Remnants of a legendary typeface have been rescued from the Thames | Hacker News Reader

Remnants of a legendary typeface have been rescued from the Thames

news.artnet.com

433 points_emacsomancer_88 comments

Doves is insanely easy on the eyes despite so much going on. There is also mebinac[1] an unauthorized contemporary take on the original doves. Mebinac doesn't leap off the page as well yet deals with modern punctuation in a more normal way.

Personally you can freely use them to great affect in your RSS reader or mail app that you read everyday.

[1] https://fontsme.com/mebinac.font

I tried looking at code in Mebinac, and was surprised at how strongly it reminded me of old screenshots of Smalltalk.

This font is beautiful, thanks for sharing.

Thanks!, hadn’t come across Mebinac. It’s quite good!

I’m also a big fan of Igino Marini’s recreation of the Fell typefaces:

The Fell Types took their name from John Fell, a Bishop of Oxford in the seventeenth-century. Not only he created an unique collection of printing types but he started one of the most important adventures in the history of typography. — https://web.archive.org/web/20240128075552/https://iginomari...

The IM Fell fonts themselves seem to live on Google Fonts these days: https://fonts.google.com/?query=Igino+Marini

I use Doves Type for… everything. One day I started to find my monomaniacal obsession a bit funny and sort of to spite myself I set every font in Firefox to Doves Type. Serif, sans-serif, monospace, no other fonts allowed, as well as the UI font by tweaking the Firefox user profile iirc.

And it was just… very good. And I kept using it.

I use Doves Type for everything, and to be able to do that on my phone I use iFont: https://apps.apple.com/is/app/ifont-find-install-any-font/id...

Or yeah I do use IBM PC VGA 9x16, IBM BIOS 8x8, and Eagle Spirit PC CGA Board Alternate 3 a little :) From the Ultimate Oldschool PC Font Pack: https://int10h.org/oldschool-pc-fonts/

I even munged together a combination of Doves Type Regular and IM Fell Great Primer Italic that matches the character scale and linespacing to both each other and to the IBM PC VGA 9x16 font at 1:1 size. The open-source FontForge did the trick!: https://fontforge.org/en-US/

(FontForge can autogenerate italics for any font. If you’re bored, I suggest loading up the classic VGA font and pressing the ITALICIZE button on ot. It’s… interesting!)

In general, on Windows I much prefer MacType’s fomt rendering: https://www.mactype.net … it’s kind of amazing that this kind of surgery is even possible.

Curious what you did in FontForge to merge IM Fell Great Primer Italics into Doves Type Regular. (As I'd very much like to use Doves, e.g., for an e-reader font, but I do want to have italics for such purposes.)

I made a native attempt in FontForge (just doing 'merge fonts'), which (unsurprisingly) didn't work.

It’s a bit of a blackout when I try to recall it, haha. I should figure it out and write it up.

If it’s useful: As far as I can recall it involved simply changing the font family to match, i.e. “Dovesfell”, and then exporting the regular and the italic. The OS font system then figures out that they belong together.

The scale is slightly different and the linespacing too. Did like a 90% rescale on one and 95% on the other? And then there was something to change in the Metrics window to make the linespacing identical.

I've become enamored of this typeface as well as of this morning, moved even if I'm honest, but I'm having trouble finding one that looks as nice as the bible page sample in the OP. I made this account here just to ask; which version have you been using, and where did you get it?

I bought it from Typespec :)

When you are looking for a font, how can you know which grapheme clusters have glyphs? Is there some classification system for fonts that let you know how complete they are?

https://wakamaifondue.com/

A discussion of other digitisations: https://web.archive.org/web/20220705070449/http://luc.devroy...

On "unauthorized" -- how would this not be public domain at this point?

Unauthorized doesn't mean it isn't in the public domain. It means there was no authorization.

Certainly -- I'm just not sure who would "authorize" it in the first place.

Interestingly, the font seems to be copyrighted because it's code...?

So if you see someone's web font, can you print it out and then redigitize it into a font and bypass copyright that way...?

As I understand, yes: but that'll be inferior to the original, since it won't have ligatures or hinting or anything. It's a lot of work for not a lot of benefit.

Copyright applies to the NAMES as trademarks, and the exact program used to reproduce them, but not to the shapes and designs, and yes, I made a lot of changes to that article to reflect current law.

I meant unauthorized in the autobiographical approval sense, the original doves creator frowned upon any modern takes of his typeface, the new designer acknowledged this and let us know.

Related. Others? I think there were others.

