Where I'm at on social media (literally and philosophically)

I now have a Bluesky account (@ImpudentStrumpet.bsky.social) and a Mastodon account (@ImpudentStrumpet@mastodon.social), and I continue to be on Twitter (@ImpStrump).

I've tried to move my follows over, but if I missed you it wasn't intentional.

 

I haven't made any decisions about which I'm going to use or how or how much. My communities seem to be leaning towards Bluesky, but its interface doesn't play nicely with my post-head-injury eyes and brain. I have discovered various web clients (turns out you can just search github for things you're wishing someone would code!) but nothing is consistently comfortable.

 

(I use a combination of GoodTwitter2 and Minimal Theme for Twitter to make Twitter comfortable. It looks like this. I'd welcome any tips on how to achieve this with Bluesky or Mastodon.)

***

I have been finding Twitter less useful lately. Either there are fewer tweets or I'm seeing fewer tweets. Quite frequently I pop in, idly scroll, and quickly reach the point where I last left off. (That hasn't happened for years!)

Previously, if I heard a weird noise or something outside, I'd search for "Yonge and Eglinton" (my neighbourhood) in Twitter and get other people tweeting about whatever weird noise I'd just heard. (e.g. they'd be saying something like "Why's there a helicopter circling Yonge and Eglinton?" and then I'd know it's a helicopter). Now, I just get real estate listings.

 

I'm finding more pornbots in my mentions and fewer actual people. I'm finding TV livetweeting hashtags less active. Basically, it's just not meeting my needs as well.

 

If my needs were just entertainment, I wouldn't be worrying about this. However, after nearly 15 years of curation, I've gotten to a point where my Twitter feed effortlessly provides me with new information that I didn't even know I needed and would never have thought to proactively seek out. 

From what it means when someone asks "What are your pronouns?" to the fact that COVID is airborne and can be mitigated with robust HEPA filtration, important things I didn't even know that I didn't even know reached me as I was idly scrolling for weather updates and puppy pictures, and I'm a more informed person better able to function in society because of it.

I fear that my social media might be doing that less well now, and I might not even notice.

(I'm also noticing something similar with Reddit - of stuff that's reaching me organically, the ratio of people who are less informed that me to people who have something to teach me is worsening.)

***

And the scary thing is how easy it would be for me to live in ignorance. I'm older. I'm established. I'm comfortable. I have leeway and credibility and social capital. I'm intrinsically pessimistic and secure in my flaws.

If someone asked me "What are your pronouns?" and I didn't know what they meant, I'd be flummoxed and baffled. Someone would explain what they meant, I'd tell them "she/her" and apologize with grace for not knowing what they meant on the grounds that I'm a milquetoast middle-aged lady. I wouldn't suffer any long-term consequences, and might even get credit for handling it with grace rather than harrumphing over it. 

And I wouldn't even realize I'd missed something - I'd just think this is yet another thing that didn't reach me until it reached me.

 

If I didn't know that the COVID protections currently being required by public health were insufficient and ended up contracting COVID as a result, I would shrug my shoulders and figure "Well, bad things happen. People get sick. Why shouldn't it happen to me?" Even if I developed Long COVID, my head injury has already taught me that life-altering medical consequences can happen for reasons completely outside your control. 

 

I'd be miserable, my life would be worse every day, and I wouldn't even realize I'd missed something - I'd think this is yet another bad thing that happens to people.

I'm sure there are already tons of things I'm missing, but I fear that the enshittification of Twitter is exacerbating it compared with if Twitter had continued to be run with the same competence as before the takeover.

***

I've also been thinking Tumblr a lot. 

Tumblr was bought out for a ridiculously high price in 2013 by disruptive management that made a lot of unpopular changes. It bled users, and was sold in 2019 at a massive loss. 

Conventional wisdom is that Tumblr is dead, but the fact of the matter is the community is still there.

I'm not on Tumblr (the interface and format never really met my needs), but I follow some people's Tumblrs in my feed reader, and a lot of Good Omens fandom happens there so I do keep an eye on it.

And the community is still there, still thriving, still being weird with their blorbos and their Goncharovs. They outlasted the occupiers and drove them off.

Maybe we can do the same thing with Twitter?

***

I've quipped that I don't call Twitter "X" for the same reason that I don't call Gdańsk "Danzig".

 

Gdańsk has, obviously, been through some shit over the centuries. Irrevocable harm was done, many did not emerge unscathed (to put it mildly).

