JacquesCarette
Remove Quantity typeclass instance of QDefinition.
Contributes to #4097
A QDefinition is not a Quantity. A QDefinition only pairs a 'quantity' with a 'defining formula'.
So what do you think a Quantity is? I agree that a QDefinition pairs a 'quantity' with a 'defining formula'. The real question: is the 'defining formula' a mere annotation, thus making a QDefinition a Quantity, or something entirely different, thus changing the very nature of QDefinition?
I finished the work, and I've also marked the PR as "ready for review" so that the CI can run. However, the PR is not quite "ready to merge" because your comment has me second-guessing.
I see your point, and I subscribe to your view on what a 'Quantity' is.
Brain dump: I think of 'QDefinition's as specific solutions to calculating a quantity. i.e., its nature is about calculation, not about defining a symbol. If a 'QDefinition' were a 'Quantity', then could we say that a 'QDefinition' is equal to a 'UnitalChunk' about them being 'Quantities'? Would that relate to them having the same symbol, UID, space, formula (the 'UnitalChunk' doesn't have a formula)?
If a 'QDefinition' provides an approximate solution to a QuantityDict, does it change the definition of the QuantityDict? Of course, I'm not talking about mutation here, but about a cloned QuantityDict that explicitly mentions it is approximated, potentially with a slightly different symbol.
I think we can also look at the "names" (I'll refrain from saying "terms" here) of these things. Wouldn't we want the names of the QuantityDict to be different from the QDefinition? We currently have this in Drasil with our IMs showing "Calculation of $x$" -- that is information embedded in the QDefinition, arguably hacked in through its internal DefinedQuantityDict and sometimes overridden by the carrier EquationalModel. For example, Projectile has:
timeIM :: InstanceModel
timeIM = imNoRefs (equationalModelN (nounPhraseSP "calculation of landing time") timeQD)
[qwC launSpeed $ UpFrom (Exc, exactDbl 0)
,qwC launAngle $ Bounded (Exc, exactDbl 0) (Exc, half $ sy pi_)]
(qw flightDur) [UpFrom (Exc, exactDbl 0)]
(Just timeDeriv) "calOfLandingTime" [angleConstraintNote, gravitationalAccelConstNote, timeConsNote]
Overall, I still tend towards a 'QDefinition' not being a 'Quantity', but I'm not entirely convinced anymore. What do you think?
I think of 'QDefinition's as specific solutions to calculating a quantity. i.e., its nature is about calculation, not about defining a symbol.
That a priori makes sense. Though I think that's what DataDefinitions are for. I think the case of QDefinition was when you are creating a new symbol, but you are also associating a way to compute it at the same time.
How many QDefinitions do we have? I think we need to look at all of them, to see what knowledge we're trying to capture with them. That's really going to be where the proper design has to stem from.
[I have no idea why you're talking about 'approximation' above.]
Overall, I still tend towards a 'QDefinition' not being a 'Quantity', but I'm not entirely convinced anymore.
Whereas I suspect that the real problem is that we are not capturing the information that we need "the right way".
How many QDefinitions do we have? I think we need to look at all of them, to see what knowledge we're trying to capture with them. That's really going to be where the proper design has to stem from.
Using rg "^[^\-].+fromEqn|mkQuantDef| ec " -ths --sort-files --stats to estimate, we have around 180:
189 matches
184 matched lines
34 files contained matches
469 files searched
13492 bytes printed
2798685 bytes searched
0.002789 seconds spent searching
0.037277 seconds
All matches (hidden).
