An outsider's view

**Magnate** · October 9, 2012, 18:21

Originally posted by raithe

Lastly, (and I'm sure this one won't be popular, but it's nonetheless true.) is a point of view not from a player perspective, but affects players, and you devs alike. Angband suffers from its codebase. It's old. Eventually when a codebase reaches a certain age the only way to keep it moving is tear it down and re-write it. Maybe not now but eventually it will have to be done. It had to be done before (x86 port, the moving to external data files etc) and it will have to be done again. I know it will suck but in the long run it would be worth it.

Quite a few of us have realised this. There have been numerous attempted rewrites over the years, but none since we had a devteam worth its salt. Perhaps this one will get there. If not, I can personally vouch for the limitations of the current codebase.

**fizzix** · October 9, 2012, 19:13

Originally posted by raithe

First off its a long slow paced game by the standards of any modern roguelike. Roguelikes seem to be all about diving, not grinding these days. That however is Angband's gameplay style and changing that changes the game, besides there are variants like Quick which address that aspect. However knowing this you need to realize that this style compounds with other UI issues to make the game feel even slower because of it's aging interface.

Long I'll grant you, but I don't think that's as much as a problem as you might think. Long games have their appeal, specifically for players who value the journey over the destination. Furthermore since Angband is much more equipment based rather than skill based, you aren't wedded to a specific character design.

I think the slow-paced comes from a long-standing perception that you need to grind to be at parity with the monsters around you. That doesn't need to happen. Clearing all levels is overkill. Still there have always been suggestions about shortening the dungeon, and I think these should be looked at very closely.

Mouse support is a big thing, and it's good to see this being addressed. Most modern roguelikes are moving toward it and also trimming down the number of keys needed. This lowers the learning curve which is important because those darn kids today don't RTFM anymore. Even for experienced players it makes laptop play easier, and switching (as I do) between many different roguelikes as well. (because I don't keep confusing command layouts) It's also needed (as Nick said) for mobile & console ports easier which draws you more players.

This is so hard for me to relate to. I use the keyboard for almost everything that I conceivably can. Nevertheless, I understand it and hopefully we'll get people to improve on Blubaron's work.

Polish is part of the problem I suppose, people like shiny these days, but it's also transparency. Most modern roguelikes have moved away from spoiler driven source code reading to a more intuitive design. I sympathize here as it's one of the reasons I hate Nethack, it makes no sense. Again this is something that's tricky because it's core to Angband's style, but realize it's there.

This is another area that we've made great strides in. No longer do you have to do hand calculate weapons to compare damage, or code-dive to find out how much damage a monster spell will do. I imagine this will only improve. I'm not sure flavors will ever disappear entirely though, but they're becoming less and less of an integral part of angband.

Bands also tend not to have much of a presence in the roguelike hubs of the net like Roguebasin. Sometimes a variant get's listed there but just as frequently not, even this site's variant page is out of date all the time. That loses you more players than you might think.

How do we improve this?

Again these are just some outside thoughts. I don't keep up with the forums here so maybe it's stuff you've already discussed, but I hope it helps.

Some of it we indeed have, but it's always good to hear new opinions.

**raithe** · October 9, 2012, 23:26

Originally posted by fizzix

Long I'll grant you, but I don't think that's as much as a problem as you might think. Long games have their appeal, specifically for players who value the journey over the destination. Furthermore since Angband is much more equipment based rather than skill based, you aren't wedded to a specific character design.

It's strictly personal preference, and I agree it's part of Angband's niche and charm. Personally I think Moria at 50 was pushing it a bit. It's that whole RPG inflation thing. Moria had 50 so we'll have 100 etc. More good sequels and offshoots have been screwed by that than I can count. However the perception of slowness in regards to length is not because it's 100 levels, it's that (like most older rogues) there's not much variation in those levels. People are used to terrain, themes and branches. V does some of this but not much so to an outsider it's 100 SIMILAR levels.

Originally posted by fizzix

I think the slow-paced comes from a long-standing perception that you need to grind to be at parity with the monsters around you. That doesn't need to happen. Clearing all levels is overkill.

I've heard this a few times, but it's usually from people who are used to Angband, or just really good at roguelikes in general. I think you guys have forgotten how hard it actually is when you don't know the game back to front. Veterans know what stats they need, what monsters do, which items are good, where to find them etc etc etc. Thus they can dive deep and fast and survive. The game has had over a decade of additions assuming existing knowledge of those skills and trying to challenge them to keep it new. Skills newer players don't have.

Originally posted by fizzix

This is so hard for me to relate to. I use the keyboard for almost everything that I conceivably can. Nevertheless, I understand it and hopefully we'll get people to improve on Blubaron's work.

It's just another paradigm shift, same as the Vi vs arrow key debate. Neither is right per se, All that's important is that the choice is available. (with the exception of tablets and other ports that don't have keyboards, there its obvious) The mouse in 3.4 is a great start. I hope other variants adopt it, but I wonder how many will since they aren't based on the current V codebase.

Another thing I forgot to mention here is auto-explore ala Crawl/Brogue. This is a fantastic way to cut down on the monotony of longer roguelikes, (and reduce R.S.I.) and with the existing Borg stuff and running code I can't see it being too hard to implement. Mostly just a good set of filters which could be stored in an ini file for easy user editing.

Originally posted by fizzix

This is another area that we've made great strides in.

True enough, This has improved a lot. There's still room for more though. The main suggestion I'd offer here is to make as much info as possible available in gameplay. Angband has a fantastic help system, but it's long and people hate reading (why I don't know, but then I come from the "here's your 5 1/4 floppy and ten pound manual era") So try to break it into critical bite size chunks displayed in the gameplay when needed. A player should only have to jump to the long manual if they want to get in depth info, not to learn to play. Tutorials aren't really an answer either because there's too much too digest, they'll still be flipping back later.

Originally posted by fizzix

How do we improve this?

That's easy. RogueBasin is a wiki, just post a release whenever there is one. Most of the other hubs seem to look there for announcements so you should get some cross-traffic from just one post.

**Derakon** · October 9, 2012, 23:41

Originally posted by raithe

Another thing I forgot to mention here is auto-explore ala Crawl/Brogue. This is a fantastic way to cut down on the monotony of longer roguelikes, (and reduce R.S.I.) and with the existing Borg stuff and running code I can't see it being too hard to implement. Mostly just a good set of filters which could be stored in an ini file for easy user editing.

From what I've read (I haven't played Crawl or Brogue), in those games it's reasonable to codify "when the player needs to make a decision", and the auto-explorer can handle things otherwise. Angband is in many ways a knowledge-management game, though, and that includes deciding when to cast detection spells and reacting to monster movement even when the monster is not visible (but can be detected by spells or telepathy). Thus practically any auto-explore mode you wrote would eventually result in someone's unearned death. That's probably the main reason why nobody's seriously spent time considering adding an auto-explore mode to Angband.

Plus, the Borg is a ravening numbskull. It's like the dancing dog -- it's not impressive in that it dances well, because it doesn't; it's impressive in that it dances at all. It's also the product of basically one developer's efforts and I don't think anyone else understands how it is written; nor is it portable to any variants because it codifies Angband's mechanics into the AI -- so when those change, the borg becomes unable to cope.

All that said, I definitely appreciate what you're saying. It's easy for a niche community to become insular and forget what it takes to attract new players, and down that road lies stagnation and eventual extinction. Angband needs new players to stay alive, just as it needs old players to provide an informed perspective on what makes the game fun.

One of the things I want to do eventually in Pyrel (the reimplementation-du-jour, in Python this time) is to add a look-by-mouse system -- when you mouse over a tile, a tooltip would pop up telling you about what's on the tile and important stats (e.g. monster name/speed/native depth/visible health; item short descriptions; terrain info). Likewise, when examining the character, you should be able to mouse over your stats to get a tooltip telling you what they influence. This kind of thing makes the game more discoverable in a very intuitive way.

**debo** · October 9, 2012, 23:50

I'm a brand-new 'bander -- I played maybe 5 Vanilla characters, and finished the game on my 3rd attempt. (The final two were mages, who pretty much insta-died.) After that, I moved onto Sil.

I'm curious about a couple of your comments:

a) Re: Angband's 'opaqueness' / needing a wiki to finish.

I felt like Angband was probably the most transparent roguelike I'd ever played. It was so simple -- get to lvl100, kill Morgoth. That's it. I posted a few times here to ask for min/maxing help, and picked up some lore along the way (e.g. Kavlax at ~DL40 is a bitch), but I always felt like I could have trucked through regardless. DCSS, on the other hand... that, to me, seems like the epitome of a wiki-roguelike. I opened it a couple times and was so brainfucked by all the options that i just got bored. And this is from a guy who played Nethack religiously.

