Article

Draw Me a Comment!

Why must a blog comment be text? Why could it not be… a drawing?¹

Red and black might be more traditional ladybird colours, but sometimes all you’ve got is blue.

I started hacking about and playing with a few ideas and now, on selected posts including this one, you can draw me a comment instead of typing one.

Just don’t tell the soup company what I’ve been working on, okay?

I opened the feature, experimentally (in a post available only to RSS subscribers²) the other week, but now you get a go! Also, I’ve open-sourced the whole thing, in case you want to pick it apart.

What are you waiting for: scroll down, and draw me a comment!

Footnotes

¹ I totally know the reasons that a blog comment shouldn’t be a drawing; I’m not completely oblivious. Firstly, it’s less-expressive: words are versatile and you can do a lot with them. Secondly, it’s higher-bandwidth: images take up more space, take longer to transmit, and that effect compounds when – like me – you’re tracking animation data too. But the single biggest reason, and I can’t stress this enough, is… the penises. If you invite people to draw pictures on your blog, you’re gonna see a lot of penises. Short penises, long penises, fat penises, thin penises. Penises of every shape and size. Some erect and some flacid. Some intact and some circumcised. Some with hairy balls and some shaved. Many of them urinating or ejaculating. Maybe even a few with smiley faces. And short of some kind of image-categorisation AI thing, you can’t realistically run an anti-spam tool to detect hand-drawn penises.

² I’ve copied a few of my favourites of their drawings below. Don’t forget to subscribe if you want early access to any weird shit I make.

Permanent Record

To:

****@fulwoodacademy.co.uk

From:

“Dan Q” <***@danq.me>

Subject:

Subject Access Request – Dan Q, pupil Sep 1992 – Jun 1997

Date:

Tue, 23 Jul 2024 15:18:07 +0100

To Whom It May Concern,

Please supply the personal data you hold about me, per data protection law. Specifically, I’m looking for: a list of all offences for which I was assigned detention at school.

Please find attached a variety of documentation which I feel proves my identity and the legitimacy of this request. If there’s anything else you need or you have further questions, please feel free to email me.

Thanks in advance;

Dan Q

To:

“Dan Q” <***@danq.me>

From:

“Jodie Clayton” <*.*******@fulwoodacademy.co.uk>

Subject:

Re: Subject Access Request – Dan Q, pupil Sep 1992 – Jun 1997

Date:

Fri, 26 Jul 2024 10:48:33 +0100

Dear Dan Q,

We do not retain records of detentions of former pupils, and we certainly have no academic records of pupils going back thirty years ago.

Jodie Clayton | Office Manager with Cover and Admissions
Black Bull Lane, Fulwood, Preston, PR2 9YR
+44 (0) 1772 719060

To:

“Jodie Clayton” <*********@fulwoodacademy.co.uk>

From:

“Dan Q” <***@danq.me>

Subject:

Re: Subject Access Request – Dan Q, pupil Sep 1992 – Jun 1997

Date:

Fri, 26 Jul 2024 17:00:49 +0100

But, but… I was always told that this would go on my permanent record. Are you telling me that teachers lied to me? What else is fake!?

Maybe I will always have a calculator with me and I won’t actually need to know how to derive a square root using a pen and paper. Maybe nobody will ever care what my GCSE results are for every job I apply for. Maybe my tongue isn’t divided into different taste areas capable of picking out sweet, salty, bitter etc. flavours. Maybe practicing my handwriting won’t be an essential skill I use every day.

And maybe I will amount to something despite never turning in any History homework, Mr. Needham!

Dan Q

Tidying WordPress’s HTML

Terence Eden, who’s apparently inspiring several posts this week, recently shared a way to attach a hook to WordPress’s get_the_post_thumbnail() function in order to remove the extraneous “closing mark” from the (self-closing in HTML) <img> element.

By default, WordPress outputs e.g. <img src="..." />, where <img src="..."> would suffice.

It’s an inconsequential difference for most purposes, but apparently it bugs him, so he fixed it… although he went on to observe that he hadn’t managed to successfully tackle all the instances in which WordPress was outputting redundant closing marks.

This is a problem that I’ve already solved here on my blog. My solution’s slightly hacky… but it works!

My Solution: Runing HTMLTidy over WordPress

Tidy is an excellent tool for tiding up HTML! I used to use its predecessor back in the day for all kind of things, but it languished for a few years and struggled with support for modern HTML features. But in 2015 it made a comeback and it’s gone from strength to strength ever since.

I run it on virtually all pages produced by DanQ.me (go on, click “View Source” and see for yourself!), to:

Standardise the style of the HTML code and make it easier for humans to read¹.
Bring old-style emphasis tags like <i>, in my older posts, into a more-modern interpretation, like <em>.
Hoist any inline <style> blocks to the <head>, and detect any repeated inline style="..."s to convert to classes.
Repair any invalid HTML (browsers do this for you, of course, but doing it server-side makes parsing easier for the browser, which might matter on more-lightweight hardware).

WordPress isn’t really designed to have Tidy bolted onto it, so anything it likely to be a bit of a hack, but here’s my approach:

Install libtidy-dev and build the PHP bindings to it.
Note that if you don’t do this the code might appear to work, but it won’t actually tidy anything².
Add a new output buffer to my theme’s header.php³, with a callback function: ob_start('tidy_entire_page').
Without an corresponding ob_flush or similar, this buffer will close and the function will be called when PHP finishes generating the page.
Define the function tidy_entire_page($buffer)
Have it instantiate Tidy ($tidy = new tidy) and use $tidy->parseString (with your buffer and Tidy preferences) to tidy the code, then return $tidy.
Ensure that you’re caching the results!
You don’t want to run this every page load for anonymous users! WP Super Cache on “Expert” mode (with the requisite webserver configuration) might help.

I’ve open-sourced a demonstration that implements a child theme to TwentyTwentyOne to do this: there’s a richer set of instructions in the repo’s readme. If you want, you can run my example in Docker and see for yourself how it works before you commit to trying to integrate it into your own WordPress installation!

Footnotes

¹ I miss the days when most websites were handwritten and View Source typically looked nice. It was great to learn from, too, especially in an age before we had DOM debuggers. Today: I can’t justify dropping my use of a CMS, but I can make my code readable.

