Regrettable phrasing

Marginalia III

Fall has been rather sudden and unfamiliar, but not unwelcome. Haven't been up to too much other than computer things:

  1. Developing some interesting analyses at work (modelling charitable giving). I'm definitely at a point where my analytical problem-solving feels R-shaped, but I've also been getting over my gripes with Python. Different jobs, different tools, and teaching/collaborating is fun.
  2. Speaking of Python, agate is phenomenal. It really does cut out a lot of the unfriendliness it comes to basic analysis in Python. Convinced a few of my colleagues who were really pushing the limits of Excel to sidle on over, and it's been smashing.
  3. Oh, and I'm writing an iOS app (it's a content blocker).

Skilful Grammar

I don't want to get into "why content blocking" yet, but I've wanted to play around with iOS and Swift for a while and this seemed like a great place to start. For the moment, I'm calling it Skilful Grammar: it was a dumb name a program gave me once, but I couldn't think of anything more right: simple keywords → useful rules.

It won't be "press once to block ads"; it won't cost anything; it might come out soon. To be frank I'm caught up on tidbits of look and feel. Building the presentation is easy; it's all the interaction fiddliness that has me hung up. That said, the app won't be a selection of switches—this will be for "professionals", if you're so inclined.

Should you have unsolicited ramblings, second opinions, or, hell, mild interest, get in touch. I honestly don't know if the way I've gone about it stands a chance in App Review—but hey, that's all part of the fun.

On Writing

"Always remember that writing is an essential expression of purpose..." the sort of horseshit you say to yourself when you're reaching for excuses after midnight because your dear friend, your gracious friend, has been silently and patiently waiting for edits for weeks now and what exactly have you been doing with all this time anyway?



Been reading a bit lately. From The Real World of Technology (Revised Edition) by Ursula M. Franklin. From pages 31–32:

The scientific method works best in circumstances in which the system studied can be truly isolated from its general context. This is why its first triumphs came in the study of astronomy.

On the other hand, the application of the general to the specific has been much less successful in situations where generalisation was achieved only by omitting essential considerations of context. These questions of reductionism, of loss of context, and of cultural biases are cited quite frequently by critics of the scientific method. We hear much less about the human and social effects of the separation of knowledge from experience that is inherent in any scientific approach. These effects are quite wide-spread and I think they can be serious and debilitating from a human point of view.

Today scientific constructs have become the model of describing reality rather than one of the ways of describing life around us. As a consequence there has been a very marked decrease in the reliance of people on their own experience and their own senses. The human sense of sight and sound, of smell and sound and taste and touch, are superb instruments. All senses, including the so aptly named "common sense," are perfectible and it's a great pity that we have so little trust in them. For instance, people know at what point an ongoing noise will give them a headache, but all too often they feel the need for an expert with a device that measures the noise in decibels. The expert then has to compare the noise level measured with a chart that indicates the effect of noise levels on the nervous system. Only when that chart and the expert say, "Yes, indeed, the noise level is above the scientifically established tolerance range," do people believe that it was indeed the noise and not a figment of their imagination that gave them persistent headaches. I'm not talking here about an either-or situation in which either personal experience or an established measuring procedure is paramount; what I am talking about is the downgrading and the discounting of personal experience by ordinary people who are perfectly well equipped to interpret what their senses tell them. I dwell on this because the downgrading of experience and the glorification of expertise is a very significant feature of the world of technology. Sometimes it is important to stress that because the scientific method separates knowledges from experience it may be necessary in case of discrepancies to question the scientific results or the expert opinion rather than to question and discount the experience. It should be the experience that leads to a modification of knowledge, rather than abstract knowledge forcing people to perceive their experience as being unreal or wrong.

Fitness Data: Kicking & Screaming

Been meaning to publish a bigger piece but got caught in the vortex of not being 100% about the result; forcing myself to blog about the topic at hand, if nothing else.

Stephanie Lee of Buzzfeed wrote a delightfully succinct summary of the state of fitness data: that there is a wealth of collection tools, but their expectations and implications are murky at best. I've been using my Fitbit for more than a year—Strava, Nike+, and Garmin Connect much longer—but never really given data extraction much more thought than their baked-in dashboards. Now, I'm interested.

And has that ever been a can of worms.

As soon as I started looking around the Fitbit API docs, I knew things were going downhill: casual mentions of Premium membership, daily totals, etc. Given that I've been somewhat-slavishly examinig my step counts sunce last year, I knew there was better data available—but how to get it? Luckily, some kind soul (ok, corynissen) put together a scraper for the Fitbit dash. Even if the data isn't reliable or meaningful, it's always a comfort to see other folks doing the same mad things as I am.

