Names["PalmerPractice`TestPackage`*"]

Or

?PalmerPractice`TestPackage`*

David's dust story metaphor is right on. There are so many things an inexperienced user can find to do wrong, that finding the one right way to do something can take a long time. In this case, I am happy to report that David's system for managing notebooks and package files in projects is now working for me. I can't say why it didn't work the first time, but it is now working for two test projects, initializing a different package file in each.

In Windows, at least, one can right click in an open Folder and create new files from a New menu. There are many types of files there but only one Wolfram file type, a Wolfram Notebook. Maybe there is some difficulty in it, but it would be really nice if WRI could figure out how to add a Wolfram Package to this menu. This would be the most convenient way to create new packages. (And notebooks also.) One might not even have a convenient Mathematica notebook open to use the Mathematica File/New menu. It's really much more convenient and logical to create new files from the folder that will contain them.

I like that idea, too.

I think Wolfram could improve their design and information.

In Windows, at least, one can right click in an open Folder and create new files from a New menu. There are many types of files there but only one Wolfram file type, a Wolfram Notebook. Maybe there is some difficulty in it, but it would be really nice if WRI could figure out how to add a Wolfram Package to this menu. This would be the most convenient way to create new packages. (And notebooks also.) One might not even have a convenient Mathematica notebook open to use the Mathematica File/New menu. It's really much more convenient and logical to create new files from the folder that will contain them.

So I usually end up creating a Mathematica Notebook, changing the suffix to .m and brushing the warning aside. But that creates the package file with the Notebook statement in it, which also needs to be brushed aside. Why is it necessary to put any cell in the new package file? When you create a new notebook it doesn't contain some default cell. Why should a package file. (Users should be aware that a package file can also contain titles, sections and text cells just like a notebook.)

There is also confusion over whether a package file should have a .m or a .wl suffix. Are they completely interchangeable? Will .m be deprecated? I'm not totally certain but I seem to recall that at some point .wl could not be used - I think it may have been as an init.wl file.

All of this creates a little dust storm that envelops the user, especially those new to writing packages. So guys. why don't you fix it and clarify it?

Oh! Well, that totally makes sense, then. I'm almost disappointed, it looked so mystifying... But I'm glad it's working for you now! Yes, you can use File -> New -> Package to create a new package file, or just create a new plain text file from your browser and give it the .m or .wl extension. Package files are just plain text, which I really like: I can read a package file with any old text editor on any device, whether or not Mathematica is installed (heck, even on my phone if I feel like it). And yet, when I do open the file in Mathematica, I get full syntax highlighting and interactive execution and I can just copy & paste code cells from a notebook to a package and vice versa.

By the way, what you did now was create a package that evaluates a Notebook object that then does stuff... It's a bit circuitous, but you know how to change that now.

I won't go into any depth on your answers to my previous diagnostic questions, the answer all look fine now that I know what happened; let me know if something came up in the process that you do want to know more about.

Gary,

I have a tutorial on my web site on setting up applications with packages. A Mathematica Style. This is, I believe the standard method WRI has set up for private packages and applications.

I implemented your application and tested that it works. It took about 20 minutes. Here is the file structure used by the application.

$UserBaseDirectory/Applications/
     PalmerPractice/
         Kernel/init.m
         DevelopmentNotebooks/TestNotebook.nb
         TestPackage.m

I've attached the TestPackage.m, TestNotebook.nb and init.m notebooks. Perhaps you could set up the folder structure and then insert the notebooks.

With this structure it is not necessary to use SetDirectory. Mathematica will automatically find everything in the application. You can load the package from anywhere. Even though you have only one package, I set it up so it could accommodate a number of packages. The package is loaded with

<< PalmerPractice`

When Mathematica see that, it automatically reads the init.m file in the application, which does the actual loading of package files.

On my Windows 7, Mathematica 10.4.1 system this all works properly and it should also work on your system.

Attachments:

I"ve sold 100's of applications with packages using this structure and they all have been able to install and use the packages. And palettes and stylesheets automatically appear on their menus and the documentation is accessible. All they had to do was to unzip into their $UserBaseDirectory/Applications folder.

So I'm a bit at a loss why this is not working for you, even with a simpler case.

Mathematica knows where to look because of one of the entries on the Path expression. Is it possible that you have somehow changed your Path definitions? Also in GlobalOptions, File Locations, NotebookPath there is an entry that points to:

FrontEnd`FileName[{$UserBaseDirectory, "Applications", Blank[], 
  "FrontEnd", "Palettes"}]

And on the Path expression there is an entry:

C:\Users\djmpark\AppData\Roaming\Mathematica\Applications

Has one of these somehow gotten changed on your computer? Maybe you could reinitialize your options by holding down Ctrl+Shift when starting Mathematica.

The simplest approach is to give packages a binary name: the application name "PalmerPractice" and the package name "TestPackage". So the BeginPackage statement would be:

BeginPackage[""PalmerPractice`TestPackage`"]

When a user just reads in an application such as:

<< PalmerPractice`

Then it looks to the PalmerPractice/Kernel/init.m file to read in all the packages you might have in the application, and do any other special initializations. That's how it's supposed to work, and how it always has worked for me and people I've sent applications. On Windows and Apple. So it is a bit of a mystery as to what is happening on your system.