Trying Typst for the first time

I've been meaning to check out Typst for over a year now. The premise was intriguing: what if we rewrote \(\TeX\) (and \(\LaTeX\)) from the ground up, discarding the parts that made sense fifty years ago and taking everything we've learned in the meantime to make a simpler, clearer markup language? (And also wrote it in Rust for blazing speed.)

I've mentioned \(\LaTeX\) and my love for it many times on this blog. I think I was first introduced to it during the spring semester of 2010, and my immediate thought was: where has this been all my life? \(\LaTeX\) is a markup language (built as an extension of the earlier \(\TeX\)) where documents written in it are compiled to output a nicely typeset version (usually in a format like .pdf or .ps). One of its big selling points is its ability to handle complex mathematics; for scientific writing in particular, which may need to include complex equations with highly specific formatting, it's a lifesaver. I've got a Javascript library called MathJax on this blog which lets me use (some) \(\LaTeX\) formatting to display mathematical output such as:

\[\text{Coulomb's Law: } F=k\frac{q_1 q_2}{r^2}\hat{r}\text{ where the Coulomb constant } k=8.988\times10^9\frac{\text{N}\cdot\text{m}^2}{\text{C}^2}.\]

I was initially drawn to \(\LaTeX\) for the beauty of its output. It was also practical: you can write pretty much any kind of document in \(\LaTeX\), and most scientific journals allow (or require) manuscripts to be submitted in it. I wrote my thesis in it, then turned it into two papers for publication.

At the same time, \(\LaTeX\) is not without its weaknesses. The underlying language \(\TeX\) is both very powerful and very complex (\(\LaTeX\) is essentially an addon for \(\TeX\) to make it easier to use), has a famously steep learning curve, doesn't natively handle UTF-8 characters, and even after years of using \(\LaTeX\) I occasionally run into bizarre issues. Perhaps the strangest was discovering that a hyperlink which spanned a page break in my thesis caused the whole document to fail with a very obscure error message – one I actually never found a real fix for despite much searching. The only advice I could find which worked was to adjust the length of the text before the hyperlink such that it no longer crossed two pages!

This is not to disparage \(\LaTeX\) – it's still worth knowing, I think, has been immensely useful for many people around the world, and it still feels a little magical to watch it take a series of abstruse text and turn it into something beautifully formatted. However, I'm cognizant of its flaws, and open to new experiences.

Enter Typst.

There's something of a pattern in software development, where after developing a piece of complex software, sometimes the best way to improve is to start over from scratch using the lessons learned rather than try to continue patching and editing it. Typst is essentially that rewrite for \(\LaTeX\), taking lessons learned from the last forty years and improving upon it (at least in theory).

Over the summer I read through the Typst tutorial, though Hurricane Lala knocking out power for two days the week before the semester started put a crimp in my plans to try writing my syllabi with it. A practical demonstration of Typst is that I sat down Sunday afternoon the day before I was due to begin teaching and attempted to write a syllabus in it (based on an existing \(\LaTeX\) template I had), and within an hour or two I'd created something which mimicked ~95% of the original's look while requiring much less text to pull off. I was seriously impressed, and then hooked.

It's a little hard to show off what Typst can do, since by design it's often used to mimic the look of \(\LaTeX\) documents. What I can say, however, is that it's much easier to use. This morning in some free time I followed the tutorial instruction to turn my syllabi into a template for future syllabi to use, something I've never even attempted in \(\LaTeX\).

It's also a real joy to use due to lightning-fast compilation. With \(\LaTeX\) you have to run a compiling command, which even for relatively short documents often takes a few seconds. Typst compiles incredibly quickly due to an incremental compilation system, allowing you to see the output of what you're typing in basically real time. And coming back to writing \(\LaTeX\) for this post, I find myself appreciating that all the more; it's amazing what a shorter feedback loops does, allowing you to more easily experiment and try things and see the result immediately.