And yet, today, it is unquestionably called Gdańsk. Not Danzig.

 

I haven't given up hope that we can do the same with Twitter.

But I've also hedged my bets and secured my pied-à-terre elsewhere.

New Twitter personal best

NBD, NBD, just Neil Gaiman taking a moment out of his busy day to personally reply to my tweet so I can make safe and informed media consumption decisions




Only in the Title credits.

— Neil Gaiman (@neilhimself) June 2, 2023

And, on further reflection, for a little bit in Episode 4. Avert your eyes when Furfur uses the clicker.

— Neil Gaiman (@neilhimself) June 2, 2023

Things They Should Invent: filter online shopping products by weight

Latest pandemic malfunction: my 18-year-old TV finally gave up the ghost.

I bought a new TV easily enough, but it's a bit too big for my existing TV stand. So I'm shopping for some kind of TV stand or table or cart or something to put it on. 

Problem: the products that catch my eye keep being heavier than I can lift. 

Under normal circumstances this isn't so much of a problem. Normally, we can have furniture delivered. Normally, it's not a huge imposition to ask someone to pop by and help me move or assemble something. 

But during a global pandemic, this isn't an option. My building's pandemic rules prohibit delivery people from coming up to apartments, instead telling them to leave the deliveries at the concierge desk and residents will bring them up. My building's pandemic rules also prohibit visitors, and public health rules are also telling me not to have contact with other households. (Sometimes public health rules let single people bubble with another household, but there are zero people in my life whose risk factors permit visiting me and aren't already bubbled with another household.) So during the pandemic, I'm limited to what I can lift myself and assemble myself.

Online shopping sites could help me with this by letting me filter products by weight, so I only see those that are light enough for me to bring up to my apartment myself and assemble myself.

The websites already have this information - it tends to be listed right under dimensions.

The websites already let you filter by various variables, such as price and size. I can already tell the website "show me all the TVs under 35 inches", so why not "show me all the TV stands under 40 pounds"?

Building on this, they really should let you filter by any characteristic that is listed on the site. Country manufactured, inseam length, number of USB ports, anything. People have all kinds of oddly specific requirements, so, especially in this pandemic era where more shopping is being done online than ever before, why not let us pinpoint exactly what we need?

The Toronto Star should print URLs next to QR codes

During the pandemic, I've been reading the epaper versions of the my newspapers rather than getting my usual home delivery, and I've noticed an annoyance: links to further information on the Toronto Star site are provided as a QR code, without a corresponding URL provided.

This is an annoyance by itself in the print version, because it only gives you the option of opening the link on a mobile device, even if you'd prefer to read on a computer.

But it's all the more annoying in the epaper version, because epaper readers are already reading on their preferred device for reading a newspaper electronically! If I'm reading on my computer like I usually do, I could, theoretically, grab my phone and scan the QR code. But what if I was already reading on my phone? Surely there are many households that don't have extra mobile devices just sitting around unused for every time you want to click a link!

If the Star would simply print URLs next to (or instead of) QR codes, everyone could access the links by the means most convenient to them, thereby maximizing the number of eyeballs on the Star's website. Using the QR code alone is inconvenient to many and impenetrable to some. There's no reason not to continue printing URLs, like they have since the advent of URLs.

The mystery of the semantically null Amazon reviews

I was recently researching a potential purchase and reading a lot of Amazon reviews to do so, because Amazon had the most reviews for the most different products.

 

And I discovered something really weird: five-star reviews that aren't actually positive reviews of the product, and very often provide no information whatsoever.

 

Examples: "I bought it as a gift from someone else." "I haven't received it yet." "I haven't tried it yet but I'm sure it's fine."

 

Why would you write a review like this when you could just . . . not?

 

 

Then the universe provided me with what might be the answer!

 

I try to avoid Amazon whenever possible because of their labour conditions, but I ended up ordering a couple of products there because I couldn't find them anywhere else. 

Each of these products had a card inserted in the packaging saying the seller would give me an Amazon gift card if I wrote a review of the product, and sent them an email with the order number and a link to the review.

One offered me a $10 gift card for a review of a $25 purchase, and the other offered me a $15 gift card for a review of a $35 purchase.

 

I can certainly see how this might incentivize people to leave a review even if they have nothing to say!

