Z+Angband dev

[Pete Mack]

One might aim to rebase with V 3.5, not V4. V4 had massive reorg. 3.5 has most of the code cleanup; the reorg is largely in name changes. But yes. Starting from V would probably be easier, since you'd just be adding or modifying procedures, not refactoring the codebase. That was a few hundred man hours.

[Derakon]

Not having looked extensively at the Z codebase in awhile, I'm still inclined to believe that you'll have better luck "re-implementing" Zangband in Vanilla than the reverse, for the following reasons:

* Vanilla has had more consistent development effort over the years. Most Zangband development halted well over 11 years ago.
* Vanilla's maintainers have made a substantial effort to make the codebase clean and easier to make into a variant, which ought to reduce variant re-implementation costs.
* Conversely, copying Vanilla features over to Zangband's current codebase would be a substantial effort, as some features would require major structural reworks.
* Perhaps most important for the long haul, re-branching Zangband off of modern Vanilla will make it easier to incorporate new features as they get added to Vanilla.

You're looking at a substantial effort either way, though. Re-implementing a variant, especially one with as much history as Zangband, is no joke.

[Nick]

Also, the older V version had tile support for X11, but the latest V only supports ascii for X11. Why were tiles dropped for X11?

I believe that happened in the process of getting V GPL-compliant, because the author of the x11 tile code couldn't be tracked down.

[droof]

The Zangband dev team were working on an improved random quest system to replace both random "kill monster" quests and static quests. But they never finished the system. So one task here is to finish that system.

To make Z modern, it's either updating Z with features from V, or porting features like the wilderness and quest system from Z to V. But Z had 11 years of development of its own. Is Z supposed to be very different from V or is porting from Z to the latest V possible without losing many important Z features?

Also, the older V version had tile support for X11, but the latest V only supports ascii for X11. Why were tiles dropped for X11?

[Derakon]

Basically, play Vanilla, notice the interface improvements, and compare that to playing ZAngband. There have been a lot of changes since ZAngband first branched off ~15 years ago.

[wobbly]

It's a while ago so I'll have to try it again to give a good answer. I think it lacked a working squelch? Stuff like that.

[droof]

I'm probably weird in that I don't prefer ascii or multiple terminals. My favorite interface is from nethack, single window and graphical tiles so I can see what I'm up against, but also tiles that are eligible even when they are as small as an ascii character.

Procgen poschengband-style quests is what I'll be aiming for. I'll take some time to hack around and see what comes of it. And I'll also check out the different town houses that Kangband and Poschengband offer, I like those a lot too.

I played even if I didn't post a character dump as did others. It's not a bad game but the interface is painful. If you fixed that it'd help a lot. Also the wilderness was a pain to navigate. It's a bunch of inconvenience to play when there's a more modern variant forked off it.

I would like to know more details, can you elaborate? What needs to happen to improve the Zangband or Z+Angband interface?

[wobbly]

Thats a shame, Just two participants? How can I submit a competition idea based on a development of Z+angband?

I played even if I didn't post a character dump as did others. It's not a bad game but the interface is painful. If you fixed that it'd help a lot. Also the wilderness was a pain to navigate. It's a bunch of inconvenience to play when there's a more modern variant forked off it.

[Nick]

There was also this: https://github.com/takkaria/angband-webterm

And this: https://github.com/takkaria/angband/tree/wip-textui2

[Gwarl]

About Kangband; it is the source for a lot of the things poschengband inherited though zangband and hengband. Town buildings, town quests, dungeon quests etc.

I like the special quests from tome2, kangband and poschengband, because they offer a special quest location, quest level and a reason to kill a monster. Much better than the random "kill monster on level" quests. If each monster had its own static quest motivation description and both quest building / location and quest level could be procedurally generated, random quests could be more interesting.

The questfile formats from Kangband through to poschengband are rather similar, with the latter being easier to create and work with. If you can hack some procgen into poschengbandlike quests I'd be interested in stealing it.

I have been thinking of using conditionals to randomise layout the way they are currently used to randomise quest rewards, so thieves hideout would have half a dozen layouts. There is a bug which defaults the random number to 0 in some situations. We have a fix for it (thankyou Sideways) but I haven't committed it yet because it'll break savefiles and I have a character on the go.

[Gwarl]

Hmm.