I could wax lyrical about other features of Typst*, but this post is getting long and I have a feeling I'll be talking about it more in the future. I find myself drawn back to it, wanting to experiment and see what else I can get it to do. And that's perhaps the greatest endorsement I can give Typst: using it is fun. I'm reminded of the heady days when I first discovered \(\LaTeX\), and I can hardly wait to write more in Typst and see what interesting and beautiful things I can create with it. Hopefully I can show some off in the future when I've learned more! A hui hou!

*Typst's syntax is generally lighter and simpler than \(\LaTeX\)'s, its error messages when something is broken are extremely helpful in pointing you to the problem, it has a powerful and flexible CSS-like functionality built-in which makes it a snap to style elements, etc.

Hurricanes and (UH) Hilo

Well, it's been an eventful month here. A week-and-a-half ago on the 15th hurricane Lala swept by and knocked out power to 60% of the entire state, the week before the fall semester started. I lost power for ~54 hours from Saturday afternoon through to Monday evening, and there are people who still haven't gotten it back. We didn't actually see too much wind or rain here in Hāmākua, but Puna and especially Kāʻu got slammed. Then tropical storm Moke passed by on Saturday, but thankfully was far enough away not to add much more than a little rain. I definitely learned a few things about disaster preparedness (namely, that having solar and a battery is the way to go, having spoken to multiple people who were able to keep their fridges running for even longer than the two days I was without electricity).

And did I mention the fall semester started at UH Hilo on Monday? I'm back to teaching with a few more classes this semester, though most of them are labs for which we have experiments already prepared. (I will be grading a lot of lab reports, though.) There was a bit of the usual start-of-semester craziness, with the building AC once again not working, the internal network acting up, the laptop I'd been using last semester giving up the ghost over the summer, and me forgetting to put in the list of rooms I needed access to for my key fob until the day before, but things have been working out. (The AC has been halfway repaired, the network seems to have sorted itself, I got my key access before the end of Monday, and the college has a loaner laptop I can use which should come this week.) I also switched up how I'm writing my syllabi in a way which should make it easier to update them (namely, using Typst instead of \(\LaTeX\), but that'll get its own post here shortly.

In brief: no real damage about the place (but hurricane seasons lasts through November), classes have started, and I'll probably get really busy next week as the labs start to kick off. But I do want to talk about Typst now that I've actually gotten around to trying it, so expect a post on that before too long. A hui hou!

Hummingbirds, sea lions, and honey bees, oh my!

I spent the final week of July in California visiting family where I had the opportunity to take a number of pictures of animals I don't see very often. Hummingbirds were active about the feeder, we visited Pier 39 in San Francisco and saw the sea lions there, and there were a lot of thirsty honey bees around in the July heat.

It's difficult to get a good shot of a hummingbird in flight with how fast they move, so my only chance was when they came up to the feeder to eat. I tried on two different days (seen here) and I'm impressed with how clear the wings came out in the first photo considering the blur they were in person. I was also impressed by how close they let me stand to get these photos. Occasionally one would fly over even closer, at about arm's reach, to hover at head-height and regard me quizzically for a few seconds, a delightful (and ever-so-slightly unnerving) experience.

We were fortunate while visiting Pier 39 that there were a number of sea lions there, since it varies. Mostly they were napping in the sun, though there were generally always one or two swimming about, jumping on or off the floating rafts, and occasionally getting into disputes.

And sometime after my last visit, my parents let a commercial beekeeper park some hives on their property. (It's a great location with all the eucalyptus and almond trees around.) I don't think I've ever really mentioned it on this blog, since it happened before I started writing it, but I used to have bees for a few years before moving out to Hawaii. I enjoyed it, though I'm a poor bee-haver; I picked a bad time to start between Varroa mites and Colony Collapse Disorder (and it hasn't exactly gotten better since then), and my hives didn't manage to survive. I figured it best left to the past when I moved, but I've always retained a sort spot for honey bees. When some fell into the pool I was filling up for them I rescued them with a leaf and let them chill on my hand while I did some macro photography. It's always funny to me how matter-of-factly they go about cleaning and drying themselves before taking to the air again.

And that's it for this post. There are plenty of cool critters here in Hawaii (including honey bees), but it's nice to see some that I haven't in a while. A hui hou!