² For a few of its extensions, some PHP developer made the interesting choice to fail silently if the required extension is missing. For example: if you don’t have the zip extension enabled you can still use PHP to make ZIP files, but they won’t be compressed. This can cause a great deal of confusion for developers! A similar issue exists with tidy: if it isn’t installed, you can still call all of the methods on it… they just don’t do anything. I can see why this decision might have been made – to make the language as portable as possible in production – but I’d prefer if this were an optional feature, e.g. you had to set try_to_make_do_if_you_are_missing_an_extension=yes in your php.ini to enable it, or if it at least logged that it had done so.

³ My approach probably isn’t suitable for FSE (“block”) themes, sorry.

Shared Email Addresses

Email Antipatterns

There are two particular varieties of email address that I don’t often see, but I’ve been known to ridicule when I have:

Geographically-based personal email addresses, e.g. OurHouseName@example.com. These always seemed to me to undermine one of the single-best things about an email address compared to postal mail – that they don’t change when you move house!¹
Shared/couple email addresses, e.g. MrAndMrsSmith@example.net. These make me want to scream “You know email addresses are basically free, right? You don’t have to share one!” Even back when most people got their email address directly from their dial-up provider, most ISPs offered some number of addresses (e.g. five).

If you’ve come across either of the above before, there’s… perhaps a reasonable chance that it was in the possession of somebody born before 1960 (and the older, the more-likely)².

You’ll never catch me doing that!

I found myself thinking about this as I clicked the “No” button on a poll by Terence Eden that asked whether I used a “shared” email address when in a stable long-term relationship.

Terence Eden (@Edent@mastodon.social) on Mastodon asks: "If you're currently in a stable, long term relationship with someone - do you have a joint email address with them?" — Of course I don’t! Why would I? Oh… wait…

It wasn’t until after I clicked “No” that I realised that, in actual fact, I have had multiple email addresses that I’ve share with significant other(s). And more than that, sometimes they’ve been geographically-based! What’s going on?

I’ve routinely had domains or subdomains that I’ve used to represent a place that I live. They’re convenient for when you want to give somebody a short web address which’ll take them to a page with directions to you and links to your location in a variety of different services and formats.

And by that point, you might as well have an email alias, e.g. all@myhouse.example.org, that forwards on email to, well, all the adults at the house. What I’ve described there is, after a fashion, a shared email address tied to a geographical location. But we don’t ever send anything from it. Nor do we use it for any kind of personal communication with anybody outside the house.

We don’t give out these all@ addresses (or their aliases: every company gets their own) to people willy-nilly. But they’re useful for shared services that send automated emails to us all. For example:

Giving a forwarding alias to the supermarket means that receipts (listing any unavailable products) g0 to all of us, and whoever’s meal plan’s been scuppered by an awkward substitution will know what’s up.
Using a forwarding alias with the household Netflix account means anybody can use the “send me a sign-in link” feature to connect a new device.
When confirming that you’ve sent money to a service provider, CC’ing one of these nice, short aliases provides a quick way to let the others know that a bill’s been paid (this one’s especially useful where, like me, you live in a 3+ adult household and otherwise you’d be having to add multiple people to the CC field).

Sure, the need for most of these solutions would evaporate instantly if more services supported multi-user or delegated access³. But outside of that fantasy world, shared aliases seem to be pretty useful!

Footnotes

¹ The most ill-conceived example of geographically-based email addresses I’ve ever seen came from a a 2003 proposal by then-MP Derek Wyatt, who proposed that the domain name part of every single email address should contain not only the country of the owner (e.g. .uk) but also their complete postcode. He was under the delusion that this would somehow prevent spam. Even ignoring the immense technical challenges of his proposal and the impossibility of policing it across the borders of every country that uses email… it probably wouldn’t even be effective at his stated goal. I’ll let The Register take it from here.

² No ageism intended: I suspect that the phenomenon actually stems from the fact that as email took off in the noughties this demographic who were significantly more-likely than younger folks to have (a) a very long-term home that they didn’t anticipate moving out of any time soon, and (b) an existing anticipation that people and companies wrote to them as a couple, not individually.

³ I’d love it if the grocery delivery sites would let multiple “accounts”, by mutual consent, share a delivery slot, destination, and payment method. It’d be cool to know that we could e.g. have a houseguest and give them temporary access to a specific order that was scheduled for during their stay. But that’s probably a lot of work for very little payoff if you’re busy running a supermarket.

Blogging Like It’s 1999

In anticipation of WWW Day on 1 August, some work colleagues and I were sharing pictures of the first (or early) websites we worked on. I was pleased to be able to pull out a screenshot of how my blog looked back in 1999!

Because I’m such a digital preservationist, many of those ancient posts are still available on my blog, so I also shared a photo of me browsing the same content on my blog as it is today, side-by-side with that 25+-year-old screenshot.¹

Dan poses in front of circa 1999 and present-day copies of his blog, both showing posts from January 1999. — The posts are in reverse-chronological order now, rather than chronological order, but the content’s all the same (even though the design is now very different and, of course, responsive!).

Update: This photo eventually appeared on a LinkedIn post on Automattic’s profile.

This inspired me to make a toggleable “alternate theme” for my blog: 1999 Mode.

Switch to it, and you’ll see a modern reinterpretation of my late-1990s blog theme, featuring:

A “table-like” layout.²
White text on a black marbled “seamless texture” background, just like you’d expect on any GeoCities page.
Pre-rendered fire text³, including – of course – animated GIFs.⁴
A (fake) hit counter.
A stack of 88×31 micro-banners, as was all the rage at the time. (And seem to be making a comeback in IndieWeb circles…)
Cursor trails (with thanks to Tim Holman)!
I’ve even applied img { image-rendering: crisp-edges; } to try to compensate for modern browsers’ capability for subpixel rendering when rescaling images: let them eat pixels!⁵

I’ve added 1999 Mode to my April Fools gags so, like this year, if you happen to visit my site on or around 1 April, there’s a change you’ll see it in 1999 mode anyway. What fun!

I think there’s a possible future blog post about Web design challenges of the 1990s. Things like: what it the user agent doesn’t support images? What if it supports GIFs, but not animated ones (some browsers would just show the first frame, so you’d want to choose your first frame appropriately)? How do I ensure that people see the right content if they skip my frameset? Which browser-specific features can I safely use, and where do I need a fallback⁶? Will this work well on all resolutions down to 640×480 (minus browser chrome)? And so on.

Any interest in that particular rabbit hole of digital history?

Footnotes

¹ Some of the addresses have changed, but from Summer 2003 onwards I’ve had a solid chain of redirects in place to try to keep content available via whatever address it was at. Because Cool URIs Don’t Change. This occasionally turns out to be useful!

