The tagging experience on Pixiv is… interesting. Images can only have a limited number of tags, so you’d think people would tend to consolidate, but no, there are thousands of tags, many of them slight variations that impair search results, and a lot that are just graffiti.
One useful descriptive term that turned up today was パイスラ (“paisura”), which at first glance looks like it might be related to paizuri (tit job), but in fact is an abbreviation for パイスラッシュ (“pai-slash”), and refers to the volume-enhancing effect of cross-body straps on clothed breasts. This also appears on Pixiv as π/.
Example images are left as an exercise for the reader.
(yes, someone in Tokyo owns paislash.com, and someone in Yokohama owns paisura.com)
I’ve never played/watched/read any Fate/whatever games/anime/manga, but Grand Order’s Mash/Mashu/Matthew Kyrielight/Kyrielite in a blue gingham dress with sweater and glasses pushes all of my buttons. The ensemble even has its own tag on Pixiv: 先輩を殺す私服 (senpai wo korosu shifuku), “clothes which kill senpai” (aka the player). A few people have tried to expand the theme to other characters, but it hasn’t taken off.
Amusingly, I typed the above at work, then went over to the break room to nuke my dinner, and three attractive asian women walked in wearing flirty little sundresses with sweaters, and they all had glasses, too.
In related news, I’ve got Pixivpy working well enough that I can extract metadata for an image based on the ID in its filename, store them in a SQLite DB along with my work-in-progress tag translations, generate HTML to post the selected images, and update my Pixiv bookmarks with whatever I just blogged. That last bit helps improve their recommendation system, another way of weeding out the chaff.
Senpai-killing body pillow (NSFW, of course)
I’ve spent 18 years trying to figure out what sort of picture to hang above the mantel. Then I stumbled across this on Sunday, and it arrived today.
Seems kind of fitting to put a volcano on top of a fireplace.
Clickbait spammy ad network Taboola keeps showing this to me (the Mill Valley bit is poor geolocation), and every time I see it, I think the rule they’re referring to is: kidnapped teens must be securely handcuffed while vehicle is in motion.
I thought everyone knew that, but maybe they don’t teach it in Drivers Ed any more?
By now some of my regulars shouldn’t need introductions, but I’ll fill in the names anyway when I get a chance.
Once upon a time, there was an actual “things you’ve marked ‘not interested’” list that could be added to, and even edited (unless, like me, yours had more than 20,000 items on it).
Now, however, your recommendation system has no memory at all. How else can I explain being offered the exact same items that I select “I’m not interested in this item” for every damn day? No, I do not want a Funko figure of Nearly-Headless Nick, and I won’t change my mind when you ask me again tomorrow. No, I do not want to read book 6 of an isekai series about a slime. No, I don’t want a Funko figure of Inigo Montoya, because I already bought the damn thing last week!
Seriously, of the 50 items in the “New Releases” you just offered me, I’ve already rejected 42 of them, some of them half a dozen times. If the buttons don’t do anything any more, just remove them and stop pretending you’re paying attention to my preferences. And understand that you’re selling less stuff to me because you’re not showing me products I might actually want.
Shipgirls. You can’t swing a loli catgirl without hitting a dozen of them. Fortunately they’re easy to classify, as the Pixiv tag まったく、 駆逐艦は最高だぜ!! (“truly, destroyers are the best!!”) makes clear. This also crosses over quite a bit with 下着艦娘 (“shipgirls in underwear”), so most of them are in the NSFW section.
As a genre, I think it’s past its peak, but I just can’t see the new isekai wave generating as much quality cheesecake. Some of the early ones have done okay with the standard fantasy tropes, but now they’re into slimes, giant spiders, and [vending machines] (https://www.amazon.com/gp/product/B0763LLK22).
I’m a big fan of the Cookie app for MacOS, which does an excellent job of scrubbing unwanted privacy-tracking cookies and other cruft from your web-browsing experience. But there’s one little problem: it can’t delete your history if you have iCloud bookmark sharing enabled, and Safari’s automatic “Remove history items” preference won’t do it either.
Meanwhile, the only Apple-supported method for clearing history nukes everything, including useful cookies like the ones that keep your bank from sending you text messages every time you try to login from a “new” browser.