Making custom coasters

Over the summer I've been experimenting with various woodworking projects, and one which I made on something of a whim was a set of coasters. I thought it'd be fun to take some thin strips of different-colored wood, glue them together (laminate them), cut out short pieces, then glue those pieces together in a pin-wheel, like this:

Woods used are whenge (black), jatoba (brown), maple (white), and purpleheart.

Here they are after I leveled out the slight differences in thickness in the wood strips and applied a finish (on top of a few cutting boards I also made):

The mineral oil in the background is for the cutting boards, for these I used a spray-on tung oil finish.

They came out looking pretty good, but I'd underestimated how much wood doing the spiral would take, so from my initial, roughly 24-inch-long piece of laminated wood strips I got a grand total of two coasters. However, they were thick enough that a friend suggested cutting them in half to make four…which I proceeded to do in possibly the dumbest, hardest, least-efficient way by using the reciprocating saw I'd acquired as part of a package deal when I bought my first power tools back in November (and hadn't used until now).

Have I mentioned I don't own a vice?

Though even if had it wouldn't have helped much, as I discovered when I built a Frankenstein-ian contraption of various clamps and pieces of wood to hold the coasters as I cut them. My attempts to secure them usually lasted a couple of seconds before the oscillating action of the reciprocating saw worked them loose resulting in a lot of awkwardly re-clamping coasters, but it somehow – incredibly – resulted in two sets of roughly-equal thickness coasters with a remarkably straight cut across them.

They're definitely not perfect (even after some more sanding and smoothing), and it was mostly a lesson in what not to do, but y'know, sometimes those are helpful. They might still work as coasters (I haven't really tried), they do look pretty, and it was just a spur-of-the-moment project, so it's all good in the end. Just part of the learning process! A hui hou!

On the difference between gasoline and electric chainsaws

Around the time I moved in last year I realized I was going to need a lawn mower, and also that (for the first time in my life) I was in a position where owning power tools made sense. I hate dealing with gasoline and internal combustion engines and the house has solar panels, so electric was the obvious way to go for a lawn mower. After some deliberation I settled on Ryobi, which fit my needs (and budget) as a DIY-er rather than a professional tradesman. I've been pretty happy with the self-propelled mower I got, along with a number of power tools for woodworking.

One thing I didn't have much need for was a chainsaw, due to there not being many large trees on my parents' property. In May, however, someone I know in Puna who'd heard I was getting into woodworking through a mutual friend reached out to say he'd had some trees come down during some stormy weather in March, and asked if I wanted some of the wood if I'd come cut it up. I said yes to free wood, and while I started out using his landlord's gas chainsaw I knew I'd eventually need one of my own to cut the wood further when I got it home. And that's how I became the owner of a 16" Ryobi electric chainsaw, and had opportunity to try both gas and electric saws recently.

The difference in volume is the obvious point of comparison; I've never been fond of what I can only term the aggressively angry whine of a gas chainsaw revved up. The electric version isn't silent, but the noise it produces is primarily from the chain moving around the blade, and it is a lot quieter – you barely need hearing protection at all.

A less obvious difference is how an electric chainsaw is always ‘on,’ essentially. With a gas chainsaw you have to explicitly start it. This can be an involved process if it hasn't been used in a while, and at minimum requires giving the starting pull-cord a couple of pretty good yanks. It's a point of friction, so you'd rather not do it too often – yet once it's running you want to minimize time spent not cutting, since it's burning gas while idling. This inherent tension makes cutting up a fallen tree tricky since it's an iterative process: you usually can only make so many cuts before needing to move the now-separated branches out of the way so you can reach more parts of the tree. (To be fair, this is based what I encountered doing it solo ­– someone to help move branches out of the way would alleviate this somewhat.)

With an electric chainsaw, however, it's trivial to make a single cut, move a branch, make a cut, etc., in ways that would be really inefficient while burning gas. If I were a professional lumberjack or arborist I'd probably still have reason to be using a gas chainsaw, but for occasional home use like cutting up a single downed tree it's a much more pleasant experience.