² Actually, the entire theme is just a CSS change, so no tables are added. But I’ve tried to make it look like I’m using tables for layout, because that (and spacer GIFs) were all we had back in the day.

³ Obviously the title saying “Dan Q” is modern, because that wasn’t even my name back then, but this is more a reimagining of how my site would have looked if I were transported back to 1999 and made to do it all again.

⁴ I was slightly obsessed for a couple of years in the late 90s with flaming text on black marble backgrounds. The hit counter in my screenshot above – with numbers on fire – was one I made, not a third-party one; and because mine was the only one of my friends’ hosts that would let me run CGIs, my Perl script powered the hit counters for most of my friends’ sites too.

⁵ I considered, but couldn’t be bothered, implementing an SVG CSS filter: to posterize my images down to 8-bit colour, for that real “I’m on an old graphics card” feel! If anybody’s already implemented such a thing under a license that I can use, let me know and I’ll integrate it!

⁶ And what about those times where using a fallback might make things worse, like how Netscape 7 made the <blink><marquee> combination unbearable!

The Elegance of the ASCII Table

Duration 15:19

Podcast Version

This post is also available as a podcast. Listen here, download for later, or subscribe wherever you consume podcasts.

If you’ve been a programmer or programming-adjacent nerd¹ for a while, you’ll have doubtless come across an ASCII table.

An ASCII table is useful. But did you know it’s also beautiful and elegant.

ASCII‘s still very-much around; even if you’re transmitting modern Unicode³ the most-popular encoding format UTF-8 is specifically-designed to be backwards-compatible with ASCII! If you decoded this page as ASCII you’d get the gist of it… so long as you ignored the garbage characters at the end of this sentence! 😁

History

ASCII was initially standardised in X3.4-1963 (which just rolls off the tongue, doesn’t it?) which assigned meanings to 100 of the potential 128 codepoints presented by a 7-bit⁴ binary representation: that is, binary values 0000000 through 1111111:

If you’ve already guessed where I’m going with this, you might be interested to look at the X3.4-1963 table and see that yes, many of the same elegant design choices I’ll be talking about later already existed back in 1963. That’s really cool!

Table

In case you’re not yet intimately familiar with it, let’s take a look at an ASCII table. I’ve colour-coded some of the bits I think are most-beautiful:

That table only shows decimal and hexadecimal values for each character, but we’re going to need some binary too, to really appreciate some of the things that make ASCII sublime and clever.

Control codes

The first 32 “characters” (and, arguably, the final one) aren’t things that you can see, but commands sent between machines to provide additional instructions. You might be familiar with carriage return (0D) and line feed (0A) which mean “go back to the beginning of this line” and “advance to the next line”, respectively⁵. Many of the others don’t see widespread use any more – they were designed for very different kinds of computer systems than we routinely use today – but they’re all still there.

32 is a power of two, which means that you’d rightly expect these control codes to mathematically share a particular “pattern” in their binary representation with one another, distinct from the rest of the table. And they do! All of the control codes follow the pattern 00_____: that is, they begin with two zeroes. So when you’re reading 7-bit ASCII⁶, if it starts with 00, it’s a non-printing character. Otherwise it’s a printing character.

Not only does this pattern make it easy for humans to read (and, with it, makes the code less-arbitrary and more-beautiful); it also helps if you’re an ancient slow computer system comparing one bit of information at a time. In this case, you can use a decision tree to make shortcuts.

Two rolls of punched paper tape. — That there’s one exception in the control codes: *DEL* is the *last* character in the table, represented by the binary number `1111111`. This is a historical throwback to paper tape, where the keyboard would punch some permutation of seven holes to represent the ones and zeros of each character. You can’t *delete* holes once they’ve been punched, so the only way to mark a character as invalid was to rewind the tape and punch out *all* the holes in that position: i.e. all `1`s.

Space

The first printing character is space; it’s an invisible character, but it’s still one that has meaning to humans, so it’s not a control character (this sounds obvious today, but it was actually the source of some semantic argument when the ASCII standard was first being discussed).

Putting it numerically before any other printing character was a very carefully-considered and deliberate choice. The reason: sorting. For a computer to sort a list (of files, strings, or whatever) it’s easiest if it can do so numerically, using the same character conversion table as it uses for all other purposes⁷. The space character must naturally come before other characters, or else John Smith won’t appear before Johnny Five in a computer-sorted list as you’d expect him to.

Being the first printing character, space also enjoys a beautiful and memorable binary representation that a human can easily recognise: 0100000.

Numbers

The position of the Arabic numbers 0-9 is no coincidence, either. Their position means that they start with zero at the nice round binary value 0110000 (and similarly round hex value 30) and continue sequentially, giving:

Binary	Hex	Decimal digit (character)
`011 0000`	`30`	0
`011 0001`	`31`	1
`011 0010`	`32`	2
`011 0011`	`33`	3
`011 0100`	`34`	4
`011 0101`	`35`	5
`011 0110`	`36`	6
`011 0111`	`37`	7
`011 1000`	`38`	8
`011 1001`	`39`	9

The last four digits of the binary are a representation of the value of the decimal digit depicted. And the last digit of the hexadecimal representation is the decimal digit. That’s just brilliant!

If you’re using this post as a way to teach yourself to “read” binary-formatted ASCII in your head, the rule to take away here is: if it begins 011, treat the remainder as a binary representation of an actual number. You’ll probably be right: if the number you get is above 9, it’s probably some kind of punctuation instead.

Shifted Numbers

Subtract 0010000 from each of the numbers and you get the shifted numbers. The first one’s occupied by the space character already, which is a shame, but for the rest of them, the characters are what you get if you press the shift key and that number key at the same time.

“No it’s not!” I hear you cry. Okay, you’re probably right. I’m using a 105-key ISO/UK QWERTY keyboard and… only four of the nine digits 1-9 have their shifted variants properly represented in ASCII.

That, I’m afraid, is because ASCII was based not on modern computer keyboards but on the shifted positions of a Remington No. 2 mechanical typewriter – whose shifted layout was the closest compromise we could find as a standard at the time, I imagine. But hey, you got to learn something about typewriters today, if that’s any consolation.

A Remington Portable No. 3 typewriter. — Bonus fun fact: early mechanical typewriters omitted a number 1: it was expected that you’d use the letter I. That’s fine for printed work, but not much help for computer-readable data.

