Difference between revisions of "ShockFrost"

From RogueBasin
Jump to navigation Jump to search
(Absolutely contentless article - please delete.)
(Really fix bugged links)
 
(6 intermediate revisions by 6 users not shown)
Line 1: Line 1:
ShockFrost is the username of an individual who posted on [[rgrd]] from ?? to ??.  
ShockFrost is the username of an individual who posted on [[rgrd]] from Oct 16 2002 to July 1 2004. In that time, he described an interesting and extremely feature-packed game, generated much enthusiasm for that game, introduced many ideas into the roguelike development community, and did not deliver a working implementation or [[prototype]] of that game.  


ShockFrost had what many considered creative and interesting ideas. These ideas did not find their way into a game by ShockFrost because the [[scope]] of the game was too much for one developer to implement.  
He is well known for his creativity, enthusiasm, and for being a useful counter-example when it comes to good [[project management]]. ShockFrost's name is usually brought up for two reasons: as a warning to developers who are focusing on increasing their [[feature]]s (hence, [[scope]]) more than actually implementing their Roguelike, and as an example of a person admired for their creativity.


Please delete entry.
== Summary ==
 
On Oct 16, 2002, ShockFrost posted a message titled [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/62489397d1a1e73/ An idea for a roguelike] to [[rgrd]], in which he outlined his game concept.
 
The general idea was that the player would start out with one character (whose class was Adventurer) in a specific city/town. The player would descend into the dungeon, and kill bad guys. When the character found and defeated a Wizard, or a Rogue, they would get a medal that matched the class that they defeated. If the player was able to take that medal to the surface, then, when a new character was started, they could choose to be that class. This meant to serve as an easy way for a novice player to get used to the game, without being turned off by the large number of classes. There were 70 or so classes listed, ranging from Viking to Glassblower. Characters could be retired, and their successors would find them living a mostly quiet life in that same town, perhaps as a merchant. If the character died, the town would live on, and another character could come along.
 
The concept was well received and in many cases highly praised, and ShockFrost was given encouragement. It is uncertain as to whether ShockFrost had serious intentions of implementing his ideas at the start, or if he simply wanted to share and get feedback on his ideas. Regardless, hints were were given by ShockFrost that he was planning to implement the idea, and development of the game was later confirmed:
 
:''"And by memory strain, I meant having a dungeon filled with over 100x100 monsters each trying to move after you take your turn. Better go get a cup of coffee. Well, I'm working on it."'' - ShockFrost, Oct 17, 2002
 
:''***HACK! * | A bug just flew down my throat. | . Good enough cue for me. | I just started coding. '' - ShockFrost, Oct 24, 2002
 
:''Well, it's getting late in the game, folks. More of this is done than you dare to imagine. But I'm not imparting anything. '' - ShockFrost, Oct 31, 2002
 
The design was expanded to include various dungeon generation schemes. By this time, many people had expressed concerns that the game was not focused enough, and that the scope had gotten out of hand. Other flaw pointed out was SchockFrost's reliance on external motivation. A clear warning sign was him polling whether readers would like to have real-time gameplay, first person perspective or 3D graphics. He promised to implement such features if there was enough demand for it.
 
Soon after first few people started answering the poll critical and cool-headed voices could be heard.
 
== Cruel to be kind ==
 
First to respond with constructive criticism was [[Copx]]. He asserted that reliance ''on "cheering" from other people for motivation'' inevitably leads to project failure. Instead, two more firmly grounded paradigms for development were suggested. The ages old ''[[start small]], grow tall'' and writing a game targeted at oneself.
 
Then, [[Kornel Kisielewicz|Kornel]] brought up a the notion of needed experience in coding to bring a roguelike game to playable state. Professional programmer could stand a good chance at completing such a task. So could a person who wrote already a RL game. In remaining cases ShockFrost's plans are far too grandiose. Kornel backed his argumentation by recounting his failed attempts and the number of [[rewrite]]s of his then main project [[GenRogue]].
 
Finally, [[Joseph Hewitt]] poured a bucket of cold water over ShockFrost by suggesting he should get an [[@]] walking around first. He strengthened further the reality of failed projects with [[DeadCold]]'s example. This time, the main argument was some things can be learned only with playtesting and that is impossible without a game to play.
 
