Hacker News new | past | comments | ask | show | jobs | submit login
Old book illustrations from the 19th and 20th centuries: an online database (openculture.com)
386 points by sohkamyung on Feb 13, 2020 | hide | past | favorite | 50 comments

I made a quick random generator that includes all 3810 images:


Maybe useful for inspiration for some sort of creative activity. Also useful if you want a list of all the image URLs (click the edit button to see them).

That is awesome, thank you for putting in some time for this!

No problem! You can re-generate the url list (in case they add more pics) with the code below. Just change `320` to the total number of pages and paste it in the developer tools console when you're on their website.

  let urls = [];
  for(let i = 1; i <= 320; i++) {
    let text = await fetch("https://www.oldbookillustrations.com/illustrations/page/"+i).then(r => r.text());
    let matches = text.matchAll(/<img width="220" height="220" src="(.+?)-220x220\.jpg"/g);
    for(let m of matches) {
      let url = `${m[1]}.jpg`;
    console.log(i, urls.length);
  console.log( urls.join("\n") );

I love this.

Thank you, this is awesome!

Best image I found was: https://www.oldbookillustrations.com/illustrations/de-groof-... ... great story, anyway. Poor bastard didn't even get a mention at https://en.wikipedia.org/wiki/History_of_aviation#19th_centu...

Shameless plug for mine and my wife's project, http://www.visualhaggard.org. We have extremely high quality scans of illustrations from many editions of H. Rider Haggard's work. Thanks for your time!

The site excellent but is serving a TLS-certificate from: * .herokuapp.com instead of * .visualhaggard.org, so everything is big red browser warnings and the search engines will punish you guys accordingly.

Beautiful stuff!

I commented elsewhere on this thread about reproductions: https://news.ycombinator.com/threads?id=cpach

Just out of curiosity, have you considered making print reproductions?

Thank you for taking a look!

You know, we hadn't considered making prints but I've always wanted to merchandise. The goal was to make an academic resource and to collect and preserve the images in large resolutions.

Also see ClipArtETC[1] from the Florida Center for Instructional Technology. They have a much larger collection, though less meticulously restored, and with a license requirement for commercial use.

[1] https://etc.usf.edu/clipart/

Why not link directly to the source?

Indeed. Had to scan halfway through the article to get to the actual site link https://www.oldbookillustrations.com/

How very strange. Why should they be censored?

Can someone explain (or find an explanation for) Phil May's ABC's? For example what should 'q' and 'Q' stand for?



perhaps "quiet" and "Queen"?

thanks! Seems reasonable :-)

the captions don't really help!

Great source of images for blogs, if they are off copyright.

I believe they are. According to the article:

> Old Book Illustrations presents itself as a scholarly resource, including a digitized Dictionary of the Art of Printing and short articles on some of the most famous artists and significant texts from the period. The site’s publishers are also transparent about their selection process. They are guided by their “reasons pertaining to taste, consistency, and practicality,” they write. The archive might have broadened its focus, but “due to obvious legal restrictions, [they] had to stay within the limits of the public domain.”

From their terms of use [1]:

> We do not try to limit the use of the Illustrations available on OBI, but we cannot guarantee these Illustrations are noninfringing, or legally accessible in your jurisdiction and your use of them is solely at your own risk. Although we do our best to offer only Illustrations that are considered public domain in most countries, copyright laws vary from one jurisdiction to another, and you agree that you are solely responsible for abiding by all laws and regulations that may be applicable to using the Illustrations. While we endeavor to provide enough information to make that process as easy as possible, we cannot guarantee that this information is accurate.

[1] https://www.oldbookillustrations.com/terms-of-use/

The image of the squirrels in the article is from Beatrix Potter's Squirrel Nutkin. I'd be really surprised (pleasantly though) if the publishers of something so famous had allowed the images to become public domain.

Published in 1903, and Potter died in 1943. The copyright would have expired in 2013 (! — i.e., 70 years after the death of the author) at the latest as far as I can tell.