The lost Doves Press typeface and its revival (2015) - https://news.ycombinator.com/item?id=20791125 - Aug 2019 (9 comments)

How the Doves Type Was Nearly Lost - https://news.ycombinator.com/item?id=12476579 - Sept 2016 (44 comments)

One man's obsession with rediscovering the Doves typeface - https://news.ycombinator.com/item?id=9951869 - July 2015 (32 comments)

Lost typeface printing blocks found in river Thames - https://news.ycombinator.com/item?id=9017307 - Feb 2015 (22 comments)

The fight over the Doves: A legendary typeface gets a second life - https://news.ycombinator.com/item?id=6964013 - Dec 2013 (12 comments)

There were many more, but these were all dang could find after pg threw them in the river.

I spittaked coffee all over my monitor just now. I know this sounds like a Reddit comment, but I couldn't help but big-up your joke. Well played.

Now that was funny. Really.

The "modernized version", available as a font file, was modernized too much.[1] It doesn't look period.

The H.P. Lovecraft Society has some 19th century fonts, if you need them.[2] Those were recovered from old documents.

[1] https://typespec.co.uk/doves-type/

[2] https://www.hplhs.org/resources.php

I'm curious what you mean by not looking "period". The HPLHS fonts frankly seem to just be poor quality, rather than old. If you look at the images of the original type, Doves appears to be quite faithful to the original. Perhaps it's worth noting that we still use typefaces remarkably similar to the Romans, particularly Times New Roman, which despite its many shortcomings retains a "modern" look by virtue of still being in use.

They are intended to be of historically appropriate quality, for use in creating period versimilitude:

> Many of these fonts have slightly rough edges or irregular shapes, to capture the feel of old lead type and bygone printing technologies

Not all documents from the time period would've had such low quality though, and not everyone would want such quality in a modern document. If you want such an effect, it's always possible to add it later, but it's rather more difficult to remove it if it's baked into the font file.

Roman type has roots in Italian printing of the late 15th and early 16th centuries, but Times New Roman's design has no connection to Rome or to the Romans.

https://en.wikipedia.org/wiki/Times_New_Roman

I'll admit I'm no typeface expert, but this seems to miss the point. Wikipedia's own page on Roman type [0] says "Roman type was modelled from a European scribal manuscript style of the 15th century, based on the pairing of inscriptional capitals used in ancient Rome with Carolingian minuscules". And visually, there's clearly an influence, though many centuries removed. My point is merely these very old typefaces remain modern looking because we still use similar ones today.

[0] https://en.wikipedia.org/wiki/Roman_type

The capital letters were indeed inspired by Roman monumental inscriptions. But all the lower case forms were taken from Carolingian designs. Humanists wanted to copy Roman forms to go back to what they saw as writing uncontaminated with medieval influence, but the texts of Roman authors they used to do so were not actually written by Romans but copied by Carolingian-era scribes. It's why its generally much easier for us to read ninth-century texts than, say, earlier (e.g., Merovingian chancery script, yikes) and later scribal hands (e.g., late medieval Gothic).

"If my grandmother had wheels, she could have been a bicycle." Serif type is based on the use of chisels to carve rock. ANY other semblance is purely speculative. The trademark for Times New Roman is owned by the British, not the Italians.

It was created by the descendants of the Romans, in the same physical location as Ancient Rome, and based on the numerous examples of letters that were still around on Roman buildings.

If that is “no connection” what exactly would a “connection” look like?

Look up the history of how Stanley Morison and Victor Lardent created Times New Roman for _The Times_.

The connection you are looking for is covered by Fra. Edward Catich in his books, and carried forward digitally in Carol Twombly's Trajan.

This has so much of what I (as an outsider) love about the UK. The love of typography & general design chops, mudlarks, art and design in public life, the spirit of enquiry and adventure and, the presence of people in the bureaucracy and elsewhere who recognize whimsy and put institutional resources behind that pursuit.

There was also a revival of the Doves type made by Torbjörn Olsson in 1994. It is no longer available, but you can find the old specimen PDF at the Internet Archive and extract the embedded fonts. The weight is a bit lighter than the Robert Green version, but also has an italic face.

https://web.archive.org/web/20121127135748/home1.swipnet.se/...

> “It is not that unusual to find pieces of type in the river,” Sandy said. “Particularly around Fleet Street, where newspaper typesetters would throw pieces in the water when they couldn’t be bothered to put them back in their cases.

Some assistant being lazy, or rushing to "finish" a task?

Or sorts that broke, or were worn out, and it was normal to toss things into the river?

Or a ritual? (Say, toss a sort into the river for the first page an apprentice sets, or when there's a press failure, or for superstition after printing very bad news?)

> it was normal to toss things into the river?

It was normal. Rivers were used for dumping the garbage. In some places they still are. I know about instances in Europe where people dump their trash in streams behind the hamlet.