That said, it's becoming absurd how complicated access to one's own fitness data is becoming. This doesn't necessarily strike me as being the same as proper medical and health records, but that seems like an awfully good model to aspire to—or even to exceed. Individual data storage and delivery is so insanely straightforward that options ought to exist at marginal cost. But, I digress.

Having resolved the conundrum of extraction, I'm merely tasked with interpretation and visualization. I know the perspective afforded by short-term trends has always been helpful, and I had hoped a bit of Feltron-esque analysis would prompt some more changes on my part. But honestly, I'm not even sure I'll find something worth sharing—this was my not-very-cunning way of forcing a choice between charting/visualisation libraries.

Wait, I did what now?

Marginalia II

  1. Feels pretty doofy to be hopping onto this particular bandwagon, but I was pleasantly surprised by the Apple Music launch today. I'm not one much for playlists—full albums, always—but every note landed for me. That, or I've caught early-onset Cool Dad Syndrome.
  2. Big news for the R community: the adults are getting involved. Of course I'm wary of Microsoft, Google, et al. throwing their weight around in the code I use daily, it's awfully nice to see some investment and interest. Rough edges abound in R and anything to stop packages breaking and dumb design decisions can only be good. (Right?)

If this, then...

IFTTT made a new channel for makers and tinkerers on the web; hook up a web endpoint and you're ready to go. Incredible. Can't count the times I've found myself looking for this sort of internet-glue.

Second Thoughts III

Finally pushed up the new version of the site. Trying to nail down whatever's going on with RSS, but otherwise pleased to get this out into the world.

Second Thoughts II

Well the new, plain-jane HTML version of the site is ready to go but I’ve had to hold off pushing the button. (Terrible summer cold knocked me out nearly a week, then traveling this weekend.) Do need to iron out RSS and some reliability/security whatsits, but I would like to announce the following: beta builds of whatever I’m working on for the site will be (sometimes) available at

At the time of writing, the new Express/Handlebars app is running intermittently. In the process of building:

  • The switch from Meteor wasn’t as painful as I had feared. I’m kind of missing some of the helpers and whiz-bang features, but no harm done.
  • Started using Vagrant for testing. Though this site is incredibly simple, it feels downright luxurious to mock up a VPS in seconds. Check out Justin Weissig’s series on Vagrant, as well as any of his other screencasts.
  • The performance and accessibility improvements are pretty astounding. Trimming the fat feels great. (Would highly recommend.)

Got some fun projects coming up; stay tuned.

Second Thoughts

I think I made a mistake.

Despite this being a busy spring and whatnot, the anxiety I’ve felt about blogging has largely evaporated: I’ve built myself a platform I enjoy using. But with the brouhaha about Facebook’s Instant Articles, it sure seems like a lot has changed very quickly.

Before, I’d mentioned that I had reservations about Meteor and those have born fruit. Much was squawked from on high with regards to the Future of Journalism™ when Instant Articles launched, but I’ve been much more interested in the meta-narrative about the web itself. Peter-Paul Koch and John Gruber brought me to the following:

  1. Instant Articles aren’t just a control play by Facebook—they’re a reaction to god-awful-slow sites.
  2. How’d the web get this slow? Developers are lazy.

Which is a long way of saying that I think I made the wrong bet. Now, don’t get me wrong—this is anything but an indictment of Meteor. I still love it. I’m still going to use it for data, work projects, and authoring here. But for an admittedly low-rent blog, it’s overkill. Fully-featured web apps have have a time and a place, but the web doesn’t need to pretend to be a native app. It should do what it’s best at: being the web.

So, I’m putting myself on blast and making some changes. Honestly not sure what to do yet. Meteor was a great ass-backwards introduction to Node, and all the modern programming features being added in ES6 are mighty tempting. But it’ll probably be something rendered on the server and sent as plain-jane HTML, with plenty of room for accessibility, performance, and whatever else comes to mind.

It won’t be perfect.

But it’ll the web.

Marginalia I

  1. I tweaked the way my desk is built—again. (For those keeping score, I’ve likened my desk’s construction to being akin to a hotrod or fancy bicycle: tinkering is much more fun than driving. Or sitting. Whatever.)

  2. I decided to toss this site up onto GitHub. I still have weird feelings about GitHub (the company) and lord knows this site is very rough around the edges, but in the interest of transparency and maintaining momentum, I’ll put it out in the open.

    DumDum is a Meteor app designed to read from collections of my blog posts and other data. I need to do more polishing and cleaning—there are some rough templates, missing features, and other bumps—but it’s certainly succeeded in helping me have more fun with blogging.