It should of course be completely unnecessary to look up the exact rules for something published that long ago, but that's the consequence of defect intellectual property laws.

Integrally available here:


2018. That would have been grandfathered into the change for 75 years after last copyright holders death.

Where did you find the 75 year period? The general consensus seems to be that her works entered the public domain in 2014 (which is 70 years counted from the year following her death, I missed a year in my previous post).




This mentions Beatrix Potter's works as entering the public domain in 2014 as well.

The copyright math is complicated, and even experts disagree commonly. Wikipedia's lists, though helpful, have been known to be disagreed with by the courts on numerous occasions.

For works published before Jan 1, 1978, it generally comes down to:

+ Lifetime of last involved author, +70, or +95 or +120 years.

+ If the work was already under copyright at that date (not everything was), then it also gets extended by 45 years

75 years came from the example given by the official documentation that attempts to explain the various acts (not definitive, but helpful) [0], and how they interact with each other:

> Example: A work that first secured federal copyright pro-tection on October 5, 1907, and was renewed in 1935, would have fallen into the public domain after October 5, 1963. The first act extended the copyright to December 31, 1965; the second act extended it to December 31, 1967; the third act extended it to December 31, 1968; the fourth act extended it to December 31, 1969; the fifth act extended it to December 31, 1970; the sixth act extended it to December 31, 1971; the seventh act extended it to December 31, 1972; the eighth act extended it to December 31, 1974; the ninth extended it to December 31, 1976; and the 1976 Copyright Act extended the copyright through the end of 1982 (75 years from the end of the year in which the copyright was originally secured)

This describes an example published in 1907, and we're talking about one published in 1909. As a rough guide, a minimum of 75 years fits works of that era. It is often actually more.

But, as I didn't put a lot of effort into my earlier statement, let's do a little more digging.

A provision in the UK's 1988 act probably means that Potter's works would have been revived for another fifty years, that is, expiring in 2038. However, that was only if Potter's estate owned the copyright - which it turns out, is not the case.

Potter left her works (and their copyright) to the National Trust upon her death, in the UK, which has a different set of copyright laws that are almost, but not quite fitting to the ones I've described above. (Which is common. The laws look the same until you try and do anything.)

The National Trust _sold_ the copyright (not just the publication right) to a company that was eventually absorbed into the Penguin Group, and so the copyright belongs to them.

This makes it extremely complicated. At one point, the National Trust held the copyright, meaning that Crown Copyright laws applied, which are generally shorter. However, the copyright was then passed on to a private company.

It is difficult to track down a date when the rights were passed from the Trust to Penguin - it may have been that the works had already expired at that point, meaning they might _mostly_ (see 'Another note' below) in the public domain, before considering the new ownership.

If however the work wasn't yet in the Public Domain, then a tweak of the UK copyright laws in 2006 may mean that they still aren't. The "artistic resale" right (literary works fall under artistic copyright in the UK), which is non-reassignable, and non-waivable, says that as long as the work is still being sold in a reasonable number, it both qualifies for royalties (in this case to Penguin), and cannot enter the public domain. However, these sales must come in units of greater than 1,000 to qualify. As a very considerable number of these books are still being sold, it appears to fall under this right.

Another note: Penguin published a previously unpublished work in 2016, containing some of her illustrations. Which, unfortunately, means that Penguin has a definite right to those works and any works that may appear to be derivatives of it. Whilst this new work might be considered separate from previous works, courts have sometimes in the past considered book series, even without overarching plotlines, to be a single work, even across extremely varied timelines. Penguin could argue all works should be linked to the copyright of the new work. (This is an odd situation. If Penguin didn't have the copyright right, and just the publication, then the work would expire in 2039, under the guidance on posthumous works given in the 1988 act. But Penguin has the actual copyright, not just publication rights.) (How does a company having the copyright work out with the general guidance of author + 70 years? It gets simplified. Publication date + 50 years is used instead.)

