diff --git a/addons/majority_judgment/MajorityJudgmentPoll.gd b/addons/majority_judgment/MajorityJudgmentPoll.gd
index 5db5766..318b9a5 100644
--- a/addons/majority_judgment/MajorityJudgmentPoll.gd
+++ b/addons/majority_judgment/MajorityJudgmentPoll.gd
@@ -107,6 +107,7 @@ func add_judgment(judgment:MajorityJudgmentJudgment) -> void:
 			(existing_judgment.candidate == judgment.candidate)
 			):
 			judgments[i] = judgment
+			return
 	judgments.append(judgment)
 
 
diff --git a/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.gd b/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.gd
index 4367bbf..a2744da 100644
--- a/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.gd
+++ b/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.gd
@@ -20,57 +20,60 @@ func set_poll(__poll:MajorityJudgmentPoll):
 	craft_nodes()
 	update_nodes()
 	
-	yield(get_tree().create_timer(3), "timeout")
+	var provider = MajorityJudgmentDemoProvider.new()
+	start_provider(provider)
 	
-	var j = MajorityJudgmentJudgment.new()
-	j.set_participant(MajorityJudgmentParticipant.make("Plume"))
-	j.set_grade(poll.grading.grades[2])
-	j.set_candidate(poll.candidates[0])
-	poll.add_judgment(j)
-	
-	var sylvain = MajorityJudgmentParticipant.make("Sylvain")
-	j = MajorityJudgmentJudgment.new()
-	j.set_participant(sylvain)
-	j.set_grade(poll.grading.grades[3])
-	j.set_candidate(poll.candidates[0])
-	poll.add_judgment(j)
-	j = MajorityJudgmentJudgment.new()
-	j.set_participant(sylvain)
-	j.set_grade(poll.grading.grades[3])
-	j.set_candidate(poll.candidates[1])
-	poll.add_judgment(j)
-	j = MajorityJudgmentJudgment.new()
-	j.set_participant(sylvain)
-	j.set_grade(poll.grading.grades[4])
-	j.set_candidate(poll.candidates[2])
-	poll.add_judgment(j)
-	
-	j = MajorityJudgmentJudgment.new()
-	j.set_participant(MajorityJudgmentParticipant.make("Sabre"))
-	j.set_grade(poll.grading.grades[5])
-	j.set_candidate(poll.candidates[0])
-	poll.add_judgment(j)
-	
-	update_nodes()
-	
-	yield(get_tree().create_timer(2), "timeout")
-	
-	var tiger = MajorityJudgmentParticipant.make("Tiger")
-	j = MajorityJudgmentJudgment.new()
-	j.set_participant(tiger)
-	j.set_grade(poll.grading.grades[4])
-	j.set_candidate(poll.candidates[0])
-	poll.add_judgment(j)
-	update_nodes()
-	
-	yield(get_tree().create_timer(2), "timeout")
-
-	j = MajorityJudgmentJudgment.new()
-	j.set_participant(tiger)
-	j.set_grade(poll.grading.grades[1])
-	j.set_candidate(poll.candidates[2])
-	poll.add_judgment(j)
-	update_nodes()
+#	yield(get_tree().create_timer(3), "timeout")
+#
+#	var j = MajorityJudgmentJudgment.new()
+#	j.set_participant(MajorityJudgmentParticipant.make("Plume"))
+#	j.set_grade(poll.grading.grades[2])
+#	j.set_candidate(poll.candidates[0])
+#	poll.add_judgment(j)
+#
+#	var sylvain = MajorityJudgmentParticipant.make("Sylvain")
+#	j = MajorityJudgmentJudgment.new()
+#	j.set_participant(sylvain)
+#	j.set_grade(poll.grading.grades[3])
+#	j.set_candidate(poll.candidates[0])
+#	poll.add_judgment(j)
+#	j = MajorityJudgmentJudgment.new()
+#	j.set_participant(sylvain)
+#	j.set_grade(poll.grading.grades[3])
+#	j.set_candidate(poll.candidates[1])
+#	poll.add_judgment(j)
+#	j = MajorityJudgmentJudgment.new()
+#	j.set_participant(sylvain)
+#	j.set_grade(poll.grading.grades[4])
+#	j.set_candidate(poll.candidates[2])
+#	poll.add_judgment(j)
+#
+#	j = MajorityJudgmentJudgment.new()
+#	j.set_participant(MajorityJudgmentParticipant.make("Sabre"))
+#	j.set_grade(poll.grading.grades[5])
+#	j.set_candidate(poll.candidates[0])
+#	poll.add_judgment(j)
+#
+#	update_nodes()
+#
+#	yield(get_tree().create_timer(2), "timeout")
+#
+#	var tiger = MajorityJudgmentParticipant.make("Tiger")
+#	j = MajorityJudgmentJudgment.new()
+#	j.set_participant(tiger)
+#	j.set_grade(poll.grading.grades[4])
+#	j.set_candidate(poll.candidates[0])
+#	poll.add_judgment(j)
+#	update_nodes()
+#
+#	yield(get_tree().create_timer(2), "timeout")
+#
+#	j = MajorityJudgmentJudgment.new()
+#	j.set_participant(tiger)
+#	j.set_grade(poll.grading.grades[1])
+#	j.set_candidate(poll.candidates[2])
+#	poll.add_judgment(j)
+#	update_nodes()
 