I didn't leave any reviews so I don't know if they actually send you the promised gift card. (Apparently my sense of "I don't want them to win!" is worth more than $25 to me.) I don't know if they only give the gift card for a five-star review (the cards didn't say anything to that effect) or if these semantically null reviewers think Amazon is like eBay and sellers will be penalized for reviews that are less than five stars.

But, in any case, there's something very, very wrong with the system if sellers are incentivized to turn over 40% of the purchase price in exchange for a review!

The Segoe UI font is easy to read with convergence insufficiency and accommodative insufficiency resulting from head trauma

Recently, for the first time since my head injury, I received a text to translate that I could actually read effortlessly!

I immediately checked what the font was, and it turns out it's called Segoe UI.  I changed my default Firefox font to Segoe UI, and now life is so much easier!

So if you're ever looking for a font that's effortlessly readable with convergence insufficiency and accommodative insufficiency resulting from head trauma, Segoe UI fits the bill!

I'm working on adding Segoe UI to my blog's style sheet so everyone can enjoy its effortless readability (just plugging it in turned out to mess up the font size, so I need to do some tweaking).

However, I'm not the audience of my blog - in fact, I am the one person in the world who has the least need to find my blog readable! - so if at any point you find the font has become less readable, leave me a comment to let me know, ideally articulating the specific problem (too big? too small? too fat? too skinny? too much of some other characteristic that I can't even fathom?)

Things They Should Invent: "you guys" gender map

Some people perceive "you guys" as masculine, even in the vocative case.

Others, such as myself, perceive it as having no element of gender.  "You guys" is a casual, inclusive vocative plural, completely unrelated to the masculine nominative singular "guy".

But I'm not here to convince you that I'm right.

I'm here to convince someone to map it.

One of the great moments of internet sociolinguistics is the Pop vs. Soda map, which shows the geographical patterns of American soft drink nomenclature.

Someone should do the same for whether "you guys" is masculine or gender-inclusive!

Based on the way people on the internet talk about the "you guys" question, I strongly suspect there's some geographical element to how it's received.  A crowdsourced mapping project, like Pop vs. Soda, could answer this question.

The technology exists, as evidenced by Pop vs. Soda. The answer would be informative, and help people better tailor their communication to various audiences. Surely there must be someone out there looking for a linguistics research project idea!

Things Twitter Should Invent: retain hashtag capitalization

When you search for or click on a hashtag in Twitter, it shows you a feed of all tweets with that hashtag.  Useful!

If you then click on the "Tweet" button while this hashtag feed is open, it populates the tweet composition box with the hashtag in question, on the assumption that you're going to tweet using the hashtag in question. Useful!

Problem: Sometimes the hashtag that populates the tweet composition box is written in all lowercase, even if the hashtag you originally searched on or clicked on was written in a combination of capital and lowercase.

This is an issue because screen readers use the capitalization in hashtags to determine when a new word starts, and writing hashtags in all lowercase makes the screen reader attempt to pronounce the hashtag as all one word. And, aside from that, #CapitalizingEachWordLikeThis is easier to read than #writingthewholethinginlowercase.

What Twitter should do: make sure that any capitalization in the hashtag clicked on or searched for is retained when populating the tweet composition box. This means people who have already made the effort to make their hashtags accessible don't have to repeat that effort every time they tweet.

While writing this blog post, I tried to determine the specific conditions under which Twitter retains the capitalization of the hashtag versus when it changes it to all lowercase, and I wasn't able to pinpoint it with any consistency. All I can tell you is sometimes it retains capitalization, and sometimes it goes all lowercase.

However, the fact that it sometimes retains capitalization means that retaining capitalization is technologically possible, so Twitter should make that happen all the time.

Things the LCBO Should Invent (or, rather, re-implement)

1. Indicate on your Favourite Products list whether each product is available in your preferred store.

2. Indicate on your Favourite Products list which products are on sale.

Both of these functions were available in the previous version of their website, but were eliminated in their last website update, and it really cramps my style!

I was thrilled when the Favourite Products list was introduced, because it made the way I shop so much easier!  Whenever I particularly enjoy a product, or hear about a new product that sounds interesting, I add it to my list.  Then, next time I'm going to buy a bottle, I simply scroll through my list, see at a glance what's both on sale and in stock at my local store, and buy whichever of those items best meets my current needs.

But since they changed their website, I now have to click through to each item to see if it's on sale, and to see if it's in stock.

The new website design seems more focused on trying to get you to order online, but that also cramps my style.

