Shoud I prefer <nav><ul><li><a-or-other-tag> or <nav><a-or-other-tag>, from an accessibility standpoint? [duplicate]

Question

Often, in books, tutorials, and some actual webpage, I see navigation bars marked up as <li>s elements in a <ul> element in a <nav> element, like this:

.site-nav {
  display: flex;
  justify-content: space-around;
  list-style-type: none;
  padding-left: 0;
}

.site-nav > li > a {
  border: 1px solid black;
  padding: 1em;
  text-decoration: none;
}

li { padding: 1em; } /* just for better appearance here on SO */

<nav>
  <ul class="site-nav">
    <li><a href="x">first</a></li>
    <li><a href="y">second</a></li>
    <li><a href="z">third</a></li>
  </ul>
</nav>

However, why shouldn't I prefer a solution like the following, where the <nav> contains directly all the contents of the <li>s from the example above?

.site-nav {
  display: flex;
  justify-content: space-around;
  list-style-type: none;
}

.site-nav > a {
  border: 1px solid black;
  padding: 1em;
  text-decoration: none;
}

<nav class="site-nav">
  <a href="x">first</a>
  <a href="y">second</a>
  <a href="z">third</a>
</nav>

One advantage I see is the lower specificity needed to target the <a> links (or whatever I'd put in their place, e.g. <button>s), but I still know too little about HTML to understand the implications of one or the other solution, especially as far as accessibility is concerned.

Answer 1

The answer is straight forward, semantics and accessibility.

If I arrive at your second example without the <ul> with a screen reader then I have no idea how many links there are.

If I arrive at the first, correctly marked up example with a screen reader I will hear something similar to "navigation landmark, list of 3 items".

This lets me know that there are 3 links in the list so I can visualise where I am in the navigation as I move between them.