[Crawl-ref-discuss] ##crawl-dev digest September 1-7

[Christoph S. <chr...@fk...>]

Since there hasn't been a ##crawl-dev digest for a while I decided to 
write one. It covers one week, September 1-7. These digests are a lot of 
work, plus it's easy to get severely backlogged if you are busy or away 
for even a few days, and sifting through IRC logs to catch up is no fun 
at all. This is too much to expect a single person to do (thanks to 
evilmike for writing so many of these!), and it is the reason these 
digests have stopped coming. So it would be good to have some kind of 
rotation. If you're a ##crawl-dev regular and feel up to doing one of 
these, just mention it on IRC and add your name and the time you want to 
cover to the "digest" learndb entry!

It has been a while since the last one, so I will summarize what these 
digests are: much of the DCSS development discussion happens in 
##crawl-dev, a chat channel on the freenode IRC network. However, not 
everyone has the time to read all of that discussion, and a summary is 
nice even for the people who were there. This is in no way an official 
development plan, just a list of DCSS development related things that 
were talked about on ##crawl-dev. If something is mentioned that just 
means it has been discussed, not that everyone agrees about it or that 
there was any actual result to the discussion. In fact I am going to 
omit most things that have lead to committed changes, you can read about 
those in the commit log. Further comments and discussion of the 
mentioned issues is welcomed and encouraged. And finally, mistakes are 
possible, I might have summarized something badly or completely missed 
important topics. Feel free to correct me.

Several things were already discussed extensively before September. This 
includes god wrath, random gods and curses. I will not go into details 
on these, because they are already covered fairly well by dev wiki 
entries and/or tavern threads. Note that all of these proposed changes 
are still deep in the design phase, so comments, criticism, and filling 
in specifics would be appreciated. Some links:
- Randgods: 
https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:god:concept:random_god
- Wrath: 
https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:god:concept:wrath
- Curses: 
https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:gameplay:curses

This mailing list was very active last week. Many of the topics were 
also discussed on IRC, but usually the result was reflected in some 
c-r-d email. A quick review of last week's topics on c-r-d (reread the 
emails for details):
- Randomly generated mechanical traps could be strongly reduced in 
frequency, almost to the point of removal. Detection of all traps could 
be done by a one-time dex check instead of skill-based search by 
waiting. It seems most people are on board with this, but there are a 
few minor issues left, mostly how to handle equipment swapping.
- John Attea (Poor Yurik) posted a retrospective on making tiles for 
DCSS on his blog, and it sparked quite a few c-r-d emails and also some 
IRC discussion. As a result ontoclasm joined the devteam, with a focus 
on handling tile submissions and art-related decisions!
- There were some emails about where and for what purposes the new runed 
doors should and shouldn't be used. Runed doors block autoexplore and 
monsters can not open them until you do. There are still some very 
problematic cases left.
- Raphael Langella (galehar) took the time to write a digest for the 
development discussion on the official DCSS forum, the Tavern. As a 
result, I don't have to cover it. Improving the memorization screen and 
changes to book ID and Ds mutations (especially scales) were the 
suggestions that generated some (positive) interest.
- Brendan Hickey (bh on IRC), who designed and implemented the current 
morphing abyss, is working on another abyss layout rewrite. He stated 
his plan for the next newabyss in a c-r-d email. Details were often 
discussed on IRC throughout the week, but I will not summarize any of 
them. The "inception" branch is frequently updated with his latest 
changes. There is also an implementable for additional abyss layout 
generators (6184 on mantis).


Now, finally the actual ##crawl-dev digest, ordered by topic. There was 
a lot of activity last week, so it's a long list:

Dungeon:
- Idea: a rare late-D timed portal vault that either sends you to a 
trove, or to pan. Probably Nemelex flavored. It is unclear whether it 
should be possible to spawn in addition to a regular trove or not.
- More (all?) portal vaults should be announced, especially baileys and 
wizlabs. The timer for unannounced ones should finally be improved. A 
simpler alternative to the long-standing implementable (5503 on mantis) 
would be to print a message and start a timer once you get closer than a 
certain distance to the portal.
- An early demon portal vault would still be cool to have. It is hard to 
do much with 5s, though. And since the terrain would not be very 
different from regular dungeon vaults, the monster set would have to 
carry the whole idea. Possibly imp ideas from 4.1 could be used.
- The current door vault sucks (it uses runed doors instead of secret 
ones, allowing for easy custom-made chokepoints). It needs to be fixed 
or maybe removed. Some other vaults using runed doors also need to be 
looked at (tar ends, due_elkab).
- Do boulder beetle trap vaults generate too early? Should there be some 
warning message?


