The Thing Growing Diary- Pre Ars Electronica 98

A Networked Version of The Thing Growing is scheduled to show at the Ars Electronica Festival Sept7th -Sept 14th.(?)

Wed	TODO make dat, make cards, make flyers compile on an impact - or clean temp in zbox(can't) finish scaling the rocks - check check all rock sounds stick the rock on the user's head when she's trapped send a message to the brain to switch to afterRoxChase - and take out the test message in the thingscene make thing run to rock thing gloat and rock suck both quieter one rock seems to dither in the air?- can't repeat this make a global body move that'll stick the thing in a good spot for its final gloat seems that though the rocks can be outrun, they always seem to stick - and maybe this section will go by too fast? check how the thingscene looks with the runafteruser and usermovingtoomuch bits in - in the cave - stand on something taller! check networking Dave says that indeed I do have to put a netTransform with the rocpath name outside one with the roc name
Tues	Got new rocks models from Sam - Dave worried they have too many polygons - oops! So far the new actions of the thing and the rocks look good The rock reset is weird and I would like the rocks to attack from the front more, but really needd to look at this au plein cave. add rock sounds back in! don't forget the .net scenes will need to be altered to match this don't think I'll get Sam's rocks in this time, my rocks at about 75 or 80% will do. Added afterRoxChase action store to brain and AFTERROCKCHASE stage to story - need to call this stage from rock when the rock has trapped the user. Don't forget to take the test button3 afterRoxChase message out of thingscene.all Kinda need a pace about/caper mode for the Thing while it reviles the user in AfterRockChase - needs not to be right behind the rock!
Mon 31 Aug	Trying to find the mergeMotions thing! in Dave's directory he says srcdave MergeMotions- alias merge ../../srcdave/mt/merge foreach i(*.tail) merge $i:r $i $i:r.tg end added new sound and action files there is an odd bug - whereby the thing scresw itself up? Dave worked on the rock movement Sam brought floppy with rock models - but when I got them they were weird
	To Do clean sound files - Sat motion capture to go with sound - Sat add sound and motion to scenefiles and -- Mon/Tues finish rock code - Mon? add new rock models - Mon? add final scene - Wed FTP and make disk backup - Wed make a flyer and cards - Thurs
Fri 28 Aug	made production notes and recorded sounds for running after and catching the user post catchuser remonstration user moving too much first recap in all flavors after rox chase ended up using coaxial cable from dat to osby and mediarecorder - using AES whatever that is
Thurs 27th Aug	Dave and I worked on the Rock stuff - its not done but getting there I don't think I'm going to get to the globalbody stuff Mustn't forget to work out the reset stuff so the sound works
Tues 25th Aug	fadeout message to during interrupt resetting dancestep if interrupted by usermovingtoomuch experiment 1 if dancestae is youdance_instruction, wedance_instruction, youdance_instruction_first, or youdance, I'm resetting to failed - so after going through the setdancestate function in analyse info the dancestate ends up being wedance or repeat dance step if dancestate is complies, recap_instruction, recap_instruction_first, failed or nomotion - reset to youdance - and then it gets spit out according to how the user complied last time if catchuser remonstration - reset to postcatchuserremonstration ->ends up as catcuser remonstration again !! xperiment 2 if dancestate is COMPLIED or RECAP_COMPLIED I resent those two states into the setDanceState function - otherwise Make a POST_USERMTM dancetsate and use it to simply reset the dancestate in every case to WE_DANCE_INSTRUCTION Maybe need a different solution if we're recapping the whole dance? - check this, it seems to be weird otherwise I like solution 2 Obviously too convoluted and in the new brain this will be addressed! !!After a reset, sometimes the sound if fucking up, repeating weirdly
Mon 24 Aug	took rest of last week off Christine et al liked it - McKay wanted to run, anyway they do get progressively bored with it - it'll be good to put in rocks - bug Dave now I have to debug after we get into sulknhide I'm still getting the catchuser stage and must not-think its fixed I'm taking the sound out of the go down - but the growl lasts for ever. Maybe the global body needs to send action a messageto stop! reset is not working - the thing is somehow at the wrong stage when we come back - but only if we go to the very end - not if I hit reset int he middle? yes - stop user. xper1 - stopped user during catch user action/during movingtoomuch actions set navon in analyse if stage_ == sulknhide also need to combine the usermovingtoomuch and catchuser things so that both will effect the thing's decision to sulk and hide plus see last Mon problems slobbersound error - comes at reset?