The following AppleScript appears to be the only way to delete just your history from iCloud:
tell application "Safari" activate close windows make new document tell application "System Events" tell process "safari" keystroke "y" using command down delay 1.0 keystroke "a" using command down key code 51 # delete end tell end tell end tell
After running it, you need to leave Safari open for a minute or two without using it, or else it will repopulate with the iCloud history from your iPhone or iPad (“ask me how I know”).
Seems this has the side-effect of causing Mobile Safari to freeze if it’s suspended on an iOS device, requiring you to force-quit it. At least, I’d never had Safari lock up on iOS so often. It’s almost like iCloud isn’t very good at this whole “sync” thing…
Bumped the delay to a full second. If you have a lot of history to nuke, it can take that long to load it all. The Mobile Safari freezes continue to be an issue after running this script, but it seems it will eventually recover on its own if you leave it running, and doing so will keep it from freezing again.
…until the next time you clean out your history, anyway. “Dear Apple, up your sync game”
While architects are known for producing high-concept designs that are impossible to live in, mass-produced subdivisions have their own quirks. I’ve previously mentioned the goofy peninsula in my kitchen that led me to search for [a decent counter-depth fridge] (/archive/2018/04/005480/). I don’t recall if I also mentioned that the pantry has no light (or power) inside, or that what looks like terrific storage space above the cabinets is ruined by code-required fluourescent uplights that I never use.
Anyway, a lot of that stuff is all on the list for the contractor I’m finally going to hire, but some upgrades just require a credit card and a web browser.
The first upgrade to the family room was the Fuji print over the fireplace.
Second was the seating. 18 years ago, I bought a dark grey couch and chair from Ikea. The machine-washable covers are nice, and they’re sturdy enough that I could keep using them, but you tend to sink into them these days, requiring some effort to get back out (a real issue for my mom).
The problem with replacing them is the way the room is laid out; it’s 14 feet wide and 12 feet deep, but the far end is consumed by a gas fireplace and a deep storage niche, and the left half of the near end needs to be kept open to allow easy traffic flow into the kitchen and dining area. That leaves the right side, and the tiled hearth sticks out nearly two feet, effectively limiting the seating space to 10x6 feet.
I wanted an L-shaped sectional, but 99% of what’s on the market are huge. If they’re under 10 feet long, then they’re at least 8 feet deep, sometimes with a chaise section instead of a proper seat with back and armrest.
I finally found this at the local Ashley, and after very carefully measuring the space and the sectional, I decided it will work:
It’s a foot longer and nearly two feet deeper than the current couch/chair combo, but it does fit, and the small “chaise” extension basically takes up the space that’s currently an end table, so the traffic flow is unchanged. I’ve ordered a slide-under tray table to provide a place for drinks at that end, and the large comfy cushions compensate for the lack of an armrest. It should show up in about two weeks.
The only real downside is that the corner space between the couch and chair is where I currently have my liquor cabinet, and now I have to find a new place for about 70 bottles of booze.
Third was the TV stand. When I bought my first BluRay player and HD TV, 37 inches provided plenty of bang for the buck, but now I want something bigger, and because a wall-mount is impractical (the left side of the room has big windows), I need a new stand that’s wide enough for a 55-65-inch TV. Here the problem isn’t height, width, or depth, it’s the fact that most of what’s on the market is either too ugly or too dark for my mostly-blonde house. I eventually stumbled across this:
My Harmony hub has two IR blasters to reach the components behind the doors, and there’s plenty of space for the FireTV, Japanese PS2, etc. I’ll have to build some new shallow shelving for the discs that don’t fit into its drawers, but that’s not hard.
That should also arrive in about two weeks, and then I can start looking for a new TV. I’d prefer the least-Smart 4K TV I can find, because the vendors have a really poor track record regarding privacy and security. If I let it access the Internet at all, it will be on an isolated VLAN from the rest of the house. I haven’t looked seriously at what’s on the market right now, so suggestions are welcome.