drasil-data/lib/Data/Drasil/Equations/Defining/Physics.hs
29:accelerationQD = mkQuantDef QP.acceleration accelerationEqn
32:velocityQD = mkQuantDef QP.velocity velocityEqn
35:newtonSLQD = fromEqn' "force" (nounPhraseSP "Newton's second law of motion")
drasil-data/lib/Data/Drasil/Quantities/Math.hs
42:piConst = mkQuantDef pi_ (dbl 3.14159265)
drasil-data/lib/Data/Drasil/Quantities/Physics.hs
145:gravitationalConstValue = mkQuantDef gravitationalConst (dbl 6.6743E-11)
149:gravitationalAccelConst = mkQuantDef gravitationalMagnitude (dbl 9.8)
drasil-data/lib/Data/Drasil/Theories/Physics.hs
39:weightQD = mkQuantDef' QP.weight (nounPhraseSP "weight") weightEqn
83:hsPressureQD = mkQuantDef' QP.pressure (nounPhraseSP "hydrostatic pressure") hsPressureEqn
100:torque = mkQuantDef QP.torque torqueEqn
113:vecMagQD = mkQuantDef QP.speed speedEqn
131:newtonSLRQD = mkQuantDef' QP.torque (nounPhraseSP "Newton's second law for rotational motion") newtonSLRExpr
drasil-example/dblpend/lib/Drasil/DblPend/DataDefs.hs
25:positionGQD = mkQuantDef velocity positionGEqn
37:positionXQD_1 = mkQuantDef xPos_1 positionXEqn_1
55:positionYQD_1 = mkQuantDef yPos_1 positionYEqn_1
73:positionXQD_2 = mkQuantDef xPos_2 positionXEqn_2
91:positionYQD_2 = mkQuantDef yPos_2 positionYEqn_2
109:accelGQD = mkQuantDef acceleration accelGEqn
121:forceGQD = mkQuantDef force forceGEqn
drasil-example/dblpend/lib/Drasil/DblPend/GenDefs.hs
38:velXQD_1 = mkQuantDef' xVel_1 (the xComp `NP.of_` (velocity `ofThe` firstObject)) E.velXExpr_1
67:velYQD_1 = mkQuantDef' yVel_1 (the yComp `NP.of_` (velocity `ofThe` firstObject)) E.velYExpr_1
88:velXQD_2 = mkQuantDef' xVel_2 (the xComp `NP.of_` (velocity `ofThe` secondObject)) E.velXExpr_2
109:velYQD_2 = mkQuantDef' yVel_2 (the yComp `NP.of_` (velocity `ofThe` secondObject)) E.velYExpr_2
130:accelXQD_1 = mkQuantDef' xAccel_1 (the xComp `NP.of_` (acceleration `ofThe` firstObject)) E.accelXExpr_1
156:accelYQD_1 = mkQuantDef' yAccel_1 (the yComp `NP.of_` (acceleration `ofThe` firstObject)) E.accelYExpr_1
177:accelXQD_2 = mkQuantDef' xAccel_2 (the xComp `NP.of_` (acceleration `ofThe` secondObject)) E.accelXExpr_2
198:accelYQD_2 = mkQuantDef' yAccel_2 (the yComp `NP.of_` (acceleration `ofThe` secondObject)) E.accelYExpr_2
drasil-example/gamephysics/lib/Drasil/GamePhysics/DataDefs.hs
42:ctrOfMass = mkQuantDef posCM ctrOfMassEqn
70:linDisp = mkQuantDef QP.linearDisplacement dispEqn
99:linVel = mkQuantDef QP.linearVelocity velEqn
117:linAcc = mkQuantDef QP.linearAccel accelEqn
135:angDisp = mkQuantDef QP.angularDisplacement angDispEqn
153:angVel = mkQuantDef QP.angularVelocity angVelEqn
171:angAccel = mkQuantDef QP.angularAccel angAccelEqn
