diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..4368ee7
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,2 @@
+/planning.pdf
+/planning.tex
diff --git a/planning.org b/planning.org
new file mode 100644
index 0000000..dc724ce
--- /dev/null
+++ b/planning.org
@@ -0,0 +1,64 @@
+#+title:How to plan your project
+#+options: toc:nil
+#+latex_class: scrartcl
+
+* What you need for your reports
+In your reports, you will need to provide some form of table or chart that shows /when/ you plan to complete each `feature'.
+
+I'm assuming here that you have already convinced yourself to take an /agile/ approach to the work, which means that you start with a /backlog/ (or list) of /features/.
+Your list of features might be something like this:
+
+| Feature               |
+|-----------------------|
+| RNN model             |
+| Data cleaning         |
+| Hyperparameter search |
+| LSTM version          |
+| Evaluation strategy   |
+|                       |
+
+You need to make sure that there is enough detail in the description that you know what each item in the list refers to.
+You will then need to organise the feature list into a priority order.
+Usually, you want to do basic or generic things, like work out your evaluation strategy, first.
+So a revised feature list might look like this:
+
+| Feature               | Relies on         |
+|-----------------------+-------------------|
+| Evaluation strategy   |                   |
+| Data cleaning         |                   |
+| RNN model             | clean data / eval |
+| LSTM version          | "                 |
+| Hyperparameter search | rnn and lstm      |
+|                       |                   |
+
+I added a "Relies on" column so that you can see the reasoning behind this ordering.
+
+The next stage is to work out how much time you have and to make a table with a column for the week beginning date.
+These will be your /sprints/, to use the agile term for them. 
+
+| When | What                               |
+|------+------------------------------------|
+| 12-7 | Evaluation strategy, data cleaning |
+| 19-7 | RNN model                          |
+| 26-7 | LSTM model                         |
+|  2-8 | Hyperparameter search              |
+
+Note that I have not left time in here for reading the literature and writing up your results.
+That's because you should incorporate such reading and writing into each sprint.
+Here are some of the things that need to go into a sprint:
+
+| How much | What                              |
+|----------+-----------------------------------|
+|      20% | Literature review                 |
+|      20% | Write what you plan to do         |
+|      40% | Do it                             |
+|      20% | Write about what you actually did |
+
+You may be tempted to produce more glamorous project Gantt charts for your final document.
+However, keep in mind that these are generally just gloss and don't provide any more information than a simple table with headings of /When/ and /What/.
+
+* What you need when you meet with your supervisor
+When you meet with your supervisor, it's a good idea to have your plan at hand.
+In the case that you are ahead of schedule, your supervisor will be encouraged to give you ideas about how to deepen the study and mine the available `bonus' points for your thesis.
+However, if things are not going well or perhaps need to be done in a different order, your supervisor might be able to help identify how to short circuit problems or get something done more quickly.
+You may feel some difficulty in admitting that things aren't going well but be assured that it is quite normal for some problems to be encountered in any study.