Online orders take multiple days to ship, and even ordering online and picking up in store takes multiple hours. In contrast, my local LCBO store is literally across the street, so I can go buy a bottle and be back home in under 10 minutes if the lines aren't too long.

Also, the LCBO website has a minimum order of $50, whereas I tend to buy only one (cheap) bottle at a time for personal use.

(Given the LCBO's mandate, perhaps they shouldn't be incentivizing buying larger quantities all at once?  I've heard that alcoholics will drink all available alcohol, so if anything the LCBO should be incentivizing buying only one bottle at a time!)

They had a fantastic feature that met my needs perfectly, and then changed it for no apparent reason!  I wish I knew how to convince them to bring it back!

Things They Should UNinvent: italics as default blockquote style

Many advice columns put reader letters in italics. This is a problem, since advice column letters are often multiple paragraphs long, and italics are more difficult to read than regular text.

Since my head injury, I've been finding paragraphs of italics so difficult that I need to switch Firefox into Reader View or turn on OpenDyslexic. (Or I just go "Ugh, blah blah whatever" and skip that column.)

Most often, the letters are in italics because that's what the style sheet does with blockquote.  Unfortunately, that makes the quoted matter difficult to read when there are multiple paragraphs of it.

I would recommend that style sheet designers instead have blockquote differentiate quoted matter with some combination of indentation, design elements adjacent to the quoted matter (I've seen large quotation marks or vertical bars used to good effect), or different font colour (while taking care to choose a colour that is also easy to read).

If there are special circumstances where certain devices can't render these effects, then those devices can come up with their own suitable way to render the blockquote tag.  But the default should be easily readable, and style sheet designers should be mindful of the fact that italics are not easily readable for all, especially when there are multiple long paragraphs.

Things They Should Invent: clothing characteristic search engine

Clothing geeks have a lot of nuanced opinions about which design features you should look for in clothing.  They'll mention things like woven vs. knit fabrics, bias cut, darts, and other fabric, pattern and construction features that I don't even understand.

The problem is it's difficult to look for clothing with these features. Even if you understand them, you can't reliably search an online clothing store for bias cut dresses, or efficiently look through a bricks-and-mortar store for bias cut apart from scrutinizing the fabric of every single garment.

Even those of us who aren't clothing geeks develop an idea over time about which design features work best on our bodies.  For example, I have learned from experience that I love modal sundresses. But typing modal into a website search bar is hit and miss (does zero results mean they have nothing in modal, or that you can't reliably search by fabric?  Or does it mean the modal sundress is labelled under the broader category of "rayon"?)  And to search for it in a real-life store, I'd have to read every single label.

Solution: a single, comprehensive search engine that does a fine, granular search of every single clothing store by clothing characteristic.

You can search for red and v-neck.  You can search for princess seams. You can search for Irish linen.  You can search by measurement, to screen out all the pants that you can't even pull up over your hips.  And it will give you the results for all the stores, or for all the stores with physical locations in your city, or all the online stores that ship to your city, or all the online stores that ship to your city with free returns.

Clothing stores' incentive to participate (and stores' and manufacturers' incentive to provide detailed information) is that this will tell potential customers that the store has the thing the customer is looking for. There are hundreds (thousands?) of clothing stores - dozens in my neighbourhood alone - and the very red v-neck or modal sundress I'm looking for could be right there in one of them, hanging in a place where passers-by can't see it.

Google Shopping could do it, I'm sure.  Currently it only seems to cover some stores, all of which appear to be major chains.  But if they could use their power and influence as the search engine of record to tell literally every store "Give us your clothing specs in granular detail, and the customers looking for what you're selling will find you".  (I'd also be thrilled to see someone non-Google do it, but Google more likely has the influence to get all the stores to participate.)

The Dyslexie font helps me read with convergence and accommodative insufficiency resulting from head trauma

The biggest problem with the visual issues that came with my head injury (for the googlers: my diagnosis was convergence insufficiency and accommodative insufficiency) is that I can no longer effortlessly skim large, dense pages of text, especially on screen.

Instead of "YAY, a new article/blog post/fanfic!", my visceral reaction is "AAAAH! Wall of text!"  It takes work to read things that would previously need no work, and all too often I'd just skip interesting articles or stories that I would otherwise enjoy, because I don't have it in me to put in the work. (Especially after a full day of translating, where I have to put in the work).