Some posters argued doomsaying "''you will fail''" was not helping SF mature. This was met with wishes by developers that they had been steered the start small way earlier. The assumption was their current projects would be much more mature. Other counter argument was the creative process of designing and thinking out possibilities should not be judged only on the basis what actually has been produced. Communication of ideas can reveal flaws and lead to better solutions. Moreover such activity is just plain fun thus worthwhile.
 
ShockFrost himself seemed unmoved by all these warnings. He claimed this is not his first major project and stated failure is not going to happen. He assured it several times saying he is going to make his game an accredited college class. There was a contradicting statement in the very same [http://groups.google.com/group/rec.games.roguelike.development/msg/7008d7609ca43597 post] though. He said if not the support of enthusiastic people the sceptical voices would have killed the project off already. At the same time he placed even more requests on community requesting help in creating a ''perfect model for the game with all the goals in mind (...)''. However, his claim was quickly challenged. He did not back up his words. Instead, a new batch of ideas was presented with new posts. Then a rough plan how main game routine would perform. Finally, he named dungeon in the game "DragonMore".
 
Other newsgroup members were simultaneously discussing the cost of creating graphics for a game in a branch of original thread. One of points agreed upon was creating (good) art takes a lot of time and is going to slow down development seriously.
 
== The glorious failure ==
 
On 13th Nov 2002 there was [http://groups.google.com/group/rec.games.roguelike.development/msg/c114ca12ea0b97ca an executable] that placed random unconnected rooms on a map. Newsgroup members were generally disimpressed with this and number of people hoping for eventual release of something playable quickly dwindled.
 
Early in January 2003 a thread [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/5bc62c128a3b1220/ "The SchockFrost memorial day"] was started. Opening topic debated was whether it was good or not to scorn would be DragonMore developer or other newcomers to rgrd with similar attitude so harshly. [[Amy Wang]] who was thought to have left the group participated in the discussion and argued strongly in favor of disillusioning overambitious novice programmers as soon as possible to increase the chance that an actual roguelike game is going to be produced. She did not manage to convince regulars much but her posting contributed strongly to making SF the icon for grandiose [[vapourware]] projects.
 
On 5th Aug 2003 ShockFrost created another {{rgrd|db86aa3f087a6789|thread}}. He started it with a post in roleplaying style. Soon group members saw through his metaphors and got him to respond directly. He asked for advices concerning bash/slash/pierce damage system for weapons and his list of twelve classes. He also asked for some 16x16 pictures and for some list of monsters. Advice was given, sparking some opinion exchange on matters of how medieval weapons were used. Next he posted long list of items he wanted to have in the game with some comments on how he envisioned skill usage by different character classes. Meanwhile concerned regulars discussed the importance of actually releasing something.
 
On 19th Mar 2004 ({{rgrd|c7b396bdb64be214|thread}}) Schockfrost dropped in presenting unrealated project to the roguelike game. It was a card game. Rules were available on his website. According to review posted by one of regulars it actually was ready to play unlike DragonMore.
 
On 11th Jun 2004 he appeared again posting single sentence: {{rgrd|6a82215f2cc8c5bb|'There has been a change'}}. Members of rgrd tried to get him to respond coherently for some time but he would continue to post replies aimed at himself in spirit of roleplaying. Finally he pointed to his website where an executable could be found. It let the user choose several character generation options (the standard character selection screen), rolled a [http://groups.google.com/group/rec.games.roguelike.development/msg/e61c35306437aa82 series of numbers] and then exited. It can no longer be found. Further in the thread he offered an explanation for posting rhetorical nonsense as he named it. The premise was posting about ideas gets him flamed and a lot of unpleasant criticism while muttering to himself avoids that but still may evoke some cheering. Which is very true but then regulars wondered why post at all. Not writing anything frees the time for development leading to playable game. The last is likely to generate a lot of feedback.
 
One interpretation of these events is that ShockFrost posted his ideas in order to get positive feedback. His ideas were very interesting, promised much, and were written in a style that generated excitement, so he did not have trouble getting this positive feedback. ShockFrost, who clearly did not have much (if any) experience in [[project management]], [[software design]], or [[programming]], underestimated the effort required to produce a working [[prototype]]. He marked his progress in terms of either idea generation or decisions made, but certainly did not mark it in terms of software design or programming: none of his writings indicate that he was familiar with software design or that he knew how he could implement any of his ideas, and the final result clearly indicates that he did not get even reasonably far in the coding.
 
