<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by William MacMillan on Medium]]></title>
        <description><![CDATA[Stories by William MacMillan on Medium]]></description>
        <link>https://medium.com/@wmacmill?source=rss-7535a2d8d372------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*NAKRXhk-Qbzd6nD_.</url>
            <title>Stories by William MacMillan on Medium</title>
            <link>https://medium.com/@wmacmill?source=rss-7535a2d8d372------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 07:34:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@wmacmill/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Nice]]></title>
            <link>https://medium.com/@wmacmill/nice-82b5309bab53?source=rss-7535a2d8d372------2</link>
            <guid isPermaLink="false">https://medium.com/p/82b5309bab53</guid>
            <category><![CDATA[mathematics]]></category>
            <category><![CDATA[cine]]></category>
            <category><![CDATA[sexuality]]></category>
            <category><![CDATA[followers]]></category>
            <category><![CDATA[numerology]]></category>
            <dc:creator><![CDATA[William MacMillan]]></dc:creator>
            <pubDate>Sun, 11 Sep 2022 18:04:35 GMT</pubDate>
            <atom:updated>2022-09-11T18:04:35.831Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/170/1*F4H5ZpH7ZL0NOI-4KoVSVA.png" /><figcaption>Source: Author</figcaption></figure><p>Nice.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=82b5309bab53" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Skeuomorphisms for Class]]></title>
            <link>https://medium.com/design-bootcamp/skeuomorphisms-for-class-4a34eb30525c?source=rss-7535a2d8d372------2</link>
            <guid isPermaLink="false">https://medium.com/p/4a34eb30525c</guid>
            <category><![CDATA[absurd]]></category>
            <category><![CDATA[automotive]]></category>
            <category><![CDATA[graphical-user-interface]]></category>
            <category><![CDATA[skeuomorphism]]></category>
            <category><![CDATA[design]]></category>
            <dc:creator><![CDATA[William MacMillan]]></dc:creator>
            <pubDate>Mon, 23 May 2022 00:03:54 GMT</pubDate>
            <atom:updated>2023-10-23T20:06:20.678Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>I keep going back and forth whether I’m inventing a landau bar or doing something consequential.</blockquote><blockquote>— Me, referencing a design decision at work.</blockquote><p>A little background knowledge is necessary to know where this essay is going. The four-in-hand knot is the name of the basic knot used for tying ties. Formally, but somewhat rarely now, men’s neck ties are known as four-in-hand ties along with the knot. A skeuomorph is a “derivative object that retains ornamental design cues (attributes) from structures that were necessary in the original” to quote <a href="https://en.wikipedia.org/wiki/Skeuomorph">Wikipedia</a>, referencing Ballas (<a href="https://archive.org/details/evolutionoftechn00basa/page/107">1988</a>). One set of skeuomorphic designs often referenced, for example, were heavily used on the first iPhone.</p><figure><img alt="A screenshot of a now dated view of the application icons from an iPhone." src="https://cdn-images-1.medium.com/max/612/1*OR_Y-K8GdHiuRjYwJvR2Mw.jpeg" /><figcaption>A literal wooden console with a test pattern for the “television” application is a great example of skeuomorphism at work.</figcaption></figure><p>Apocryphally, the knot, tie, and name come from the one used by coachmen to bind together the reins from two or more horses. That is to say, etymologically, it was the knot used to tie “four [reins] in hand.” It appears most likely it’s actually from the Four-In-Hand Club in London, a social club for members of the elite who enjoyed recreational driving (of horse-drawn carriages) which predates the naming of the tie knot four-in-hand; its members tied their ties in this manner.</p><p>More often than not (no pun intended), the carriage of choice for members of this club was a lightweight carriage that typically seats four, designed such that the riders could easily see around them and be seen in their fancy clothes — the <a href="https://en.wikipedia.org/wiki/Landau_(carriage)">landau carriage</a>.</p><figure><img alt="A pciture of an antique landau carraige with the shells closed." src="https://cdn-images-1.medium.com/max/1024/1*YhXPq89gHg4sqzK6ooOjEg.jpeg" /><figcaption>This is a landau carriage. A very fancy one.</figcaption></figure><p>The landau can be enclosed, for weather or if the riders prefer, with collapsible fabric shells that meet in the middle and are supported by metal framing. This framing is most visible on the outside of the carriage from a pair of curved rods called landau rods that are linked together by a landau joint. Fully extended, they form a smooth “S” shape, with a round joint mid-way down the length. They look like an integral sign, basically — additional useless side note, integral signs look like an S because they are for “summation.” If a landau is cut down from four seats to two (i.e. a <em>coupe</em>) it is referred to as landaulet.</p><p>In the case of a conversion, the rear half folding section was most commonly permanently affixed to be up rendering the outside bars largely decorative. The sight of the bars on a landaulet, in any case, came to be associated with the upper class and refinement; you had to be rich enough to either have built custom or converted a landau into a landaulet. The bodywork of a car is referred to as a coach and the specialists who build or modify only the exterior and interior bodywork are referred to as coachbuilders. This is no accident. Historically, as horseless carriages rose in popularity, coachbuilders switched from making (regular) carriages to horseless ones. <a href="https://en.wikipedia.org/wiki/Coachbuilder#Horse-drawn_origins">Some are still in operation to this day</a>. Originally, automobile manufacturers only made the <em>mobile</em> parts of the carriage. Absent horses, the mobile parts of the carriage are capable of autonomous propulsion when combined with an internal combustion engine. Hence <em>automobile</em>.