198:chasles = mkQuantDef' velB (nounPhraseSP "Chasles' theorem") chaslesEqn
218:impulseV = mkQuantDef QP.impulseV impulseVEqn
249:reVelInColl = mkQuantDef initRelVel reVelInCollEqn
269:coeffRestitution = mkQuantDef QP.restitutionCoef coeffRestitutionEqn
289:kEnergy = mkQuantDef QP.kEnergy kEnergyEqn
303:momentOfInertia = mkQuantDef QP.momentOfInertia momentOfInertiaEqn
320:potEnergy = mkQuantDef QP.potEnergy potEnergyEqn
drasil-example/gamephysics/lib/Drasil/GamePhysics/GenDefs.hs
70:accelGravityQD = mkQuantDef' QP.gravitationalAccel (nounPhraseSP "Acceleration due to gravity") accelGravityExpr
150:impulseQD = mkQuantDef' QP.impulseS (nounPhraseSP "Impulse for Collision") impulseExpr
drasil-example/gamephysics/lib/Drasil/GamePhysics/IMods.hs
48:transMotQD = mkQuantDef accj transMotExpr
94:rotMotQD = mkQuantDef angAccj rotMotExpr
drasil-example/gamephysics/lib/Drasil/GamePhysics/TMods.hs
35:newtonTLQD = mkQuantDef' force_1 (nounPhraseSP "Newton's third law of motion") newtonTLExpr
86:newtonSLRQD = mkQuantDef' torque (nounPhraseSP "Newton's second law for rotational motion") newtonSLRExpr
drasil-example/glassbr/lib/Drasil/GlassBR/DataDefs.hs
37:hFromtQD = mkQuantDef minThick hFromtEq
48:loadDFQD = mkQuantDef lDurFac loadDFEq
63:glaTyFacQD = mkQuantDef gTF glaTyFacEq
75:standOffDisQD = mkQuantDef standOffDist standOffDisEq
86:aspRatQD = mkQuantDef aspectRatio aspRatEq
97:eqTNTWQD = mkQuantDef eqTNTWeight eqTNTWEq
108:calofDemandQD = mkQuantDef demand calofDemandEq
drasil-example/glassbr/lib/Drasil/GlassBR/IMods.hs
53:riskQD = mkQuantDef riskFun ((sy sflawParamK $/
67:strDisFacQD = mkQuantDef stressDistFac strDisFacEq
87:nonFLQD = mkQuantDef nonFactorL nonFLEq
102:dimLLQD = mkQuantDef dimlessLoad dimLLEq
117:tolPreQD = mkQuantDef tolLoad tolPreEq
129:tolStrDisFacQD = mkQuantDef sdfTol $ ln (ln (recip_ (exactDbl 1 $- sy pbTol))
142:probOfBreakQD = mkQuantDef probBr (exactDbl 1 $- exp (neg $ sy $ risk ^. output))
152:calofCapacityQD = mkQuantDef lRe (sy (nonFL ^. output) $* sy (glaTyFac ^. defLhs) $* sy loadSF)
162:pbIsSafeQD = mkQuantDef isSafePb (sy probBr $< sy pbTol)
174:lrIsSafeQD = mkQuantDef isSafeLR (sy lRe $> sy demand)
drasil-example/glassbr/lib/Drasil/GlassBR/TMods.hs
30:lrIsSafeQD = mkQuantDef' isSafeLoad (nounPhraseSP "Safety Load") lrIsSafeExpr
45:pbIsSafeQD = mkQuantDef' isSafeProb (nounPhraseSP "Safety Probability") pbIsSafeExpr
drasil-example/glassbr/lib/Drasil/GlassBR/Unitals.hs
160:dimMax = mkQuantDef (unitary "dimMax"
164:dimMin = mkQuantDef (unitary "dimMin"
168:arMax = mkQuantDef (vc "arMax"
172:cWeightMax = mkQuantDef (unitary "cWeightMax"
176:cWeightMin = mkQuantDef (unitary "cWeightMin"
180:sdMax = mkQuantDef (unitary "sdMax"
184:sdMin = mkQuantDef (unitary "sdMin"
