I write cooking recipes for publication. Documentation is a lot like writing recipes. Recipes are for all types of cooks, beginners and experts. Most of the documentation I find is sparse, lacking details, written for the person who wrote the code. It is hard to write directions for people who don’t know how to use or do the coding. I have been trying to expand the documentation for some Docker containers and I can’t find a new user that can tell me what is missing. I hear everything is find from experts.
Suggestions for requirements for documentation from recipes submission requirements.
Title, something instead of documentation.
Tell us about your recipe. Needs to be a little more than it is a Docker container.
Ingredients. Let people know what is needed. Software, libraries, how much memory needed. Give people an idea if they even can do the task with what system they have. Do not send people to the grocery store for a missing ingredient or to borrow a cup of sugar from a neighbor. This is really important for when people come to a half day workshop only to find out they can’t participate.
Prep method, for recipes what is needed is stove top, oven, grill, blender. What are we doing? Baking? Spinning up a web server, database. Be clearer are we frying or blending?
After clearly stating what we are doing, write out the directions. Include prep and cook time. For example it takes three hours to code this. Build time is half a day. Recipes state what occasion Holiday they are for. Also what category. Github tags make this easy, use them. Course Breakfast, lunch, Dinner. Quantity served. From these ideas Think the documentation thru to more useful documentation. Stack Overflow should not be your only documentation.
On git hub hackoregon/data-science-pet-containers are very useful Docker containers. Tools included are PostgreSQL, PostGIS, Anaconda and Rstudio. Add data and you are all set for exploring and developing. If you have any problems file an issue on git-hub. I am looking for what is missing from the directions.
What if you have a lot of old data files in SPSS form and you do not have SPSS software. You are in luck if you have Rstudio. Current versions of Rstudio load SPSS data files without needing a special library for the task. There are libraries for this task.
To load SPSS data:
Open RStudio, on menu bar choose File, scroll to Import Data set, choose from SPSS.
A window opens to load the data.
Enter file name or url into file/url box.
Preview data in preview window. If it looks okay push the import button. And the data is loaded into Rstudio ready for use.
Cirque du Soleil Kurios show has an act where they mirror a table. It is amazing to see people upside down mirroring a table.
R programming language has several packages for doing tables with R. Basic has a function called table. Which is good enough. Sometimes you want more. At a meeting last night someone said pander was the best package. Someone else said that they liked htmltable better. Also there is xtable and tables. tables was written by someone to be like SAS PROC TABULATE. Many choices, pick out the one that you understand the directions and meets your publishing needs. Better depends on your point of view.
I had some messy data to turn tidy. Column of data that needed to be separated into two columns. All the directions where obscure and not helpful. Try searching for a regular expression on the web.
One of the things I was puzzled over was \\.+ found out it meant gosub(). Much easier to search on. Delimiter was another puzzling thing until I realized that I could treat it the same as when I read csv files. This is the R code that worked.
I am taking a DevOps class. We are using vagrant. Saturday I lost my box. I typed vagrant up on the command line in a terminal window and nothing happened. I was thinking it would pop up like web servers do.
The command that I was missing was vagrant ssh. This command ssh (secure shell) into the virtual box.
vagrant provision command allows you to make changes and add things like games to your virtual box.