 var profiles_nodes := Array()
 func craft_nodes():
@@ -115,8 +118,6 @@ func update_nodes():
 	var tally : MajorityJudgmentPollTally = get_poll().tally()
 	var candidate_index = 0  # as they were initially written, before the sort
 	for candidate in get_poll().get_candidates():
-#		var candidate_tally := tally.get_tally_of_candidate(candidate)
-#		var merit_profile = candidate_tally.merit_profile
 		var profile = profiles_nodes[candidate_index]
 		if tally:
 			var merit = tally.get_tally_of_candidate(candidate).merit_profile
@@ -178,3 +179,39 @@ func create_merit_profile_scene(gradation_size:int):
 	mps.set_poll(get_poll())
 	mps.craft_nodes(gradation_size)
 	return mps
+
+
+var provider:MajorityJudgmentAbstractJudgmentsProvider
+
+func start_provider(__provider):
+	self.provider = __provider
+	var connected = self.provider.connect(
+		"judgment_emitted",
+		self, "__on_judgment_emitted"
+	)
+	self.provider.start_providing()
+
+
+var known_participants := Dictionary()  # id => Participant
+
+func get_or_create_participant(identifier:String) -> MajorityJudgmentParticipant:
+	if not known_participants.has(identifier):
+		known_participants[identifier] = MajorityJudgmentParticipant.make(identifier)
+	return known_participants[identifier]
+	
+
+func __on_judgment_emitted(author_identifier, grade_index, candidate_index):
+	# Data comes from userland, best be careful here
+	
+	
+	var j = MajorityJudgmentJudgment.new()
+	j.set_participant(get_or_create_participant(author_identifier))
+	j.set_grade(poll.grading.grades[grade_index])
+	j.set_candidate(poll.candidates[candidate_index])
+	poll.add_judgment(j)
+	
+	update_nodes()
+
+
+
+
diff --git a/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.tscn b/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.tscn
index 5470076..f3212d2 100644
--- a/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.tscn
+++ b/addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.tscn
@@ -16,7 +16,7 @@ anchor_bottom = 1.0
 
 [node name="ProfilesContainer" type="VBoxContainer" parent="CenterContainer"]
 margin_left = 312.0
-margin_top = 99.0
+margin_top = 100.0
 margin_right = 712.0
-margin_bottom = 499.0
+margin_bottom = 500.0
 rect_min_size = Vector2( 400, 400 )
diff --git a/addons/majority_judgment/providers/MajorityJudgmentAbstractJudgmentsProvider.gd b/addons/majority_judgment/providers/MajorityJudgmentAbstractJudgmentsProvider.gd
index d446355..c43ced3 100644
--- a/addons/majority_judgment/providers/MajorityJudgmentAbstractJudgmentsProvider.gd
+++ b/addons/majority_judgment/providers/MajorityJudgmentAbstractJudgmentsProvider.gd
@@ -1,13 +1,42 @@
 extends Reference
 class_name MajorityJudgmentAbstractJudgmentsProvider
+#class_name MajorityJudgmentAbstractProvider  # perhaps this? (short = sweet)
 
 
-# Abstract class for Judgment Providers
+# Abstract class for (Judgment) Providers
 # The job of the children is to emit the judgment_emitted signal.
+# 
+# Example providers (ideas):
+# - Test/Demo providers
+# - Twitch chat commands (like "A3B2C0")
+# - Youtube Live chat commands (if possible)
+# - IRC chat commands (low, as a bot would be better)
+# - Discord, somehow
+# - Riot, Jami, etc.
+# - Heck, any videoconferencing software with a workable API
+# 
+# So far, providers don't care about whether the candidate or grade indices
+# are within correct range ; that work will be done by the listener to the
+# "judgment_emitted" event, and it will do its best, and ignore bad data.
+# 
+# See MajorityJudgmentLinearResultsControl.gd, that's where a (the) listener is.
 
