Allin Cottrell schrieb:
You have a point. I'm not totally clear on how all the specific
menu
items should best be distributed between the Data, Sample and Variable
headings. The layout has basically "just grown" to its current form,
and I'd be glad to hear suggestions for rationalizing it.
Ok, here are my suggestions for the "evolution" of the menu structure,
reshuffling some menu entries
between the "data", "sample", "variable" menus to hopefully
make it more intuitive. (Of course,
sometimes it might be the case that I simply haven't understood what the respective
command really
does... In those cases, please correct and enlighten me.)
I) "Data" menu: I interpret that as relating to the loaded dataset as a whole,
so one may think of
renaming that to "Dataset" instead. Given that interpretation, the following
items should be moved
elsewhere:
1. graph specified vars (-> "sample")
2. Multiple scatterplots (-> "sample")
3. everything from summary statistics to differences of variances should also go to
"sample"
II) "Sample" menu: in my view, all commands that act on or modify data in the
current sample range
(as opposed to the entire dataset) should come here. (see what was moved from
"data" in I) Therefore
remove:
1. Dataset structure / compact data /expand data (-> "data(set)")
2. restructure panel / transpose data (also -> "data(set)")
[3. not sure about the missing-value and case marker things, don't know enough about
them)
III) "Variable" menu: interpreted as variants of univariate analsis, so move:
1. find (-> "data(set)")
2. define new variable ( "data(set)")
[3. again not sure about the missing-value issue, why here _and_ in "sample"?)
So what do you guys think? Of course experienced users would have to adapt, but I would
argue the
changes are relatively mild and make more sense in the longer run.
cheers,
sven