Monsters:
- It would be nice if Nergalle's tile changed during DDoor. It's very 
easy to miss. Recoloring could work in console.
- There aren't any naga uniques. It would be cool to have one. Ideas 
include using Cheibriados, polearm reaching, summons, statue form, and 
having a snake pet that the naga would somehow interact with (cf. the 
twins, Grum).
- Should spiny frogs be spiny? If not, should they be renamed? Should 
they (or maybe some other enemy) get retaliation to unarmed attacks (for 
example poison)? Can something be done with porcupines to make them 
relevant? Maybe the spines formula could be changed to do damage based 
on the damage you did (possibly different for different weapon types). 
If spines are changed, Hell Sentinel balance should be looked at.
- Is monster pain mirror too hard to notice?
- The abyss monster set needs work. Filling the design gaps in some of 
due's abyss monster ideas (see part of the dev wiki entry "Abyss 
Creatures") and implementing them would help.
- Plain mummy death curses suck. They are relevant only in zigs, and 
they are always annoying. Either because you have to jump through hoops 
to kill mummies, or because you have to uncurse your whole inventory. 
Ideas for changing them include working only or mostly on equipped 
items, frequency diminishing the more of your items are cursed, and 
simply replacing them with toned down versions of proper mummy curses. 
The last idea is probably the best one. If equipment curses were somehow 
made interesting (see dev wiki link above), cursing items could be a 
monster spell.
- Non-plain mummy curses work well, except for guardian mummies in 
ossuaries. Those should probably be removed.


Summons:
- Some things about friendly summons not attacking out of line of sight: 
How to explain it to the player (in spell descriptions, maybe the 
manual)? Do summons from charmed enemy summoners need special handling? 
Will it hurt immersion (the game world seems to be sentient and against 
you)?
- Summons nerf brainstorming: even with a summon cap, summons would 
still be completely throwaway and spammable. So each summon spell could 
draw from a limited pool of creatures. For example you could have a pool 
of 20 butterflies and summon five per cast. So you would reach the cap 
after four castings. If a summon times out, it is returned to the pool. 
But if it dies it is not returned, and the cap is reduced. Possibly 
damaged summons could also be damaged when resummoned. After some time 
the pool is replenished (time may not be the best thing to use).
- Crawl Light neostairs code could be used to implement allies following 
you across levels even if not adjacent, and for interlevel recall if 
desired.


Items:
- Deck of dungeons sucks and should be removed. Interesting cards 
(dowsing) could be moved to other decks. Nemelex deck weightings for 
sacrifices would have to be redone, and should probably just be 
resplaced by fixed weights for the different decks instead.
- Should sabres be renamed to something more appropriate for a short 
blade, which are basically assassination weapons in crawl?
- All fruit could be condensed into a single "fruit" item, possibly with 
different colors. Would require reviewing fruit vaults.
- Should books ID when picking them up? Should burning books give more 
Trog piety? Should you be told that you already have that book if you 
find a duplicate? Should there be some way to check how many books a 
spell you have is in (for example when reading a book, and/or on the 
proposed MM screen)?
- Is item destruction a good mechanic? Should strategic consumables be 
turned into misc items?
- Things that increase MP should use a consistent name (possibilities: 
magic, mana, power, energy). Currently for example there are potion of 
magic, staff of power and crystal ball of energy.
- The staff of energy could be absorbed into the staff of channeling. 
Staff of enchantment should simply be removed, it is not interesting. A 
hexes enhancer on some other item might be good.
- Scrolls of remove curse could work like scrolls of identify. This 
would make them slightly less abundant, and let you uncurse things you 
aren't wearing. It would also change scroll ID a bit by adding another 
item-targeted scroll. Also, uncursing wouldn't have to be specialcased 
for Ashenzari. On the other hand, there is a card that can mass-curse 
items (could be nerfed), and while plain mummy curses are like they are, 
nerfing ?RC would just encourage people to jump through the silly 
mummy-killing hoops even more. Changing plain mummy curses would help here.