188:stressDistFacMin = mkQuantDef (vc "stressDistFacMin" (nounPhraseSP "minimum value for the stress distribution factor")
191:stressDistFacMax = mkQuantDef (vc "stressDistFacMax" (nounPhraseSP "maximum value for the stress distribution factor")
408:constantK = mkQuantDef sflawParamK $ dbl 2.86e-53
409:constantM = mkQuantDef sflawParamM $ exactDbl 7
410:constantModElas = mkQuantDef modElas $ dbl 7.17e10
411:constantLoadDur = mkQuantDef loadDur $ exactDbl 3
412:constantLoadSF = mkQuantDef loadSF $ exactDbl 1
drasil-example/hghc/lib/Drasil/HGHC/HeatTransfer.hs
47:htTransCladCool = fromEqn "htTransCladCool" (nounPhraseSP
61:htTransCladFuel = fromEqn "htTransCladFuel" (nounPhraseSP
drasil-example/pdcontroller/lib/Drasil/PDController/DataDefs.hs
26:ddErrSigDefn = mkQuantDef qdProcessErrorFD ddErrSigEqn
51:ddPropCtrlDefn = mkQuantDef qdPropControlFD ddPropCtrlEqn
72:ddDerivCtrlDefn = mkQuantDef qdDerivativeControlFD ddDerivCtrlEqn
95:ddCtrlVarDefn = mkQuantDef qdCtrlVarFD ddCtrlEqn
drasil-example/pdcontroller/lib/Drasil/PDController/Unitals.hs
118:odeAbsTolConst = mkQuantDef dqdAbsTol (dbl 1.0e-10)
119:odeRelTolConst = mkQuantDef dqdRelTol (dbl 1.0e-10)
drasil-example/projectile/lib/Drasil/Projectile/DataDefs.hs
24:speedIXQD = mkQuantDef ixVel $ sy iSpeed $* cos (sy launAngle)
25:speedIYQD = mkQuantDef iyVel $ sy iSpeed $* sin (sy launAngle)
drasil-example/projectile/lib/Drasil/Projectile/GenDefs.hs
37:rectVelQD = mkQuantDef' projSpeed (nounPhraseSent $ foldlSent_
61:rectPosQD = mkQuantDef' projPos (nounPhraseSent $ foldlSent_
85:velVecQD = mkQuantDef' velocity (nounPhraseSent $ foldlSent_
102:posVecQD = mkQuantDef' position (nounPhraseSent $ foldlSent_
drasil-example/projectile/lib/Drasil/Projectile/IMods.hs
44:timeQD = mkQuantDef flightDur E.flightDur'
83:landPosQD = mkQuantDef landPos E.landPosExpr
116:offsetQD = mkQuantDef offset E.offset'
126:messageQD = mkQuantDef message E.message
drasil-example/projectile/lib/Drasil/Projectile/Unitals.hs
47:tol = mkQuantDef (vcSt "tol" (nounPhraseSP "hit tolerance") (autoStage vEpsilon) Real) (perc 2 2)
drasil-example/sglpend/lib/Drasil/SglPend/DataDefs.hs
27:positionIXQD = mkQuantDef QP.ixPos positionIXEqn
43:positionIYQD = mkQuantDef QP.iyPos positionIYEqn
60:frequencyDDQD = mkQuantDef QP.frequency frequencyDDEqn
75:angFrequencyDDQD = mkQuantDef QP.angularFrequency angFrequencyDDEqn
89:periodSHMDDQD = mkQuantDef QP.period periodSHMDDEqn
drasil-example/sglpend/lib/Drasil/SglPend/GenDefs.hs
45:velocityIXQD = mkQuantDef' xVel (the xComp `NP.of_` (velocity `ofThe` pendulum))
72:velocityIYQD = mkQuantDef' yVel (the yComp `NP.of_` (velocity `ofThe` pendulum)) $ express E.velocityIYExpr