== Ideas brought forward by ShockFrost ==
Despite not producing a working roguelike, ShockFrost contributed many unique and interesting ideas.
 
* [[Class]] and [[race]] unlocking to prevent new player from choosing a potentially difficult combination from the start.
** To unlock new class it is necessary to defeat NPC unique of that class, take a medal it drops and bring it to surface.
** Optionally one may unlock and bring up most of skills associated with a class to high proficiency. This would allow to change class in middle of character's adventuring life make it available.
** Exception are prize classes like King. These can be only gained access to by [[wish]]ing.
 
* To attain possibility to play a race one needs to [[genocide]] them. Kill all hobgoblins to play one but you will not see them from now on...
** ...unless you discard hobgoblin medal from game options. Then they will respawn but you lose ability to be one.
 
* Character classes previously not thought of, discarded easily or rarely included.
** Glassblower: Creates containers of various sizes. Vial of levitation could be very useful for bypassing a single pit with poisoned spikes but to cross a river without getting wet a standard issue bottle is needed. Want to levitate for an hour? Chug that gallon pitcher of potion but watch out for nasty side effects of overdosage.
** Sage: Get [[experience]] merely for [[exploration|exploring]]?
** Miner: Access to riches but also dangers such as gas pockets.
** Tailor: Soft armor maker. Requires corpses to tailor something from it. Freshly killed carrion ensures best quality armor.
** Shaman: A spellcaster without his own reserve of [[mana]]. Instead his magic is fueled by supplies of a tile he is standing on.
 
* Character retirement places them in starting city as useful NPCs. May sell stuff for player or be taken as an ally for a
 
* An actual use for [[attributes|charisma]] attribute making it not entirely a [[dump stat]]. Retired characters need to have at least basic social skills or they will get in trouble and end up killed by some other town inhabitant.
 
* Special [http://groups.google.com/group/rec.games.roguelike.development/msg/fa650e03db89d018 room varieties]: brood rooms, hive rooms, pits, lairs and vaults.
 
* Dungeon [http://groups.google.com/group/rec.games.roguelike.development/msg/86f1e454be2bb3b8 design idea] based on strongest monster. The layout would be altered to that monster's theme.
 
* Arachtar monster. Really, just a drider with different name. This was pointed out to him.
 
* Level generation based on primary [http://groups.google.com/group/rec.games.roguelike.development/msg/04df4a6b3bbd92cd? focus]. This focus may be:
** An [[out of depth]] wandering monster, potentially possesing an excellent item. Signs warning about such monster everywhere on the level.
** Significantly higher creature population factor of otherwise ordinary level.
** A hive of monsters. It is a cross between a monster vault from [[Crawl]] and monster pit from [[Angband]] with a twist. Representants of selected creature type are present scattered about the floor.
** A lair. Whole area is influenced by <creature> lair theme. Monster leaders with escorts wander randomly.
** A brood of monsters. Chosen specimen will dominate the level almost completely and defines layout of the area. A hidden room containing brood leaders is placed where young are likely to be found.
** An available medal. Kill the (always hostile) holder to gain access to new race/class.
** Human party. Small group of humans is hiding somewhere nearby. They might have a leader. Usually amount of their treasure is large. Level is slightly changed to indicate their presence.
** A notable [[shop|store]]. Modifies only part of the area it is generated in.
** No food level. Planned to be more difficult to explore and contain little comestibles.
** Dark level.
** [[Boss]] and minions.
** Treasure room. Generated with varying amount of guards. Might be of different flavors:
*** "Ordinary" treasure room.
*** Treasury: only money of equivalent valuables.
*** Mundanes: great amount of usually common items.
*** Disposables: food and single use items.
*** Rarities: ''special ingredients''? Was not explained thoroughly.
*** Single item: usually an artifact.
*** Jewelry.
:: ShockFrost said the best treasure would be kept in rooms hidden carefully and to find them one would have to search thoroughly every level. Actually that one might not be a good idea at all.
 
* Some vague [http://groups.google.com/group/rec.games.roguelike.development/msg/6b66cc842a8c6524 ideas] about dungeon generation but no algorithms.
 