Mon	Today Christine et al are coming so I want to see how this works for them. thing height/thing emerge not in face/need to stop the stop moving? check in CAVE - user stop? checks that action exist at very beginning if the CATCHUSERREMONSTRATION then goes to RECAP, the soundbite must contain instructions to that effect. the Thing needs to lose patience and go to sulknhide mode for now it will go if the person runs more than 5 times I need to make the gradations of running away - first time etc etc. Problems - reset dancestate to what? -forcibly stop user? something is wrong with globalbody get in face(calcing height from floor but we're getting userpos from 0? - how are we usually getting thing's height?)
Fri 14th Aug	If the user runs and runs and runs - do I want to forcibly stop!? and for how long and how to reset? - the length of one action? - I think so - I don't want the user able to run straight through the thing - can I do this with collision detection? Cannot get the userMovingTooMuch check to work only when I want! Also I am able to interrupting the CATCHUSER - which is bad It was getting out of the USERMOVINGTOOMUCH stage_ too soon - so I reset the stage_ afte the usermoving2much action is finished Maybe I'll just make some stand in sound bites for each action store so I can more easily tell what's going on. - check
Thurs 13th Aug	NB: Sounds to record soundbites for 5 new action stores music - drum sound plus keyboard sound(diff ways) Second "RunAfter" Functionality is to check to see if the user is moving away while the thing is trying to teach her to dance In the app I will put a checkUser() - (maybe the beginning of the userInfo class - yes I made it the start of that class, called by the brain app every 1/2 sec, it will compare userPosThis with UserPosLast, if the distance is greater than x, userMovingTooMuch_ = 1 Also in app - inside else if(brainstate_ == DO_ACTION) if (stage_ == TEACHDANCE) if(userMovingTooMuch_ & dancestate ! = NEXTDANCESTEP/REPEATDANCESTEP/WE_DANCE { soundSource ->stop(); // need to fade out? brainstate = ACTION_DONE} In analyse Info - put before userRunningAway if clause:- if (userMovingTooMuch_) setStage(USERMOVINGTOOMUCH) add this stage definition to the story class Make a makeUserMovingTooMuchDecision() - call from makeDecision three categories and actionStores initialMovingOffence ->pick action from Store repeatMovingOffence ->ditto persistantMovingOffence->ditto offenceNum_ setStage(TEACHDANCE) dancestate = FAILED - (first solution) Make a check in makeTeachDanceDecision() else if (dancestate_ == WE_DANCE) {if (danceMemoryList->getNum()>0) action = "get it out of list" else, pick a dance action from that action store} may also need the check for the recap
Wed	Need to count how many times this happens and then have the thing sulknhide because of it, but don't want to implement this check yet, cos it will also be effected by the other functionality - 2 Need to check and fix my brain and localbody changes!!! ie look at it with the sound on - its not doing many things it should be doing - its doing many things that it should not be doing - and there is no health in it!!! But it is a start! Also need to think - what is the networked user doing during all this? - mood changing I guess Let's add a catchuserduration parse option to tgGlobalBody the danceClock is getting reset which I don't want!- added an onlyOnce Think we should simply go to RECAP without the RECAP_INSTRUCTION, becuase those instructions don't quite fit and instead add the news that we're going to start from the beginning to the REMONSTRATION- check Big problems - check - getting to the right place - in front of and facing the user - during the catching phase/duration issue/ why getActionByEweight is not working. (happy not hap!)
Tues 11th August	Adapting Brain to add Two "runAfter" functionalities The thing will not "stay close" to the user while it dances, so the user can run-away. I need to Thing to react to this. The user will be able to navigate while the thing is teaching - the thing needs to react if the user is driving about and not dancing Functionality 1 Need to figure out why the new actionStores aren't working with getActionByEweight!!! Also need to check who is winning the duration battle - I'd say it was obvioulsy the action! So somehow the have to effect the duration tha the action says it is - impossible now Dave has hidden it inside the action! Add a RUNAFTER storyStage_ In Brain - analyse info function - check the user's distance and orientation from/to thing - fill a userRunningOff_ variable. If userRunningOff - setStage(CATCHUSER) Make a makeRunAfterDecision function - which: picks a runafter action (with duration appropriate to distance?) must make a run-after-actionStore - check make sounds and actions for this actionStore sets storyStage_ back to TEACHDANCE sets danceState_ to POST_CATCHUSER In setDanceState add If danceState_ == POST_CATCHUSER, danceState = CATCHUSER_REMONSTRATION If danceState_ == CATCHUSER_REMONSTRATION, danceState_ = RECAP_INSTRUCTION In makeTeachDanceDecision add if danceState_ == CATCHUSER_REMONSTRATION, action= run_after_remonstration_actionStore_->getActionByEweight(eweight_) must make run_after_remonstration_actionStore_- check make sounds and actions for this actionStore How can I check this? - on a button? Put in a catchUser actionStore and a catchUserRemonstration and use button to change storyStage once we are in the Teachdance stage Also fix GlobalBody runafter Must fix reset of GlobalBody - check the danceClock is getting reset which I don't necessarily want!