</p><p>Today’s auto manufacturers are direct descendants of coachbuilders. In an attempt to give their cars an air of refinement, manufacturers decided to wrap the rear quarter panel in leather (or vinyl embossed to look like leather) and then apply an appliqué to give the appearance of the original, now twice removed, functional framing of the carriages they used to make.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4a34eb30525c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/skeuomorphisms-for-class-4a34eb30525c">Skeuomorphisms for Class</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Prime number reasons you should short TSLA]]></title>
            <link>https://medium.com/@wmacmill/prime-number-reasons-you-should-short-tsla-24deb4941c9f?source=rss-7535a2d8d372------2</link>
            <guid isPermaLink="false">https://medium.com/p/24deb4941c9f</guid>
            <category><![CDATA[tesla]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[automotive]]></category>
            <category><![CDATA[political-economy]]></category>
            <dc:creator><![CDATA[William MacMillan]]></dc:creator>
            <pubDate>Sat, 21 May 2022 16:01:53 GMT</pubDate>
            <atom:updated>2022-05-21T18:41:00.386Z</atom:updated>
            <content:encoded><![CDATA[<h3>Prime number reasons TSLA is a short</h3><p>It’s because you have to build your own compiler.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*trlNb4pMrRmWawLUkmDxbA.jpeg" /><figcaption>The 3 was supposed to be the “E”. smdh.</figcaption></figure><p>So, when you build cars, you buy parts from suppliers. These parts are rated for environmental conditions (and are insurable based on those ratings). This includes processors and integrated circuits. The rating process takes a pretty long time and is not free. As a result, the computing hardware on cars is typically quite out of date. As the rigor of the environmental demands rises, this certification process (and the resulting product development when things don’t meet the specs) is increasingly costly, complex, and time consuming.</p><p>Now, if you want to run computing on a car that’s somehow tied to what the car is actually doing, it needs to be compiled and using a pretty low-level language so it’s fast enough. This means using C, Rust, Ada, Assembly, and that’s about it. Remember, however, that the chips and instruction sets they are using are, at best, many years antiquated in addition to potentially being entirely non-commodity hardware. This means you’ll have to build a compiler targeted for the hardware, no matter what. Who’s going to send you a binary? Ford? Good luck. Tesla has had to, of course, build their own compiler for their own hardware. Sophisticated, no?</p><p>Building a compiler that targets specific hardware, even for fairly custom and optimized and low-level implementations of things is less difficult than you might think. Essentially, you need a syntax, a compiler-compiler, and an instruction set for the target hardware. The manufacturer will give you the instruction set (i.e. the specs for assembly on that processor). There are a number of compiler-compilers available that are very good and open source. After that, you just need a formal grammar to feed the lexical parses. Here is <a href="https://docs.python.org/3/reference/grammar.html">Python’s grammar specification</a>. Here is the <a href="https://www.lysator.liu.se/c/ANSI-C-grammar-y.html">grammar specification for C</a>. And for <a href="https://github.com/rust-lang/wg-grammar">Rust</a>.</p><p>You may also now be perceiving that every auto manufacturer has been going through this exact same process to the exact same ends, now for <em>decades</em>. From the early ’90s, at a minimum.</p><p>In 2021, Tesla shipped 936,222 cars. A pretty remarkable figure, considering that they’re basically the only new auto maker that’s had any lasting power in the last century, give or take. Roughly speaking, Toyota may not have done a whole lot better as they shipped about 1.9 million cars in 2021. In the first quarter of 2021, that is. And, of course, this was down year for Toyota (and not for Tesla). Pre-pandemic, their most recent peak had been over 9 million per annum.</p><p>At the end of the day, Tesla is far less innovative than the popular consciousness gives them credit for — and they’re likely not giving them credit the areas where they <em>h</em>ave<em> </em>innovated. In terms of <em>making cars</em>, Tesla is only now even remotely approaching the efficiency and product quality of established auto makers. Their capital efficiency is <em>terrible </em>(this is worthy of it’s own discussion). The unfortunate reality is that, as an auto maker they still have to…make cars. The firms that have been making cars for 120 years are, somewhat unsurprisingly, better at making cars.</p><p>Their real innovation, in my book, was the vertical integration of making batteries into the auto manufacturing process. This allowed them to much more rapidly iterate on design and the manufacturing process for the car itself than would have even been remotely possible with a typical supplier arrangement. A comparable achievement from the past would be the vertical integration of the manufacturing process by Ford at the River Rouge plant. Henry Ford, along with being ragingly antisemitic, was acknowledged as being innovative.