I'm going on holiday

[Magnate]

Nearly didn't get a holiday this year but managed to squeeze in a week in October. Looking forward to getting away from the insane demands of Diablo 3 and World of Warcraft. Sorry, I mean of work, DIY and gardening.

Last year I did a bunch of work on item generation which ultimately resulted in the creation of v4. The year before that I rediscovered my love of Sangband so did almost no hacking. This year there is a very real possibility that I'll manage to build Sil for Linux and get hooked ... but in case that doesn't happen, I thought I'd ask people what they would do in my situation.

Here's the V world as I see it:

1. V3.4.0 is almost out (honestly, any day now - the tag is up on github and everything). Immediately after a release is the best time for major V development, because people haven't started finding gnarly bugs in the new release yet (and anyway there's a release manager whose job is to manage the 3.4.x releases - hi fizzix!), so we can get on and try interesting things for 3.5. But the problem here is that it isn't clear to me what, if anything, should be the driving goals for 3.5 - we've reached the point in the roadmap where the only two big outstanding goals are increasing difficulty (which we hope we've done with 3.4.0) and splitting off the UI properly (which I don't think I could do if I tried). The advent of v4 means that changes for 3.5 will be more conservative, at least until something is proven in v4 and ported over. So my love of the unproven will naturally lead me to work on v4 instead.

2. v4 now has new items, new combat, new rooms and new traps. I wouldn't mess with the latter two, but the new combat system needs finishing off (archery, throwing, maybe martial arts?) and of course there's always room for improvement with item generation. But the main thing I was going to do here was rewrite effect handling, because I decided that much more flexibility with effects was necessary before I 'finished' item gen (and rewrote randarts). I did a ton of thinking about this earlier in the year, and got to a point where I was pretty much ready to start hacking - but then Derakon started Pyrel. [That isn't strictly true because WoW intervened, but allow me some narrative licence.]

3. I honestly think Pyrel is the future of Angband (no pressure Derakon!). Put simply, it's in a nicer language, and it doesn't suffer from the hundreds of "you wouldn't start from here" problems you have with the V codebase. I don't think Derakon has made any major design mistakes so far (not that I'd necessarily know if he had, but I'm reasonably confident), and now that the code is public I could brush up my python and make a serious contribution. But it is an awful long way from playable, and I don't know if the things that would be Derakon's priorities are things I'd find interesting to work on.

So what would you do if you were me, and had a week of coding time interrupted only by babies, beer and beaches?

[Derakon]

Well, you know my opinion. For what it's worth, my strategy regarding Pyrel development has been to work on the things that will be needed by the largest number of things that have not yet been implemented. So long as the features you want to implement have had all of their pre-requisites implemented already, you should be clear to start work on your pet project without anything changing out from under you.

I was going to write a longer post here, but thought it'd be better to go into the main Pyrel thread, so I'll just say enjoy your holiday!

[fizzix]

Thanks for this summary. I've been pondering these questions a lot also.

Here's the V world as I see it:

1. V3.4.0 is almost out (honestly, any day now - the tag is up on github and everything). Immediately after a release is the best time for major V development, because people haven't started finding gnarly bugs in the new release yet (and anyway there's a release manager whose job is to manage the 3.4.x releases - hi fizzix!), so we can get on and try interesting things for 3.5. But the problem here is that it isn't clear to me what, if anything, should be the driving goals for 3.5 - we've reached the point in the roadmap where the only two big outstanding goals are increasing difficulty (which we hope we've done with 3.4.0) and splitting off the UI properly (which I don't think I could do if I tried). The advent of v4 means that changes for 3.5 will be more conservative, at least until something is proven in v4 and ported over. So my love of the unproven will naturally lead me to work on v4 instead.