Mon 10th August	Continuing with the Globalbody I need the getinface and stay close to have setting which are different for cave and idesk and can be put in from the text file. RunAfter - needs to check the distance from the user to calculate the duration BUT what about the local action's duration??? - needs to come to user's face! Not to the front of CAVE. I want a jumpaUP and a runIn Circle/runAroundUser/1/2 circle around user Put in do nothing for all the dance motions - so user can sneak away! -> this will lead to changes in the brain, and runAfter? Add a disappear into Ground thing? - Check it! - had to disable followground for this function to work. WHy the crash? - cos there are no rocDCS's in the scne file Need to: Put the rocDCS's in the plainscene - check! Put in error checking in the code - check! Check turnaway thoroughly - check! I've made the navOn in script_blowup2 EMERGE instanteanous so I can check whether the turnaway works when theings running to various rocks - seems to. Now I've made it 25 secs - should be until the first dance motion. also put back all the actions! and actionStores - check
Fri 7th August	calculated distance to seach rock calculated nearest rock and made that the destination made a TURNAWAY state so the thing will orient to the rock, then run off and hide under it (maybe I should put a little jump in the TURNAWAY) Added a getRunToRocduration, proportional to the distance. Tested with the usual thingscene and got a core dump when I try the turnaway but not the gotoroc - whappen? bad system call I can call it from the scenefile but not from the brain? getinface -OK stayclose -OK turnback -OK donothing -crashes turnaway -crashes The turnaway problem seems to be in the getTurnAwayDestination() at line 407 I moved the thingscene to thingsceneTEMP! What I need to do - is to run this in the CAVE and on the idesk and refine the calls to the global body - maybe figure out a jump thing?
Thurs 6th August	Today turning attention to the globalBodyDCS and trying to rationalize it. Making a getUserInfo, calculateDestination and doMove function. Now making all the individual destination functions getStayCloseDestination getRunAfterDestination getTurnBackDestination getGetInFaceDestination The RunAfter and StayClose destinations are actually the same so perhpas instead I just need a different duration, more time needed if the user if further away Also added a setState function Next I made a function that finds all the rocks in the scene To-do - calculate the distance to each rock, select the nearest, calculate a sensible duration to get there and getRunToRocDestination Hmm - I also have to make the Thing Turn the right amount before it leaves, maybe I can add to the turn function? or somehow have the turn happen faster than the rest?
Wed 5th August	testing functions doRunNearer - ok doPause -OK doMiss - OK doStop - OK doTrap -OK doRelease -OK doScatter - ? doAvoid -OK followGround -OK on smaller slopes initialized all soundNode pointers to NULL! Make RunNearer - make destination in front of user. I think I can simplify since I'm reusing all the same stuff, the difficulty is that the RunNearer runs a certain distance + vector over time, whereas the doTrap and doMiss, have a destination that they are aiming for so there is a difference. Make Rox always oriented to user Might need them to all go to an initial position Dave will help and make them run on hermite curves - hooray Talked to Sam about the models - maybe he can also help me to beef up the textures for the thing itself.
Tues 4th August	I'm starting to work on the section where the rocks chase the user, because I would like to include that in the version that goes to Ars. Today I fixed various reasons that the rox class crashes and started to fix various of its functions! I also put in some print outs so I can check what each roc is doing. I'm not sure if I have to really rewrite the code in order for the user to get the sense of the rocks ganging up, chasing and trapping her; or if I can use the class more or less as is with clean-ups, and judicious use of all the variables (speed, pause time etc). However, I have decided that the rox should "aim" a little infront of the user, so that it always attacks from the front, and seems to be cutting the user off - I guess different rocks could have different in front distance to aim at? The problem in the CAVE is that you can't seee behind you, so the rocks need to attack from in front and at the side. I also believe that I want the Rock always oriented to the user, so that it grows and gulps straight at her and doesn't turn as it gulps. I further believe that checking the globalBody?, anyway some other code may help me. Perhaps they should both be similarly written - a destination point selected and then a function that just runs for it? Don't forget the models that Sam made. Once the movement is better I also need to fix the morpher - that jitter part is not working especially, Dave fixed it for Margaret D - so I should get her copy. To test this I need to have the user running/moving and it still looking good!