* An assortment of magical [http://groups.google.com/group/rec.games.roguelike.development/msg/362a177fe0e68154 item categories]:
** Balm: Applied to skin for immediate effect. Potency depends how much bare skin has been exposed at the time of use. Armor needs to be taken off to use those most effectively.
** Essence: A perfume that has two possible uses. Breathe it for a moderate strength short term effect or splash it in your face for longer duration and stronger effect at the cost of overdose penalties kicking in.
** Marble: A glass sphere containing a prepared spell. Throw it to make it break and release its magic.
** Draught (archaic for draft): To be quaffed.
** Powder: Gives choice how much of powder to use. Sprinkle a bit on armor/weapon to magically enhance it. May be thrown at monsters.
** Gem: A resource. May be crushed to create a powder or used in alchemical reaction.
** Reagent: Might be ash, water, bone, poison sacs and other such items. Themselves without an obvious use but valued by magic users who employ them to strenghten their spells..
** Orb: An amplifier for certain magic shool. Staves in [[DCSS]] work this way.
** Totem: Enables or enhances certain powers. Need skill in specific magic shool to use them.
** Rod: Contains magical spell but relies on user's energy to cast it.
** Staff: Contains energy for casting spells from a certain shool.
** Wand: Contains spell and required energy to release it but unlike two item types above this one has limited charges. Considered to be an accesory unworthy of a real wizard because it requires almost no skill to use.
** Runestone: Contain only a part of spell and need contact with bare hand to activate. Non-magicians can use them with a little effort but only skilled users can combine multiple runestones to create more complex magic. These stones may detonate when overused or combined with other stones containing conflicting energies. Breaking such stone usually has no effect.
** Filament: A scroll that needs to be burned to work.
 
* Some [http://groups.google.com/group/rec.games.roguelike.development/msg/ae137bc4983ab6a4 party mechanics]:
** Each party member has a guts meter. If it depletes they chicken out. Close to morale system but with minor difference. Characters of exact same class badly influence each other because they want to fill the same role in party.
** Loot gained needs to be split fairly. Mages want books while warriors want weapons. Characters may trade items.
** Player character may earn favor with the group by attacking monsters currently fought by the party, sharing loot belonging to himself and disarming traps. With score good enough other heroes might resurrect the player if he falls.
 
* [http://groups.google.com/group/rec.games.roguelike.development/msg/4dcfa327c67e15b4 Tarrasque monster].
 
* Item damage system somewhat similar to what [[NetHack]] uses but with heavy maintenance burden placed on player. ([http://groups.google.com/group/rec.games.roguelike.development/msg/14ce2c6ac8ff6009 1] [http://groups.google.com/group/rec.games.roguelike.development/msg/e607a0851698aafd 2] [http://groups.google.com/group/rec.games.roguelike.development/msg/025fb55b8ee998a5 3] [http://groups.google.com/group/rec.games.roguelike.development/msg/3239514f4fedb11a 4])
 
== Ideas contributed by rgrd members in responses to SchockFrost ==
 
* The Taarg, a [[monster]] [http://groups.google.com/group/rec.games.roguelike.development/msg/98b3ee16f214c693 race] described by Timo Viitanen
: Supplement [http://groups.google.com/group/rec.games.roguelike.development/msg/cdbac837fbf1e365 here].
* Log Golems described by R. Alan Monroe:
In cultivated forest areas you could have Log Golems. A jointed
creature with small, sturdy sawn log pieces for body & limbs. I bet
there are loads of natural materials you could animate as Golems,
other than the traditional stone.''
* Weapons of psychic or spiritual damage instead of physical. (Amy Wang)
 
== Reference ==
Usenet threads of relevance:
* [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/62489397d1a1e73/ An idea for a roguelike] October 2002
* [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/5bc62c128a3b1220/ ShockFrost memorial day] January 2003
* [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/db86aa3f087a6789/ The Darkest Hour] August 2003
* [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/c7b396bdb64be214/ *yawns* Hiya, how is everyone? I did a thing, come see.] March 2004
* [http://groups.google.ca/group/rec.games.roguelike.development/browse_frm/thread/6a82215f2cc8c5bb/ *lurks deep in the shadows*] June 2004
 
[[Category:Developers]]

Latest revision as of 19:27, 29 May 2013

ShockFrost is the username of an individual who posted on rgrd from Oct 16 2002 to July 1 2004. In that time, he described an interesting and extremely feature-packed game, generated much enthusiasm for that game, introduced many ideas into the roguelike development community, and did not deliver a working implementation or prototype of that game.