    As of this moment, I’m keeping my authoring methods a secret—I’ve a number of half-formed thoughts of where I’d like to go next, and need to do a fair bit more work before showing it off.

  3. As I mentioned a while back, I attended Theorizing the Web 2015 a few weeks back, and what a wonderful thing. Really, I’ve hardly not been able to think about it since. The discussions of surveillance and the social roles of algorithms were simply oustanding, and brought a great deal else in the world into new light.

Adventures in Pi

Last fall I bought a Raspberry Pi for kicks and really loved it—a delightfully simple introduction to Linux proper and more command-line tools. That said, I really didn’t do much more with it than say, “Hey! I did a computer thing!” And despite fondness for Brian X. Chen’s neighbourhood camera, taping my little neck of the woods was entirely too creepy.

And so I wandered off to noodle on other projects, casting a mild grimace at the Pi gathering dust on the shelf. That is, until I found Johnny-Five.

Simply put, Johnny-Five is a fun Node library for interacting with hardware across all sorts of platforms. And it works terribly well! Aside from stumbling onto the fact that servos need to be calibrated after aimless puzzling, I was able to slap together a simple bot for raising and lowering blinds over a day or two. It’s exactly the sort of fun I’ve been looking for in small weekend projects.

The blinds control flow/script is here as a gist—comments and suggestions are more than welcome. I’ve sketched out some automation/API features but I’m holding off until I’m a bit more confident in the actual hardware bits. (If the power goes out, a number of bits and bobs will go a-clattering to the ground.)


I’m not very good at blogging. Writer’s block has always been a bit of a thing with me, but even more so when writing expressly for myself. I rebuilt this site in the Spring of 2014, with the full intention of blogging—I’d noodled with Tumblr and other platforms for writing, but could never get over the underlying sense of “euch.” I wanted to own my work. So I built myself a platform; but it still didn’t seem right.

Last October, Brent Simmons wrote about blogging engine options. While his home-grown site had served perfectly well, it felt stale and didn’t meet all his needs. So he decided to make something fun to learn with, to fool around with; to write an “app with an audience of one”.

And all of the sudden it all seemed to click into place.

— Which is really just a long way of telling how I ended up building this site. It’s a Meteor app running on a virtual private server. Why Meteor?

  • The Meteor Development Group seems to be doing a bang-up job, and there’s an active development community (even here in Cincinnati).
  • The prospect of one language everywhere—even if it’s JavaScript—is awfully tempting given that an awful lot of my work ends up written in JS. I have a handle on the basics; I want to get better.
  • It’s modular. Meteor encompasses a mess of smaller, interconnected projects that can be swapped out and extended as necessary. And, incorporating existing Node packages couldn’t be simpler.

Meteor’s proven itself several times over to me. Various work and personal projects have been a snap to slap together with Meteor, and although I’m starting to feel some of the rough edges, I don’t plan on changing anytime soon. As I mentioned above, I’m not fond of others owning my writing or work; so if this site has any sort of mission, it is to remedy that. Once I have something more tangible than a nebulose notion I’ll be writing about all that here.

Having said that, the site is far from perfect. As is the case with virtually every JavaScript front-end framework, things like accessibility, performance, and progressive enhancement are non-entities. I’d like for that to change—either with Meteor or without it. And, I’ll be honest. There’s plenty else that needs polish.

— So, why do all this? Because I forgot I am my most important audience. Of course I’m writing here to share thoughts and ideas, but that depends entirely on my own experience. All the technical bits and bobs have been a blast, and I’m delighted to say that programming and writing here is a thrill. Which matters. My writer’s block isn’t born of what I might have to say; it’s that I might not say it well enough.

I am young and naïve and prone to mistakes, but that shouldn’t get in the way of fun.


Well it’s certainly been a while. A recap:

  • I got an honest-to-goodness, full-time job. I work with data. It’s nice.
  • I learned R for the data stuff. It’s a funny language, but the community’s good and there’s (mostly) no weird, Python-ish whitespace.
  • I also got a whole lot better at stuff on the web. Started fiddling with all kinds of JavaScript frameworks[1] and interactive visualisations for work.

So, a year of learning. And it’s been great! I needed some time to meet people and soak in some new ideas. But, I’ve missed blogging. I have a bad habit of sitting on things until I’m convinced they’re perfect—which never happens—so I’m trying to get better at putting things out there and letting you figure it out.

