I see the value in the author's mission to create a replacement for pangrams, but it sounds less like he replaced pangrams (on which I've always been sold as being a tool for content creators to quickly visualize a font with loose approximations of common use cases) and more like he came up with a quicker letterproof. http://famira.com/article/letterproef
Almost like unit testing a font more efficiently rather than using a full letterproof for that process. Which is pretty cool and probably fills a need, though maybe not the same need that pangrams address more broadly.
--
I'm not an expert on this topic and am prepared to have what I've learned in the last five minutes fact-checked by someone who actually does this for a living.
For a very long time, pangrams were used to demonstrate typefaces because they were the shortest string that included every letter. I think it was perceptive of him to realize that they weren't optimized for showing letters in their typical settings.
Either one is still lacking when evaluating a font. While seeing every letter is helpful, knowing how different letter combinations look adjacent to each other is also important (a j in the middle of a word for example).
One of my favs (apropos for the HN crowd): I would code it differently by sending the ajax request to the backend server and randomize the number there :Grinning face with smiling eyes:
#magento #php #laravel
For those interested Hoefler and typography.com were covered in Netflix’s Abstract (last episode of season 2 I think). I really enjoyed it and learned a few things about typography in general.
Very interesting article, but set up against a strange straw man.
First, the usefulness of a full page of text for critically analyzing a font is obviously extremely useful for type designers looking for flaws to fix, or perhaps a national newspaper or magazine selecting a new body type that must be absolutely flawless. This is indeed a great text for proofing fonts.
But "quick brown fox" was never intended for proofing fonts. When the author says pangrams are "singularly useless" for that, he's right, but they're very good for something else.
Mainly, browsing fonts in a catalog to get the full personality at a single glance, on a single line. Above all what kind of "a", "e", "f", "g", "j", "l", "t", and "w" it has, since these (in non-italic) generally have the most variation. To check if there are any letters the font does something "weird" with. And just get an overall basic sense of look and spacing.
So setting the paragraphs he comes up with, as a replacement for pangrams, just feels like the entirely wrong way to frame it.
(All that said, I have insane respect for the man -- his designs set the absolute top bar.)
Mmmhh - I’m not a font designer, but just someone who chooses fonts for websites on a regular basis.
And I’ve often found myself frustrated about not finding a good representative page of text, allowing me to evaluate fonts better than just single sentences.
So this is a hugely welcome addition to my toolkit - definitely replacing pangrams for me and big thanks to Mr. Hoefler for making it publicly available.
I think the parent argued that both serve a purpose. When you're dealing with a large selection of fonts a small pangram is more useful to help you narrow the list down - until it becomes small enough that you can switch to one of these font proofs to evaluate the final few fonts.
Websites like Google Fonts should absolutely use one of these font proofs for the dedicated font pages - or when simply comparing two fonts, but use a shorter pangram when comparing several fonts at once.
I think there is a general problem of fonts being treated as if they were all equivalent when in reality there are fonts which can only be sanely used as display fonts, and body text fonts which admittedly you might use as both. The computer has lumped them all together in a way a type setter would never do. So, yes, different approaches for different uses.
As to choosing a font, having had a typographer, a graphic designer and a book dealer as 3 of my brother's, it is a job which I can hardly find the courage to do. I also regularly wish others shared my self doubt and left it to a professional, or at least just followed conventional wisdom. A font you notice is a bad font.
In short, Frere-Jones accused Hoefler of intellectual property theft and not honoring the terms of the verbal agreement outlining their business partnership. I'm not sure how the case ended up.
I’m not a designer or typesetter either and I find that sampling the typeface and font via a simple representative sentence to be helpful. How wide is the “m”, is the “a” single storey or double, αre the legs of the k offset or simple v?
> (All that said, I have insane respect for the man -- his designs set the absolute top bar.)
I have insane respect for the typefaces that are sold under his name. Unfortunately, the most iconic designs were taken from his former partner Tobias Frere-Jones, and are now sold without the appropriate credit. That is not worthy of respect.
Titles, or more likely time. I submitted this a couple weeks ago and was just notified it was in the "second chance" pool, so I resubmitted. Second time around, it made it to the front page with the same exact title.
Or maybe people are just more into typography news now as opposed to two weeks ago...
Sometimes you just need large quantities of text to see how it really fills a space for proofing. What about using good old lorem ipsun text or the copious number of ipsum variations that generate natural-sounding sentences in almost every imaginable combination?
I am missing words with 'ij' in the proof text. If they do not look roughly similar and with their dots on the same height, the font looks pretty bad in Dutch (ij is a frequent combination, and often even taught as one letter). Should a proof text care about a few of such language specific cases?
I was hoping to find a good example text here, but this still just covers a fraction of ascii. In 2020, you have to expect das große ß und من اليمين الى اليسار and perhaps even ө, ү, ң (which should be the same height as о, у, and н[1]). So far [2] is the only page I've found with some representative samples.
I guess you mean Unicode, not ascii. It covers just a fraction (about half of it) of ascii, but not for those reasons.