Letters

Like the numbers, the letters get a pattern. After the @-symbol at 1000000, the uppercase letters all begin 10, followed by the binary representation of their position in the alphabet. 1 = A = 1000001, 2 = B = 1000010, and so on up to 26 = Z = 1011010. If you can learn the numbers of the positions of the letters in the alphabet, and you can count in binary, you now know enough to be able to read any ASCII uppercase letter that’s been encoded as binary⁸.

And once you know the uppercase letters, the lowercase ones are easy too. Their position in the table means that they’re all exactly 0100000 higher than the uppercase variants; i.e. all the lowercase letters begin 11! 1 = a = 1100001, 2 = b = 1100010, and 26 = z = 1111010.

If you’re wondering why the uppercase letters come first, the answer again is sorting: also the fact that the first implementation of ASCII, which we saw above, was put together before it was certain that computer systems would need separate character codes for upper and lowercase letters (you could conceive of an alternative implementation that instead sent control codes to instruct the recipient to switch case, for example). Given the ways in which the technology is now used, I’m glad they eventually made the decision they did.

Beauty

There’s a strange and subtle charm to ASCII. Given that we all use it (or things derived from it) literally all the time in our modern lives and our everyday devices, it’s easy to think of it as just some arbitrary encoding.

But the choices made in deciding what streams of ones and zeroes would represent which characters expose a refined logic. It’s aesthetically pleasing, and littered with historical artefacts that teach us a hidden history of computing. And it’s built atop patterns that are sufficiently sophisticated to facilitate powerful processing while being coherent enough for a human to memorise, learn, and understand.

Footnotes

¹ Programming-adjacent? Yeah. For example, geocachers who’ve ever had to decode a puzzle-geocache where the coordinates were presented in binary (by which I mean: a binary representation of ASCII) are “programming-adjacent nerds” for the purposes of this discussion.

² In both the book and the film, Mark Watney divides a circle around the recovered Pathfinder lander into segments corresponding to hexadecimal digits 0 through F to allow the rotation of its camera (by operators on Earth) to transmit pairs of 4-bit words. Two 4-bit words makes an 8-bit byte that he can decode as ASCII, thereby effecting a means to re-establish communication with Earth.

³ Y’know, so that you can type all those emoji you love so much.

⁴ ASCII is often thought of as an 8-bit code, but it’s not: it’s 7-bit. That’s why virtually every ASCII message you see starts every octet with a zero. 8-bits is a convenient number for transmission purposes (thanks mostly to being a power of two), but early 8-bit systems would be far more-likely to use the 8th bit as a parity check, to help detect transmission errors. Of course, there’s also nothing to say you can’t just transmit a stream of 7-bit characters back to back!

⁵ Back when data was sent to teletype printers these two characters had a distinct different meaning, and sometimes they were so slow at returning their heads to the left-hand-side of the paper that you’d also need to send a few null bytes e.g. 0D 0A 00 00 00 00 to make sure that the print head had gotten settled into the right place before you sent more data: printers didn’t have memory buffers at this point! For compatibility with teletypes, early minicomputers followed the same carriage return plus line feed convention, even when outputting text to screens. Then to maintain backwards compatibility with those systems, the next generation of computers would also use both a carriage return and a line feed character to mean “next line”. And so, in the modern day, many computer systems (including Windows most of the time, and many Internet protocols) still continue to use the combination of a carriage return and a line feed character every time they want to say “next line”; a redundancy build for a chain of backwards-compatibility that ceased to be relevant decades ago but which remains with us forever as part of our digital heritage.

⁶ Got 8 binary digits in front of you? The first digit is probably zero. Drop it. Now you’ve got 7-bit ASCII. Sorted.

⁷ I’m hugely grateful to section 13.8 of Coded Character Sets, History and Development by Charles E. Mackenzie (1980), the entire text of which is available freely online, for helping me to understand the importance of the position of the space character within the ASCII character set. While most of what I’ve written in this blog post were things I already knew, I’d never fully grasped its significance of the space character’s location until today!

⁸ I’m sure you know this already, but in case you’re one of today’s lucky 10,000 to discover that the reason we call the majuscule and minuscule letters “uppercase” and “lowercase”, respectively, dates to 19th century printing, when moveable type would be stored in a box (a “type case”) corresponding to its character type. The “upper” case was where the capital letters would typically be stored.

Good Food, Bad Authorisation

I was browsing (BBC) Good Food today when I noticed something I’d not seen before: a “premium” recipe, available on their “app only”:

I clicked on the “premium” recipe and… it looked just like any other recipe. I guess it’s not actually restricted after all?

Just out of curiosity, I fired up a more-vanilla web browser and tried to visit the same page. Now I saw an overlay and modal attempting¹ to restrict access to the content:

It turns out their entire effort to restrict access to their premium content… is implemented in client-side JavaScript. Even when I did see the overlay and not get access to the recipe, all I needed to do was open my browser’s debugger and run document.body.classList.remove('tp-modal-open'); for(el of document.querySelectorAll('.tp-modal, .tp-backdrop')) el.remove(); and all the restrictions were lifted.

What a complete joke.

Why didn’t I even have to write my JavaScript two-liner to get past the restriction in my primary browser? Because I’m running privacy-protector Ghostery, and one of the services Ghostery blocks by-default is one called Piano. Good Food uses Piano to segment their audience in your browser, but they haven’t backed that by any, y’know, actual security so all of their content, “premium” or not, is available to anybody.

I’m guessing that Immediate Media (who bought the BBC Good Food brand a while back and have only just gotten around to stripping “BBC” out of the name) have decided that an ad-supported model isn’t working and have decided to monetise the site a little differently². Unfortunately, their attempt to differentiate premium from regular content was sufficiently half-hearted that I barely noticed that, too, gliding through the paywall without even noticing were it not for the fact that I wondered why there was a “premium” badge on some of their recipes.

Recipes probably aren’t considered a high-value target, of course. But I can tell you from experience that sometimes companies make basically this same mistake with much more-sensitive systems. The other year, for example, I discovered (and ethically disclosed) a fault in the implementation of the login forms of a major UK mobile network that meant that two-factor authentication could be bypassed entirely from the client-side.

These kinds of security mistakes are increasingly common on the Web as we train developers to think about the front-end first (and sometimes, exclusively). We need to do better.

Footnotes

¹ The fact that I could literally see the original content behind the modal was a bit of a giveaway that they’d only hidden it, not actually protected it in any way.