Yes, 3.4.0 will be out as soon as rephial is updated. Should be any day now. I agree that there will not be many major changes to 3.5. Certainly the combat changes of v4 won't go in until we get to 4.0. Here are some things that could very well be included in 3.5

• trap changes (see here in addition to what's already in v4)
• bron's immunity changes
• nomad's new rooms (i wanted this in 3.4.0, but no one was willing to try to implement it and my attempt failed)

2. v4 now has new items, new combat, new rooms and new traps. I wouldn't mess with the latter two, but the new combat system needs finishing off (archery, throwing, maybe martial arts?) and of course there's always room for improvement with item generation. But the main thing I was going to do here was rewrite effect handling, because I decided that much more flexibility with effects was necessary before I 'finished' item gen (and rewrote randarts). I did a ton of thinking about this earlier in the year, and got to a point where I was pretty much ready to start hacking - but then Derakon started Pyrel. [That isn't strictly true because WoW intervened, but allow me some narrative licence.]

I have not played v4 as much as I would have liked, but it does seem like there's plenty of room to tweak the item generation. The first thing I would do is cut the display by a decimal place, so something that's +53 is now +5. That resolution was overkill. The second is some extensive tweaking.

There are big problems in how spellbooks work in v4. Forcing a mage to kill kavlax to get MB5 is not fair at all, and way too much scumming is required to get them in the normal way. I have ideas for this, all of which are hackish, but might be worth discussion.

And of course, the biggest ticket item that hasn't been attempted yet in v4 has been the curses revamp.

3. I honestly think Pyrel is the future of Angband (no pressure Derakon!). Put simply, it's in a nicer language, and it doesn't suffer from the hundreds of "you wouldn't start from here" problems you have with the V codebase. I don't think Derakon has made any major design mistakes so far (not that I'd necessarily know if he had, but I'm reasonably confident), and now that the code is public I could brush up my python and make a serious contribution. But it is an awful long way from playable, and I don't know if the things that would be Derakon's priorities are things I'd find interesting to work on.

I feel this way also. However, there's a trap here where we wait for Pyrel and progress stagnates in the meantime, and people lose interest. However, if contributing to Pyrel is something you think you can do, this might be of use.

So what would you do if you were me, and had a week of coding time interrupted only by babies, beer and beaches?

One thing that I might recommend is to play 3.4.0, play v4, play Sil, FA, tome4, halls of mist and whatever other variants hit your fancy. There are a lot of good ideas in these games that are worth considering implementing.

I have to run to a meeting now, but there's a lot of stuff to ponder about here, and feedback from everyone about what direction they would prefer the V/v4/pyrel branches to head towards would be very useful.

[Magnate]

nomad's new rooms (i wanted this in 3.4.0, but no one was willing to try to implement it and my attempt failed)

I took a look too, but I don't know the dungeon gen code very well and it was more work than I had the capacity to do at the time - sorry.

There are big problems in how spellbooks work in v4. Forcing a mage to kill kavlax to get MB5 is not fair at all, and way too much scumming is required to get them in the normal way. I have ideas for this, all of which are hackish, but might be worth discussion.

Definitely. A quick fix is to choose someone easier than Kavlax for MB5 - Castamir? Lokkak? But also happy to have a wider discussion.

And of course, the biggest ticket item that hasn't been attempted yet in v4 has been the curses revamp.

Yes. That's one of those things where there have been so many threads and ideas that it's hard to know where to start. It needs someone to come in and Just Do It, like Gabe did with traps. Any fresh implementation would be a good starting point for further development.

I feel this way also. However, there's a trap here where we wait for Pyrel and progress stagnates in the meantime, and people lose interest. However, if contributing to Pyrel is something you think you can do, this might be of use.

Having seen Derakon's list of needed features, I fear that I lack the code design skill to contribute this early. It's one thing to make significant improvements to something that already exists - or even a total rewrite of part of an existing codebase - but it's a different thing altogether to design and code something from the ground up. Then there's the fact that I don't really do UI coding in C, let alone in python. I am seriously thinking about having a go at character generation, but it's likely to be a dog's breakfast.

