The most upvoted answer is obsolete today
I would recommend the exact opposite, see step by step with reasons:
good:
<a id="myLink" href="javascript:MyFunction();">link text</a>
It depends, might be good, because crawlers follows href targets and if there is any meaningful content produced by MyFunction()
(dynamic link), it is followed more likely than in the click event, which may have multiple or none listeners.
bad:
<a id="myLink" href="#" onclick="MyFunction();">link text</a>
#
means meaningless link, crawlers are often interested only in first x links, so it can prevent them to follow some meaningful links later in the page.
worse:
<a id="myLink" href="#" onclick="MyFunction();return false;">link text</a>
Same as previous plus return false
prevents following the link. If some other scripts want to add another listener and update the target (say to redirect via proxy), they can't without modifying the onclick (okay, it's just a minor setback as such use cases are rather theoretical).
worst:
Use jQuery or other similar framework to attach onclick handler by element's ID.
$('#myLink').click(function(){ MyFunction(); return false; });
jQuery is outdated in 2020+ and should not be used in new projects.
Events in href
The href attribute handler doesn't get the event object, so the handler doesn't implicitly see which link was the source. You can add it in onclick handler, which fires before the href is followed:
<a href="javascript:my_function(event2)" onclick="event2=event">
JS based link
</a>
<script>
function my_function(e) {
console.log(e.target); // the source of the click
if(something) location.href = ...; // dynamic link
}
</script>