Anyway, we'll see what comes of all this cypress wood I've acquired. The trees were clearly healthy and just went down in the storm, as they were still sopping wet when I cut them up even months after falling. From what I've read they'll need to dry for some time before I can really use the wood, but I'm looking forward to eventually making use of them. Future projects, perhaps. A hui hou!

Happy Tau Day 2026!

Happy Tau Day everyone! It seems like hardly any time at all since last year's Tau Day, which is probably a reflection of how utterly, overwhelmingly, busy I've been most of the time between now and then, only finally slowing down this past month-and-a-half. This slowdown (and some non-rainy weather at last) has allowed me to finally start getting to some projects around the house and yard that I identified soon after moving in, so that's been nice.

I don't have anything particularly tau-related to discuss, except for the fact that this week I started looking into Typst. Much like its inspiration \(\LaTeX\) Typst seeks to be a mark-up language that can be used for writing pretty much anything, including scientific or technical papers with complex formatting requirements or mathematical equations. I'll have more to say about it in the future when I've had more time to actually try it (so far I've only read through the tutorial), but it looks interesting. I've heard journals are starting to accept manuscripts typeset in Typst, so it seems to be catching on in the broader scientific world.

We'll see where that goes, though. A hui hou!

Celestial conjunctions and telluric temblors

Two nights ago was the closest approach of Venus and Jupiter in a conjunction, and despite the very rainy weather this summer the sky was actually clear enough just after sunset for me to get a few photos. Not entirely clear, mind you, but I think the ragged clouds drifting across the sky add to its effect.

Venus and Jupiter in conjunction. Captured with my Pixel 7 Pro.

Venus is the brighter of the two bright lights on the right, with Jupiter on the right. After the Sun and Moon Venus and Jupiter are the third- and fourth-brightest objects in the sky, so their conjunction is particularly impressive. Above and to the left of them you can faintly see Messier 44, the Beehive Cluster in Cancer.

Last night we also had an earthquake just as I was dropping off to sleep of about magnitude 4.7, with an epicenter about 23 km to the east. I realize now that I haven't mentioned the larger earthquake (magnitude 6.0) which happened back on May 22nd over on the western side of the island. (The largest earthquake I've felt since the magnitude-6.2 tremor I felt within a month of moving back to Hilo in 2021.) Both of these earthquakes felt pretty long; it's always hard to estimate time during an earthquake, but I've lived through enough that I feel pretty confident that these were on the longer side of ones I've experienced. Neither produced any damage here thankfully, though the larger one did some damage on the western side of Hawaiʻi island.

According to the USGS the large one was fairly deep, at 22.6 km; deep enough that it likely wasn't volcano-related, but was instead due to the ocean crust buckling and adjusting to the weight of the island. Last night's was also pretty deep, at 39.2 km, so I wonder if it's simply the crust adjusting on this side in response. As I like to say, life's never boring when you live on a volcano in the middle of the Pacific ocean! A hui hou!

End-of-semester reflections

Now that the semester is over I've had some time to rest and relax the past fortnight, and reflect on my experience this semester. And I think…I might enjoy teaching.

It's a question everyone's been asking me, but not a straightforward question for me to answer. The superposition of emotions that makes up how I feel about it refuses to yield to observation and collapse to a single, simple, state. Mostly, I don't feel anything – I don't dislike it, but I also don't feel any particular deep yearning towards it either. (Though in fairness this is also how I feel about programming, which I like well enough as a profession.)

The clearest indication came near the end of April, when I saw and applied for a software developer position in astronomy on this island. It was similar to positions I've held and enjoyed before, and had I applied last year I've have been quite excited for it – yet as I was applying, I felt an unexpected pang of…regret, I suppose, that I wouldn't be around to watch the students I've begun to get to know this year develop over the course of their undergraduate journey.

We'll see if it goes anywhere, of course. I'm on the schedule to teach several classes next semester so I have no idea how or if that would work if I got hired. I find myself thinking a dual arrangement might be nice, where I teach a few classes on the side going forward. But this is all speculation for now.