Semi related - the UK pouring sewage into its waterways has been front page news of late. It’s up to 3.6 million hours of sewage discharge per year.

https://www.bbc.com/news/explainers-62631320

Franch put a lot of effort into rerouting old swears that dumped into the seine ahead of the up coming Olympics so they could hold events in the river

When you say the UK, what you mean is a group of corrupt private companies that find it more convenient and cost friendly to dump raw sewage rather than correctly process it. I'm fairly sure the majority of people in the UK would be in favour of nationalising such companies and instead dumping their executives into the river instead [1].

[1] https://www.youtube.com/watch?v=7V-Nwk9E1bs

Such as...in the Danube river O_O

https://phys.org/news/2020-09-brown-danube-belgrade-sewers-t...

One notable instance was “Bubbly Creek” in Chicago where the slaughterhouses dumped so much refuse in a tributary to the Chicago River that the water bubbled from the decay of the trash. The riverbed there is still polluted with toxic chemicals.

It's STILL normal for people to litter and toss junk in rivers and culverts

A bit hard to belive they would toss types in the water out of laziness given these types would be relatively expensive and the typesetters would need them to keep their job.

The particular case seem to be a deliberate act of sabotage which sound more believable.

Throwing worn-out or broken types and other tools in the river would probably be common though.

The recovery of the Doves typeface from the Thames was discussed on HN in 2015, so this story goes way back.

https://news.ycombinator.com/item?id=9017307

Yes, very old news.

hek free

Added to https://news.ycombinator.com/item?id=40271786. Thanks!

Thames New Roman

You win the Internet today :-)

Whoever downvoted this has no soul.

  >pun

  >you win the internet

  >downvote

It is against the rules to tell you which rule you are breaking. hint hint.

Don't disrespect the downvote-a-bot !

What's you expecting a soul for?

I came looking for this and was not disappointed. (Though I also appreciate that the top comment is more thoughtful, a credit to HN's priorities.)

We've had centuries of embankment works along the Thames¹, a fair bit concentrated around the areas you'd expect to find type like this². There must be a phenomenal amount of history that was purposely covered around there. Given the scale of the works you'd have to imagine there is a significant chunk of non-London history to be found there too(the scale of granite imports from Cornwall being an obvious example).

I'm less optimistic about the possibility of more large scale digs though, as the Golden Jubilee bridge history³ points out the area is an also an exciting zone for stumbling in to unexploded ordnance and you always seem to be within few metres of a tube line or Victorian sewer.

[It is the reason I love those plucky Crossrail⁴ developers who've felt the anger from the havoc they've left across London over the few past decades. We get incredible large scale engineering works to lust over, coupled with really wacky archaeological digs tagging along for the ride.]

¹ https://en.wikipedia.org/wiki/Embanking_of_the_tidal_Thames

² https://en.wikipedia.org/wiki/Thames_Embankment - Both the "home" of the type in Hammersmith and Fleet were the targets of embankment work in the 19th century

³ https://en.wikipedia.org/wiki/Hungerford_Bridge_and_Golden_J...

⁴ https://en.wikipedia.org/wiki/Crossrail

On a trip to London and having heard of mudlarking, I walked in one of the ‘beaches’. I immediately found an old belt belt buckle and about 20 stems from old clay pipes.

My father found a 17th century cork screw.

There must be an absolute wealth of finds along its banks.

Curious as to why this refers to recovering the type being important to creating a digital version of the typeface, when lower in the article it shows that there is a surviving bible. Couldn't that have already been used to reproduce the font?

Due to irregular spreading of the ink when printing, the shapes on the page are not perfect representations of the type shape, so the true shape of the metal form has to be inferred from comparing multiple printed samples.

There are digital reproductions of old typefaces that try to reproduce the actual weight on the page, but they seem to be not very popular with modern designers unless they are going for a deliberately archaic look.

I'd expect the original designers to know and consider that effect, and make the types slightly thiner, so the printed version looks as intended.

They did and that is a huge point of contention in the revival of classic typefaces. In the 1970s, there was a massive push to digitize existing founts, but the type companies did it by tracing the metal rather than the prints. The result was digital fonts that printed much lighter than the original metal type. Most digitizations of early 20th-century typefaces you can find have this problem.

By the late 1970s, people began to pay more heed to the actual printed shapes. I like early 20th-century typographic style and am always on the lookout for good type reproductions, but there are two other factors that come into play. One is that a font designed to look a certain way when press-printed won't look quite the same coming out of your laser printer. The other is that modern taste is for thinner lines. When I use a revival of a classic type, I want it to look at it did back when, but apparently I'm in the minority.

The worst symptom of this imo is the inclusion of ink traps in digital fonts. I think they look awful, especially when blown up to poster size.