After that comes the real fun: a new receiver and speakers. My old 5.1
surround system is a Kenwood
purchased way back when to drive my top-of-the-line laserdisc
player. It’s held up surprisingly well, but yeah, it’s time. For the
receiver, I’m thinking overkill, specifically the
Denon AVR-X3400H for its large
and versatile set of inputs, and ability to upscale the analog inputs
(which is why I crossed out the S940H; I’m going to look at some other
decent brands to see if I can find something that doesn’t cost quite
so much more). For speakers, honestly I haven’t a clue; the Kenwood
came with a set.
I’m taking a sort-of Underpants Gnomes approach to selecting cheesecake themes on Pixiv.
- Collect Cheesecake.
Step 2 in the Pixiv Underpants Gnome Plan is “pick a theme I’ve already downloaded enough images for”. In this case, that’s everything containing the kanji 狐 except where it’s used in a proper name, or: 狐耳 狐娘 キャス狐 狐 狐耳巫女 #キャス狐 アサ狐 乳狐 狐っ娘 狐巫女. That produces a set of 107 images that I already like, and there’s even a pair of red half-rims for an offering.
For amusement, here’s what my theme-discovery process ends up looking like, using the offline DB I build with the API:
% sqlite3 -separator " " pixiv-girls.sqlite3 <<EOF | head -20 select count(i.id),i.tag,t.en from illust i left outer join trans t on i.tag = t.jp where i.tag not like "%users入り" group by tag order by count(id) desc EOF 1010 オリジナル Original 597 女の子 Girl 484 魅惑の谷間 Charming_Cleavage 462 おっぱい Oppai 329 Fate/GrandOrder 258 R-18 242 水着 Swimsuit 237 FGO 230 猫耳 Cat_Girl 228 極上の乳 Superb_Breasts 220 艦これ KanColle 210 尻神様 Butt_Goddess 194 おへそ Belly_Button 194 艦隊これくしょん 167 魅惑のふともも Captivating_Thighs 156 マシュ・キリエライト Mashu_Kyrielight 134 巨乳 Big_Breasts 125 CLIPSTUDIOPAINT 116 下着 Underwear 114 アイドルマスターシンデレラガールズ Idolmaster_Cinderella_Girls
Once I’ve picked a set of tags, I have a little Bash script that does boolean searches on the SQLite DB and opens all the matching images in Preview.app, so I can do things like “(catgirl or foxgirl) and big boobs and not loli”.
The tagging still isn’t 100%, though, because tags are applied to illustration IDs, and it’s common for users to post multiple “pages” of images. Sometimes these are actual manga pages, but it’s also used for “grab-bag of sketches”, “N variations of the same drawing”, “N characters from same series”, and the ever-popular “progressively-more-explicit scene”. So, if a series includes a busty fox-girl, a loli catgirl, and a glasses-wearing robot, the ID will have all of those tags. Similarly, if the final image in a sequence is the aftermath of a gang-bang, the wholesome first image gets all the explicit tags as well.
Cheesecake sites come and go, but most of the old reliable ones I used to link to just went poof one day, and most of the others ended up hidden behind login pages (Tumblr has gotten very aggressive about flagging adult content, and also shutting down too-popular sites).
Here’s my current rotation:
Giga-Kojimblr: I’ve lost count of how many times Kojimblr’s been shut down on Tumblr, but this is the current one. Excellent taste, reliable name-tagging, no endless scrolling, frequent updates. Bad: sure to disappear and change names again.
Gaisenmon: generally decent taste, usually includes names, no endless scrolling, frequent updates. Bad: too much Nemo, no tagging, occasional burst of candid-schoolgirl photos.
SheCool: frequently-updated archive of recent magazine shoots, including nudes. Well-tagged and categorized, one adult ad in sidebar. Bad: some picture links take you to product pages rather than larger images.
Big Boobs Japan: excellent taste, not restricted to either big boobs or Japan, multiple new shoots per day, all named, romanized, and tagged. No real negatives, apart from a bunch of social-media icons splattered everywhere (including a prominent “save to Pinterest” link, which I’d rather die than do).
Gazō Navi: raunchy, lots of porn pics, lots of porn ads, lots of garbage linkposts. Updated every few days, but maybe only a quarter of the posts are worth clicking on. More and more I visit this less and less, because actual photo galleries that aren’t just copied from one of the above sites only show up at most once a week.
