/* ==========================================================================
   Burbank palette.

   Structure is roofreplacementnewark.com's, unchanged -- this is the fifth
   site on that layout, after Newark, Cambria Heights (which ship the same
   palette as each other), Canoga Park and Baldwin Park. Only these tokens
   differ, so the five do not share a colour fingerprint.

   GENERATED by tools/palette.py, not eyeballed. Ink hue 179, primary hue
   229, secondary hue 199. Measured against every other site on this
   layout: the closest any of the three hues comes to another site's same role
   is 33 degrees (floor 30), and the closest accent in plain RGB distance is
   70.9 units (floor 60).

   Distance is measured on HUE for the dark bands, not RGB, because RGB
   distance is useless at the extremes: every --ink on this layout is
   near-black and every --bone near-white, so RGB called everything a
   collision. Hue is what a reader actually recognises.

   Eyeballing failed twice before this script existed: Canoga Park's secondary
   landed 18.8 RGB units from Newark's and Baldwin Park's did the same thing
   again. Changing the primary is not enough -- a shared secondary is still a
   shared colour.

   The hues are also restricted to ranges that read as trade colours. Left to
   maximise distance alone the search picked aubergine ink with a purple CTA,
   which is the furthest corner of the wheel from the other four sites and
   completely wrong for a contractor asking a homeowner for $17,000.

   SECTION SEPARATION. Every site on this layout before Baldwin Park shipped
   --sand at roughly 96% of white -- 1.14:1, a tint nobody can see, which is
   why the sections read as one slab. Here --sand is 1.31:1 against --bone,
   with a second tone for alternation, and the generator darkens it until that
   is true rather than trusting a nominal lightness.

   All fourteen checks pass:
     text on bone            12.95:1  (needs 7.00)
     text on sand             9.90:1  (needs 7.00)
     text on sand-2           8.26:1  (needs 7.00)
     muted on bone            7.29:1  (needs 4.50)
     muted on sand            5.58:1  (needs 4.50)
     muted on sand-2          4.65:1  (needs 4.50)
     white on brick          11.46:1  (needs 4.50)
     white on brick-d        14.64:1  (needs 4.50)
     white on ink            15.09:1  (needs 4.50)
     white on ink-3           7.36:1  (needs 4.50)
     ember on ink (large)     9.56:1  (needs 3.00)
     brick-l on ink           4.75:1  (needs 4.50)
     sand vs bone             1.31:1  (needs 1.25)
     sand-2 vs sand           1.20:1  (needs 1.15)
   ========================================================================== */
:root {
  --ink:      #0d2b2a;
  --ink-2:    #193e3d;
  --ink-3:    #2c5e5d;
  --brick:    #182e91;
  --brick-d:  #0d1f6d;
  --brick-l:  #7a8cdc;
  --amber:    #9fd6ef;
  --amber-d:  #3ea9da;
  /* The hero H1 emphasis. Same value as --amber on purpose: a separate
     third accent on this layout reads as clutter, and the two earlier sites
     that tried one both ended up with a near-collision to fix. */
  --ember:    #9fd6ef;
  --bone:     #f9fbfb;
  --sand:     #cbe2e1;
  --sand-2:   #b3d0d0;
  --line:     #c1d7d7;
  --line-2:   #afc5c5;
  --text:     #1a3332;
  --muted:    #3a5959;
}