93:accelerationIXQD = mkQuantDef' xAccel (the xComp `NP.of_` (acceleration `ofThe` pendulum))
122:accelerationIYQD = mkQuantDef' yAccel (the yComp `NP.of_` (acceleration `ofThe` pendulum)) $ express E.accelerationIYExpr
183:angFrequencyQD = mkQuantDef' angularFrequency (angularFrequency `the_ofThe` pendulum) $ express E.angFrequencyExpr
219:periodPendQD = mkQuantDef' period (NP.the (period `ofThe` pendulum)) $ express E.periodPendExpr
drasil-example/ssp/lib/Drasil/SSP/DataDefs.hs
42:intersliceWtrFQD = mkQuantDef watrForce intersliceWtrFEqn
65:angleAQD = mkQuantDef baseAngle angleAEqn
84:angleBQD = mkQuantDef surfAngle angleBEqn
102:lengthBQD = mkQuantDef baseWthX lengthBEqn
115:lengthLbQD = mkQuantDef baseLngth lengthLbEqn
132:lengthLsQD = mkQuantDef surfLngth lengthLsEqn
149:slcHeightQD = mkQuantDef midpntHght slcHeightEqn
168:normStressQD = mkQuantDef totNormStress normStressEqn
179:tangStressQD = mkQuantDef tangStress tangStressEqn
191:ratioVarQD = mkQuantDef scalFunc ratioVarEqn
207:convertFunc1QD = mkQuantDef shrResC convertFunc1Eqn
225:convertFunc2QD = mkQuantDef mobShrC convertFunc2Eqn
247:resShearWOQD = mkQuantDef shearRNoIntsl resShearWOEqn
269:mobShearWOQD = mkQuantDef shearFNoIntsl mobShearWOEqn
297:nrmForceSumQD = ec nrmForceSum (inxi intNormForce $+ inxiM1 intNormForce)
300:watForceSumQD = ec watForceSum (inxi watrForce $+ inxiM1 watrForce)
303:sliceHghtRightQD = ec sliceHghtRight (inxi slopeHght $- inxi slipHght)
306:sliceHghtLeftQD = ec sliceHghtLeft (inxiM1 slopeHght $- inxiM1 slipHght)
drasil-example/ssp/lib/Drasil/SSP/GenDefs.hs
194:normShrR = mkQuantDef intShrForce nmShrRExpr
drasil-example/ssp/lib/Drasil/SSP/IMods.hs
61:fctSftyQD = mkQuantDef' fs factorOfSafety fctSftyExpr
385:nrmShrForQD = mkQuantDef normToShear nrmShrFExpr
611:crtSlpIdQD = mkQuantDef fsMin crtSlpIdExpr
drasil-example/ssp/lib/Drasil/SSP/TMods.hs
42:factOfSafetyQD = mkQuantDef' fs factorOfSafety factOfSafetyExpr
84:mcShrStrgthQD = fromEqnSt' (shrStress ^. uid) (nounPhraseSP "Mohr-Coulumb shear strength")
114:effStressQD = fromEqnSt' (effectiveStress ^. uid) (nounPhraseSP "effective stress")
drasil-example/swhs/lib/Drasil/SWHS/DataDefs.hs
40:waterMassQD = mkQuantDef wMass waterMassEqn
54:waterVolumeQD = mkQuantDef wVol waterVolumeEqn
70:tankVolumeQD = mkQuantDef tankVol tankVolumeEqn
81:balanceDecayRateQD = mkQuantDef tauW balanceDecayRateEqn
97:balanceDecayTimeQD = mkQuantDef eta balanceDecayTimeEqn
109:balanceSolidPCMQD = mkQuantDef tauSP balanceSolidPCMEqn
122:balanceLiquidPCMQD = mkQuantDef tauLP balanceLiquidPCMEqn
135:ddHtFusionQD = mkQuantDef htFusion htFusionEqn
154:ddMeltFracQD = mkQuantDef meltFrac meltFracEqn
172:aspRatQD = mkQuantDef aspectRatio aspRatEq