When I was researching vision therapy, I stumbled upon the fact that many child vision therapy patients are initially diagnosed as dyslexic. Apparently the difficulties with reading look the same to external observers, and, since the children have never been able to read, they can't tell the difference between "pulling the letters into focus and keeping them there is hard" and "reading is hard".

I remembered seeing something about a font designed for people with dyslexia. Even though I'm not dyslexic, maybe it could also help me?

It turns out there's a browser extension called OpenDyslexic that lets you toggle the Dyslexie font on and off.  So I gave it a try, my eyes instantly relaxed, and I could once again effortlessly skim and absorb everything.

When I hit a wall of text, I just toggle it on, and suddenly it's back to being effortless to read!  It doesn't change the formatting like Firefox's reader mode does, it's the same layout and design as the regular webpage - just with this funny-looking font instead.  It even works with dynamic, constantly-updating pages like Twitter!

Because Dyslexie is funny-looking, I don't like to use it all the time. When the combination of font, layout, spacing and colours is such that I don't struggle to read, I actually find Dyslexie intrusive. Fortunately, the OpenDyslexia browser extension makes it effortless to toggle on and off, so I don't have to choose.

I don't know if the improvements I experience with OpenDyslexia are specific to my post head trauma convergence insufficiency and accommodative insufficiency, or if the design of the Dyslexie font can make dense text more skimmable to anyone. But if you struggle with walls of text for any reason, it might be worth giving OpenDyslexia a try. It's free, it takes seconds to install, and you can switch if off instantly if it doesn't help.

Why I'm not happy with the Weather Network latest website redesign

Weather Network 7-day forecast

The default page for each city is the 7-day forecast, shown to the right. (Click to embiggen. The long and narrow shape is the result of Firefox's awesome screenshot function, which allows you to take a screenshot of the full page, rather than just what appears on screen.)

At the top of the page are the current conditions.  That part's good - that's exactly the information I'm looking for.

However, the next thing I'm looking for is the short-range forecast, which isn't there.  There are two small boxes below the current conditions giving a brief summary of the next two 12-hour periods (labelled "Tonight" and "Tuesday" in this screenshot), but that isn't sufficient information. At a minimum, I'm also looking for humidex/windchill (labelled as "Feels like" in these screenshots) and probability of precipitation (labelled "POP" in these screenshots), but they don't have that information on the default page for the short-range forecast. They just have those stingy, inadequate summary boxes with way too wordy a description and way too little quantitative information.

I do want to see the long-range forecast on the main page as well, and it's right there in a format that makes me happy, just below the row of news videos.  But without a proper short-range forecast, there's a gap in the information provided.

Weather Network 36-hour forecast

The short-range forecast can be found on the 36-hour page, shown to the left.  (Click to embiggen).  And all the information I'm looking for is right there, in a format that makes me happy, in the table just below the row of news videos.

However, the current conditions at the top are incomplete. They  have the sky condition with the temperature and humidex, but that's it. No wind speed, humidity, air quality, UV, etc.

This is a problem, because now I have to have two tabs open to get all the information I want, especially when I have weather-sensitive outdoor plans, or in shoulder seasons where I have to make multiple decisions throughout the day about heating/air conditioning, windows open/closed, blinds open/closed to keep my home comfortable.

For example, I'm currently trying to find a good time to wash my windows.  To do this, I need to know the current temperature, humidex, wind, humidity and sunset time, all of which are in the current conditions on the main 7-day page, but not all of which are on the 36-hour page.  I also need the temperature, POP, and wind for the next couple of days, all of which are on the 36-hour page but not the main 7-day page.  So what was a simple at-a-glance task with the Weather Network's old design now requires two tabs.

The best thing the Weather Network could do to fix this is remove the two small boxes ("Tonight" and "Tuesday" in the 7-day screenshot) from the 7-day page, and remove the row of news videos. Then they should put the 36-hour chart from the 36-hour page in their place.  This would give us the same at-a-glance skimmability we had on the old website.

If it really is important to separate 7-day and 36-day, the second most useful thing the Weather Network could do is put full current conditions on the 36-day page. This would provide a single-page at-a-glance of the information that updates most frequently throughout the day, and whose updates are most immediately relevant.  (In other words, if the overnight forecast changes, that becomes relevant to me far earlier than if the forecast four days from now changes.)