Another thing I've had time to reflect on was attending the graduation ceremony for UH Hilo the day before my birthday. I left in the middle of the semester when I finished up my Ph.D. at Swinburne, so my last time being part of a ceremony was my undergraduate commencement in 2012. This was also my first time attending as faculty, and I found it a profoundly different experience.

In some strange way, watching the doctoral students being awarded their hoods and degrees gave me a sense of closure. Back in 2021 I literally went from submitting my thesis Sunday afternoon to starting my job at Gemini the next morning, so I had zero time off in-between. That wasn't really the end of it, though, since I then spent most of the next six months turning my thesis into publishable papers, working on evenings and weekdays (with a small break in December for my thesis defense a few days before Christmas). I was pretty burnt out for a long time, and while that's probably representative of most grad students at some level I never really had much in the way of closure; my papers finally got published, and I just…didn't need to be spending most of the time I wasn't working thinking about the process anymore.

So I found it surprisingly emotional to be sitting among the other faculty at the graduation ceremony two weeks ago, watching my fellow newly-minted doctors receiving the rewards for their years of toil. Sitting with the faculty dressed in our formal regalia (wearing a borrowed finery in my case), I felt a connection to the scholars around the globe who for hundreds of years have sat in ceremonies like these, welcoming new learners to their ranks in fellowship. It was a really nice feeling being recognized by the chancellor, both for being a UHH alumnus and as part of the faculty. I try to stay humble and not let my accomplishments go to my head, but it's gratifying to have the many years of hard work I put in bettering myself be recognized.

So what's next for me? As mentioned I'm on the schedule to teach again in the fall, but I don't have anything officially going on over the summer. I have no lack of ideas for projects and things to try, however, so I hope to have ideas for posts before too long. After a little break I've started up on some woodworking projects, including constructing some jigs to make future projects easier. (One of the fun things I'm discovering about woodworking is the ability to simply make your own tools to accomplish things.) We'll see what I get up to, though! A hui hou!

Living my prime years

Another year, another birthday! And what a year it's been. Last year around my birthday I was working with NEON on the slopes of Mauna Loa (about which I gave a talk this week at the Hilo photography club meeting). This year I've just wrapped up finals weeks at UH Hilo after my first semester teaching, and attended the graduation ceremony yesterday. Attending graduation as faculty was an interesting experience and I have some thoughts about it, but they're still crystallizing so they'll come in a future post.

For today, I was thinking earlier that I'm celebrating my 37th birthday, and how that's back to being a prime number after 36. And then I thought about phrases like “being in one's prime,” or “being in the prime of life,” and started wondering: just how many years have I spent at a prime age?

Now, what does that very hastily-conceived notion mean? It's essentially just how many of your preceding years of life corresponded to prime numbers. We have to a little careful about birthdays, though; in essence, birthdays count years of age since the day of your birth, so your first birthday comes at the end of your first year of life. You are then “1 year old” during your second year of life until your second birthday. If we count the years which correspond to a prime number as “prime years”, we can compute, at each birthday, how many of your preceding years were “prime years” and what fraction that makes with your total number of years lived, as in this formula:

\[\text{prime years fraction}_{\text{at age }n}=\left.\frac{\text{prime years}}{\text{total years}}\right]_{\text{at age }n}\]

Confused yet? Let's start from your first birthday, which celebrates your first year outside the womb. One is not a prime number, so your total number of prime years lived to this point (out of a total of one) is zero; your prime years fraction is thus \(0/1=0\). At your second birthday, you have lived a total of two years; the first one is not prime, but your second year corresponds to two, which is prime, so you have one prime year out of two, or \(1/2=0.5\). At your third birthday, you have lived one non-prime year and two prime years (since two and three are prime), for a fraction of \(2/3=0.66\dots\)

And it's not hard to see that this is as good as it gets – at your fourth birthday you'll have lived two prime and two non-prime years so the fraction returns to 0.5, and since prime numbers only ever get farther apart your fraction of prime years begins a slow slide downwards. One interrupted by upwards jumps at each prime birthday, to be sure, but the overall trend is obvious. I found this sufficiently interesting to write up a quick Python program and plot what it would look like for ages 1–100, and here it is:

You can see the trend I've described: it starts at zero at your first birthday, shoots upwards the next two years, then begins a jagged descent. Since 37 is prime I've just jumped up to a prime fraction of 0.324, or, another way of putting it, just under one-third of my lived years have been prime (numbers).