I would add digits and more punctuation. Question marks, for example, can be cases where type designers let loose their creativity.
I think the sample also is biased towards English. It’s missing letter combinations such as “sch” and “cht” that aren’t uncommon in German (or French, which has the wonderful word “schtroumpfs” since 1958), or the pairs “au” or “oe”, used in Dutch, all of which may have ligatures.
That said, I already have my covid project. All state of the art optical flow algorithms have problems seeing thin branches. You might have seen videos of Skydio drones crashing into branches, wires, or water. My goal is to fix that.
I started designing a new AI architecture in February and the main detection training finished mid-May. By now, I'm building the post-processing for guessing plausible fillings for occluded areas.
Plus, I'm now working from home, so my basement is already overheating from all those GPUs. I should probably pause AI for a while or I'll have to move the icebox to another floor.
53 comments
[ 3.3 ms ] story [ 119 ms ] threadAlmost like unit testing a font more efficiently rather than using a full letterproof for that process. Which is pretty cool and probably fills a need, though maybe not the same need that pangrams address more broadly.
--
I'm not an expert on this topic and am prepared to have what I've learned in the last five minutes fact-checked by someone who actually does this for a living.
But I'm still confused as to why pangrams were mentioned, is my point.
New job: fix Mr. Gluck's hazy TV PDQ
https://twitter.com/JackOfQuills/status/984290288259788800
One of my favs (apropos for the HN crowd): I would code it differently by sending the ajax request to the backend server and randomize the number there :Grinning face with smiling eyes: #magento #php #laravel
First, the usefulness of a full page of text for critically analyzing a font is obviously extremely useful for type designers looking for flaws to fix, or perhaps a national newspaper or magazine selecting a new body type that must be absolutely flawless. This is indeed a great text for proofing fonts.
But "quick brown fox" was never intended for proofing fonts. When the author says pangrams are "singularly useless" for that, he's right, but they're very good for something else.
Mainly, browsing fonts in a catalog to get the full personality at a single glance, on a single line. Above all what kind of "a", "e", "f", "g", "j", "l", "t", and "w" it has, since these (in non-italic) generally have the most variation. To check if there are any letters the font does something "weird" with. And just get an overall basic sense of look and spacing.
So setting the paragraphs he comes up with, as a replacement for pangrams, just feels like the entirely wrong way to frame it.
(All that said, I have insane respect for the man -- his designs set the absolute top bar.)
And I’ve often found myself frustrated about not finding a good representative page of text, allowing me to evaluate fonts better than just single sentences.
So this is a hugely welcome addition to my toolkit - definitely replacing pangrams for me and big thanks to Mr. Hoefler for making it publicly available.
Websites like Google Fonts should absolutely use one of these font proofs for the dedicated font pages - or when simply comparing two fonts, but use a shorter pangram when comparing several fonts at once.
As to choosing a font, having had a typographer, a graphic designer and a book dealer as 3 of my brother's, it is a job which I can hardly find the courage to do. I also regularly wish others shared my self doubt and left it to a professional, or at least just followed conventional wisdom. A font you notice is a bad font.
In short, Frere-Jones accused Hoefler of intellectual property theft and not honoring the terms of the verbal agreement outlining their business partnership. I'm not sure how the case ended up.
[0] https://www.theverge.com/2014/9/29/6868945/the-worlds-bigges...
I have insane respect for the typefaces that are sold under his name. Unfortunately, the most iconic designs were taken from his former partner Tobias Frere-Jones, and are now sold without the appropriate credit. That is not worthy of respect.
Just for fun, the JavaScript one-liner I used to check:
https://news.ycombinator.com/from?site=typography.com
Or maybe people are just more into typography news now as opposed to two weeks ago...
[1] https://github.com/mozilla/voice-web/issues/1141#issue-33511...
[2] http://www.madore.org/~david/misc/unitest/
I would add digits and more punctuation. Question marks, for example, can be cases where type designers let loose their creativity.
I think the sample also is biased towards English. It’s missing letter combinations such as “sch” and “cht” that aren’t uncommon in German (or French, which has the wonderful word “schtroumpfs” since 1958), or the pairs “au” or “oe”, used in Dutch, all of which may have ligatures.
1. You run statistics on a large text corpus to determine the most likely letter combinations that a font needs to display correctly.
2. You train a skip-word model on a large text corpus to evaluate if a sentence makes sense or not.
3. You train an AI to produce random sentences that match the word-pair-probabilities from #1 while being natural sentences according to #2.
=> You should end up with natural-sounding text with proper grammar that still contains all the important character pairs for font testing.
https://arxiv.org/abs/1910.12604
That said, I already have my covid project. All state of the art optical flow algorithms have problems seeing thin branches. You might have seen videos of Skydio drones crashing into branches, wires, or water. My goal is to fix that.
I started designing a new AI architecture in February and the main detection training finished mid-May. By now, I'm building the post-processing for guessing plausible fillings for occluded areas.
Plus, I'm now working from home, so my basement is already overheating from all those GPUs. I should probably pause AI for a while or I'll have to move the icebox to another floor.