If they really, really, really can't do either of those things, one very simple thing they definitely can do is put humidex/windchill information in those two inadequate short-term boxes on the 7-day page ("Tonight" and "Tuesday" in the screenshot.)  They have the information, it appears in every other place in the forecast that mentions temperature, and there's room in the boxes.  I have no clue why they chose to omit it in that one very specific location, but that would be easily remedied.

And if they want a bold, innovative option, they could let users customize their own homepage, with the forecasts and data of their choice.  This would have the additional benefit (from the Weather Network's point of view) of incentivizing users to create accounts and stay logged in.  They've been trying for ages to convince me to create an account and I haven't seen the need to, but I'd do it in an instant if that were the price of admission for all the at-a-glance information I want on one page.  The technology exists - iGoogle did it in 2005!

Things They Should Invent: Stuff You've Already Tried filter

On my old computer, I had to do a clean reinstall of Firefox. A couple of issues subsequently cropped up, and when I googled around the new issues, I kept finding advice to do a clean reinstall. That's what caused the issues in the first place!  (With the combination of the new computer and the new version of Firefox, the issues are now moot.)

When my old computer died, it simply wouldn't power up. Pressing the power button had exactly the same result as not pressing the power button. The troubleshooting of first resort in this case is a power reset: unplug the computer, remove the battery, hold down the power button for about 30 seconds to drain any residual electrical charge, then plug in the power adapter only and try again.

I tried that several times, and it didn't work.

And my attempts to google for the next steps in troubleshooting were stymied by interference from instructions for a power reset. I found a single reference to replacing the CMOS battery (haven't tried that yet because the age of the computer and the low likelihood of success made me prioritize getting a new computer), but, even with my advanced google-fu (and trying other search engines as well), I couldn't get away from the pervasive suggestion of a power reset to the rest of whatever the appropriate troubleshooting protocol would be.

Our internet usage patterns are increasingly being spied on. Couldn't they at least make use of this data to give us the option of filtering out the stuff we've already tried.

Profiled

Last spring, I experienced thigh chafing for the first time in my life.

Due to my disproportionately long inseam and dislike of the current trend of tight pants, it turns out that on most of my pants the gusset fell below the bit at the top of my thighs that was chafing, meaning that the legs of my pants couldn't serve as a barrier to protect me against the chafing. What few pants I owned that did have a high enough gusset were made of unpleasantly rough or unbreathable material, which may have even made matters worse.

I clearly had immediate need of softer pants, and they probably needed to be more fitted so the gusset would stay right at the very top of my thighs and protect the area being chafed. But the last thing I wanted to do when every step was painful was go pants shopping!  So I went to multiple websites with generous return policies and ordered multiple pairs of yoga pants, one of each plausible pair in my usual size and one a size smaller.

Over the next few days, a wide selection of yoga pants arrived at my door. I tried them on and kept everything that worked for me.  It turned out my idea of going a size down was unnecessary (I hadn't bought new yoga pants in years and they're more fitted now than they were last time I shopped for them), so I returned everything in the smaller size and some of the things in the larger size.

Shortly after that happened, I started getting coupons and offers and recommendations for maternity wear.  I guess I triggered an algorithm somewhere - frantically shopping for yoga pants and opting for the larger size in every case is totally something a pregnant lady would do! 

This was all about nine months ago.  And now I'm getting coupons and offers and recommendations for baby things!  Even though I haven't bought yoga pants or maternity wear or anything comparable in the meantime, apparently online shopping algorithms are the kind of people who count months.

I wonder how long this will persist for? Will I be getting offers for toddler things for a few years, followed by back-to-school offers and high school graduation offers?  Will they start trying to sell me those conception monitors if I don't shop like a pregnant lady for a few years, on the grounds that my non-existent child should have a sibling?

Maybe I should use Privacy Mode when googling for baby gifts just in case...

How Google can solve the "post-truth" problem in one easy step

Google searches contain the option to refine results time posted. On the results page, click on "Tools", then click on the little drop-down arrow next to "Any time".

This means that Google maintains "last updated" metadata for the pages it crawls.  Which means that Google can sort results by date.

Google can use this power to combat the "post-truth" problem with one easy step: allow users to sort search results from oldest to newest.  That way, the very first instance of a particular combination of keywords will be right at the top.

This will make it a lot easier to see when a story or an alleged fact has been fabricated out of whole cloth, because the first result (or, at least, the first result that actually refers to the thing in question) is very recent and originates from the person making the false statement.

