I completely disagree.
Intermingling PHP and HTML is one of PHP strengths. The processing/executing difference you’re describing is almost always negligible due to how PHP is optimized (specifically for this kind of thing - outputting HTML to the browser sometimes).
Seriously, compare this to the 4chan image board view we really aren’t looking at the same source code. In comparison, the WordPress function is blocked, purposeful, together. It’s a single CSS block output all at once. On the otherhand, in the linked 4chan discussion board PHP file, it echos as strings, broken up by multiple conditionals, and is difficult (even from an IDE perspective on highlighting) to tell where a block starts and where it ends (again due to it being echo’d in strings, and broken by conditionals). Trying to modify this blocked CSS is going to be wayyyyyyyyyyyyyyyyyyyy easier than trying to modify a bunch of printed HTML strings broken up by multiple nested conditionals. Plus it’s just straight-up easier to read and straight-forward to understand what the function does right away.
To harp on this even more, one of the benefits of blocking HTML in this way is IDE highlighting. In your example, if you were to pop that into a modern IDE like VSCode, it’ll highlight tags and allow collapsing like a normal HTML doc. It’ll probably even highlight the CSS as expected. On the other-hand, by echoing / printing HTML strings, IDEs aren’t going to highlight these things as HTML since they’re PHP strings, and in the case of the imageboard, it’s going to struggle finding matching open/end tags due to PHP strings and broken conditionals. I’d much prefer the WordPress example over echo / printing multiple lines of HTML strings (this is really a pet-peeve of mine).
I can’t think of a single system that doesn’t “stop PHP executing” at some point to output HTML in some way. Maybe an app that dynamically pulls it’s views in through JS I guess.
For comparison to future readers, this is just a small portion of the imageboard which goes on like this for another 600 lines:
if( $resno ) {
$closed = $log[$resno]['closed'] || $log[$resno]['archived'];
if( !$stripm ) {
$msg .= '<div class="navLinks mobile">
<span class="mobileib button"><a href="/' . BOARD_DIR . '/" accesskey="a">' . S_RETURN . '</a></span> <span class="mobileib button"><a href="/' . BOARD_DIR . '/catalog">' . S_CATALOG . '</a></span> <span class="mobileib button"><a href="#bottom">' . S_BOTTOM . '</a></span> <span class="mobileib button"><a href="#top_r" id="refresh_top">' . S_REFRESH . '</a></span>
</div>';
}
if( !$stripm ) $msg .= '<div id="mpostform"><a href="#" class="mobilePostFormToggle mobile hidden button">' . S_FORM_REPLY . '</a></div>';
} else {
if( !$stripm ) $msg .= '
<div class="navLinks mobile">
<span class="mobileib button"><a href="#bottom">' . S_BOTTOM . '</a></span> <span class="mobileib button"><a href="/' . BOARD_DIR . '/catalog">' . S_CATALOG . '</a></span> <span class="mobileib button"><a href="#top_r" id="refresh_top">' . S_REFRESH . '</a></span>
</div>
<div id="mpostform"><a href="#" class="mobilePostFormToggle mobile hidden button">' . S_FORM_THREAD . '</a></div>';
}
Formatting and structure is important for a readable framework / project and longevity. 4chan had none of this - Moot took the bag and ran and whoever took it over just left left the PHP standards/organization in 2003.
What you’re talking about is semantics. At a base level, whether you use a templating engine, include / require, or just straight up mix HTML / PHP - PHP “stops execution” to output to the browser. The few exceptions to this that I can think of is if it’s instead handing off that responsibility to JS or some other frontend processor.
Templating engines are cool. They make it easier to separate your views from logic. It makes interloping more straight-forward and possibly more maintainable (though, not always - Engines don’t save from bad practices), but I do not agree that it’s defacto. I think the strength of PHP is it’s ease to just jump into it and get something working, right “out of the box”. The ease of mixing PHP and HTML is a boon from an entry level aspect. Low entry level leads to wider adoption, leads to more discussions, more volunteers for FOSS, more bug reports, more more more.
I could create a vanilla PHP application that organizes views just as well without a templating engine which could be understood by someone with baseline PHP knowledge - that’s good thing. It’s inherit to PHP and I won’t need to worry about keeping any templating library updated or ported to a new engine. In my made-up vanilla app, I wouldn’t do what 4chan did in their views, but I may do what WordPress does in your example because, used sparingly, in an organized application, it’s not that big of a deal. For the most part though, I do like to keep my HTML views and my PHP logic separate in an MVC kind of way either through templating or just straight up includes.