion didn't even involve him showing his own code to anyone.