It would also be an incredibly useful feature to have in Google's Reverse Image Search. Often I do a reverse image search to find the origin of an image that's circulation, but the fact that even Google's relevance algorithm tends to favour novelty means I get pages and pages of results from social media. If we could easily show the oldest instances of an image first, we could quickly identify cases where someone is posting "This is what's happening right now" when really it's an image taken in a different country several years ago.

Google already has this data, as evidenced by the fact that it allows you to refine results by time posted. Any computer can sort by date. All Google has to do is put an "Oldest First" option on its interface, and everyone will be able to fact-check with a single click.

Things They Should Study: do political positions correlate with attitudes towards politically-incompatible celebrities?

Sometimes the celebrities I follow on Twitter get people telling them to shut up about politics and stick to entertainment.

This is something I find difficult to understand. 

I do see why someone might not want incompatible political opinions turning up in their Twitter feed.  But what I don't understand is why you'd want to keep following someone once you know that they hold these incompatible opinions.

When someone has incompatible politics (by which I don't mean simply that I don't agree with them, but rather that I see their position as outright harmful and/or cruel) I'm not able to respect them enough to be a fan of them. I cease to be interested in their day-to-day life and thoughts, and most likely in their work as well.  Even if for some reason I do maintain interest in their work (for example, perhaps if one member of an ensemble cast for a major fandom has incompatible politics) I no longer have any desire to hear from them as an individual, just to see the finished work.

It would be interesting to study this on a broader level.  Are there any patterns of the political opinions or affiliation of people who want to continue following politically-incompatible celebrities but not hear about their politics, as compared with people who lose interest in politically-incompatible celebrities, as compared with people who can cheerfully continue following a celebrity without regard for their incompatible politics.

They could also study whether there are patterns in real-life relationships as opposed to celebrity-fan relationships, but I find the celebrity-fan relationship particularly interesting because it's unidirectional. If a parent holds political opinions you consider harmful, there's an element of "How can you bring a child into the world and then work politically to make the world a worse place?" But the celebrity has no loyalty or attachment to the individual fan and the fan adores the celebrity, so it's an interesting and unique dynamic.

Things I did invent!

For years and years, I've been telling the universe to invent things for me. This week I shut up and invented the things for myself!

1. Since I discovered the Toronto Fire Active Incidents page years ago, I've gotten in the habit of checking it whenever I hear a siren, just to see what's going on. However, not all sirens are the fire department.  So I was going to write a Things They Should Invent that someone should merge the Toronto Fire Calls map and the Toronto Police Calls map (as well as ambulance data, if it is available) into a single "What's that siren?" map.

Making a map isn't in my immediate skill set, but people who are smarter than I am have already turned these data streams into twitter feeds. So I made my very first twitter list, which shows all police and fire calls in near-real time (there's about a 5 minute delay). So now when I hear a siren, I just pull up my list and within moments the answer to my question will appear.

(Although if anyone is feeling ambitious or creative, I still think a map would be a better interface).

2. There was some visible sediment in the reservoir of my coffee maker.  Neither running vinegar through the machine nor rinsing it out would budge it, so I figured it needed to be scrubbed. Unfortunately, since it's only a 4-cup coffee maker, the reservoir is small enough that I can just barely get my hand in and couldn't move it around in the way I needed to to scrub the sediment. A bottle brush wasn't soft enough, and that sponge-on-a-stick thing that's like a bottle brush but with a sponge was too bulky. I thought a q-tip would be about the right size and texture, but I couldn't get my hand in properly to manipulate it the way it needed to be manipulated.

I was going to write a Things They Should Invent of extra-long q-tips for these kinds of cleaning challenges, but then I had an inspiration:

I took a cotton ball (the kind you use to remove makeup or nail polish), stuck it on the end of a fork like it's a meatball, and used that to scrub the inside of the coffee maker reservoir.  The cotton was the right texture, the fork gives it the kind of stiff support you need for scrubbing, and the fork was long enough that I could manipulate the movements of the cotton fully because my hand could be outside the reservoir. The whole thing was perfectly clean in about 20 seconds!


I've never before been able to actually make one of my Things They Should Invent, and this week I made two in one week!

Research/Journalism Wanted: what's up with the people who didn't see it coming?

This post is about the information that reaches people (including me) organically, without them making any effort to find it, as opposed to the full set of all information available.  While reading this, you may find yourself thinking "But you don't have all the information! You're just talking about the subset of information that reached you organically!" Yes, and that is exactly what this post is about.

