<p><span class="h-card" translate="no"><a href="https://functional.cafe/@PaniczGodek" class="u-url mention">@<span>PaniczGodek</span></a></span> I have indeed seen &quot;naive interpreters&quot; that just operate directly on the s-expression level, effectively re-parsing it over and over. I view this as a bad implementation that missed (or intentionally decided to pass on, for some other reason) the benefits of parsing. But that&#39;s then just an implementation without a parser. (Or it has a &quot;weak&quot; parser — one that checks for errors but doesn&#39;t produce an AST like a &quot;strong&quot; parser.)</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@film_girl" class="u-url mention">@<span>film_girl</span></a></span> this is the most Twitter of all posts.</p>

<p><span class="h-card" translate="no"><a href="https://types.pl/@krismicinski" class="u-url mention">@<span>krismicinski</span></a></span> <span class="h-card" translate="no"><a href="https://mathstodon.xyz/@jonmsterling" class="u-url mention">@<span>jonmsterling</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@basus" class="u-url mention">@<span>basus</span></a></span><br />The major limitation for making any of this faster than the lowest numbers there is that (a) the system is mostly written in Scheme/Racket (unlike Python or JS) so there&#39;s a bunch of code to load that&#39;s not in the binary and (b) that code goes in the heap, not in the code segment of an ELF file, so loading it isn&#39;t free as it is with mmap() for C code. </p><p>I can understand that 200 ms is a lot for eg an interactive shell command, but it&#39;s not that big otherwise.</p>

<p><span class="h-card" translate="no"><a href="https://mathstodon.xyz/@zhxsdm" class="u-url mention">@<span>zhxsdm</span></a></span> <span class="h-card" translate="no"><a href="https://mathstodon.xyz/@j2kun" class="u-url mention">@<span>j2kun</span></a></span> <br />That is neat!</p><p>I&#39;m confused: there&#39;s a project there called Enso, but it doesn&#39;t seem to be William Cook&#39;s Enso, but also seems very similar to it? Wonder what&#39;s going on with that.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@film_girl" class="u-url mention">@<span>film_girl</span></a></span> you did something considered rude. There was a reaction. Deal with it.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> &#39;m a 4th-year Business Administration student at Bilkent University, here for my MAN432 Consumer Behavior project. Our goal is to understand why users choose Mastodon and learn about your experiences.</p><p>Your privacy is important—insights will only be used for academic purposes. I’d love to hear:</p><p>How does it compare to other platforms in your experience?<br />What brought you to Mastodon?</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@gamingonlinux" class="u-url mention">@<span>gamingonlinux</span></a></span> awh cmon why not x11</p>

<p><span class="h-card" translate="no"><a href="https://mathstodon.xyz/@olynch" class="u-url mention">@<span>olynch</span></a></span> I have no clue what a &quot;homoiconic&quot; language is, so I have no idea whether Julia is or is not one. (That was the point of the article.)</p>

<p><span class="h-card" translate="no"><a href="https://mathstodon.xyz/@mudri" class="u-url mention">@<span>mudri</span></a></span> No, those come downstream of what I&#39;m calling the parser and/or are incorporated into it. In crisper terms, those are all very context-sensitive, and the complexity hierarchies in the article should make it easy to see why this is different.</p>