diff --git a/src/typescript/control.component.ts b/src/typescript/control.component.ts
index cedc2ea..c270bad 100644
--- a/src/typescript/control.component.ts
+++ b/src/typescript/control.component.ts
@@ -145,7 +145,7 @@ export class ControlComponent implements OnInit {
         this.service.stopTest().subscribe()
     }
 
-    private beforeText = `# To set an output:\noutputs['SystemOnOff'] = 1\n# Any output will be automatically set AFTER the code finishes running\n\ndef someLongFunction(x):\n    return x\n\n# States\ns['x0'] = 23\ns['ux'] = 48\ns['Kp'] = 0.4\ns['Ki'] = 0.3\ns['Kd'] = 0.05\ns['someLongFunction'] = someLongFunction\n# Use it to set variables that will be available in the next block/run`
-    private controllerText = `# The state is kept between runs\nsomeLongFunction = s['someLongFunction']\n\n# The selected inputs are always available\nx = inputs['SomeSelectedInput']\n\nu1 = someLongFunction(x)\n\n# np and math are already imported!\nu = math.sin(u1)\n\noutputs['TheOutput'] = u\n\n# Inputs and outputs are automatically logged.\n# If you want to log something else (to generate graphs)\n# use the log dictionary.\nlog['u1'] = u1`
+    private beforeText = `# Use this scope to set constants and do any startup need.\n# For example, it can be used to turn on subsystems, do heavy one-shot math, or\n# to define constants.\n\n# Some global variables are set automatically. They are:\n# inputs -- a dictionary whose keys are the inputs define above.\n# outputs = dict() -- the key is the output name, the value what should be set.\n# s = dict() -- This is a state variable. It persists across scopes and runs.\n# dt: float -- The sampling interval\n# np: np -- import numpy as np\n# math: math -- import math\n\n# To set an output, use the outputs dictionary:\noutputs['SystemOnOff'] = 1\noutputs['Pump1OnOff']  = 1\noutputs['Pump1PC']     = 100\n\n# To read an input, you must first check its checkbox above. After that it will\n# be accessible through the inputs dict.\ny = inputs['Level1CM']\n\n# Use the s dict to define constants and save values to use on next iterations.\ns['Kp'] = 1\ns['Ki'] = 2\ns['Kd'] = 3\n\n# You can also use the s dict to save functions, but be careful not to rely on\n# global variables, as they won't be available in the next scope/run. All\n# variables are lost between scopes/runs, except for s.\n\n# The timer only starts ticking AFTER this scope has finished, so you can put\n# long running logic here, like advanced controller matrix pré-multiplication.\n# Function declaration has no impact on performance, as the code is compiled\n# once and only rerun, so there's no problem in putting them in the next scope.`
+    private controllerText = `# This scope will be run every dt seconds. If it runs for longer than dt\n# seconds, it will run on the next multiple of dt seconds since the start of the\n# test.\n\n# All variables from the Before section will be present here, plus:\n# t: float -- Time since the test started, in seconds.\n# log: dict() -- Will log the value for the key as if it was an output or input.\n\n# For example, let's recover the PID constants defined in Before:\nKp, Ki, Kd = s['Kp'], s['Ki'], s['Kd']\n\n# Let's calculate the error. It's the setpoint minus the reading\nerror = 1 - inputs['Level1CM']\n\n# Let's log the error. Now we get a nice live graph for it too.\nlog['error'] = error\n\n# You know how to do PID, let's just close the loop here:\noutputs['Pump1PC'] = error\n`
     private afterText = `# Use it to finish before exiting. Like shutting the power off.\noutputs['SystemOnOff'] = 0\n# state and log are not available here`
 }