 
 signal judgment_emitted(
 	author_identifier,  # String, unique per author (aka participant)
-	candidate_index,    # int
-	grade_index         # int (0 == REJECT, up to the grading size minus one)
+	grade_index,        # int (0 == REJECT, up to the grading size minus one)
+	candidate_index     # int, position in the original array of candidates
 )
+
+
+# Called by the scene managing the poll's lifecycle.
+# See MajorityJudgmentLinearResultsControl.gd
+# Meant to be overridden (put your logic here instead of _init)
+func start_providing():
+	pass
+
+
+# Meant to be overridden.
+# In here you can stop your timers, close your network connections, etc.
+func stop_providing():
+	pass
diff --git a/addons/majority_judgment/providers/MajorityJudgmentChatCommandJudgmentsProvider.gd b/addons/majority_judgment/providers/MajorityJudgmentChatCommandJudgmentsProvider.gd
index 6c25a73..da94184 100644
--- a/addons/majority_judgment/providers/MajorityJudgmentChatCommandJudgmentsProvider.gd
+++ b/addons/majority_judgment/providers/MajorityJudgmentChatCommandJudgmentsProvider.gd
@@ -2,17 +2,38 @@ extends MajorityJudgmentAbstractJudgmentsProvider
 class_name MajorityJudgmentChatCommandJudgmentsProvider
 
 
+# Base class for chat command providers to extend.
+# Is able to process_chat_command().
+#
+# Command Syntax
+# --------------
+#
+# <candidate><grade>
+#
+# such as
+#
+# A2
+#
+# means give grade #2 (passable) to candidate A.
+#
+# Grades (when there are 6):
+#
+# 0. TO_REJECT .0
+# 1. POOR      .1
+# 2. PASSABLE  .2
+# 3. GOOD      .3
+# 4. VERY_GOOD .4
+# 5. EXCELLENT .5
+#
+#
 # Command examples
 # ----------------
-# 
+#
 # A0 B2 c0 D1
-# 
+#
 
 
 func process_chat_command(author_identifier:String, chat_command:String):
-	var candidate_index := 0
-	var grade_index := 0
-	
 	var regex = RegEx.new()
 	regex.compile("(?<candidate>[a-zA-Z]{1})(?<grade>[0-9]{1})")
 	
@@ -20,15 +41,14 @@ func process_chat_command(author_identifier:String, chat_command:String):
 	if results:
 		for result in results:
 			var candidate_string : String = result.get_string('candidate')
-			var grade_string : String = result.get_string('grade')
-			
-			grade_index = str2var(grade_string)
 			candidate_string = candidate_string.to_upper()
-			candidate_index = ord(candidate_string) - ord("A")
+			var candidate_index = ord(candidate_string) - ord("A")
+			var grade_string : String = result.get_string('grade')
+			var grade_index = str2var(grade_string)  # FIXME: safer, please
 			
 			emit_signal(
 				"judgment_emitted",
 				author_identifier,
-				candidate_index,
-				grade_index
+				grade_index,
+				candidate_index
 			)
diff --git a/addons/majority_judgment/providers/MajorityJudgmentDemoProvider.gd b/addons/majority_judgment/providers/MajorityJudgmentDemoProvider.gd
new file mode 100644
index 0000000..67d82af
--- /dev/null
+++ b/addons/majority_judgment/providers/MajorityJudgmentDemoProvider.gd
@@ -0,0 +1,20 @@
+extends MajorityJudgmentChatCommandJudgmentsProvider
+class_name MajorityJudgmentDemoProvider
+
+
+func start_providing():
+	yield(App.timer(1), "timeout")
+	process_chat_command("Stup", "A0B1C3")
+	yield(App.timer(1), "timeout")
+	process_chat_command("Flip", "A0B2C1")
+	yield(App.timer(1), "timeout")
+	process_chat_command("Flip", "A1")
+	yield(App.timer(1), "timeout")
+	process_chat_command("Clap", "A3B4")
+	yield(App.timer(1), "timeout")
+	process_chat_command("Clap", "C5B3")
+	yield(App.timer(2), "timeout")
+	process_chat_command("Neo", "B5")
+	yield(App.timer(1), "timeout")
+	process_chat_command("April", "A4")
+
diff --git a/core/App.gd b/core/App.gd
index c6f3ff7..87623e4 100644
--- a/core/App.gd
+++ b/core/App.gd
@@ -42,5 +42,7 @@ func start_poll(poll):
 	var _done = get_tree().change_scene("res://addons/majority_judgment/nodes/MajorityJudgmentLinearResultsControl.tscn")
 	yield(get_tree(), "idle_frame")
 	get_tree().current_scene.set_poll(poll)
-	
 
+
+func timer(duration:int) -> SceneTreeTimer:
+	return get_tree().create_timer(duration)