What parts/aspects of Angband specifically do you feel are opaque?

b) Diving requiring knowledge of the game

To be honest, Vanilla was so damned easy that when I picked a reasonably newbie-friendly character (high-elf rogue), I dove like crazy to 3000' and didn't have much trouble staying alive. After that, I started clearing levels just because I was addicted

This has probably changed in 3.4 (I played 3.3), but I imagine for certain builds that newbies are likely to build (e.g. high-elf warrior), diving is pretty simple and sort of a natural thing to try.

I've since decided that Vanilla is a little too nebulous and repetitive for me, and that I much prefer 'dense' experiences like Sil. However, I totally understand why Angband fans love the game -- it is a long, sprawling, straight-up dungeon crawl, without much complication. Some people just like that

God knows I couldn't put it down it the first few times I played.

**raithe** · October 10, 2012, 00:05

Hmm, point taken. Actually the auto explore in crawl was pretty stupid for a long time, making it work was more a matter of changing how and what this got displayed in the message line and extensive filter writing. It's still far from great. Brogue is different, first its a dead simple game and second as I recall it planned to have the explore in there from day one.

My main thing with it in Angband is all the long twisty go nowhere passages and the size of the maps. Actually I be just as happy with a floating large map so I could keep an overveiw of the level. (one that was actually readable like Angband TK's not the mess off blocks map you get in V) Cryptband has a fantastic automap window

TK did the floating mouse over info thing very well. Personally, I find it to be one of the better UI's I've seen in a roguelike. (once you set to ascii display

of course it might look pretty slick with shockbolt's tiles)

I guess that what puzzles me, most of the slick stuff already exist for Angband, they just got stuck to a single variant and never went anywhere. TK had the UI, Crypt's map, UN's iso display etc etc... but they never make back home.

**raithe** · October 10, 2012, 00:31

Originally posted by debo

I'm a brand-new 'bander -- I played maybe 5 Vanilla characters, and finished the game on my 3rd attempt. (The final two were mages, who pretty much insta-died.) After that, I moved onto Sil.

I'm curious about a couple of your comments:

<snip> And this is from a guy who played Nethack religiously.<Snip>

That's why, right there.
I divide roguelike players into two groups Those get/like nethack and those who don't. It's like in 3rd ed D&D, you had the guys who hung out in the character optimization board, and everyone else. Some people intrinsically "get" stat heavy min max stuff like that and can milk it for all its worth without a second thought. Some don't

Originally posted by debo

What parts/aspects of Angband specifically do you feel are opaque?

Me? At this point not too much but I've been playing it quite a while. My response of "makes no sense" was referring to Nethack, specifically it ID system. (if you can call it that) I was speaking in general terms based on the responses from the other people I've introduced to roguelikes. If you'd like a list I can have my wife do a couple quick runs later, she loves roguelikes but never really cared for *bands or nethack much for precisely that reason: she doesn't read help files. She's great at most other rogues though, so obviously there's some difference there.

Diving has a different meaning here and was the wrong choice to use on these forums, apologies, so let me clarify. In most modern roguelikes the progress is linear, you clear 1 got to 2 etc, Diving (ever downwards). bands don't work that way, (well Ironband does but...) you jump up and down, hover at certain levels for various reasons etc. That's not the kind of thing people who played other roguelikes are going to get naturally by and large.

And again, I'm not saying at all that the things that make Angband Angband should be changed, just pointing out where stumbling blocks might be for attracting new players.

**Derakon** · October 10, 2012, 00:49

Originally posted by raithe

I guess that what puzzles me, most of the slick stuff already exist for Angband, they just got stuck to a single variant and never went anywhere. TK had the UI, Crypt's map, UN's iso display etc etc... but they never make back home.

The *Tk series never got much traction, for whatever reason; otherwise, the lack of sharing of UIs largely boils down to a) the core codebase not being very flexible (which makes porting UI work a major effort instead of "just apply this patch and you're done"), and b) UI work not being something most devs are especially interested in. Which is a shame, because a good UI makes a world of difference when it comes to attracting new players.

**ekolis** · October 10, 2012, 01:01

Originally posted by Derakon

b) UI work not being something most devs are especially interested in

What kind of devs have you been talking to?

Maybe it's just me who's weird, but I actually like doing UI work - I'm not all that great at graphic design or anything, but just creating UIs gives me this visceral thrill of "I made the computer do something!" SO much more than creating an algorithm that works behind the scenes and just runs, not displaying anything...