He is well known for his creativity, enthusiasm, and for being a useful counter-example when it comes to good project management. ShockFrost's name is usually brought up for two reasons: as a warning to developers who are focusing on increasing their features (hence, scope) more than actually implementing their Roguelike, and as an example of a person admired for their creativity.

Summary

On Oct 16, 2002, ShockFrost posted a message titled An idea for a roguelike to rgrd, in which he outlined his game concept.

The general idea was that the player would start out with one character (whose class was Adventurer) in a specific city/town. The player would descend into the dungeon, and kill bad guys. When the character found and defeated a Wizard, or a Rogue, they would get a medal that matched the class that they defeated. If the player was able to take that medal to the surface, then, when a new character was started, they could choose to be that class. This meant to serve as an easy way for a novice player to get used to the game, without being turned off by the large number of classes. There were 70 or so classes listed, ranging from Viking to Glassblower. Characters could be retired, and their successors would find them living a mostly quiet life in that same town, perhaps as a merchant. If the character died, the town would live on, and another character could come along.

The concept was well received and in many cases highly praised, and ShockFrost was given encouragement. It is uncertain as to whether ShockFrost had serious intentions of implementing his ideas at the start, or if he simply wanted to share and get feedback on his ideas. Regardless, hints were were given by ShockFrost that he was planning to implement the idea, and development of the game was later confirmed:

"And by memory strain, I meant having a dungeon filled with over 100x100 monsters each trying to move after you take your turn. Better go get a cup of coffee. Well, I'm working on it." - ShockFrost, Oct 17, 2002
***HACK! * | A bug just flew down my throat. | . Good enough cue for me. | I just started coding. - ShockFrost, Oct 24, 2002
Well, it's getting late in the game, folks. More of this is done than you dare to imagine. But I'm not imparting anything. - ShockFrost, Oct 31, 2002

The design was expanded to include various dungeon generation schemes. By this time, many people had expressed concerns that the game was not focused enough, and that the scope had gotten out of hand. Other flaw pointed out was SchockFrost's reliance on external motivation. A clear warning sign was him polling whether readers would like to have real-time gameplay, first person perspective or 3D graphics. He promised to implement such features if there was enough demand for it.

Soon after first few people started answering the poll critical and cool-headed voices could be heard.

Cruel to be kind

First to respond with constructive criticism was Copx. He asserted that reliance on "cheering" from other people for motivation inevitably leads to project failure. Instead, two more firmly grounded paradigms for development were suggested. The ages old start small, grow tall and writing a game targeted at oneself.

Then, Kornel brought up a the notion of needed experience in coding to bring a roguelike game to playable state. Professional programmer could stand a good chance at completing such a task. So could a person who wrote already a RL game. In remaining cases ShockFrost's plans are far too grandiose. Kornel backed his argumentation by recounting his failed attempts and the number of rewrites of his then main project GenRogue.

Finally, Joseph Hewitt poured a bucket of cold water over ShockFrost by suggesting he should get an @ walking around first. He strengthened further the reality of failed projects with DeadCold's example. This time, the main argument was some things can be learned only with playtesting and that is impossible without a game to play.

Some posters argued doomsaying "you will fail" was not helping SF mature. This was met with wishes by developers that they had been steered the start small way earlier. The assumption was their current projects would be much more mature. Other counter argument was the creative process of designing and thinking out possibilities should not be judged only on the basis what actually has been produced. Communication of ideas can reveal flaws and lead to better solutions. Moreover such activity is just plain fun thus worthwhile.

ShockFrost himself seemed unmoved by all these warnings. He claimed this is not his first major project and stated failure is not going to happen. He assured it several times saying he is going to make his game an accredited college class. There was a contradicting statement in the very same post though. He said if not the support of enthusiastic people the sceptical voices would have killed the project off already. At the same time he placed even more requests on community requesting help in creating a perfect model for the game with all the goals in mind (...). However, his claim was quickly challenged. He did not back up his words. Instead, a new batch of ideas was presented with new posts. Then a rough plan how main game routine would perform. Finally, he named dungeon in the game "DragonMore".

Other newsgroup members were simultaneously discussing the cost of creating graphics for a game in a branch of original thread. One of points agreed upon was creating (good) art takes a lot of time and is going to slow down development seriously.

The glorious failure

On 13th Nov 2002 there was an executable that placed random unconnected rooms on a map. Newsgroup members were generally disimpressed with this and number of people hoping for eventual release of something playable quickly dwindled.