In the wake of Brexit, my twitter feed has been showing me examples of people who voted Leave but were unaware of the consequences. I was rather surprised by this, because I was aware of those same consequences, and I haven't even been actively following the issue!  The information reached me with no effort on my part (and, in fact, despite my having mentally categorized it as To Disregard), but it didn't reach people who actually got to vote in this referendum, and would have voted differently if they'd had this information.

Someone should do research and/or journalism about these people. What did they think was going to happen? Where did they get that idea from? Were they given incorrect information, or just not given all the correct information they needed? Why didn't the information they missed reach them?

And, perhaps most importantly, how close did they the information get to reaching them? Was a friend of a friend on a social network posting the information they needed? Was it in the newspaper they read but on a boring page they just skimmed over?  Or were they nowhere near it and would have needed to drastically revamp their media consumption practices and/or voting research to have reached it.

After interviewing as many of the people who didn't see it coming as possible, the researchers/journalists should publish the results, highlighting any patterns they noticed.  This would serve two purposes: helping regular people see information consumption patterns that correlate with being less informed than one would like, and helping people who are trying to spread information or raise awareness see how to reach the people who would like to be more informed but don't even know it yet.

As a random made-up example, suppose 68% of the people who were misinformed got their incorrect information from their hairdresser. Then people would know that you should question/snopes/factcheck political information provided by your hairdresser, no matter how brilliant she is about doing your hair.  Or, suppose 68% of people who didn't get the information they wanted were two degrees of social media separation from that information. Knowing that, people might retweet links to political information that they normally wouldn't retweet because they think it's glaringly obvious.

And this isn't just a Brexit thing. Similar postmortems should be conducted for all elections, and for any other undertaking where they can find a significant number of people who didn't see it coming.  For Brexit we're hearing the morning after about the people who didn't see it coming, but the turnaround isn't always this fast. They should follow up after six months or a year, find people who didn't see it coming, and figure out why.

There's something wrong when the desired information doesn't reach people who will be voting in a referendum, even though that same information organically reached a random foreigner who is deliberately disregarding information on the issue. Investigating exactly how this happened is probably the first step to making the problem go away.

Warning: LinkedIn could machine translate your profile without your permission!

I was shocked to receive an email from LinkedIn cheerfully announcing "We have translated parts of your public profile into German!"

This is a major problem for me for two reasons: because it's machine translation, and because it's German.

Machine translation is a problem because I'm a translator. The presence of anything that sounds like machine translation in my profile is harmful to my professional credibility, because it makes it look like I can't translate and/or can't judge what constitutes good translation.

German is a problem because I've never worked professionally in German. My professional experience is in the domestic official languages market, so if I were to have another language in my profile, it should be French.

(The French isn't already there because I'm not actively networking, so I maintain a very minimal LinkedIn profile - just enough information for people who already know me to distinguish me from my doppelnamers.)

My bare-bones English-only profile makes me look like an Anglophone who isn't actively using LinkedIn. Many multilingual people who don't make full use of LinkedIn have unilingual profiles, so it doesn't parse as significant. But an English and German profile makes me look like someone who is actively seeking work in English and German, but considers French not sufficiently relevant to bother with.  That would be off-putting to people looking for the kind of French-English translation that is my bread and butter! An English-German profile marks me as irrelevant to the official languages market before you even look at the content of the profile, while the content of the profile renders me irrelevant to any non-desperate client on the English-German market.

I'm fully aware of the argument for having a fully fleshed-out multilingual profile, and I made a deliberate choice not to do so at this time.  However, I did not make a deliberate choice to have a machine-translated profile or to have an English-German profile, and it's assholic of LinkedIn to impose that on me.  That would be like if they noticed that I don't have a photo in my profile, so they did a google image search for my name and inserted the first result, or if they noticed that I haven't put where I went to high school so they populated that field with data scraped from classmates.com. Only this is worse, because people may well evaluate a translator's translation skills based on the quality of the translation in her profile.


If LinkedIn does in fact have German-speaking users who want to access English-language profiles via machine translation, the option to machine translated should appear on their interface, like it does (to varying degrees of success) with Facebook and Twitter.  They should see my profile as I wrote it, with a little "Translate into German" link that they can click on, making it apparent that the German machine translation is the result of their having clicked on this link, not in any way something I wrote.


If LinkedIn has also machine translated your profile, or if you want to prevent it from doing so in the future, you can opt out on your public profile settings page. Specific instructions can be found here.