Control Measure fragments: separate the drawn content from the construction scaffolding

[ctgnz]

Every other fragment holds exactly one content element. Control Measure fragments hold two or three,
because they carry illustrative scaffolding alongside the content:

shape	files
<g>#main + <g>#template	534
<g>#main + <g>#template + <g>#example	6
<pattern>#SVGID_2_ + <g>#main + <g>#template	3
<polygon>#main + <g>#template	2

545 of the 573 Control Measure fragments, and they are the only main icons in either tree that are not
single-root.

That is consistent with what these files are: GIS construction samples showing how a control measure is
drawn on a map, rather than icons composed into a symbol. APP-6E 8.1.3 exempts them from the composition
rules, and #59 gave them FREE_CANVAS for the same reason.

Why it matters now

Injecting fragments into the generated classes needs a rule for "which element is the content". Everywhere
else that rule is trivial. Here it needs either an exception, or these files tidied so the content stands
on its own.

Options

1. Split the scaffolding out - keep #main in the fragment and move #template/#example to a
   separate file or directory, so the fragment holds only what is drawn.
2. Mark it rather than move it - keep the files as they are and rely on #main being the content,
   making that part of the documented rule. Cheaper, but leaves the fragments carrying material no
   consumer ever renders.
3. Leave them out of injection entirely, since they are exempt from composition anyway.

Worth deciding deliberately - these are 545 files and whichever way it goes should be a single mechanical
pass, verified with SvgFingerprint the way #71 was.

Note the 3 files with a <pattern> also hold the only live internal id reference in the tree
(url(#SVGID_2_)), so they need care whichever option is taken.

Preparation for the 2.0.0 fragment injection work.

[ctgnz]

The rest of #78 is done in #96. These are the 26 that still have no template group, for the call on whether to fix now, leave, or split out.

By entity group

entity group	with template	without
Protection Areas	25	9
Space Debris	0	6
Sustainment Areas	5	3
Fires Areas	65	2
Maneuver Areas	31	2
Maritime Control Areas	8	2
Airspace Control (Corridors) Areas	19	1
Maneuver Lines	22	1

Nineteen other entity groups have a template on every fragment.

Space Debris is the only group where none of them has one - 0 of 6. That reads as a coherent class rather than an omission: those six are the simplest fragments in the tree, #main alone with no example either, and a point of debris has no anchor points to construct it from.

The other 20 are lone exceptions inside groups where between 5 and 65 siblings all carry a template, which looks more like something missed than something decided.

The files

fragment	shape	#main draws	label
25141300	g#main + g#example	1 path, 1 text	Maneuver Lines / Airhead Line
25150100	g#main	1 path	Maneuver Areas / Area
25151900	g#main + g#example	1 path	Maneuver Areas / Penetration Box
25170800	g#main	1 circle, 1 text	Airspace Control (Corridors) Areas / Base Defense Zone
25200401	g#main + g#example2	1 ellipse, 1 text	Maritime Control Areas / Ship Area of Interest / Ellipse/Circle
25200402	g#main + g#example2	1 rect, 1 text	Maritime Control Areas / Ship Area of Interest / Rectangle
25240808	g#main	1 path, 1 text	Fires Areas / Area Targets / Bomb Area
25242000	g#main	1 path, 1 text	Fires Areas / Terminally Guided Munition Footprint (TGMF)
25271000	g#main	1 path, 2 polygon, 2 text	Protection Areas / Unexploded Explosive Ordnance (UXO) Area
25271700	g#main	6 path, 1 polygon, 1 text	Protection Areas / Biological Contaminated Area
25271701	g#main	6 path, 1 polygon, 2 text	Protection Areas / Biological Contaminated Area / Toxic Industrial Material
25271800	g#main	6 path, 1 polygon, 1 text	Protection Areas / Chemical Contaminated Area
25271801	g#main	6 path, 1 polygon, 2 text	Protection Areas / Chemical Contaminated Area / Toxic Industrial Material
25271900	g#main	6 path, 1 polygon, 1 text	Protection Areas / Nuclear Contaminated Area
25272000	g#main	6 path, 1 polygon, 1 text	Protection Areas / Radiological Contaminated Area
25272001	g#main	6 path, 1 polygon, 2 text	Protection Areas / Radiological Contaminated Area / Toxic Industrial Material
25272101	g#main	2 path, 2 rect, 2 text	Protection Areas / Minimum Safe Distance Zone / Multiple Strike
25310500	g#main	1 path, 2 text	Sustainment Areas / Regimental Support Area
25310600	g#main	1 path, 2 text	Sustainment Areas / Brigade Support Area
25310700	g#main	1 path, 2 text	Sustainment Areas / Division Support Area
25350101	g#main	2 circle, 4 path	Space Debris / Human Made Space Debris / Human Made Space Debris, Small
25350102	g#main	2 circle	Space Debris / Human Made Space Debris / Human Made Space Debris, Medium
25350103	g#main	2 circle, 4 path	Space Debris / Human Made Space Debris / Human Made Space Debris, Big
25350201	g#main	2 circle, 4 path	Space Debris / Natural Space Debris / Natural Space Debris, Small
25350202	g#main	2 circle	Space Debris / Natural Space Debris / Natural Space Debris, Medium
25350203	g#main	2 circle, 4 path	Space Debris / Natural Space Debris / Natural Space Debris, Big

Four of them do carry an example group (25141300, 25151900, 25200401, 25200402), so those are missing only the template. The remaining 22 are #main on its own.

Worth noting for whichever way this goes: writing a template means drawing new content - T/AS labels and PT anchor arrows placed to suit each measure - rather than moving existing content around. That puts it in a different category from the rest of #78, all of which was verifiably non-visual.

[ctgnz]

Confirmed. Against the rule as agreed — exactly one <g>#main, zero or one <g>#template, zero or more <g>#example*, plus any <defs> that main references:

tree	free canvas fragments	failing
jmsfx-standard	577	0
jmsfx-hallux	577	0

The check covers all four conditions, not just the group ids: exactly one main, at most one template, every other root group named example or example+digits, and nothing at the root that is not a group or a <defs>.

The shapes actually present:

count	shape
338	main + template + example
177	main + template
22	main
14	main + example + template
6	defs + main + template + example
5	main + template + example + example2
5	main + template + example1 + example2
2	main + example
2	main + example2
2	example + main + template
1 each	four more orderings, including the two with example1–example4

Order varies — 16 fragments put an example group before main or template. That changes paint order, so it stayed out of a pass that was meant to be provably non-visual. Worth a look sometime, but it is not a conformance failure under this rule.

The rule is now written down in svg/README.md, whose Details section had been a TBD placeholder for "the internal composition of the SVG files" since the fork. It records why the template is optional — areas have at least three user-placed control points so there is no fixed geometry to draw, and the Space Debris fragments are whole symbols in the way an ordinary icon is — so the decision survives the issue being closed.

If it would be useful, this check is small enough to bind to the build the way FragmentNormaliser --check is, so a fragment that drifts off the shape fails rather than being noticed later. Happy to add it here or as its own issue.