² I can see why they’d think that: personally, I didn’t even know there were ads on the site until I did the experiment above: turns out I was already blocking them, too, along with any anti-ad-blocking scripts that might have been running alongside.

Vmail via FreshRSS

It’s time for… Dan Shares Yet Another FreshRSS XPath Scraping Recipe!

Vmail

I’m a huge fan of the XPath scraping feature of FreshRSS, my favourite feed reader (and one of the most important applications in my digital ecosystem). I’ve previously demonstrated how to use the feature to subscribe to Forward, reruns of The Far Side, and new The Far Side content, despite none of those sites having “official” feeds.

Vmail is cool. It’s vole.wtf’s (of ARCC etc. fame) community newsletter, and it’s as batshit crazy as you’d expect if you were to get the kinds of people who enjoy that site and asked them all to chip in on a newsletter.

Totes bonkers.

But email’s not how I like to consume this kind of media. So obviously, I scraped it.

Recipe

Want to subscribe to Vmail using your own copy of FreshRSS? Here’s the settings you’re looking for –

Type of feed source: HTML + XPath (Web scraping)
XPath for finding news items: //table/tbody/tr
It’s just a table with each row being a newsletter; simple!
XPath for item title: descendant::a
XPath for item content: .
XPath for item link (URL): descendant::a/@href
XPath for item date: descendant::td[1]
Custom date/time format: d M *y
The dates are in a format that’s like 01 May ’24 – two-digit days with leading zeros, three-letter months, and a two-digit year preceded by a curly quote, separated by spaces. That curl quote screws up PHP’s date parser, so we have to give it a hint.
XPath for unique item ID: descendant::th
Optional, but each issue’s got its own unique ID already anyway; we might as well use it!
Article CSS selector on original website: #vmail
Optional, but recommended: this option lets you read the entire content of each newsletter without leaving FreshRSS.

So yeah, FreshRSS continues to be amazing. And lately it’s helped me keep on top of the amazing/crazy of vole.wtf too.

A Pressure Cooker for Tea

Duration 5:15

Podcast Version

This post is also available as a podcast. Listen here, download for later, or subscribe wherever you consume podcasts.

I’m not a tea-drinker¹. But while making a cuppa for Ruth this morning, a thought occurred to me and I can’t for a moment believe that I’m the first person to think of it:

What about a pressure-cooker, but for tea?²

Hear me out.

It’s been stressed how important it is that the water used to brew the tea is 100℃, or close to it possible. That’s the boiling point of water at sea level, so you can’t really boil your kettle hotter than that or else the water runs away to pursue a new life as a cloud.

That temperature is needed to extract the flavours, apparently³. And that’s why you can’t get a good cup of tea at high altitudes, I’m told: by the time you’re 3000 metres above sea level, water boils at around 90℃ and most British people wilt at their inability to make a decent cuppa⁴.

It’s a question of pressure, right? Increase the pressure, and you increase the boiling point, allowing water to reach a higher temperature before it stops being a liquid and starts being a gas. Sooo… let’s invent something!

I’m thinking a container about the size of a medium-sized Thermos flask or a large keep-cup – you need thick walls to hold pressure, obviously – with a safety valve and a heating element, like a tiny version of a modern pressure cooker. The top half acts as the lid, and contains a compartment into which you put your teabag or loose leaves (optionally in an infuser). After being configured from the front panel, the water gets heated to a specified temperature – which can be above the ambient boiling point of water owing to the pressurisation – at which point the tea is released from the upper half. The temperature is maintained for a specified amount of time and then the user is notified so they can release the pressure, open the top, lift out the inner cup, remove the teabag, and enjoy their beverage.

This isn’t just about filling the niche market of “dissatisfied high-altitude tea drinkers”. Such a device would also be suitable for other folks who want a controlled tea experience. You could have it run on a timer and make you tea at a particular time, like a teasmade. You can set the temperature lower for a controlled brew of e.g. green tea at 70℃. But there’s one other question that a device like this might have the capacity to answer:

What is the ideal temperature for making black tea?

We’re told that it’s 100℃, but that’s probably an assumption based on the fact that that’s as hot as your kettle can get water to go, on account of physics. But if tea is bad when it’s brewed at 90℃ and good when it’s brewed at 100℃… maybe it’s even better when it’s brewed at 110℃!

A modern pressure cooker can easily maintain a liquid water temperature of 120℃, enabling excellent extraction of flavour into water (this is why a pressure cooker makes such excellent stock).

A mug of tea held by the handle. — It’s possible that the perfect cup of tea *hasn’t been invented yet*, owing to limitations in the boiling point of water.

I’m not the person to answer this question, because, as I said: I’m not a tea drinker. But surely somebody’s tried this⁵? It shouldn’t be too hard to retrofit a pressure cooker lid with a sealed compartment that releases, even if it’s just on a timer, to deposit some tea into some superheated water?

Because maybe, just maybe, superheated water makes better tea. And if so, there’s a possible market for my proposed device.

Footnotes

¹ I probably ought to be careful confessing to that or they’ll strip my British citizenship.

² Don’t worry, I know better than to suggest air-frying a cup of ta. What kind of nutter would do that?

³ Again, please not that I’m not a tea-drinker so I’m not really qualified to comment on the flavour of tea at all, let alone tea that’s been brewed at too-low a temperature.

⁴ Some high-altitude tea drinkers swear by switching from black tea to green tea, white tea, or oolong, which apparently release their aromatics at lower temperatures. But it feels like science, not compromise, ought to be the solution to this problem.

⁵ I can’t find the person who’s already tried this, if they exist, but maybe they’re out there somewhere?

Building a Secret Cabinet

We’ve recently had the attics of our house converted, and I moved my bedroom up to one of the newly-constructed rooms.

To make the space my own, I did a little light carpentry up there: starting with a necessary reshaping of the doors, then moving on to shelving and eventually… a secret cabinet!

I’d love to tell you about how I built it: but first, a disclaimer! I am a software engineer, and with good reason. Letting me near a soldering iron is ill-advised. Letting me use a table saw is tempting fate.

Letting me teach you anything about how you should use a soldering iron or a table saw is, frankly, asking for trouble.

A spirit level on an unfinished shelf, under a window and beneath an uncarpeted floor. — Knowing that I’d been short on shelf space in my old bedroom, I started work on fitting shelves for my new bedroom before the carpet had even arrived.

