Category: Kumihan/Japanese Typography

Japanese font and typography news and market analysis

New Maps for Japan nowhere in sight in big Look Around expansion (Updated)

Apple started rolling out Look Around expansions in Tokyo (including Tokyo Islands), Chiba, Shizuoka, Kyoto and Osaka on 2023-05-12 covering most if not all of the public roads in those prefectures, along with capital city or select city and district additions for Aichi, Hyogo, Gunma, Okayama, Fukuoka and Kumamoto prefectures.

This is the largest Look Around expansion in Japan since the August 2020 ‘Tokyo Olympics’ kickoff launch. Yet redesigned ‘New Maps’ with Apple proprietary cartography seen in America and recent European rollouts remain an elusive goal after five years of Apple Maps image collection. There has been discussion on social media about when redesigned Apple Maps are coming to Japan, but most people are confused about how much is really mapped at this point. Even Justin O’Beirne initially listed Japan as a 2023 new maps candidate then wisely removed it.

The confusion is understandable. People assume entire prefectures have been mapped when this is not the case, easy to do if you don’t live here and imagine that Japan is a small, easy to map island country. You can only appreciate how big Japan is and how circuitous the roads are by driving them.

Apple has a simple formula for image collection in Japan: Cities + Districts = Prefectures. They don’t bother with other classifications (prefectures, villages, towns, etc.), only cities and districts. Apple starts with select key cities in a prefecture then gradually adds less populated cities and districts over several mapping seasons until an entire prefecture is mapped. This could be changing.

The entire Apple Maps Image Collection 2023 schedule for Japan was pulled just as it was due to start on February 1 then reposted in early March with new dates running three months shorter than the original schedule and a new collection area covering all unmapped sections of Aichi Prefecture. Yamanashi Prefecture is also on the schedule and will be the first prefecture Apple is mapping in one season (March~October 2023). Perhaps this is a new faster paced image collection strategy, we shall see.

To explain the situation, I have come up with (what else?) a map that hopefully clearly shows the piecemeal image collection in easy to understand terms of what’s mapped and what’s not. Highlighting every city and district is way too much detail so I chose color coded prefectures in 3 categories:

  1. Mostly Mapped Prefectures
    ‘Mostly mapped’ because Apple only maps public roads (city, district, prefecture, national). In Japan there are lots of publicly used local community maintained roads classified as ‘private roads‘ that Apple does not map (nearly 40% of all roads in my city), that Google does. This means there are significant dead spots in areas ‘completely covered’ by Apple Maps image collection vans and backpacks and this has lots of implications not only for redesigned map cartography, Look Around, AR Walking Directions, etc., but overall quality and usefulness.
  2. Partially Mapped Prefectures
    Major metropolis areas that include surrounding parts of multiple neighboring prefectures: Greater Tokyo (Tokyo, Kanagawa, Saitama, Chiba, Ibaraki), Greater Osaka (Osaka, Kyoto, Nara, Shiga, Hyogo), Greater Nagoya (Mie). For some reason Apple has not mapped traditional greater area region prefectures like Gifu (Greater Nagoya) and Wakayama (Greater Osaka).
  3. Selectively Mapped Prefectures
    Capital cites in Hokkaido, Miyagi, Niigata, Ishikawa, Ehime, Kumamoto prefectures. Capital and/or select cities in Okayama, Hiroshima, Fukuoka, Kagawa, Mie.

As the graph clearly shows, Apple Maps image collection in Japan has a long way to go: by the end of 2023 out of 47 prefectures, 8 will be completely mapped, 4 will be partially mapped, 13 selectively mapped, 22 will remain completely unmapped. The gradual expansion means redesigned proprietary New Maps and the services connected with them, Detailed City Experiences, 3D landmarks, etc., are not coming in 2023, 2024 or even 2025, not unless Apple greatly accelerates its image collection pace. Those features cannot happen until Apple image collection vans map all of Japan, including private roads as Google does. If New Maps and the features that depend on them magically appear before the job is done, it will mean one thing: the map data isn’t Apple’s and isn’t proprietary. They have to get the missing pieces from someone.