Does the carpet match the drapes?
A number of people at work had this pop up in LinkedIn:
Many of them cheerfully connected to her. When it hit folks in IT this morning, though, everyone responded with a cheery “WTF?!?”. No one by that name has ever worked here, in any department, and certainly not the dates she claims. We flagged it as phony, and the profile has now been deleted.
I was curious to find out where the photo came from, since it had that artificial posed look commonly found in stock photos and catalog shoots.
I cropped the screen-capture down to a head-and-shoulders shot, but Google Image Search came up empty, so I tried TinEye, who matched it with a small copy of this photo on price-hunt.com. I fed that URL to Google, and it came back with an Amazon India product page that had a larger copy and several more.
Kudos to TinEye, since I only gave it a ~120x120 copy of the top third of the picture.
Whenever I’m visiting family in Dayton, I will hoover up at least one Cassano’s or Marion’s pizza. Donatos and Ron’s are other chains that make Dayton-style, but it’s pretty much unknown outside the region, and often confused with St. Louis style.
The primary characteristics are a very thin crust (yeast dough, not cracker!), sauce and toppings all the way to the edge, salt and either cornmeal or semolina flour dusted on the bottom of the crust, and cut into small squares. Basically, if it takes more than two bites to eat a piece, you cut it too big. The difference between Cassano’s and its descendants is mostly based on when their founders stopped working for Vic and opened their own stores, although it seems Cassano’s has increased the amount of salt they use on the crust over the years, to the point that some people find it excessive (not me!).
In theory, you can get frozen Cassano’s shipped to you, but it’s not a real business for them, so they charge a ridiculous amount to get it to California. It would actually be cheaper to have a case of six shipped to my parents and then re-shipped by UPS.
So I’ve been working on making my own.
I, um, kind of want to watch this version.
I’ve been running a few releases behind on GNU Emacs, because the last
time I upgraded, they removed the
count-lines-region function, for
no apparent reason.
Unfortunately, I recently upgraded with Homebrew, and the latest dumbfuckery to ruin my day is making non-regexp searches match on one or more Unicode whitespace characters when you type a literal space character. This includes newlines.
As a result, out of the box it is no longer possible to find either an
errant space, or multiple spaces where there should only be one. You
need to Google and discover how to shut it off in your
(setq isearch-lax-whitespace nil) (setq search-whitespace-regexp nil)
Names as I get a chance, as usual.
[TinEye to the rescue! Google image search couldn’t find a name for this cutie, but TinEye managed to match it to a tiny thumbnail (89x120) of the uncropped image on a defunct Russian site, and I was able to pick out the name “Azumi” and the kanji “川島”; sure enough, she’s former model and AV actress 川島和津美.]
Someone found this in one of our conference rooms this morning:
The entire wall is a whiteboard.
The projector screen is not.
(side note: Photoshop’s content-aware erase does a surprisingly-good job of redacting the text in the boxes…)
“Are your platelets cute enough?”
NSFW, because fan artists who spot a loli are on her like white on rice.
Today’s retroactive Pixiv category is おへそ (oheso), because while a shapely ankle may once have turned heads, an exposed midriff increases the odds that other interesting bits may come to light.
I is confused:
So which is it, 48 hours or five weeks? Are these pretzel bites ticking like a time bomb? Or is Pretzilla just encouraging the buyer to pig out?
[side note: Safeway has also started carrying Cheesecake Factory brown bread. For some reason, they stock both of these items near the in-store Starbucks at all the locations I’ve been in recently.]
So, I stumbled across the link that takes you to the old Items you’ve marked “Not interested” page, and by golly, my 50,000+ entries are still there, and are still being updated with new entries.
You’re just not using them to filter items from the current “Recommended for you” tiles.
In fairness, there’s a bit of load associated with slurping in 50,000+ ASIDs every time I visit the site, but you could at least use the most recent 100, and either make it possible to bulk-edit the list or simply wipe it. The old recommendation used to be “create a new Amazon account”, but that’s not practical now, with Prime, Kindle, FireTV, etc.