One thing that I might recommend is to play 3.4.0, play v4, play Sil, FA, tome4, halls of mist and whatever other variants hit your fancy. There are a lot of good ideas in these games that are worth considering implementing.

I have two problems with this. One is that I tend to obsess about one game at a time, and I'm likely to get hooked by one and not play the others. The bigger problem is that we have tons of good ideas to implement already, so finding more ideas seems the wrong priority.

[kaypy]

There are big problems in how spellbooks work in v4. Forcing a mage to kill kavlax to get MB5 is not fair at all, and way too much scumming is required to get them in the normal way. I have ideas for this, all of which are hackish, but might be worth discussion.

Is this from theory or experiment? Because my last paythrough went the opposite way- I decided that I had the gear and HP to take out Kavlax and dove down a few levels in a hurry so I could get that spellbook.

Of course my default playspeed is a little on the slow side...

[fizzix]

Of course my default playspeed is a little on the slow side...

By the time I come across kavlax with a mage, I generally do not have enough HP to survive one gravity breath.

[kaypy]

How not enough are we talking? Would swapping in a few Con items and quaffing a !Toughness do the trick? A few wands of drain life will put Kavlax out of commission well before the timeout...

On a related note, is unique autodrops supposed to be the preferred method of getting books? I thought it was mainly there for the no-artifact games?

Should we be talking less about tweaking the fairness of the autodrops, and more about tweaking the artifact rarity so that you get the books naturally at the right times?

[fizzix]

I rarely have found a drain life by the time I reach kavlax. I would say I usually have around 120-200 HP with a mage at that point. Gravity breath is 300+. You don't really get stat gains kicking in until 18/50 at least. I won't hit 300+ HP until very deep in the dungeon (including max CON boosting items, which become a priority very quickly.)

Note: I haven't played v4 in a while, as I've been focusing on v3.4

[Derakon]

On a related note, is unique autodrops supposed to be the preferred method of getting books? I thought it was mainly there for the no-artifact games?

It's not supposed to be, but given current artifact rarities it functionally is. It's been awhile since I played, I admit, but I think I can comfortably say that of my last three caster characters, one of them found one of their class-appropriate books "naturally" (i.e. without autodrops).

[takkaria]

Here are some things that could very well be included in 3.5

• trap changes (see here in addition to what's already in v4)
• bron's immunity changes
• nomad's new rooms (i wanted this in 3.4.0, but no one was willing to try to implement it and my attempt failed)

What are the immunity changes again? I think even just these changes would make a good 3.5.

[fizzix]

What are the immunity changes again? I think even just these changes would make a good 3.5.

Immunities did not prevent physical damage from melee attacks.

See here for the thread and description of the changes.

There was talk of putting it in 3.4 but I felt it was too large of a gameplay change too late in the release cycle to put it in.

[kaypy]

I rarely have found a drain life by the time I reach kavlax. I would say I usually have around 120-200 HP with a mage at that point. Gravity breath is 300+. You don't really get stat gains kicking in until 18/50 at least. I won't hit 300+ HP until very deep in the dungeon (including max CON boosting items, which become a priority very quickly.)

Um. Gravity maxes out at 200... K can dish out 216 for some of his other breaths, but only when unhurt. Of course, if you don't have any decent way to hurt him...

Maybe I was just insanely lucky, but drainlife isn't that far out of depth and I had some by the time I reached K. Still not exactly easy to find, but compared to an artifact?

It's not supposed to be, but given current artifact rarities it functionally is.

I was going to suggest trying to simply up the drop prob on the books, but they are already kind of ludicrous there[1]... How high can that be ramped up[2]? Does that probability affect the overall drop rate of artifacts, or does it just determine which one will drop[3]?

<wanders off to dig through the code...>

[1] edit: or at least they would be if they weren't stuffed into a 1 byte variable immediately after being loaded

[2] edit: it can be ramped up about a quarter as far is it is trying to be

[3] edit: just what type of artifact, given an a priori make_artifact call