Early in January 2003 a thread "The SchockFrost memorial day" was started. Opening topic debated was whether it was good or not to scorn would be DragonMore developer or other newcomers to rgrd with similar attitude so harshly. Amy Wang who was thought to have left the group participated in the discussion and argued strongly in favor of disillusioning overambitious novice programmers as soon as possible to increase the chance that an actual roguelike game is going to be produced. She did not manage to convince regulars much but her posting contributed strongly to making SF the icon for grandiose vapourware projects.

On 5th Aug 2003 ShockFrost created another thread

. He started it with a post in roleplaying style. Soon group members saw through his metaphors and got him to respond directly. He asked for advices concerning bash/slash/pierce damage system for weapons and his list of twelve classes. He also asked for some 16x16 pictures and for some list of monsters. Advice was given, sparking some opinion exchange on matters of how medieval weapons were used. Next he posted long list of items he wanted to have in the game with some comments on how he envisioned skill usage by different character classes. Meanwhile concerned regulars discussed the importance of actually releasing something.

On 19th Mar 2004 (thread

) Schockfrost dropped in presenting unrealated project to the roguelike game. It was a card game. Rules were available on his website. According to review posted by one of regulars it actually was ready to play unlike DragonMore.

On 11th Jun 2004 he appeared again posting single sentence: 'There has been a change'

. Members of rgrd tried to get him to respond coherently for some time but he would continue to post replies aimed at himself in spirit of roleplaying. Finally he pointed to his website where an executable could be found. It let the user choose several character generation options (the standard character selection screen), rolled a series of numbers and then exited. It can no longer be found. Further in the thread he offered an explanation for posting rhetorical nonsense as he named it. The premise was posting about ideas gets him flamed and a lot of unpleasant criticism while muttering to himself avoids that but still may evoke some cheering. Which is very true but then regulars wondered why post at all. Not writing anything frees the time for development leading to playable game. The last is likely to generate a lot of feedback.

One interpretation of these events is that ShockFrost posted his ideas in order to get positive feedback. His ideas were very interesting, promised much, and were written in a style that generated excitement, so he did not have trouble getting this positive feedback. ShockFrost, who clearly did not have much (if any) experience in project management, software design, or programming, underestimated the effort required to produce a working prototype. He marked his progress in terms of either idea generation or decisions made, but certainly did not mark it in terms of software design or programming: none of his writings indicate that he was familiar with software design or that he knew how he could implement any of his ideas, and the final result clearly indicates that he did not get even reasonably far in the coding.

Ideas brought forward by ShockFrost

Despite not producing a working roguelike, ShockFrost contributed many unique and interesting ideas.