drasil-example/swhs/lib/Drasil/SWHS/GenDefs.hs
61:htFluxWaterFromCoilQD = mkQuantDef htFluxC htFluxWaterFromCoilExpr
75:htFluxPCMFromWaterQD = mkQuantDef htFluxP htFluxPCMFromWaterExpr
drasil-example/swhs/lib/Drasil/SWHS/TMods.hs
84:sensHtEQD pc eqn desc = fromEqnSt'' "sensHeat" np desc (symbol sensHeat) (sensHeat ^. typ) eqn
drasil-example/swhs/lib/Drasil/SWHS/Unitals.hs
474:consTolAux = mkQuantDef consTol $ perc 1 5
477:tankLengthMin = mkQuantDef (unitary "tankLengthMin"
481:tankLengthMax = mkQuantDef (unitary "tankLengthMax"
485:fracMinAux = mkQuantDef fracMin $ dbl 1.0e-6
487:arMin = mkQuantDef aspectRatioMin $ dbl 0.01
488:arMax = mkQuantDef aspectRatioMax $ exactDbl 100
491:pcmDensityMin = mkQuantDef (unitary' "pcmDensityMin"
495:pcmDensityMax = mkQuantDef (unitary' "pcmDensityMax"
500:htCapSPMin = mkQuantDef (unitary "htCapSPMin"
504:htCapSPMax = mkQuantDef (unitary "htCapSPMax"
509:htCapLPMin = mkQuantDef (unitary "htCapLPMin"
513:htCapLPMax = mkQuantDef (unitary "htCapLPMax"
518:htFusionMin = mkQuantDef (unitary "htFusionMin"
522:htFusionMax = mkQuantDef (unitary "htFusionMax"
527:coilSAMax = mkQuantDef (unitary' "coilSAMax"
532:wDensityMin = mkQuantDef (unitary' "wDensityMin"
536:wDensityMax = mkQuantDef (unitary' "wDensityMax"
541:htCapWMin = mkQuantDef (unitary' "htCapWMin"
546:htCapWMax = mkQuantDef (unitary' "htCapWMax"
552:coilHTCMin = mkQuantDef (unitary' "coilHTCMin"
558:coilHTCMax = mkQuantDef (unitary' "coilHTCMax"
565:pcmHTCMin = mkQuantDef (unitary' "pcmHTCMin"
571:pcmHTCMax = mkQuantDef (unitary' "pcmHTCMax"
578:timeFinalMax = mkQuantDef (unitary' "timeFinalMax"
drasil-example/swhsnopcm/lib/Drasil/SWHSNoPCM/DataDefs.hs
18:waterVolumeQD = mkQuantDef wVol waterVolumeEqn
drasil-lang/lib/Language/Drasil/Chunk/Eq.hs
10: fromEqn, fromEqn', fromEqnSt,
11: fromEqnSt', fromEqnSt'', mkQDefSt, mkQuantDef, mkQuantDef', ec,
109:mkQDefSt u n s symb sp (Just ud) e = fromEqnSt u n s symb sp ud e
110:mkQDefSt u n s symb sp Nothing e = fromEqnSt' u n s symb sp e
113:mkQuantDef :: (Quantity c, MayHaveUnit c) => c -> e -> QDefinition e
114:mkQuantDef c = mkQDefSt (c ^. uid) (c ^. term) EmptyS (symbol c) (c ^. typ) (getUnit c)
118:mkQuantDef' :: (Quantity c, MayHaveUnit c) => c -> NP -> e -> QDefinition e
119:mkQuantDef' c t = mkQDefSt (c ^. uid) t EmptyS (symbol c) (c ^. typ) (getUnit c)
drasil-lang/lib/Language/Drasil.hs
111: , QDefinition, fromEqn, fromEqn', fromEqnSt, fromEqnSt', fromEqnSt''
112: , mkQDefSt, mkQuantDef, mkQuantDef', ec
358:import Language.Drasil.Chunk.Eq (QDefinition, fromEqn, fromEqn', fromEqnSt,
359: fromEqnSt', fromEqnSt'', mkQDefSt, mkQuantDef, mkQuantDef', ec,
[I have no idea why you're talking about 'approximation' above.]
Trying to explain why I was talking about 'approximation' shows me that I was just confused. We can safely ignore that question.
Whereas I suspect that the real problem is that we are not capturing the information that we need "the right way".
I certainly agree with that, too.
I think part of the issue is that we don't know what the Quantity constraint means. Does it mean "is a quantity"? Or does it mean "represents a quantity"?
-- | A Quantity is an 'Idea' with a 'Space' and a 'Symbol'.
-- In theory, it should also restrict to being a part of 'MayHaveUnit', but that causes
-- all sorts of import cycles (or lots of orphans).
class (Idea c, HasSpace c, HasSymbol c) => Quantity c where
Whatever we decide on, the comment needs to be updated.
That being said, what differences between "represents a quantity" and "is a quantity" do you see?
I can vaguely see a difference: "is a quantity" means that the data types can be treated as if they were quantities/variables about specific dimensions, and "represents a quantity" means that data types strictly represent quantities, "quantities" are defined by those specific features (Idea, HasSpace, HasSymbol, May HaveUnit), and might impose extra restrictions about those dimensions (a quasi-type).
I'm not sure if I like the existence of Quantity at all. Why shouldn't quantities (variables) always be represented by the same type (QuantityDict at the moment, but DefinedQuantityDict preferred because it has the prose descriptions as well)?
Nor do we really know what QDefinition is supposed to embody. So it is really hard to know of those instances should be there or not.
&
I think of 'QDefinition's as specific solutions to calculating a quantity. i.e., its nature is about calculation, not about defining a symbol.
That a priori makes sense. Though I think that's what DataDefinitions are for. I think the case of QDefinition was when you are creating a new symbol, but you are also associating a way to compute it at the same time.
Hmm, now that's a surprising twist to me (that this is what DataDefinitions and QDefinitions were previously for). In that case, would QDefinition be a chunk at all? The comment for DataDefinition would also be a bit odd.
-- | A data definition is a 'QDefinition' that may have additional notes:
-- the scope, any references (as 'DecRef's), maybe a derivation, a label ('ShortName'), a reference address, and other notes ('Sentence's).
data DataDefinition where
DDE :: SimpleQDef -> DDPkt -> DataDefinition
DDME :: ModelQDef -> DDPkt -> DataDefinition
data DDPkt = DDPkt {
_pktUID :: UID,
_pktST :: ScopeType,
_pktDR :: [DecRef],
_pktMD :: Maybe Derivation,
_pktSN :: ShortName,
_pktS :: String,
_pktSS :: [Sentence]
}
^ 'term' and 'abrv' are inherited from the QDef (which also inherits this information from its internal DefinedQuantityDict), but doesn't inherit the defn -- only the QDef has a defn that describes what the formula is (except this formula is
Similar to our recent discussion about IMs, GDs, and TMs wrapping "theories" with display/presentation information, I think DDs might be the same, except purely specialized for QDefinitions.
Taking a step back on all of these and rethinking their relationships is probably a good idea right about now...
'5 apples' is a quantity (of apples); '(n+k+3) apples' represents a quantity of apples. One is immediate knowledge, one is future knowledge. That's not the only difference, but it's a big one.
Whenever you're not sure between 'is' and 'represents', go stare at This painting.
I'm not sure if I like the existence of Quantity at all. Why shouldn't quantities (variables) always be represented by the same type (QuantityDict at the moment, but DefinedQuantityDict preferred because it has the prose descriptions as well)?
Because whatever a 'Quantity' is, it should be a promise of certain pieces of information being available. The exact representation should not matter. That a particular representation may in fact contain additional information most definitely should not matter.
Our important internal concepts should be 'information promises' and not concrete representations.
In all cases, the most important thing (whether it's QDefinition, DefinedQuantityDict or DataDefinition) is "what information does this structure promise to provide". At use-site, we then can use exactly those items that contain at least all the information that is needed to function properly.