The problem is not the ink traps but the lack of optical scaling.

most serious fonts have separate "display" variants for such applications, whereas the version accounting for ink spread is called "text" or "book".

It’s also a matter of printing technology, letterpress, vs offset. The latter tends to have less ink spread. It’s also a matter of printing on dry vs damp paper (letterpress works best on slightly dampened sheets of paper which contributes to the ink spread). Then there are things like subtle curves that don’t digitize well (digital Optima is a poor approximation of the original analog letterforms).

This question on Graphic Design StackExchange is very informative: https://graphicdesign.stackexchange.com/questions/142622/pri...

We actually have this. Obviously not this particular font. My family were all printers and I've sort of inherited a huge cabinet full of old school typefaces all carved out of some special kind of hard wood - pear wood - all over 100 years old. Absolutely 0 idea what we can do with it, but it's all hand made and very cool. Felt pertinet to share lol

Maybe drop Robert Green (the man behind this article) an email on: https://typespec.co.uk/custom-font-services/

Klim Type Foundry [1] may also be worth a contact — they’ve been inspired by woodcut type before (e.g. [2] [3]), so I wonder if they might be interested in knowing about this.

[1] https://klim.co.nz/

[2] https://klim.co.nz/blog/maelstrom-design-information/

[3] https://klim.co.nz/blog/manuka-design-information/

Interesting! While I'm no Kris Sowersby or Robert Green, I have a side-side project to carve my own wood type (I've got a stack of Beech rounds drying in the shed) when I finally get a chance. It'd be cool to see some photos of your collection. If you're in the States, places like Hamilton Type and Hatch Show Print have tonnes of old an quirky display faces that they still use regularly.

I remember the earlier story about the disposal and Robert Green's obsession with reviving it back in 2013 in The Economist[1]—at that time, "Intrepid fans have occasionally tried to recover pieces of the type from the river, but no one has ever found any"—so it's good to hear that the story didn't end there.

[1]: https://www.economist.com/christmas-specials/2013/12/19/the-... (paywalled; https://archive.is/XfK1x)

PSA for the inspiration for this font, the great Nicolas Jenson, who around 1470 had pretty much perfected the latin typeface.

Later, more famous types, such as Caslon or Garamond, are just variations on this.

Not really. Jenson’s typeface (modern revivals include Montype’s Centaur and Adobe Jenson) retained a lot of calligraphic features which figure less in later faces. Garamond is much more of a type design than lettering in type and Caslon is wholly typographic in its nature.

You are technically correct. Even though that's the best kind of correct, I'd see these later developments as mostly natural refinements of the medium, that didn't really alter the fundamentals.

From Jenson until the first Grotesk fonts I don't think there was anything large one-time leap, but rather a sequence of gradual evolutions.

https://web.archive.org/web/20240507064635/https://news.artn...

Archive link, this site won't load at all for me.

I'm wryly curious why fonts are among the odd things that really get the goat of us turbo-nerds on forums like HN.

When you spend most of your day staring at text on a screen, the minutia of how that text looks like become very important.

The centuries-old artistry of mass reproduction of thought has many wonderful minutiae!, as high technology often does.

It has just the right balance of technology, art, history, and trivia fun facts. Makes it one of the best topics for us nerds.

Also, programmers spend a huge fraction of their time reading. Reading code, reading docs, reading reading reading. Fonts are important for us from an ergonomic point of view (and it's also a matter of taste and aesthetics!)

Agreed, same reason the Monaspace font 1.1 release made news: https://news.ycombinator.com/item?id=40267120

That's cool. I admit hearing that story and thinking, "Is that how it happened? could a diver find it?" Apparently, they could! Great work on someone seeing it through.

I’m left wanting to hear more about the motivation for dumping the type in the first place. What kind of swindle was suspected? Did the partner try to reconstruct the type?

https://youtu.be/e8harWbZN6U?si=4D5ZDCn2WLlciy5T&t=1002

There’s a great episode of the Futility Closet podcast about the Doves Type and the dispute that lead to it being dumped into the Thames:

https://www.futilitycloset.com/2017/09/04/podcast-episode-16...

You might also like "Zilvertype" which is from the dutch font school of roughly the same time. https://www.alphabettes.org/zilvertype/

Wild they didn’t include an example of the actual typeface in the article.

Also, I’m curious how there were 500,000 pieces in the typeface.

Nicola White documents here interesting mudlarking adventures on youtube, I recommend it! https://youtu.be/rVxncipNvvY?si=1DGluOHT8T5fRNfE

Again? I swear this happened about a decade ago. Yeah, here it is: https://metro.co.uk/2015/03/15/lost-typeface-rediscovered-al...

EDIT: It's the same typeface.

Thames Old Roman?

Reply on news.ycombinator.com