It would be very helpful if the error reported the size of matrix that was being dimensioned, and the dimnames provided.

e.g. Error: The algebra 'thresholdsMatrix' in model 'ACE' generated the error message: length of 'dimnames' [1] not equal to array extent: You are giving dimnames to a 3:4 matrix, but provided 2:4 dimnames:
c("X","X")

It is illegal to create variables other than mxModel components inside an mxModel. Therefore things like this fail (with errors that will be cryptic to users):
lboundList = c(rep(-3,times=nVar), rep(.0001,times=nVar*(maxThresholds-1))),
mxMatrix("Full", maxThresholds, nVar, free=T, lbound = lboundList, values=.5, name="baseAndIncrements"),

Because OpenMx tries to add the items of lboundList to the model:

e.g.

> tableFitStatistics(twinSatFit, c(twinSatSub1Fit, twinSatSub2Fit))
          Name    ep -2LL    df   AIC      diffLL diffdf p   
Model 1 : twinSat 6  1599.78 2282 -2964.22 -      -      -   
Model 2 : twinSat 4  1602.96 2284 -2965.04 3.18   2      0.2 
Model 3 : twinSat 3  1605.63 2285 -2964.37 5.85   3      0.12

> mxCompare(twinSatSub1Fit,c(twinSatSub2Fit))
     base comparison ep minus2LL   df   AIC diffLL diffdf     p
1 twinSat         4     1600 2284 -2970     NA     NA    NA
2 twinSat    twinSat  3     1610 2285 -2960   2.67      1 0.102

*NOTE*: minus2LL is 1600 and 1610

But in the actual models (accessed via summary) the values are 1602.957 and 1605.626

> summary(twinSatSub1Fit)
# observed statistics:  2288 
# estimated parameters:  4 

> summary(twinSatSub1Fit, SaturatedLikelihood=1599.781)

It gets the χ² but not the df between the sub-model and the saturated model.

Do we need a parameter for summary to specifiy "sat model estimated parameters"?

ideally, one could simply do

> summary(subFit, satFit)

And summary would summarise the second model recursively then grab $Minus2LogLikelihood and $estimatedParameters

Cheers,
t

Given these defined variables, I expected that the values and lbounds for the matrix to be created below would be populated repeatedly until the desired number of rows/columns based on nrow and ncol was obtained, but it appears that the dimensions of the matrix are determined by the values and lbound arguments.

> mxMatrix( type="Full", nrow=nth, ncol=ntv, free=TRUE, values=c(-1,.5,.5), lbound=thLBound, name="Thre")
FullMatrix 'Thre'

@labels: No labels assigned.

@values

it would be very helpful to add a call to summary Names and say something like

Error: The data object 'MZ.data' contains an observed matrix that is not of type 'double'
details:
type(MZ.data) is that is fine|I am expecting a matrix or dataframe
MZ.data contains 2 variables as follows:
twin1: type logical
twin2: type integer

Often multivariate scripts won't get started when start values are specified with one value (because matrix is not invertable). I wrote a helper that jiggles starts to avoid this, but it would be nice if mx did this itself, if it can't get started.

This would also take away a lot of frustration from users whose scripts won't run, when the only problem is that the start values are identical across one or more matrices.