</p><p>N.b. no. 1: ruggedizing (the formal industry term for it) things for outer space is the most challenging. Any space vehicle being launched right now, today, is, at best, using hardware from the mid-2000s. Likely the 90’s, or even earlier. This is part of why the James Webb telescope was insanely expensive and delayed.</p><p>N.b. no. 2: Tesla, in their infinite wisdom, decided they knew better and tried using parts that were not as ruggedized against environmental conditions as other auto manufacturers typically use. And, as a result, they got hit with a ton of warranty claims for things like cracked screens on the big dash-mounted tablet in the 3. To their credit (and I’m being sincere here, honestly), they realized their mistake and have been changing the parts and the manufacturing process. This is why I, at the end of the day, rate Tesla as a buy.)</p><p>Another interesting side note: did you know that Tesla is basically running Linux and Ethernet inside the car? The different microcontrollers in the car all have Ipv4 private network addresses. It’s a somewhat customized version of Linux, but it’s not radically altered. SpaceX is also running Linux, more or less, on their spacecraft, but theirs is waaaaaaaaaaay more custom as they’ve had to make it completely hard real-time.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=24deb4941c9f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Procedural Anomalies in SAS]]></title>
            <link>https://medium.com/@wmacmill/procedural-anomalies-in-sas-cde25d1d1be2?source=rss-7535a2d8d372------2</link>
            <guid isPermaLink="false">https://medium.com/p/cde25d1d1be2</guid>
            <category><![CDATA[graduate-school]]></category>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[sass]]></category>
            <dc:creator><![CDATA[William MacMillan]]></dc:creator>
            <pubDate>Thu, 04 Jun 2020 17:42:10 GMT</pubDate>
            <atom:updated>2020-06-04T17:53:52.999Z</atom:updated>
            <content:encoded><![CDATA[<p>SAS is a data analysis software (and language). It’s much more common in health care and actuarial work than most other spaces in data analytics. With the rise of data science and the accompanying popularity of R and Python, SAS has really fallen out of vogue and the popular consciousness. Compared to these other pieces of software SAS is quite monolithic and enterprise-friendly. When you first encounter it, you’re much more likely to learn to use the graphical user interface (GUI) versus learning the SAS language itself — which is actually quite verbose and somewhat hard to use.</p><p>Like R or Python, SAS has libraries, but are referred to in their nomenclature as “procedures.” These are usually abbreviated to “proc” when written, and pronounced as such, but with a hard “c”. When using the GUI, you’ll refer to the work you might do as using the procedure. Even when, functionally, you’re just opening a window in the interface and looking at table summaries of data.</p><p>A frequently used “proc” is Frequencies. Similar to the “procedures” label, it’s commonly abbreviated to “freq.” If you’re doing table summaries of data, dealing with data broken into categories, and some forms of linear modeling, you’re likely to be using the Frequencies procedure. It’s a very workhorse procedure to use.</p><p>When I started graduate school, I hadn’t ever used that software. In undergraduate, I was likely best described as a metalhead that took philosophy and politics courses. The extent of my data analysis software experience was with SPSS (a whole other piece of software worth a post). I knew nothing about SAS.</p><p>The bit of math we did learn my first semester came mostly in pen and paper form was not particularly intense. There was a little bit of exposure to some data analysis tools, but that didn’t include SAS. I only started to use that software when I took a graduate course from the Statistics Dept. at the University of Iowa. Being a graduate course in statistics, everyone (else) in the room had significant exposure to SAS at that point. Nearly all had been statistics undergraduates, and many had also worked in industries that used the software as well.</p><p>(Side note: if you want a sense of my level of hubris and/or ignorance at the moment, consider that the first time I <em>ever,</em> in my <em>entire life</em>, took a college course from a statistics department, it was a post-masters, Ph.D. class. This was my second semester of graduate school. It was about only two months after I had, more or less, taught myself the basics of calculus and linear algebra.</p><p>I still have never taken a full, formal calculus course, at any level, high school or otherwise. And I took that stats course <em>for a grade</em>, not pass/fail. If memory serves, I got a B+. How am I still alive?)</p><p>So, the first day of class comes. The professor is running through typical first day things from the syllabus and dully talks about how they’ll be using SAS (which, again, everyone but me knows intimately). The sounds coming into my brain through my ears were roughly like this:</p><blockquote>bzz bzz bzz…mostly we’ll use a lot of prock freak…</blockquote><p>and, just completely involuntarily busted out a huge laugh followed by the stares of the rest of the students.</p><p>I did get a good internet handle out of it, at least, and went by “the_proc_freak” for a good year and a half online.</p><h4>Addendum:</h4><p>I have a complicated personal history with SAS. I’ve had two jobs where using SAS nearly full time was a requirement of the position. I’d say the word “fired” is maybe a bold term to use here. But I did end up walking out of the building both times, not on a Friday, and with a box of the stuff from my desk under my arm.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cde25d1d1be2" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>