Hello, I'm Sean and I'm a programmer in the UK. I've cobbled together this page to let people know how to contact me and learn a little more about me.

The best way by far to get in touch is to use a kind of instant messaging system called IRC. I'm in the #sbp channel on the irc.freenode.net server. You can use a web interface, or download a native IRC client for your platform. IRC was invented by a Finnish guy in the 1980s, so you know it's bound to be great.

Alternatively you could email me, at sean at the domain miscoranda.com. Unfortunately I'm deluged with email so I rarely spot the many wonderful and interesting things that people email me, and IRC is a much better option. Although cryptography apps aren't very user friendly, even my own, you could also use something like Tox to chat with me privately.

My main research interest as of writing this page is open computing based on language oriented environments. We're talking systems like Genera, STEPS, InterimOS, and to some extent Oberon, TempleOS, and EmacsOS. For the programming language I'd like to design something very radical, instead of being stuck with languages like the C, Python, and Shell that I normally use. Maybe we should all just use FORTH.

Ideally I think that programs should be formally verified, but current tools such as Coq, Idris, and Z3 don't give programmers what they need to do this verification easily. Data should be persistent, safe, and undoable. At the hardware level I'm really enthusiastic about RISC-V chips and the possibilities they should bring for more pervasive open computing.

Since using InterimOS and the like is infeasible, I'm a fan of Arch Linux. To make X more palatable, I use Openbox with the Arc and Paper GTK theme, and Noto Sans 9 and Hack 10 as my main fonts. I don't like the digit "0" in Hack, but apart from that it's a really decent monospaced font. Eventually I'd like to use Wayland.

I'm a big fan of decentralisation in technology, so I abhor DNS and the CAs that we use for the secure web. This centralising tendency even affects codebases, so browsers (like banks) are now too big to fail and can do anything they like, even if that means deprecating and removing isindex, gopher, MIDI playback, and even http. Although this page should have reached you sealed using simple mathematics, can you verify that seal for yourself? We could at least fight DNS and CAs if more people used hidden services on Tor, but Tor is not widely used even though it also provides the further advantage of anonymity. Don't get me started about DRM.

Although I do like HTML, and am very interested in its history and handmade vernacular, I don't like bloat and the JavaScript fuelled Win98-isation of the web. Therefore I tend to write very spartan webpages. This one for example is generated from a lightweight markup format that I created, and uses only the P and A elements. I've made sure that no links are more than 80 characters long, for compatibility with IBM's 1928 punch card format which is still a widespread convention in modern terms thanks to their common lineage from the VT100.

The lack of headings probably make this page a little harder to read, but many centuries ago things were written in scriptio continua style, which is ALLMAJUSCULEANDNOSPACES. Now we have hypertexts that we can click or even just touch to load in new hypertexts from around the world instantly, so be grateful!

Typography in HTML is awful, though. Even if I included something like Gentium or Linux Libertine using WOFF2, no browsers support the Knuth-Plass algorithm for hyphenation so you can't use nicely justified paragraphs. As for symbols, gaiji, ruby, rubricated diacritics, mathematics, automatically generated footnotes, and the like, those things are still a few decades off. Once pixels give way to vectors through greater density or better devices, at least typefaces can really shine.

Representing music is also difficult. I'm often tinkering with songs and would like to display the results. I've experimented with using SoX to generate beautiful spectrograms in the melodic range, with some success, but automatic microtonal harmonic analysis is some way off.

Talking of web bloat, this page is getting quite large: over 10 log nats, uncompressed, where a byte is about five and a half nats. Even zstd can't get it below 9 log nats. Chuck Moore could write a whole colorForth space probe in that amount of information!