Melee combat mechanics:
- Proposal: cleaving for axes. The discussion is summarized very well on 
the dev wiki ("Fighting Reform", axes section). The common ground is 
that while fighting multiple enemies you would be able to do damage to 
multiple enemies at once. Any implementation would require playtesting 
to see if it is useless or overpowered, so probably something should 
just be implemented and put in a branch or directly into trunk, then 
tweaked if the numbers turn out to be bad. There is a patch!
- Weapon moves should probably not depend on skills or stats, to ensure 
they are interesting both in the early and the late game.
- The effect of stats on melee damage and accuracy should be increased.
- The effect of accuracy on the actual hit rate should be smoother.
- The strength requirement for armor is not great. Two problems:
   1) the range of requirements is too large, most importantly dragon 
and plate are too far apart. Elliptic proposes 6+2*EVP instead of the 
current 3*EVP formula.
   2) Hard breakpoints suck, maybe armor skill should have some effect 
on the strength requirement or more strongly reduce the impact of not 
having enough strength.
- Weapon noise is based on post-AC damage, which makes no sense. Also, 
weapon noise is way too complex and should not depend on so much stuff.
- Would sidestepping (EV bonus against missiles for walking at an angle 
or zig-zagging towards the enemy) and charging (damage bonus for walking 
towards the enemy in a straight line) be good or terrible in DCSS? A 
slightly different version of sidestepping is in Crawl Light, so people 
who are interested can just try it there. Elliptic mentioned some 
serious problems of these moves on the dev wiki ("Fighting Reform", 
special moves section): autofight breakage, encouraging of kiting, and 
general annoyance. Different versions of sidestepping or charge could be 
interesting as god abilities.

Tiles:
- John Attea's blog post sparked some discussion about animation in 
DCSS. However, there is no agreement on whether it would be a good 
addition. Animated tiles would certainly look better, but there are a 
lot of problems: it might look too busy, and if the wrong things are 
animated they might distract from important information (this might be 
an issue with animated torches for example). Also, for example adding 
animations to monsters would be a huge amount of work. DCSS has great 
art contributors, so this might actually be possible. But along the 
(very long) way some monsters would be animated and some wouldn't be, 
which would look extremely weird. Adding capability for proper animation 
would be nice, since it might get people to try interesting stuff. Used 
very sparingly, animation could be useful for highlighting important 
things. edlothiol just added optional realtime animations to trunk 
webtiles, so once the change arrives on CSZO you can check out how it 
looks. They are off by default.
- Should there be filter options to go with tile scaling? It might be 
easy to do with OpenGL.
- Having something like webtiles menus in local tiles would be nice, 
because seeing monster list and stats area while in menus is awesome. 
The Android port has some code that could be used for that.
- The local tiles monster list sucks and the webtiles one is much better.


Miscellaneous:
- What to do with sound messages that don't have a silence check because 
they are too important to hide (example: portal announcements)? 
Reflavor, hide when silenced, or do nothing?
- Some kind of advanced interface help would be useful.
- Regarding the tavern proposal to forbid the eating of unclean chunks 
for most races: nausea isn't really that bad. One reason for that is 
that there is a lot of permafood, so the annoyance of eating optimally 
is far from necessary. Reducing early-game food by this much would lead 
to a lot of unfun starvation deaths.
- Should high elves be removed? No, they are different enough from all 
other species, and apts could be adjusted to make them more different. 
Renaming some elven species might be nice, though.
- God wrath duration could work like it does for Ash for all gods: get 
enough experience for x levels to end the wrath. This would be a simple 
way to prevent the waiting out of wraths. Wrath effects would still be 
based on time, and not reduce wrath duration. That would allow people to 
get out of wrath with fewer effects by braving a difficult area. Scaling 
wrath duration and severity with max piety or total gained piety or some 
other metric would be good in addition of course.
- Draining should be improved. For example it could temporarily reduce 
skills (earn exp to remove).
- Stats (str dex int) should be coarser, to make the levelup bonuses 
stand out more and improve readability (smaller numbers). This worked 
very well for skill aptitudes. Proposal:
   - Starting stats and all stat bonuses are halved, and their effect 
doubled to compensate. Rounding up or down for odd numbers has to be 
decided on a case by case basis.
   - Levelup stat picks come at levels 6n+3, starting at 3. Stat prompts 
require pressing an uppercase letter because they would actually be 
relevant.
   - Racial stat gains on levelup are randomly scheduled at char 
creation, similar to Demonspawn mutations.
   - Stat drain and sustab will also need to be looked at.
   - Balance changes (like increasing or decreasing effective stat 
bonuses on equipment or adjusting the power of the stats themselves) 
should be done separately.
- Having an online repository for third party lua scripts would be nice. 
There could be a section on the dev wiki for them. All it would take is 
someone setting it up, and anyone can do that.
- The "how to access help and various other stuff" message spam on 
casting a spell should be reduced. 4 messages for aiming a spell are too 
much. It could be reduced to two lines if the help messages were shortened.


                                              --- alefury