The plot kind of reminds me of the nuclear binding energy curve in atomic physics, but I was also intrigued with where it seemed to be going. I thought it might be converging towards 0.2, but running it to 1000 dispelled that notion:

At this point I suspected it was converging towards zero at infinity, but my naïve implementation of prime-checking in my code was starting to take noticeably longer to run at higher upper limits. For a limit of ten thousand it still took under two seconds to run, but at a hundred thousand it took over four minutes:

I'm sure I could implement a more sophisticated method of prime-checking and push it to higher numbers*, but ultimately, checking whether a number is prime or not is a difficult problem (and one on which much of modern computer security depends). The overall trend is clear, mathematically, so for an idle musing like this I'm satisfied with what I got.

Anyway, I hope you found that at least a little interesting. We'll see where life takes me this year. Currently my summer is wide open (I've got a lot of ideas for things to do) and I'm on the schedule to teach again in the fall, but who knows where things might end up going. A hui hou!

*Edit: I certainly could, as I was checking all the way up to \(n/2\) instead of merely \(\sqrt{n}\).

On the semantics of uniqueness

I've mused occasionally for a while about the semantics of the concept of “uniqueness”. I've seen it asserted that uniqueness is essentially a boolean quality: something is either unique or it isn't. There's no comparative or superlative form of unique: uniquer, uniquest (although having written that, those sound pretty cool!). And the first two definitions from Merriam-Webster seem to agree with this: “being the only one : sole; being without a like or equal : unlike anything or anyone else : unequaled”.

But when I find myself pondering the concept, it feels like there should be degrees of uniqueness. For one thing, if we're talking about real objects, then every thing larger than a single molecule is technically unique at some scale. But that renders the concept rather useless as a word. If have a set of five “identical” rubber ball, four white balls and one blue one (⚪⚪⚪⚪🔵), it makes sense to describe the blue one as unique compared to the set of all of them, even if they technically are not all perfectly identical (and thus are all technically “unique”). So there's some tolerance built-in, within which we consider things to be identical for the sake convenience. That's fine – we live in the real world, not an ideal one.

Now consider a set of five balls, all of different colors. ⚪🔵🟢🟠🟣 It makes sense to describe each one (within the set) as being unique: they're all different from each other member of the set in terms of color, which is a concept we generally attach importance to. (Though I just realized I haven't considered color-blind people in this example; I should remember that for the future.)

But what if we add an elephant to that set? ⚪🔵🟢🟠🟣🐘

Every item in the set is still technically unique – but it certainly feels as if the elephant is somehow more unique than the other items. It's more different from all of them than any of them are from each other. Remove any one item from the set and the diversity (in pretty much any metric) will be at a global minimum when that item is the elephant. Calling any of the balls unique – while technically true – seems somewhat understated in the face of the literal elephant in the room set.

I was prepared to argue this point more, until as I was looking unique up in the online Merriam-Webster dictionary I noticed a third definition, and a usage guide. The third definition reads “very special or notable : unusual”, and the usage guide notes that:

Unique is often cited as a word that should never be modified by terms like somewhat or very. The thinking is that unique properly only describes what is unequaled or otherwise distinct from all others. Just as something cannot be more "only" than another, it cannot be more unique than another. This logic fails, however, when we consider that unique can also mean "unusual" or "rare," as in "a unique opportunity" or "a unique feature." In these cases, phrases like "very unique" are standard.

 “Very”, I would argue, is just an intensifier adverb like “more” is, so I think it's ultimately fine to consider something more or less unique; it's expressing something about the ratio of how similar it is to other things in whatever set we're including it in vs. how similar those other set members are to each other. (You could probably come up with a mathematical definition for this in terms of set theory, but I'm not sure I know enough about it to do so.) But that's just my take on it, and it is, ultimately, semantics. Feel free to chime in in the comments! A hui hou!