I’m typically in the camp of “tools don’t matter”, but the previous version of this site never felt quite right; publishing a post meant moving files wrapped up in YAML over FTP to my server. Not bad, just… kludgy. I’ll do a full write-up at some point when I’ve had a chance to tidy up some of my mess, but this incarnation comes much closer to meeting my own needs. Which means I now don’t have an excuse to be on here.

Anyways, if you’re still listening, thanks; and if you’re new, well, also thanks—and say hi[2].

  1. Every time a JS framework is born, an angel gets its wings.  ↩

  2. I’m going to Theorizing the Web! If you’re around, I’m hash-modding “Here Comes Every Body” (Studio B #b5) on Saturday, April 18th.  ↩


So I changed a few things.

After getting the first version of the site up, I was already unhappy. While not terrible, it was very much the product of my early lessons in Sass and Bourbon. And of course, life intervened and I had to put down development for longer than I would have liked.

So, a week ago, I finally cleared my docket and threw myself back into the mix. This redesign is the happy result. I'm still chewing through my portfolio — obscenely high standards notwithstanding — but hope to share some of it in the very near future. And, I'm happy to say that I'm ready to announce a few other irons in the fire.

Coming Soon(-ish)

With a wholesale redesign of the site, I'm very behind. Some updates on what I announced last post:

  • Standards-compliant code. I rewrote a ton of PHP and Sass for the redesign, and it is dirty. I'll be validating everything and optimising a ton of the final code.
  • Open-sourced code. I still don't grok git. Advice and pointers are welcome. As for hosting the code: given relatively recent events, I'm not wild about GitHub. Suggestions for alternatives are also welcome.
  • CSS transitions and animations. Though these were shelved for this redesign, I already have plans for various progressive enhancements.
  • Portfolio. It's real. It exists. And it's coming very, very soon. Promise.

And, as I mentioned above, I have a few new ideas I wanted to get out in the open:

  • Dynamic homepage. A few designer-types have said some very clever things about hosting your own presence and data online, and have given animus to a few ideas. I'm still in the early-thinking stage, but hope to be showing off something in the next few months.
  • Directory of stuff I like. This has been rattling around in my head for a long time. I always loved things like the Whole Earth Catalog and Kevin Kelly's Cool Tools, and the past few years have seen a serious upswing in curation/recommendation websites (e.g. The Wirecutter). While I don't intend to launch a reviews site, I've always wanted the space to wax poetic about my favourite widgets and media.

With luck, I'll have more cool stuff to show you soon. See you on the other side.

Hello, world.

My name is Matt Policastro. I am a designer. This is my website. I'm fairly certain I've made other "hello world" posts elsewhere, it only seemed right for my new, self-hosted blog. It also only seemed right to talk about how this all came about, and what's coming next. Shall we?


I began in print design at The College of Wooster. As a student I was involved in all kinds of activities and groups and sort of landed into the role of making posters and printed materials. That said, I was also a huge geek and had always dabbled with computers and the web. That dabbling finally reached critical mass in October of 2013.

Working as a freelancer, I was awash in terrible websites and uneven online presences. Burdened with a sense of insufferable curiousity and the print designer's obsession with control, I threw myself to the web to learn as much as I could. And I found beautiful things. With the guidance of brilliant writers, podcasters, and designers, I realised I needed to build something of my own.

And now — after myriad prototypes, landing pages, and experiments — I have this. Welcome.

The Site Itself

Everything was written in Coda 2 and Byword using Kirby as the CMS and Bourbon/Bourbon Neat/Bourbon Bitters by thoughtbot for Sass/CSS writing.

Everything on the site is built mobile-first and fully-responsive. Though the site is certainly still a rough draft, what you see today is the foundation for everything I do moving forward. This site functions as my home on the web, and I can't wait to show you all what I have planned. Luckily, I can tell you about some of it.

Coming Soon

As mentioned above, some things are still in rough shape, while others are missing altogether. Here's some of the things I have planned for the near future:

  • Standards-compliant HTML. Even though I'm a young designer still learning the ropes of code, I plan on keeping things as accessible and up-to-date with the latest and greatest.
  • Open-sourced code. I'm glad to share any tricks or tools I create in the process of building and maintaining the site. On the other hand, I'm still figuring out how git works.
  • CSS transitions and animations. CSS is an amazing tool, and it only gets more powerful with things like Sass. Expect things to get an awful lot fancier in the near future.
  • Portfolio. And but of course, my actual portfolio. This site is my home online, and I've love to show off my work. But only when it's as beautiful as it can be.


Despite the many, many missteps I've taken on the road to the site you see, I'm really quite proud of what you see. And it's only going to get better.

Talk again soon.

– Matt