I still think Indoor Station Maps with AR directions are in the works (likely for WWDC23) but Japan maps are stuck with 3rd rate GeoTechnologies supplied cartography and will be for some time to come. And believe me, the pretty looking map details don’t match the ground truth outside of metropolitan areas. In short, Apple Maps in Japan has been, and will remain, the all bets off outliner for redesigned maps and associated features.

Not that it matters much for iPhone users in Japan: 80% of Japanese iPhone users don’t use Apple Maps. Mind share and market share has long been dominated by Google Maps and Yahoo Japan Maps. That is not going to change. Take the current state of Look Around for example. Say you are a real estate business and you link available houses and rental apartments with an online mapping service. With Google Maps you are assured that potential buyers can use Street View to examine everything because Street View covers all areas including private roads. Not so with Look Around. It’s for this reason I think Apple Maps Business Connect will be a tough sell in Japan when so many potential business locations are literally off the Look Around map. Until things change in a big way, going with Google Maps for business is a no brainer.

2023-04-28 Update
Apple updated existing pre-expansion Look Around areas with 2022 image data, including backpack images of station areas. Re-mapped driven areas appear to be limited to central areas and underground roadways such as the Yamanote Expressway. New features include:

Look Around in-station areas
Main stations now include backpack collected images for station entrance areas (outside the transit gates) and connecting corridors to shopping areas and stations, but no station indoor maps or AR directions. This first appearance of indoor Look Around images in Japan suggests more is coming: the current implementation is hard to use as it does not have the ability to switch between layered above and below ground Look Around views. Lots of backpack image collected underground stations mapped in 2022 have yet to appear.

POI Look Around
Look Around areas with Point of Interest (POI) labels as also expanded to most Look Around JP areas incorporating 2022 collected images but is mostly limited to central areas in Tokyo, Nagoya, Kyoto, Osaka. Tokyo for example is mostly limited to wards ‘inside’ or ‘along side’ the Yamanote Line, though Tokyo areas with 2022 refreshed imagery are likely to have them. Look Around #2, #3, #4 (Hiroshima, Niigata, Sapporo, Kanazawa, etc.) also have POI Look Around in central areas. Look Around areas without POI labels have labels for major bridges, roads, stations and other important navigation landmarks.

2023-05-12 Update
Apple is rolling out the Look Around #5 expansion. All of it seems loaded but not fully optimized. Shizuoka expansions are not showing as available, for example, even though Look Around views can be seen by zooming in from adjacent available areas.

2023-05-15 Update
Kumamoto City Look Around has been taken offline, there was a large hole in the central famous Kumamoto Castle area that needed to be fixed. As with the Shizuoka additions, it may take a while before optimizations and the rollout are completed and available.

The gallery below shows new Look Around views.

All of Chiba Prefecture has Look Around
Maebashi City, Gunma Prefecture
Atami City Shizuoka Prefecture
Himeiji City, Hyogo Prefecture
Maizuru City Kyoto Prefecture
Kameoka City Kyoto Prefecture
Okayama City
Kitakyushu City
Kumamoto City
Nagoya 2022 backpack image collection shows station outlines
Shinjuku Station
Shinjuku Station
Ikebukuro Station
Tokyo Station
Tokyo Station
Shinagawa Station
Shinagawa Station
Nagoya Hisaya-odori Station
Kyoto Station
Osaka Station
POI Look Around Shizuoka
POI Look Around Nagoya
POI Look Around Osaka
POI Look Around Kyoto
  • Look Around #1 August 2020: Greater Tokyo, Greater Osaka, Greater Nagoya
  • Look Around #2 January 2021: Fukuoka City, Hiroshima City, Nara (Greater Osaka), Takamatsu City
  • Look Around #3 May 2021: Sendai City, Kanazawa City
  • Look Around #4 May 2022: Sapporo City, Niigata City, Shizuoka City, Akashi City
  • Look Around #5 May 2023: Tokyo Metropolis including Tokyo Islands, Chiba Prefecture, Shizuoka Prefecture, Kyoto Prefecture, Osaka Prefecture, Maebashi City, Takasaki City, Okayama City, Kita-Kyushu City, Kumamoto City

Related posts
When will Japan get Apple’s New Maps? Part 1 and Part 2: the private road problem

2022 mapping season: the basis for Look Around #5 expansion in 2023

The original Apple Maps image collection Japan schedule for 2023 posted in January, pulled, then reposted 2023-03-30 with new collection dates (March 30~October 30) and new collection areas for Aichi Prefecture. This will be the basis for Look Around #6 expansion in 2024.

WWDC outlook for a San Francisco C+J+K Apple Font

This is not a WWDC23 prediction, but at some point Apple will certainly unveil a variable San Francisco CJK (Chinese-Japanese-Korean) system font to match the rest of the Apple SF font family, and it will be unveiled at WWDC. I’m not a fan of the CJK name and the mental baggage that comes with it because it’s one of those western concoctions that deal with pesky Asian cultural differences by sweeping those differences under the rug of indifference. Like using a leaf blower to blow and hide dirt everywhere instead of removing the dirt by neatly sweeping the mess into a bag. It’s all Chinese right?

Wrong. Chinese is merely the start point for centuries of cultural evolutions and written language aesthetics that are distinctly different for each language. There are Kanji created in Japan that have migrated the other way. Cultural flow is never one way. CJK is a kind of snub intended to keep the cultural flow one way by neatly collapsing important differences as ‘CJK styles’ for the convenience of westerners who can’t be bothered to understand what those differences are. Just as western based baseline font technology can’t reproduce high quality vertical kanji layout, all in one CJK designs can’t reproduce high quality typography across languages. One hopes Apple is spending the money and time to get those differences right for each language group, C+J+K if you will, because it’s not easy.

What will a SF C+J+K design instead of an all in one CJK design look like? Hiragino Sans GB is a good font to examine as it represents an early Apple lead effort to create a mixed Japanese and Chinese design, best described as a “Simplified Chinese version of the Hiragino typeface…designed to make Simplified Chinese characters look good in Japanese texts, and vice versa.” When I talked with one of the key Hiragino designers, Osamu Torinoumi, in 2009 about the Hiragino Sans GB bundled in Snow Leopard, he explained that one design does not fit all.

Hiragino Sans Comparison 2
The modifications for Simplified Chinese characters in Hiragino Sans GB from the original Hiragino character designs are circled in red

We (JIYUKOBO and Screen) visited Beijing Hanyi Keyin Information Technology Co. in December 2007. The top designer is a young woman, Ms. Zhong. We couldn’t talk to each other because of the language barrier and didn’t know if we had the same design sensibility so she started pulling out the hand drawn templates for one of their designs and we went through them one by one. I would point out the design problems and she would nod her head in agreement and after a while I realized we both thought alike.” JIYUKOBO sent all the original Hiragino design data to Hanyi Keyin through Screen and they adapted the designs for China.

“We worked with the Adobe GB 1-4 character set (29,064 glyphs) at 2 weights. Basically we had to finish one weight in 6 months. One year for the entire project. At first we only thought we would be there as backup, but Screen kept passing us all the questions from Beijing. It turned out to be a lot more work than we anticipated.”

“One of the major differences is that Chinese design demands that Gothic (sans serif) characters mimic handwritten style. This means the character should be slightly off center within the virtual body. Even after the project was over I still didn’t understand the difference between Japanese and Chinese “Kokoro” glyphs which the Chinese designers insisted were different.”

If one of the top font designers in Japan cannot understand the differences between Japanese and Chinese “Kokoro” glyph designs, I doubt Apple designers will be able to figure it out on their own. I hope for the best but all too often ‘all in one’ CJK font designs sweep those kinds of important differences under the rug.

Japanese Typography and Font Posts

This is a collection of long form Japanese typography posts. They were written as stand alone pieces, so there is some background explanation overlap, always a weak point of the blog format.

Refurbishing classic Japanese text for the digital age

One of my favorite work tasks is bringing classic Nichiren Shu Japanese texts into the digital age so they can be translated easily or republished using the latest print technologies for paper and ebooks. Before a title goes into production there are essential steps of obtaining the basic text in digital format, if any exists, and exploring archives for definitive published Showa era sources to double check digital text integrity. Exploration is spelunking into the past to find people connected with the original production process, however remotely, and tease out helpful details: are there any production materials, was it all analog, is there any digital content to work with, and so on.

When helping to bring Senchu Murano’s wonderful Lotus Sutra English language translation back to life, I was heart broken to learn that after months of searching, the original production materials had been destroyed when the printer closed the business only a year earlier. Fortunately there was already a team working on recreating all the English text (over 120,000 words with lots of transliterated, diacritical heavy Pali vocabulary) in desktop computer word processing software. However, when the project finally entered into the primary layout stage I quickly discovered that Murano, or the kumihan typesetters of the 1974 1st edition, had used a number of non-standard Kanji characters in the glossary section, aka Gaiji.

The glossary from The Lotus Sutra 3rd edition

Fortunately I knew the designers of the Hiragino Japanese macOS system font and they introduced a former apprentice who did outside contract font design work. After a careful review he found 15 gaiji characters, unique regular kanji variations not included in the Hiragino Gothic Pro N extended character set and created them for Lotus Sutra 3rd edition.

The 15 Gaiji characters created for the Lotus Sutra 3rd edition.

One thing I learned from the gaiji creation process is that the line between a quirky Japanese kanji design of a regular character and a real gaiji can be very fine. It’s not always an easy black or white call. There is also the publishing history to consider, what was the original intention? Did latter editions swap out complex kanji with simplified versions due to the transition from analog production, and because the early electronic layout production systems were so limited? These are all important points to consider when porting classic Japanese texts to modern production system software.

I was reminded of this with a new project recreating a 31 day chant book of Nichiren Shonin’s Minobu Letters. Fortunately Okazawa san was available to do another fine comb review of our materials.

We found that the original Showa text kanji, which is considered the definitive source, had been changed in the Heisei version. Upon further investigation I discovered the Heisei text had been reproduced on a proprietary Panasonic electronic typesetting device that had limited character sets, and was obsolete. The Panasonic device Japanese fonts used i the book were also somewhat quirky. They looked different enough to consider them gaiji-like, but in the end after comparing everything to Showa printed books, we realized they were just quirky simplified designs of the Panasonic device. Not the original intention.

The happy end here is that the default macOS Hiragino Gothic Pro N extended character set has all our production needs covered. And it’s a great design that travels very well.

I highly recommend it.

Fixing the Apple Maps Point of Interest content problem with Apple Business Connect

One of the long term challenges with Apple Maps is improving the Point of Interest (POI) content. It’s a problem that remains even as Apple rolls out ‘New Maps’ based on their proprietary collected image data. Justin O’Beirne has covered it from the US angle, I have posted about the messy Japanese POI situation many times. Despite the Apple Maps image collection effort around the globe, the quality of POI content has not improved. It is all over the map compounded by the inability of the Apple Maps system to filter and intelligently juggle multiple POI sources. Apple is stuck with 3rd party POI content from Yelp, Foursquare, TripAdvisor, Tabelog and countless others that Apple doesn’t ‘own’: they don’t collect it, they don’t edit it. Until now.

Today Apple rolled out Apple Business Connect. Eddie Cue:“We created Business Connect to provide Apple users around the world with the most accurate information for places to eat, shop, travel, and more.” Whew, good thing because people who use Apple Maps always complain about Yelp: the content is out of date, ancient reviews don’t reflect reality, or worse, the reviews are gamed by bots, hacks or ‘kakikomi butai’ (post entry battalions) in China or North Korea.

Don’t laugh, a Japanese Korean friend once told me about the computer class curriculum at his Korean school in Japan. The teacher would announce the class assignment of the day: writing and posting glowing product reviews of Korean products on various review sites. The old Unification Church in Japan was notorious for employing a virtual ‘post to order’ kakikomi butai operation that paid by the character. This is why I never believe in crowdsourced anything. To me it’s mostly fake or manipulated, with little oversight by stupidity or design. Most Americans seem to believe in it still but crowdsourced content is risky and trouble prone: Yelp and even Tabelog have had to address periodic content scandals online and in court.