Building a secret cabinet wasn’t part of my plan, but came about naturally after I got started. I’d bought a stack of pine planks and – making use of Ruth’s table saw – cut them to squarely fit beneath each of the two dormer windows¹. While sanding and oiling the wood I realised that I had quite a selection of similarly-sized offcuts and found myself wondering if I could find a use for them.

Dan drinks a 0% alcohol beer in front of three upright planks of wood. — The hardest part of sanding and oiling wood on the hottest day of the year is all the beer breaks you have to take. Such a drag.

I figured I had enough lumber to insert a small cabinet into one of the bookshelves, and that got me thinking… what about if it were a secret cabinet, disguised as books unless you knew where to look. Or to go one step further: what if it had some kind of electronic locking mechanism that could be triggered from somewhere else in the room².

Magpie decal 'perched' on a light switch. — There are other ways in which I’ve made my new room distinctly-“mine” – like the pair of magpies – but probably the secret cabinet is the most-distinctive.

Not wanting to destroy a stack of real books, which is the traditional way to get a collection of book spines for the purpose of decorating a “fake bookshelf” panel³, I looked online and discovered the company that made the fake book spines used at the shop of my former employer. They looked ideal: carefully shaped and painted panels with either an old-school or contemporary look.

Buuut, they don’t seem to be well-equipped for short runs and are doubtless pricey, so I looked elsewhere and found the eBay presence of Beatty Lockey Antiques in Loewstof. They’d acquired a stack of them second-hand from the set of Netflix’s The School for Good and Evil.⁴

(By the way: at time of writing they’ve still got a few panels left, if you want to make your own…)

I absolutely must sing the praises of Brad at Beatty Lockey Antiques who, after the first delivery of fake book fronts was partially-damaged in transit, was super quick about helping me find the closest-available equivalent (I’d already measured-up based on the one I’d thought I was getting) and sent a replacement.

The cabinet is just a few bits of wood glued together and reinforced with L-shaped corner braces, with a trio of thin strips – made from leftover architrave board – attached using small brass hinges. The fake book fronts are stuck to the strips using double-sided mounting tape left over from installing a bathroom mirror. A simple magnetic clasp holds the door shut when pushed closed⁵, and the hinges are inclined to “want” the door to stand half-open, which means it only needs a gentle push away from the magnetic catch to swing it open.

I mounted a Raspberry Pi Zero W into a rear corner inside the cabinet⁶, and wired it up via a relay to what was sold to me as a “large push-pull solenoid”, then began experimenting with the position in which I’d need to mount it to allow it to “kick” open the door, against the force of the magnetic clasp⁷.

This was, amazingly, the hardest part of the whole project! Putting the solenoid too close to the door didn’t work: it couldn’t “push” it from a standing start. Too far away, and the natural give of the door took the strain without pushing it open. Just the right distance, and the latch had picked up enough momentum that its weight “kicked” the door away from the magnet and followed-through to ensure that it kept moving.

A second solenoid, mounted inside the top of the cabinet, slides into the “loop” part of a large bolt fitting, allowing the cabinet to be electronically “locked”.

A Raspberry Pi Zero, relay, and solenoid assembly on the bottom outside edge of the inside of the cabinet. — I seriously must’ve spent about an hour getting the position of that little “kicker” in the bottom right *just right*.

Next up came the software. I started with a very simple Python program⁸ that would run a webserver and, on particular requests, open the lock solenoid and push with the “kicker” solenoid.

#!/usr/bin/python
#
# a basic sample implementation of a web interface for a secret cabinet
#
# setup:
#   sudo apt install -y python3-flask
#   wget https://github.com/sbcshop/Zero-Relay/blob/master/pizero_2relay.py
#
# running:
#   sudo flask --app web run --host=0.0.0.0 --port 80

from flask import Flask, redirect, url_for
import pizero_2relay as pizero
from time import sleep

# set up pizero_2relay with the two relays attached to this Pi Zero:
r1 = pizero.relay("R1") # The "kicker" relay
r2 = pizero.relay("R2") # The "locking bolt" relay

app = Flask(__name__)

# GET / - nothing here
@app.route("/")
def index():
  return "Nothing to see here."

# GET /relay - show a page with "open" and "lock" links
@app.route("/relay")
def relay():
  return "<html><head><meta name='viewport' content='width=device-width, initial-scale=1'></head><body><ul><li><a href='/relay/open'>Open</a></li><li><a href='/relay/lock'>Lock</a></li></ul>"

# GET /relay/open - open the secret cabinet then return to /relay
# This ought to be a POST request in your implementation, and you probably
# want to add some security e.g. a 
@app.route("/relay/open")
def open():
  # Retract the lock:
  r2.off()
  sleep(0.5)
  # Fire the kicker twice:
  r1.on()
  sleep(0.25)
  r1.off()
  sleep(0.25)
  r1.on()
  sleep(0.25)
  r1.off()
  # Redirect back:
  return redirect(url_for('relay'))

@app.route("/relay/lock")
def lock():
  # Engage the lock:
  r2.on()
  return redirect(url_for('relay'))

Don’t use this code as-is on any kind of open network, obviously. Follow the comments for some tips on what you’ll need to change.

Once I had something I could trigger from a web browser or with curl, I could start experimenting with trigger mechanisms. I had a few ideas (and prototyped a couple of them), including:

A mercury tilt switch behind a different book, so you pull it to release the cabinet in the style of a classic movie secret door.
A microphone that listens for a specific pattern of knocks on a nearby surface.

I had far too much fun playing about with crappy prototypes.

An RFID reader mounted underneath another surface, and a tag on the underside of an ornament: moving the ornament to the “right” place on the surface triggers the cabinet (elsewhere in the room).
The current design, shown in the video above, where a code⁹ is transmitted to the cabinet for verification.

I think I’m happy with what I’ve got going on with it now. And it’s been a good opportunity to improve my carpentry, electronics, and Python.

Footnotes

¹ The two dormer windows, wouldn’t you guarantee it, were significantly different widths despite each housing a window of the same width. Such are the quirks of extending a building that the previous occupier had previously half-heartedly tried to extend, I guess.

² Why yes, I am a big fan of escape rooms. Why do you ask?

³ For one thing, I live with JTA, and I’m confident that he’d somehow be able to hear the silent screams of whatever trashy novels I opted to sacrifice for the good of the project.

⁴ As a bonus, my 10-year-old is a big fan of the book series that inspired the film (and a more-muted fan of the film itself) and she was ever-so excited at my project using real-life parts of the set of the movie… that she’s asked me to make a similar secret cabinet for her, when we get around to redecorating her room later in the year!