**debo** · October 10, 2012, 01:36

Originally posted by raithe

That's why, right there.
I divide roguelike players into two groups Those get/like nethack and those who don't. It's like in 3rd ed D&D, you had the guys who hung out in the character optimization board, and everyone else. Some people intrinsically "get" stat heavy min max stuff like that and can milk it for all its worth without a second thought. Some don't

Me? At this point not too much but I've been playing it quite a while. My response of "makes no sense" was referring to Nethack, specifically it ID system. (if you can call it that) I was speaking in general terms based on the responses from the other people I've introduced to roguelikes. If you'd like a list I can have my wife do a couple quick runs later, she loves roguelikes but never really cared for *bands or nethack much for precisely that reason: she doesn't read help files. She's great at most other rogues though, so obviously there's some difference there.

Diving has a different meaning here and was the wrong choice to use on these forums, apologies, so let me clarify. In most modern roguelikes the progress is linear, you clear 1 got to 2 etc, Diving (ever downwards). bands don't work that way, (well Ironband does but...) you jump up and down, hover at certain levels for various reasons etc. That's not the kind of thing people who played other roguelikes are going to get naturally by and large.

And again, I'm not saying at all that the things that make Angband Angband should be changed, just pointing out where stumbling blocks might be for attracting new players.

Yeah I just got sort of confused because I thought DCSS was being cited as a counterexample, but reading back it looks like I got that wrong. I liked Nethack but hated DCSS because I got tired of the Nethack style of play. I've made that argument to other DCSSers before and they were all like, "No way man, DCSS is totally straightforward" and then proceed to link me to thirty wiki pages lol

**debo** · October 10, 2012, 01:38

(The best thing about my nethack years, though, is that I came home drunk about a year ago and my girlfriend was giving me a hard time about something or another... I kicked the sink in the kitchen, and when she asked me what the hell I was doing, I said "I'm hoping a succubus will come out, because you are so not fun right now")

**getter77** · October 10, 2012, 02:18

My harassment work on getting variants to get themselves listed up proper on Roguebasin of late has been nigh-universally successful---the "issue" as it currently stands there mainly just boils down to the variant scene being somewhat condensed at the moment compared to the apparently myriad golden ages when the like of this and ToME 2 were in the spotlight of the overall more devoid Roguelike scene outright. There's some likely to be good stuff brewing like Beleriand and there was allegedly going to be another Un release back for ARRP 2012 before Doull got eaten by the Brogue Grue....otherwise just a great many variants that have either gone quiet, are in transition, or in stasis and in the worst cases abandoned.

In terms of masses of people poking things with sticks and fashioning nifty bits, the current frontrunner is Tales of Maj'Eyal (ToME 4)---so another avenue to attack might be the why and how to make that a communally closer race going forward.

Steamband coming back to life...apparently any day now these past many weeks...strikes me as something that will help a fair bit once word gets around providing there are some nice gains in the mix when it does---the new Hengband getting an English release would also probably stir some folks.

I'm keeping The Roguetemple in the loop as best as possible as there are things to loop with, 3.4 will surely help set the stage for a great many good things to come as it had been simmering for so long----gotta keep at the audience mindshare.

**thapper** · October 10, 2012, 10:44

Originally posted by raithe

Actually I be just as happy with a floating large map so I could keep an overveiw of the level. (one that was actually readable like Angband TK's not the mess off blocks map you get in V)

A bit off topic, but this is so true! The TK minimap was perfect (as I remember it). We constantly have people pointing out that big tiles and small screens can cause instadeath from off-screen breaths/spells. The TK-type mini map, accurate in that is shows everything that is there (all monsters/items etc displayed) but made so small that you don't see what monster/item it is, would be great. It worked really well for me when I (re-)started *banding with ZangbandTK a few years back.

**Timo Pietilä** · October 10, 2012, 11:20

Originally posted by thapper

A bit off topic, but this is so true! The TK minimap was perfect (as I remember it). We constantly have people pointing out that big tiles and small screens can cause instadeath from off-screen breaths/spells. The TK-type mini map, accurate in that is shows everything that is there (all monsters/items etc displayed) but made so small that you don't see what monster/item it is, would be great. It worked really well for me when I (re-)started *banding with ZangbandTK a few years back.

You can set a term-window in vanilla to show map. If you choose small enough font and use large enough term window you can see everything, it would just be really tiny.

Two options for map: map and overhead view. IIRC names feel reversed to what they actually do. "overhead view" shows "compressed" map, "map" shows everything, but only what fits into screen trying to keep you at the center of the screen (and there is a bug that causes part of the screen not to refresh properly, so it might be rather confusing). IMO "overhead view" is better especially if you use small enough font that it doesn't need to compress things to fit everything in the screen.

Or you can always just use very large main screen if you have large enough monitor.

An outsider's view

An outsider's view

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment