Challenge 2: Group a Checkbox Set — Possible Solution
====================================================================

<fieldset>
  <legend>Notification preferences</legend>
  <label><input type="checkbox" name="notify" value="email"> Email me updates</label>
  <label><input type="checkbox" name="notify" value="sms"> SMS me updates</label>
  <label><input type="checkbox" name="notify" value="push"> Push notifications</label>
</fieldset>

WHY THIS WORKS
--------------
- The previously "nearby but unconnected" visual heading "Notification
  preferences" becomes a real <legend> element, wrapped inside a
  <fieldset> containing all three checkboxes — this is the exact
  structural fix the chapter describes: turning a visual-only
  relationship into a PROGRAMMATIC one that assistive technology can
  actually detect.
- A screen reader user navigating into this group now hears "Notification
  preferences" announced BEFORE (or alongside) each individual
  checkbox option — giving essential context ("this is a set of
  notification-related options") that simply wasn't available before,
  when the heading and checkboxes had no structural connection despite
  sitting near each other visually.
- Each checkbox STILL has its own individual <label> ("Email me
  updates," etc.) — the fieldset/legend fix operates at the GROUP
  level (announcing what the group as a whole represents), while each
  individual label continues to identify what that SPECIFIC checkbox
  does. Both levels of labeling are necessary and serve different
  purposes; fieldset/legend doesn't replace individual labels, it adds
  group-level context on top of them.
- Using checkboxes here (rather than radio buttons, as in the chapter's
  own contact-method example) is appropriate since the scenario
  describes "select all that apply" style options (a user could
  reasonably want both email AND SMS updates) rather than a single
  mutually exclusive choice — the same fieldset/legend wrapping applies
  correctly to either input type, since the grouping technique is about
  providing shared context, not about which specific input type is
  being grouped.