So Apple is taking charge of its own POI content. Over the past year Apple Maps has rolled out POI ratings and picture uploads linked the user Apple ID, wisely omitting reviews and limited to places to eat and drink, places to shop and places to stay. So Apple now controls both the POI upload content pipeline and the ratings pipeline. The biggest challenge will be how well Apple manages the POI content swap out process. Is 3rd party POI content automatically swapped out when Business Connect POI is uploaded and Apple verified? More importantly, how exactly does Apple verify Business Connect content? There certainly isn’t an Apple army of ground truth experts roaming around. The proof will be in the content verification and management, and will take time to find out the results. There is also the Eddie Cue mentioned ‘places to travel and more’ stuff that isn’t addressed by Apple Business Connect. We’ll find out about that in time as well I guess, but at least the Apple Maps team finally has a game plan to solve their POI content problems.

Some Japanese Points of Interest allow photo uploads and streamlined thumbs up/down ratings

The curse of Japanese PostScript fonts will live on: outline your outlines

I have not used Adobe Illustrator much the past few years and certainly don’t use it enough to justify buying a Creative Suite subscription that only lasts 12 months. Recently a localization project came in where I needed to edit the original Illustrator file data text. The printer sent me their Illustrator print files and I blithely opened the file with a name that ended with ‘OL’.

As soon as I clicked the body text I realized what OL meant: outline. All the text in a 2 page document with lots of text had been converted to outlines via the Illustrator convert text to outline feature. I couldn’t edit anything. I contacted the printer and received a backup file with the text intact that had not been converted to outlines.

Japanese service bureaus and printers require all Illustrator file text be converted to outlines (left image), they usually will not accept regular text data (right image)

I reflected on this basic Japanese designer practice of converting all Illustrator file text to outlines before sending work files to the printer. It took me back to my days setting up some of the first Japanese PostScript DTP production lines for print companies in Shizuoka. Any printer or high end print service like Lithmatic (a great service company by the way) always requests designers to submit Illustrator work files with all the text data pre-converted to outlines. I hate doing this because it strips away all the font hinting. Font hinting is now only thought of as a screen display thing, but printer font hinting was necessary back in the days of 300~600 DPI PostScript laser printers.

Maybe printer font hinting is no longer necessary in this era of high resolution CTP (computer to plate) on-demand small print runs. Even so, to my eye, stripping out the font hinting reduces the Japanese typographic quality of smaller printed kanji text with their complex glyph strokes. Why is it necessary in this age of PDF workflows to even bother converting text to outline anymore?

It all goes back to the many original sins of the first Adobe Japanese PostScript fonts, the biggest sin being they could not be downloaded to the printer on a job basis…they had to reside on the printer. And they were not cheap: ¥300,000 a pop (back in 1990 when that kind of price was a lot heavier on the wallet) for a single unlimited resolution Japanese PostScript printer font. Not only that, early Japanese PostScript print drivers sucked. They were slow and print jobs would often trip up the RIP job with a memory error or something arcane. Like it or not, print job managers learned to read voodoo tea leave PostScrip error codes to decipher problems, fix the Illustrator file and run the job again. Late work nights were common for production staff.

Usually it was just easier to convert text to outlines which was the godsend feature that arrived with Illustrator v5 along with Japanese Adobe ATM. Instead of buying expensive printer fonts and dealing with incomprehensible PostScript output errors, it was easier (and cheaper) for print service bureaus to require all Illustrator file text data be converted to outlines. This was a time when Illustrator was the workhorse choice for DTP designers in Japan.

All of the PostScript problems were eventually fixed with OpenType fonts and PDF workflows, PostScript fonts themselves will officially die on January 2023. But the PostScript font damage done in Japan will never be fixed. There’s just too much legacy data out there, both in data files, and printer fonts still installed on high end output devices. And Morisawa will always provide legacy OCF fonts for their Passport customers that need them, no matter what Adobe says.

PostScript fonts may be going away, but the ghosts of PostScript fonts, the fine art of outlining Illustrator text data, will be haunting Japan for a very long time.