As Nick said the windows interface is great. I guess people who use linux also tend to prefer ascii?

The windows port is generally preferred over an SDL one because of the flexibility of the subwindows which are offered. The GCU ports tend to be rather inflexible on how they do it and the SDL ports are generally just the GCU output translated to pictures.

Poschengband has the most elaborate system for specifying subwindow parameters to be fit into a single window (like the GCU/SDL ports do) but the syntax is arcane and it's done from the command line prior to game start.

I think X11 can do multiple subwindows so you can see what that's about without using micro$oft products.

There's also NotEye which is a graphical frontend designed to be plugged in to curses roguelikes. PWMAngband uses it.

Finally, if you are able and willing to write JavaScript we can pipe the output through the browser and script the layout to draw appropriately onto a canvas element. Probably not that much harder than configuring NotEye. With a bit of effort and understanding we could even get subwindows going running the thing in an embedded browser (e.g. electron)

I want to do it myself but I am short on time lately.

There was also this: https://github.com/takkaria/angband-webterm

[droof]

I want to comment on some of these.

Thank you, I highly appreciate your feedback. What variant would you recommend I should look into for a "best" example SDL interface?

I would like to have both Angband's 'h' quick shoot and Unangband's 'h' item-based interaction keys in Zangband. I find them both very convenient.

Kangband's town features look very similar to those in Poschengband, do they share the same source? Where do the casino's originate from?

The static worlds and stories of tome2 and Unangband offer more variety and personality than Zangband's random world. But maybe Z's random world can be expanded to offer more interesting procedurally generated landscapes, procedurally generate more feature-rich town houses like those found in Kangband and offer multiple town themes to give each town more personality and a reason to explore the world without expecting more of the same. Townfolk could sometimes leak rumours and monsters can speak based on AI actions similar to friendband. Some monsters could have static witty dialogue that suits them.

I like the special quests from tome2, kangband and poschengband, because they offer a special quest location, quest level and a reason to kill a monster. Much better than the random "kill monster on level" quests. If each monster had its own static quest motivation description and both quest building / location and quest level could be procedurally generated, random quests could be more interesting.

If Z+Angband offers sound and music, that's great. If it doesn't work, I'm not keen on fixing separate interfaces. That's where the SDL interface could come to the rescue, one interface to fix and test everything. If I fix it for linux, it'll be fixed for windows too without neglecting one or the other.

[johnretroreload]

Competition response

We had one recently, it didn't prove very popular.

http://angband.oook.cz/competition.php?showcompo=210

Thats a shame, Just two participants? How can I submit a competition idea based on a development of Z+angband?

[Gwarl]

I was also thinking of a zangband competition with prizes, like a games expo or gamedev presentation to really play tribute to the community and fans. Just a thought.

We had one recently, it didn't prove very popular.

http://angband.oook.cz/competition.php?showcompo=210

[johnretroreload]

Nice activity and discussion here, I'm liking the ideas being suggested.
I agree on the random and procedural is what made zangband unique even when wilderness came in for me.
Nice work droof on cleaning the code.
The Quests also made it special for me, I was thinking of more sub-quests with random item drops and even armour sets for bonus? Plus an extra final boss in a special area.
I've noticed on Z+Angband that sound + music doesn't seem to function despite wav and midi in place.
For controls, I still use roguelike myself being oldschool.

I was also thinking of a zangband competition with prizes, like a games expo or gamedev presentation to really play tribute to the community and fans. Just a thought.

Okay, I patched it into 2.7.5, just in case.

I also tried a little harder compiling Z+Angband. The maintainer used windows and forgot linux, looks like. I had to remove a couple of carriage returns from the configure file and add a missing cmd7.o to the makefile.std. Now it runs. I'll try it out.

I like the idea of keeping everything as true procedural and random as possible, not pre-designed. That's what draws me to Zangband. I'll drop a couple of ideas for a Z variant here:
- adding missing procedural unangband rooms for more dungeon variation
- SDL interface (looking at Sangband as example interface)
- optional haggling instead of on / off at birth
- random interesting quests
- skills instead of classes
- chatty townfolk and monsters for added personality and AI decision insight
- story picked random from "7 plots", each with chatbot-style variations
- optional brogue-style inventory-based item interactions as alternative to single-purpose keybindings
- easier doomrl-style shooting controls
- multiple town themes, but still procedural generated