update t-test message and add info to readme on compile time

Makefile

@@ -33,6 +33,9 @@ seedsdir=$(shell head -n1 analysis/allseeds.txt)
 #! allseeds.txt
 allseeds=$(shell tail -n+2 analysis/allseeds.txt)
 
+#! Count how many seeds there are
+n_seeds=$(words $(strip $(allseeds)))
+
 # Check to make sure there are no hyphens in seed names, they'll muck things up
 ifneq ($(findstring -,$(allseeds)),)
 $(error Hyphen in one or more seed name(s))
@@ -41,6 +44,9 @@ endif
 #! What are the groups in this analysis?
 groups=$(patsubst group-%.txt, %, $(wildcard group-[[:alpha:]]*.txt ))
 
+#! Count how many groups there are
+n_groups=$(words $(strip $(groups)))
+
 # Check to make sure there are no hyphens in group names, they'll muck things up
 ifneq ($(findstring -,$(groups)),)
 $(error Hyphen in one or more group name(s))
@@ -52,6 +58,9 @@ contrasts=$(foreach g1,$(groups), \
 			$(foreach g2,$(groups), \
 				$(filter-out $(g1)-$(g1),$(g1)-$(g2) ) ))
 
+#! Count how many contrasts there are
+n_contrasts=$(words $(strip $(contrasts)))
+
 # If a covariate file is given, add the flag to the covariate variable and
 # store the number of covariates for later. Otherwise, store the number of
 # covariates as 0.
@@ -74,12 +83,16 @@ endif
 #! Will be echoed any time makefile is interacted with.
 ifndef PAIRED
 pairflag=
-$(info Doing an UNPAIRED t-test)
+$(info t-test is set to UNPAIRED)
 else
 pairflag=-paired
-$(info Doing a PAIRED t-test)
+$(info t-test is set to PAIRED)
 endif
 
+# Give the user a message while the makefile compiles.
+$(info Preparing your makefile ... $(n_groups) groups, \
+		$(n_contrasts) contrasts, and $(n_seeds) seeds)
+
 ################################################################################
 # Single-group analysis
 

README.md

@@ -9,6 +9,20 @@ Below is a general guide to setting up this pipeline for any project.
 We assume that you will clone this repository and, in that directory,
 follow these instructions.
 
+This script creates a number of recipes based on the number of groups and seeds.
+The number of recipes it has to create is based on:
+
+    seeds * groups + seeds * contrasts
+
+where `contrasts` is:
+
+    seeds^2 - seeds
+
+So, the time it takes make to prepare rises with the square of groups *and*
+linearly with the number of seeds. As these increase, make might take a while to
+get ready to be invoked, and might look like it's hanging. Try `make
+test-contrasts` to see how long it takes to compile.
+
 *Important*: Do not use a hyphen (`-`) in group, seed, or covariate names.
 Or anywhere, really. The system relies in many places on hyphens being used
 exclusively to identify contrasts (e.g. `patient-control`). I have built in a
@@ -308,7 +322,8 @@ one-core execution to complete.
 #### 8.2 Parellelizing on the grid engine
 
 Check that the grid engine is set up properly for your user/machine. See
-[here](http://faculty.washington.edu/madhyt/using-the-gridengine-to-process-data-quickly/) for instructions.
+[here](http://faculty.washington.edu/madhyt/using-the-gridengine-to-process-data-quickly/)
+for instructions.
 
 You can submit the jobs on the queue with `qmake`: