Hi All --
I'd like to conduct some interviews and gather a little information from as many as you as possible about how you:
- How do you organize your JMP demos (data tables, journals, scripts, projects, folders etc.)?
- How do you deal with a request for building a new demo?
- How do you deal with things like version control and testing?
- How do you deal with things like duplicate data sets, scripts and other issues like this?
- How do you build new journals when asked to prepare a demo?
Anyone that is willing to help, leave a note in this thread and we can set up a 30-45 min interview about your SoP.
Just to start the conversation out, here is a look into my organization scheme. It is far from perfect but works reasonably well for me. I should note that I use a Mac, and share files between Mac and Windows. I often demo on Windows through Parallels (as I think many of us do here).
1. I have a shared (Mac/Windows) folder on my Mac HD called "Data Sets". I try to put all of my data sets and demos in here. I really am aiming to not have any JMP tables outside this folder eventually, but I am not there yet completely.
In this folder, I have a crude numbering system that is approximately in chronological order. The first 6 folders are big containers to house data sets related to each of the Application Areas, but not associated with a working Demo. Then I have a number of other buckets of folders I can put things into relating to JMP version-specific stuff, personal data sets, customer data sets etc.
I also have a "To Sort" for things I want to deal with later. That folder is rather full and should be sorted...
The real meat of my library is in the "WorkingDemos" folder. These are projects where I can be reasonably confident that If I open up a journal it will work. Within each working demo, I try to keep all of the files associated with the demo in that folder. This often means makes multiple, local copies of a specific data set.
When I need to build a new project, I will add a new numbered folder and start from a journal that I know will be reasonably similar to the one I am creating. I largely build the journal with copy/paste in JMP and copy data sets and scripts into the folder, as well as any images I have.
I often struggle with things like version control or doing a similar demo later in time. Should I build a new project? Should I modify the old project? I usually lean toward the later, as this keeps the total number of projects to a minimal, and I often refine broad topics like "JMP as an analytic hub" based on new examples I have etc.
I really wished I had a better why of doing content management of data sets, but I tend to like the local copy of everything and stripping out the absolute path and keeping the local. Now, Brady's journal packager makes life a lot easier, but that still doesn't solve the CMS-needs I want around this problem.
(This relates to the Project work we are doing for JMP 13).
