search

skiadas / PanthR

Statistics front-end and webserver with R connection
1 stars 2 forks source link

Top-Level Menu #5

Open altermattw opened 11 years ago

altermattw commented 11 years ago

I know this will be something that takes several iterations.
Criteria:

  1. Frequent use
  2. Major branch of activity
  3. Intuitive
  4. Corresponds to existing stats software (not necessary, but will speed the learning of new converts)
    • File
    • Import (includes capturing from clipboard)
    • Save
    • Data
    • Compute new variable (includes transformations)
    • Recode variable (a subset of compute new?)
    • Rename variable
    • Merge data
    • Subset (or filter) - includes "remove cases with missing data" feature
    • Reshape
    • Make a copy
    • Plot
    • Analyze
    • Descriptives
    • Frequencies
    • Cross-tabulation
    • Correlation
    • Reliability
    • One sample test
    • Two sample test
    • Paired sample test
    • ANOVA
    • Linear model
    • Logistic model
    • Help
skiadas commented 11 years ago

I agree with 'Import' as a catch-all for both "Open" and "Read Text Data/Clipboard". The dialog on the next step could clear that up.

I think "Save" should be "Export". The dataset is "saved" on the server them moment it is created/changed. What we are talking about is a local copy on the computer.

Possible formats:

This all might be a job for the import tool. I.e. it should not just import various formats, but export them as well.

Under File should have an option for New. We should also allow the possibility of creating an individual variable, instead of a whole dataset. In particular, this "New" option would include the option of creating for example a variable consisting of 1000 values drawn at random according to a normal distribution

Regards Data menu: Rename variable I feel is the kind of thing you could just double-click on the variable to do, in any of the specific places it appears. Recode variable should probably be part of either "new/compute" if you create a new variable out of it, or it should be a contextual feature of the variable's edit window. I mean when we have the view to show the variable in the main window, there could be a lot of actions related to it, starting from the options in "Define Variable Properties" in SPSS to things like recode variable. We do need a separate "recode variable" dialog perhaps, but not sure we need a separate menu item for it. Subset and "Make Copy" are the same, unless you meant "Subset" in the sense of "select cases" in SPSS, without actually changing the data but creating a filter.

Analyze: I'm thinking menu-wise we should split them into many fewer groups, and some of those choices in that list would be options once you are in that menu. So I am thinking top-level we would have 3 groups: Tables, Graphs, Tests. Every analysis is essentially producing one of these. My thinking is each of those would have a standard "Editing Dialog" that you can use to set it up, and in that dialog a prevalent role would be played by a pulldown menu choosing the type of thing (type of test or graph etc). Then that dialog window would be changing accordingly, to facilitate the kinds of inputs. j

skiadas commented 11 years ago

Could we put "merge" under the "reshape" umbrella? Or are they far apart to warrant their own place.

Should "Workspaces" go into a new menu group, or placed under File.

We also need a Settings menu. Unless we stick that under Help.

Btw, currently the menu that appears in the mockup can be set up in the file source/views/navbar.jade The format is simple enough, you can add "li" items where we need new items. Or we can keep the discussion going here.

altermattw commented 11 years ago