⁵ If I did it again, I might consider using a low-powered electromagnetic lock to hold the door shut. In this design, I used a permanent magnet and a pair of latch solenoids: one to operate a bolt, the second to “kick” the door open against the pull of the magnet, and… it feels a little clumsier than a magnetic lock might’ve.

⁶ That double-sided mounting tape really came in handy for this project!

⁷ Props to vlogger Technology Connections, one of whose excellent videos on the functionality of 1970s pinball tables – maybe this one? – taught me what a latch solenoid was in the first place, last year, which probably saved me from the embarrassment of trying to do this kind of thing with, I don’t know, a stepper motor or something.

⁸ I’m not a big fan of Python normally, but the people who made my relays had some up with a convenience library for them that was written in it, so I figured it would do.

⁹ Obviously the code isn’t A-B; I changed it temporarily for the video.

Ladybird Browser

I’ve been playing with the (pre-Alpha version of) Ladybird, and it fills me with such joy and excitement.

Browser diversity

Back in 2018, while other Web developers were celebrating, I expressed my dismay at the news that Microsoft Edge was on the cusp of switching from using Microsoft’s own browser engine EdgeHTML to using Blink. Blink is the engine that powers almost all other mainstream browsers; all but Firefox, which continues to stand atop Gecko.

The developers who celebrated this loss of rendering engine diversity were, I suppose, happy to have one fewer browser in which they must necessarily test their work. I guess these are the same developers who don’t test the sites they develop for accessibility (does your site work if you can’t see the images? what about with a keyboard but without a pointing device? how about if you’re colourblind?), or consider what might happen if a part of their site fails (what if the third-party CDN that hosts your JavaScript libraries goes down or is blocked by the user’s security software or their ISP?).

But I was sad, because – as I observed after Andre Alves Garzia succinctly spelled it out – browser engines are an endangered species. Building a new browser that supports the myriad complexities of the modern Web is such a huge endeavour that it’s unlikely to occur from scratch: from this point on, all “new” browsers are likely to be based upon an existing browser engine.

Engine diversity is important. Last time we had a lull in engine diversity, the Web got stuck, stagnating in the shadow of Internet Explorer 6’s dominance and under the thumb of Microsoft’s interests. I don’t want those days to come back; that’s a big part of why Firefox is my primary web browser.

A Ladybird book browser

Spoof cover for "The Ladybird Book of The Browser" — I actually still own a copy of the book from which I adapted this cover!

Ladybird is a genuine new browser engine. Y’know, that thing I said that we might never see happen again! So how’ve they made it happen?

It helps that it’s not quite starting from scratch. It’s starting point is the HTML viewer component from SerenityOS. And… it’s pretty good. It’s DOM processing’s solid, it seems to support enough JavaScript and CSS that the modern Web is usable, even if it’s not beautiful 100% of the time.

They’re not even expecting to make an Alpha release until next year! Right now if you want to use it at all, you’re going to need to compile the code for yourself and fight with a plethora of bugs, but it works and that, all by itself, is really exciting.

They’ve got four full-time engineers, funded off donations, with three more expected to join, plus a stack of volunteer contributors on Github. I’ve raised my first issue against the repo; sadly my C++ probably isn’t strong enough to be able to help more-directly, even if I somehow did have enough free time, which I don’t. But I’ll be watching-from-afar this wonderful, ambitious, and ideologically-sound initiative.

#100DaysToOffload

Woop! This is my 100th post of the year (stats), even using my more-conservative/pedant-friendly “don’t count checkins/reposts/etc. rule. If you’re not a pedant, I achieved #100DaysToOffload when I found a geocache alongside Regents Canal while changing trains to go to Amsterdam where I played games with my new work team, looked at windows and learned about how they’ve been taxed, and got nerdsniped by a bus depot. In any case: whether you’re a pedant or not you’ve got to agree I’ve achieved it now!

Science Weekend

This weekend was full of science.

Research

String of large electromagnets used for steering an electron beam. — If you wanna bend a stream of electrons travelling at nearly the speed of light, you’re gonna need a lot of big magnets.

This started on Saturday with a trip to the Harwell Campus, whose first open day in eight years provided a rare opportunity for us to get up close with cutting edge science (plus some very kid-friendly and accessible displays) as well as visit the synchrotron at Diamond Light Source.

Dan with a child in front of beamlines coming out of the Diamond Light Synchrotron ring. — It’s hard to convey the scale of the thing; turns out you need a big ol’ ring if you want to spin electrons fast enough to generate a meaningful amount of magnetobremsstrahlung radiation.

The whole thing’s highly-recommended if you’re able to get to one of their open days in the future, give it a look. I was particularly pleased to see how enthused about science it made the kids, and what clever questions they asked.

For example: the 7-year-old spent a long time cracking a variety of ciphers in the computing tent (and even spotted a flaw in one of the challenge questions that the exhibitors then had to hand-correct on all their handouts!); the 10-year-old enjoyed quizzing a researcher who’d been using x-ray crystallography of proteins.

Medicine

And then on Sunday I finally got a long-overdue visit to my nearest spirometry specialist for a suite of experiments to try to work out what exactly is wrong with my lungs, which continue to be a minor medical mystery.

Dan holds a piece of medical apparatus to his mouth. — “Once you’ve got your breath back, let’s fill you with drugs and do those experiments again…”

It was… surprisingly knackering. Though perhaps that’s mostly because once I was full of drugs I felt briefly superpowered and went running around the grounds of the wonderfully-named Brill Hill Windmill with the dog until was panting in pretty much the way that I might normally have been, absent an unusually-high dose of medication.

For amusement purposes alone, I’d be more-likely to recommend the first day’s science activities than the second, but I can’t deny that it’s cool to collect a load of data about your own body and how it works in a monitorable, replicable way. And maybe, just maybe, start to get to the bottom of why my breathing’s getting so much worse these last few years!

A Stupid Joke About Elephants

Duration 2:48

Podcast Version

This post is also available as a podcast. Listen here, download for later, or subscribe wherever you consume podcasts.

You’ve probably been taught that you can tell the difference between African Elephants and Indian Elephants by looking at their head and ears. The larger African Elephants have a rounded cranium and big ears (with a shape somewhat like the continent of Africa itself!), whereas the smaller Indian Elephants have a two-lobed skull and diminutive ears that tuck tidily alongside their heads.