  • Class and race unlocking to prevent new player from choosing a potentially difficult combination from the start.
    • To unlock new class it is necessary to defeat NPC unique of that class, take a medal it drops and bring it to surface.
    • Optionally one may unlock and bring up most of skills associated with a class to high proficiency. This would allow to change class in middle of character's adventuring life make it available.
    • Exception are prize classes like King. These can be only gained access to by wishing.
  • To attain possibility to play a race one needs to genocide them. Kill all hobgoblins to play one but you will not see them from now on...
    • ...unless you discard hobgoblin medal from game options. Then they will respawn but you lose ability to be one.
  • Character classes previously not thought of, discarded easily or rarely included.
    • Glassblower: Creates containers of various sizes. Vial of levitation could be very useful for bypassing a single pit with poisoned spikes but to cross a river without getting wet a standard issue bottle is needed. Want to levitate for an hour? Chug that gallon pitcher of potion but watch out for nasty side effects of overdosage.
    • Sage: Get experience merely for exploring?
    • Miner: Access to riches but also dangers such as gas pockets.
    • Tailor: Soft armor maker. Requires corpses to tailor something from it. Freshly killed carrion ensures best quality armor.
    • Shaman: A spellcaster without his own reserve of mana. Instead his magic is fueled by supplies of a tile he is standing on.
  • Character retirement places them in starting city as useful NPCs. May sell stuff for player or be taken as an ally for a
  • An actual use for charisma attribute making it not entirely a dump stat. Retired characters need to have at least basic social skills or they will get in trouble and end up killed by some other town inhabitant.
  • Special room varieties: brood rooms, hive rooms, pits, lairs and vaults.
  • Dungeon design idea based on strongest monster. The layout would be altered to that monster's theme.
  • Arachtar monster. Really, just a drider with different name. This was pointed out to him.
  • Level generation based on primary focus. This focus may be:
    • An out of depth wandering monster, potentially possesing an excellent item. Signs warning about such monster everywhere on the level.
    • Significantly higher creature population factor of otherwise ordinary level.
    • A hive of monsters. It is a cross between a monster vault from Crawl and monster pit from Angband with a twist. Representants of selected creature type are present scattered about the floor.
    • A lair. Whole area is influenced by <creature> lair theme. Monster leaders with escorts wander randomly.
    • A brood of monsters. Chosen specimen will dominate the level almost completely and defines layout of the area. A hidden room containing brood leaders is placed where young are likely to be found.
    • An available medal. Kill the (always hostile) holder to gain access to new race/class.
    • Human party. Small group of humans is hiding somewhere nearby. They might have a leader. Usually amount of their treasure is large. Level is slightly changed to indicate their presence.
    • A notable store. Modifies only part of the area it is generated in.
    • No food level. Planned to be more difficult to explore and contain little comestibles.
    • Dark level.
    • Boss and minions.
    • Treasure room. Generated with varying amount of guards. Might be of different flavors:
      • "Ordinary" treasure room.
      • Treasury: only money of equivalent valuables.
      • Mundanes: great amount of usually common items.
      • Disposables: food and single use items.
      • Rarities: special ingredients? Was not explained thoroughly.
      • Single item: usually an artifact.
      • Jewelry.
ShockFrost said the best treasure would be kept in rooms hidden carefully and to find them one would have to search thoroughly every level. Actually that one might not be a good idea at all.
  • Some vague ideas about dungeon generation but no algorithms.
  • An assortment of magical item categories:
    • Balm: Applied to skin for immediate effect. Potency depends how much bare skin has been exposed at the time of use. Armor needs to be taken off to use those most effectively.
    • Essence: A perfume that has two possible uses. Breathe it for a moderate strength short term effect or splash it in your face for longer duration and stronger effect at the cost of overdose penalties kicking in.
    • Marble: A glass sphere containing a prepared spell. Throw it to make it break and release its magic.
    • Draught (archaic for draft): To be quaffed.
    • Powder: Gives choice how much of powder to use. Sprinkle a bit on armor/weapon to magically enhance it. May be thrown at monsters.
    • Gem: A resource. May be crushed to create a powder or used in alchemical reaction.
    • Reagent: Might be ash, water, bone, poison sacs and other such items. Themselves without an obvious use but valued by magic users who employ them to strenghten their spells..
    • Orb: An amplifier for certain magic shool. Staves in DCSS work this way.
    • Totem: Enables or enhances certain powers. Need skill in specific magic shool to use them.
    • Rod: Contains magical spell but relies on user's energy to cast it.
    • Staff: Contains energy for casting spells from a certain shool.
    • Wand: Contains spell and required energy to release it but unlike two item types above this one has limited charges. Considered to be an accesory unworthy of a real wizard because it requires almost no skill to use.
    • Runestone: Contain only a part of spell and need contact with bare hand to activate. Non-magicians can use them with a little effort but only skilled users can combine multiple runestones to create more complex magic. These stones may detonate when overused or combined with other stones containing conflicting energies. Breaking such stone usually has no effect.
    • Filament: A scroll that needs to be burned to work.
  • Some party mechanics:
    • Each party member has a guts meter. If it depletes they chicken out. Close to morale system but with minor difference. Characters of exact same class badly influence each other because they want to fill the same role in party.
    • Loot gained needs to be split fairly. Mages want books while warriors want weapons. Characters may trade items.
    • Player character may earn favor with the group by attacking monsters currently fought by the party, sharing loot belonging to himself and disarming traps. With score good enough other heroes might resurrect the player if he falls.
  • Item damage system somewhat similar to what NetHack uses but with heavy maintenance burden placed on player. (1 2 3 4)

Ideas contributed by rgrd members in responses to SchockFrost

Supplement here.
  • Log Golems described by R. Alan Monroe:
In cultivated forest areas you could have Log Golems. A jointed 
creature with small, sturdy sawn log pieces for body & limbs. I bet 
there are loads of natural materials you could animate as Golems, 
other than the traditional stone.
  • Weapons of psychic or spiritual damage instead of physical. (Amy Wang)

Reference

Usenet threads of relevance: