Rethinking Compiler Design

Speaking as someone who (in some small way) helps with compiler dev, while a lot these ideas sound pretty cool, I can pretty comfortably say that these ideas would make compiler engineering even harder! (though I'd love to be proven wrong by some brave soul)

Articles like this always make me wanna weep. Not because of their contents but because nobody bothered to put a date on (or more precisely, before) those thoughts. As it is, this article is only for archaeologists and historians who like to guess from when it might be. And the “last edited” date is no replacement or excuse.

<p><pre><code> last edited September 9, 2008</code></pre>

<p><pre><code> - Compiler internals cleanly exposed </code></pre> This exists in clang!<p>I&#x27;m not sure I get this one:<p><pre><code> - Constancy as an attribute </code></pre> Does it mean read-only variables such as &quot;const&quot; variables in C&#x2F;C++?