But suppose you don’t manage to get a glimpse at the front end of the elephant as it passes you. What hope is there of identifying the species? Well: you can look at its back!

African Elephants, it turns out, have a concave back, whereas Indian elephants have a convex back (a bit like a hump)!

I was having difficulty sleeping one night during the UK‘s current heatwave, so naturally I opted to practice my newfound ability to distinguish elephant species by their spines. Indian, Indian, African, Indian, African, African… etc.

And then I came across this one:

African Elephant backs are concave. Indian Elephant backs are convex. But what does it mean when you see a flat elephant’s back?

It turns out…

…

…that’s a grey area.

Special Roads

Duration 7:19

Podcast Version

This post is also available as a podcast. Listen here, download for later, or subscribe wherever you consume podcasts.

Sometimes I’ve seen signs on dual carriageways and motorways that seem to specify a speed limit that’s the same as the national speed limit (i.e. 60 or 70 mph for most vehicles, depending on the type of road), which seem a bit… pointless? Today I learned why they’re there, and figured I’d share with you!

To get there, we need a history lesson.

As early as the 1930s, it was becoming clear that Britain might one day need a network of high-speed, motor-vehicle-only roads: motorways. The first experimental part of this network would be the Preston By-pass¹.

Construction wouldn’t actually begin until the 1950s, and it wasn’t just the Second World War that got in the way: there was a legislative challenge too.

When the Preston By-pass was first conceived, there was no legal recognition for roads that restricted the types of traffic that were permitted to drive on them. If a public highway were built, it would have to allow pedestrians, cyclists, and equestrians, which would doubtless undermine the point of the exercise! Before it could be built, the government needed to pass the Special Roads Act 1949, which enabled the designation of public roads as “special roads”, to which entry could be limited to certain classes of vehicles².

Monochrome photograph showing a sign at the entrance to the Preston By-pass, reading: 'Motorway. NO L-drivers, mopeds, motorcycles under 50cc., invalid-carriages, pedal-cycles, pedestrians, animals'. — The original motorways had to spell out the regulations at their junctions.

If you don’t check your sources carefully when you research the history of special roads, you might be taken in by articles that state that special roads are “now known as motorways”, which isn’t quite true. All motorways are special roads, by definition, but not all special roads are motorways.

There’s maybe a dozen or more non-motorway special roads, based on research by Pathetic Motorways (whose site was amazingly informative on this entire subject). They tend to be used in places where something is like a motorway, but can’t quite be a motorway. In Manchester, a couple of the A57(M)’s sliproads have pedestrian crossings and so have to be designated special roads rather than motorways, for example³.

Now we know what special roads are, that we might find them all over the place, and that they can superficially look like motorways, let’s talk about speed limits.

The Road Traffic Act 1934 introduced the concept of a 30mph “national speed limit” in built-up areas, which is still in force today. But outside of urban areas there was no speed limit. Perhaps there didn’t need to be, while cars were still relatively slow, but automobiles became increasingly powerful. The fastest speed ever legally achieved on a British motorway came in 1964 during a test by AC Cars, when driver Jack Sears reached 185mph.

In the late 1960s an experiment was run in setting a speed limit on motorways of 70mph. Then the experiment was extended. Then the regulation was made permanent.

There’ve been changes since then, e.g. to prohibit HGVs from going faster than 60mph, but fundamentally this is where Britain’s national speed limit on motorways comes from.

You’ve probably spotted the quirk already. When “special roads” were created, they didn’t have a speed limit. Some “special roads” were categorised as “motorways”, and “motorways” later had a speed limit imposed. But there are still a few non-motorway “special roads”!

Putting a national speed limit sign on a special road would be meaningless, because these roads have no centrally-legislated speed limit. So they need a speed limit sign, even if that sign, confusingly, might specify a speed limit that matches what you’d have expected on such a road⁴. That’s the (usual) reason why you sometimes see these surprising signs.

As to why this kind of road are much more-common in Scotland and Wales than they are anywhere else in the UK: that’s a much deeper-dive that I’ll leave as an exercise for the reader.

Footnotes

¹ The Preston By-pass lives on, broadly speaking, as the M6 junctions 29 through 32.

² There’s little to stop a local authority using the powers of the Special Roads Act and its successors to declare a special road accessible to some strange and exotic permutation of vehicle classes if they really wanted: e.g. a road could be designated for cyclists and horses but forbidden to motor vehicles and pedestrians, for example! (I’m moderately confident this has never happened.)

³ There’s a statutory instrument that makes those Mancunian sliproads possible, if you’re having trouble getting to sleep on a night and need some incredibly dry reading.

⁴ An interesting side-effect of these roads might be that speed restrictions based on the class of your vehicle and the type of road, e.g. 60mph for lorries on motorways, might not be enforceable on special roads. If you wanna try driving your lorry at 70mph on a motorway-like special road with “70” signs, though, you should do your own research first; don’t rely on some idiot from the Internet. I Am Not A Lawyer etc. etc.

ARCC

In the late ’70s, a shadowy group of British technologists concluded that nuclear war was inevitable and secretly started work on a cutting-edge system designed to help rebuild society. And thanks to Matt Round-and-friends at vole.wtf (who I might have mentioned before), the system they created – ARCC – can now be emulated in your browser.

3D rendering of an ARCC system, by HappyToast.

I’ve been playing with it on-and-off all year, and I’ve (finally) managed to finish exploring pretty-much everything the platform currently has to offer, which makes it pretty damn good value for money for the £6.52 I paid for my ticket (the price started at £2.56 and increases by 2p for every ticket sold). But you can get it cheaper than I did if you score 25+ on one of the emulated games.

Most of what I just told you is true. Everything… except the premise. There never was a secretive cabal of engineers who made this whackballs computer system. What vole.wtf emulates is an imaginary system, and playing with that system is like stepping into a bizarre alternate timeline or a weird world. Over several separate days of visits you’ll explore more and more of a beautifully-realised fiction that draws from retrocomputing, Cold War fearmongering, early multi-user networks with dumb terminal interfaces, and aesthetics that straddle the tripoint between VHS, Teletext, and BBS systems. Oh yeah, and it’s also a lot like being in a cult.

Needless to say, therefore, it presses all the right buttons for me.

If you enjoy any of those things, maybe you’d like this too. I can’t begin to explain the amount of work that’s gone into it. If you’re looking for anything more-specific in a recommendation, suffice to say: this is a piece of art worth seeing.