However, despite all this, for now, Penguin have said they won't enforce their copyright on the works. Copyright in the UK is not (today) a right that can be waived, so it doesn't automatically mean Public Domain. However it does probably mean that you don't have to have a lawyer looking over the situation.

[0] https://www.copyright.gov/circs/circ15a.pdf

Phew. I wonder if sensible copyright reform will happen in my lifetime. Either that, or we go back to a society that accepts intellectual property infringement of works older than a couple of decades as noble and necessary (cf. the attitude towards SciHub).

How did you get her death plus 75 years though? The example quoted works from the date of publication, which is 1903, so the longest possible copyright would be death of Beatrix Potter plus 70 (ignoring the informative, but weird dystopian legal possibility of Penguin making the case for owning the rights to all of her works in 2020).

Somebody can riff on "Fuzz Against Junk" again maybe!


Brilliant, I was looking for something similar for an old project a few years ago. Thanks for sharing

Kind of off-topic, but uh, is there a name for that suns-and-moons on a starry night background pattern? It's used in a lot of places, with all kinds of variations, but I haven't been able to find a name for it (to use in searches, etc.)

Would this be what you are referring to? https://en.wikipedia.org/wiki/Flammarion_engraving

Cheers! That's the kind of thing, and likely the inspiration/source for the patterns I'm talking about. I have two examples on my desk right now, as it happens. One is a kind of potpourri sachet of herbs and the other is a little circular box. Common elements are a dark Navy blue background studded with little stars (both as tiny dots and little five- and six-pointed star shapes) with large suns and moons with faces.

Searching on "Flammarion engraving" just brings up endless versions of that one image. But at least I have a reference now! Thank you. :-)

Here's a keeper. I did not colorize it myself but it's close to what I would have done,


While where on the topic of line-art:

One thing that would be really awesome IMO would be to make faithful reprints of old illustrations. E.g. the picture with the daemon and the snake (which reminds me of Gustave Doré’s works). Produced with high-quality ink/paper.

Does anyone have any ideas regarding how to do that? E.g. what kind of resolution does the scans need to be, etc.

Would a plotter be able to do it? The fact that the scans are bitmaps might be an obstacle since I imagine that maybe a bitmap can’t be sent directly to the plotter.

300dpi at whatever physical size you want. Printing at 1 inch x 1 inch? 300px x 300px.

Hm. Is 300 dpi sufficient? Sounds a little low to me but I could be wrong. And then I guess one has to ”clean” the scanned image quite a bit in order to get good results.

It also would be really cool if one could ”carve” the lines to the paper. Might be possible with a plotter and the right kind of stylus.

A while a go I bought an artwork that actually was cut out from an issue of Harpers’ Magazine c:a 1860 or so. One thing that’s really cool about that line art is that the printing technique made the lines embossed to the paper. Beautiful!

300dpi is typical for most things that go from graphic designer to professional printer. It should be sufficient for most things.

If you're looking for a way to turn a bitmap into a vector type thing, for a plotter or laser or so on, then potrace or the bitmap->vector converter inside Inkscape should be able to do it. They work well for line art.

Great, thanks for the advice!

Is there a limit of images from a single book? Checking various images I wasn't able to find a book with more than 7 images.

Now we know where Drew gets all his images for http://marriedtothesea.com/

Is there a torrent or a zip for all the raw scans? My google-fu is not really helping me in this :/

Apparently they have an API available to their Patreon donators. [0][1]

At $15/month you get access to a single zip file of all the raw scans.

[0] https://www.oldbookillustrations.com/#patlogin

[1] https://www.patreon.com/oldbookillustrations

a script-fu scraper is probably being written as we speak ;)

The Terms of Use ask you not to scrape https://www.oldbookillustrations.com/terms-of-use/

more specifically, they do not approve of putting heavy load on their web server, so a polite scraper could be used

This is how Memes are born.

Ha! I am working on a few now.

Can't access, has blocking cookie banner. Is this related to Liam Quin's https://www.fromoldbooks.org/ (Liam makes frequent appearance on markup conferences as barefootliam)?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact