(no subject)

Friday, 8 May 2026 07:46 pm
gregh1983: (Default)
I had forgotten, until it happened, that Monday was actually a pretty important day in local news: the Pittsburgh Post-Gazette is under new ownership. Six months ago, the National Labor Relations Board found that the family company that owned the paper had bargained in bad faith over the terms of employees' health insurance. The union that had been on strike for three years went back to work. Ownership's response to the decision was to announce the closing of the paper for May. (To me this felt classically like the rich kid grabbing up his toys and storming home because he can't stand having to play nicely with others.) Well, the idea of a metro area the size of Pittsburgh not having a daily newspaper was horrifying all around, but there didn't seem to be any obvious answer to that problem. Finally, just last month, it was announced that the Post-Gazette would be sold almost entirely intact to a non-profit organization out of Baltimore that started a newspaper there a couple of years ago.

With Monday as the official change-over date, I found afterwards that Sunday's opinion section was largely given over to farewell messages from the outgoing editorial staff (viz. the editorial board collectively and Brandon McGinley, Adriana Ramirez, and David Shribman individually). It seems that this new Venetoulis Institute, as a non-profit, neither endorses candidates for office nor advocates for policy positions. The first part of this I understand as standard procedure for a 501(c)(3) organization, but the second half seems to go needlessly far. I donate monthly to quite a number of non-profits that advocate for policy positions all the time: it's exactly why I donate to them. A major newspaper that doesn't carefully weigh in on important public topics is missing part of its job.

Except that the Post-Gazette is still apparently going to have opinion columnists, just not off in their own section of the paper...? Here's Brandon McGinley again, datelined Monday itself, announcing the start of a new series that promises to provide political commentary and policy analysis, ha ha. So perhaps we're going to be OK after all. Either way, I'm really glad that I don't have to cancel my subscription or find another main source of local news. Having the PG operated by a journalism-specific non-profit is worlds better than having the paper sold to unfeeling private equity or a profit-squeezing billionaire — or, you know, shut down altogether! I hope it lasts and does really well.

The other roller coaster we've been on recently has been the weather. We're definitely getting April and May out of order. Last month was the warmest April we've ever had on record in Pittsburgh. I find eight or nine days (i.e. 30 percent of them) in the official data where the high temperature exceeded the 90th percentile for the day. But this summer-like warmth has been interspersed with notable frosts and freezes, especially away from the heat of the city, and this month we're looking at a seriously colder than normal trend. The airport shows two days in April and (so far) four out of eight in May when the overnight low fell below the 10th percentile.

Apparently this pattern is wreaking havoc with fruit crops all over the state, not to mention my own yardwork and that pesky little weather station that I keep behind my right eye. Every time I think I should be hardening off my tomato seedlings to plant out in the garden, I find idiotic temperatures like 2°C creeping around our upcoming forecast. Nothing's growing. Nor have I been able to make good progress on my several other outdoor projects, thanks to the combination of cold and rain and other work to do. The last straw was when I had a double-dip migraine during the first part of the week, which effectively kept me from wanting to do anything at all. Perhaps that's also playing into my sense of overall lethargy, though, that goes beyond the weather. We'll get to that next time.

(no subject)

Friday, 24 October 2025 07:42 pm
gregh1983: (Default)
Ahoy there! The bevy of this week's posts — including a few more abortive ones that I haven't fleshed out yet — will have to wait a bit, because we've reached a more important writing milestone around here. This is Statistics Day, my annual commemoration of the evening, now 21 years ago, that I sat down and coded up my first online journal post. We have been through a lot since then, yes? To the nearest integer age, half of my life's been documented here by now. Even the numerical accounting of exactly how much we've been through is turning into a hefty object in its own right. As usual it's below, with the new update for the past year way down at the bottom. Equally as usual, statistics don't include this post.

Giant statistics table )

Our latest results come as quite a shock, I must admit: if Alan had been watching, he would have caught from across the room the notably surprised look on my face at the moment when the number "150" appeared on my on-screen calculator. I knew I was headed for a pretty big writing year, but not as big as all that! (My private guess beforehand, crystallized from "more than 2009-10 and 2019-20," was 100,000 words in 125 posts.) In fact, I pounded out more words during the past year than in all but this journal's second year, and I almost took down that record as well. Only the first three years come out ahead in terms of number of posts. The difference between now and then remains my run-on verbal-diarrhea average entry length, which at 755 words still represents about a 30 percent bonus over those early days.

I'm not sure if I would have made the connection on my own, but in retrospect my lifestyle during the past year does kind of correspond to how it used to be in 2005-06. They could both be categorized as relatively free-form — the round-the-clock chaotic social life of an undergrad-mad new master's student 19 years ago, and the choose-your-own-adventure pastimes of an unemployed middle-aged scientist now. Lack of a fixed schedule is a great excuse, apparently, to spend a lot of time typing out my random thoughts. I'd also suspect it leads to there being more interesting content to write about.

Statistics Day is my typical opportunity to solicit reader feedback (if any) on any desired or undesired topics of future posts, linguistic analyses to run on my accumulated text, etc. Feel free to drop a comment. While I'm no longer quite unemployed, I can't imagine that there's going to be any dearth of things to rant and/or rave about as time goes by in the medium term. I have another big road race coming up, Alan and I have a San Francisco trip booked for early December, and I'd sure love to go skiing somewhere exciting in the mid-winter. In a more abstract sense, though I've intentionally avoided politics in here for the past 11 months, it wouldn't be foreign to my interests to cook up an essay or two on some topical line.

(no subject)

Saturday, 11 January 2025 06:13 pm
gregh1983: (Default)
Here's a fun little Pittsburgh weather factoid. Around 7 p.m. on New Year's Day, the temperature fell below 0°C; it hasn't been above since. I mention this now, 10 days later, because it's just possible that our streak ends tomorrow: the current forecast is for exactly 0°, with 2° to follow for Monday. For me this sub-zero stretch is reassuring more than anything else, since it's supposed to be cold during the middle of winter. The season is much easier to enjoy when it's not bouncing all around like a crazed superball. On the other hand, a quick check of the data also reveals that this kind of persistent freezing is actually not all that common. Since I've lived in Pittsburgh, we've had a 10-day stretch of below-zero temperatures on only six other occasions:
  • Jan. 20 to 30, 2022 — I called it "properly snowy again" on the 24th.

  • Dec. 25, 2017, to Jan. 7, 2018 — Well-remembered in my mind now, and well-documented at the time, as the 12 Days of Freezing (actually 14).

  • Jan. 5 to 14, 2011 — Looks like I was too absorbed by my upcoming thesis proposal to make any specific weather-related comment.

  • Dec. 13 to 28, 2010 — Ditto: my mind was full of the thesis proposal and holiday travel during this longest cold stretch.

  • Jan. 2 to 13, 2010 — Five of us "braved the frozen slopes of Pittsburgh" for a snow- and icicle-filled urban hike on the 10th.

  • Jan. 29 to Feb. 11, 2007 — "That just means it won't melt on the way home," "The weather has been the topic of conversation around here for the past few days," and "Yesterday it was above freezing for the first time in more than a week, and it felt like spring."

Check this space in January 2028 or 2029 for the next installation, I guess!

The weather news out of Los Angeles this week is much less amusing. I've been astonished by the pictures and videos appearing in The New York Times for the past several days. Such incredible destruction almost defies feeling — these are scenes we've come to associate more with places like Basra or Aleppo or Dnipro or Gaza than with a major American cosmopolitan city. I myself have hiked in the hills of Griffith Park and sat on the beach at Santa Monica, not too far from the scenes of the Sunset and Palisades fires. As an intellectual comparison of scale, The Times says the total area burned is now larger than the city limits of San Francisco, Pittsburgh, Boston, or Miami. But I'm finding the images of even one neighborhood or one property totally reduced to ashes much more striking.

(no subject)

Thursday, 9 January 2025 02:03 pm
gregh1983: (Default)
Recovery is definitely in progress, thank goodness! Tuesday felt like an important milestone in that regard: I was able to stay awake and out of bed in one cohesive chunk like normal during the day, and I slept essentially solid regular hours overnight as well. It's been much less important to be accompanied at all times by my own personal box of Kleenex, which is a great relief to the horribly peeling skin all around the bottom of my nose. Sufficient energy for the gym would be an ideal resource to regain next — today is my 10th consecutive day without a visit, and I'm still feeling half-and-half about attempting the task.

The past nine days have felt like such a weird suspension of life. Five years' anti-COVIDing may have under-prepared my immune system to the point that I expect I haven't experienced many colds this heavily before. Seven months' unemployment means I've never had the chance to nurse it so thoroughly before either. The combination ends up being strangely a bit nihilistic. I haven't been well enough to undertake my usual programme of doing nothing all day, so I've instead just been sitting around in bed doing nothing all day. And my prolonged absence from the scene is causing undone nothing to accumulate into a backlog of nothing. Right!

So a bit of a mixed mood, perhaps, as I attempt to jump-start my way back into... something for the new year. Given that the only fraction of 2025 when I haven't felt some form of crappy was the little hour between midnight and 1 a.m. before I went to bed on New Year's Eve, I haven't put together any serious thoughts on my personal hopes, plans, or goals yet. I did however have dreams three nights in a row now about getting job offers or being employed. Perhaps that's a sign of where to start. Or perhaps I should warm up to the task via a light full-body workout first.

(no subject)

Wednesday, 18 December 2024 07:24 pm
gregh1983: (Default)
Long delayed, but it's finally here! In this post we're going to re-visit the LABSE embeddings (i.e. mappings into some high-dimensional "semantic" space) that I created for each of this journal's 64,000 or so sentences. Last time, we tried exploring the embedding space a bit by picking out random sentences and looking at the other sentences closest by. Today, a natural extension to that line of thinking: let's cluster the whole set of sentences and see what kind of higher-level groups emerge.

Setup and procedure )

Example clusters )

Analysis and thoughts )

(no subject)

Monday, 28 October 2024 02:32 pm
gregh1983: (Default)
Fun 20th-anniversary journal analysis will now commence! Today's post should be a good learning experience all around: we're going to be exploring what connections a neural network finds among my varigated scribblings of the past two decades.

I won't get into the guts of model architecture or matrix math, but here's a high-level sketch of what I mean. These days it's easy to grab a pre-trained sentence "embedder" that represents a line of text as (say) 768 floating-point numbers, which you can think of as a vector or a point in (say) 768-dimensional space. Based on how the embedding models have been trained, it usually turns out that similar sentences — for whatever that might mean — end up with similar embeddings. Numerical similarity, at least, can be quantified and ranked... which means we can examine whether the model's formal definition of closeness really does reveal anything deep about human syntax or semantics.

So, for a bit of fun, I used a third-party Python module to automatically break the text of my first 2268 journal posts into sentences, and then I obtained the embedding of each one according to LABSE. Given that data, I picked out a number of journal sentences at random and used cosine similarity to identify the three other sentences whose embeddings are closest to each one. Now let's see some results! I'm going to necessarily select and walk through a few examples that I think are interesting, rather than bulk-present all 200 that I computed, but I hope that won't bias our conclusions too much.

Examples and discussion )

We could go on with this sort of analysis forever, but perhaps this post is long enough. Other types of slicing and dicing are still coming soon!

(no subject)

Friday, 2 August 2024 11:29 pm
gregh1983: (Default)
One of the truisms of modern life: to be alive is to experience tech problems. If the reverse is also true, then my vital signs from yesterday must be very strong indeed!

This is to say that I wanted to write a little Python script that would take a string and write out a LaTeX table showing that string's byte-by-byte representation in a character encoding of your choice. To do this, I ended up having to copy seven different object libraries from one directory to another on my desktop. I wanted put my script in a Git repository that I run out of my old home directory at CMU, you see, and it turns out that the CMU UNIX servers are now only accessible via VPN. My download of the required VPN client didn't seem to come with all the libraries it really needed; meanwhile, the copies that I found already on my system were locked away in Ubuntu's horrid new "Snap" structure, where the VPN client wasn't looking for them. Once I had my connection to CMU working, for good measure I also made a misconfiguration in my Git repo that needed further troubleshooting.

So, you know, an average sort of Thursday afternoon in the tech world... These are the kind of ever-deepening slogs I totally don't miss having to do on a tight deadline as part of a real job! So much time and effort spent on getting set up to work, to the point that you run out of hours for the work itself.

Another benefit of being unemployed today: when Kayak Pittsburgh has a last-minute 25 percent discount, you can go out and get it. Apparently it's some kind of organizational hedge against bad weather. I heard someone at the rental tent explaining that this particular discount code comes up whenever the chance of rain is above 50 percent. Certainly it got me to spend $30 at a time when I would have otherwise wasted the entire afternoon flopping listlessly at home, unsure of whether to go biking or catch up on necessary yardwork or sit inside reading a book or watch another random YouTube video, etc. Instead I booked a two-hour kayak rental from Sharpsburg starting at 4:00.

I had just enough time to kayak down to 40th Street and back, which feels reasonable but not particularly fast in a hard shell. For some reason I felt like I was having a really hard time steering the boat efficiently. It didn't seem like it was the fault of the wind, so maybe because I hadn't been back in a rental kayak for two years? The paddle is notably longer than the one I'm now used to; I expect the heavier, lengthier boat responds differently as well. Aside from that internal trouble, it was a pleasantly quiet day on the water — choppy from the wind, but not much boat traffic. I did come home unusually wet from splashing myself, even after the half-hour bike ride to nominally dry off.

Coming after my run yesterday, that's two solid days in a row of non-gym cross-training, or some kind of fancy term like that. If my old physical adage is coming true again, that I can only be good at one pursuit at a time, my body seems to have decided that the pursuit this year is running. It's been many, many years since my progress has been this easy. Last week I found a 10.4-km course that starts and ends at a city parking kiosk, which allows me the chance to time my run to the minute level if I assume that all the kiosk clocks are centrally synchronized. My first result at this distance was 59 ± 1 minutes — a fair 15 percent slower than I was a decade ago, but not as painfully slow as I was afraid I might be. Yesterday I ran the same route again in 57 ± 1 minutes, after a low- to mid-effort to go faster. No real knee trouble so far. It seems pretty clear that I've got the requisite distance for the Great Race already, and I have something like eight weeks to work further on the speed.

(no subject)

Friday, 17 May 2024 06:56 pm
gregh1983: (Default)
Tonight the origins of Greg's Journal are "Celebrating 25 Years," as they might say in the display ad. It was on May 17, 1999, that I made the first handwritten entry in this series of periodic observations on life that has continued right up until the present. Earlier this week, when I first noticed that the significant anniversary was coming up, I was getting all excited to review my journal's history — to recall my early inspirations, to quote from the original passage of 1999, to detail the moves to various typed and online formats over the years, etc. Turns out my past self has beaten me to the punch! I can do no better than to link to my similarly commemorative entry of May 17, 2019, where the whole history was already laid out five years ago.

"Periodic" is certainly the word for my pre-Internet journal. During the five and a half years when I wrote on paper, I averaged less than three entries a month and never reached more than 53 in any consecutive 12-month period. Two- and three-month gaps without a single update were pretty common up until 2002, although on the other hand I also strangely exerted myself to the tune of 16 entries during August 1999. It makes for a rough record: I certainly regret not having my early undergrad days more fully documented, for example.

Still, to have my life chronicled for a quarter of a century now is pretty impressive! A kid born the day I declared "I'm actually going to try to write a journal again" could this week be graduating from law school.

The language scientist in me eagerly awaits the day when I will have finished typing up all my handwritten material, such that I can run stats and analysis over an unbroken corpus of personal text. (Note to future self: Sorry it's taking so long. I currently haven't touched that file since October, and I'm only through 57 of the 216 or 217 total entries.) Until such far-off time as that may occur, we will of course continue the electronic journal as per normal. Greg's Online Journal will be celebrating 20 years in October!

(no subject)

Tuesday, 7 May 2024 12:33 pm
gregh1983: (Default)
Time out today for a little book review, since I just finished reading "The Great Depression: A Diary" last night and wanted to jot down some notes about it before I forget. At a high level, this book does exactly what it says on the tin: it contains "personal notes on the panic of 1929" as written down by Youngstown lawyer Benjamin Roth starting in June 1931. Roth meant to publish his journal during retirement, though he died in 1978 without getting around to it. The text was finally put out in 2009 — under the auspices of Roth's son and with a good sprinkling of editor's notes by a modern financial columnist — as a sort of topical tie-in to the then-in-progress Great Recession.

So we've got a 37-year-old professional, a journal, an interesting historical period, and a Northeast Ohio setting. Sounds like the kind of book I would like! Indeed, I found it quite an interesting read. Roth's entries are frequently brief and occasional, such that we're able to move through an entire turbulent decade (June 5, 1931, to December 31, 1941) in a very readable 252 pages. The only downside to this economy is that Roth tends to arrive at the same few conclusions over and over again; such thoughts are understandably months or years apart for the writer, but irritatingly mere minutes apart for the reader. Several times I wanted to yell at the page: "Yes, I know your preferred investment advice... Yes, I know you're worried about inflation..."

On the other hand, what really shines through in the quick movement of time is just how uncertain life in the '30s must have felt. It's so relatable to our modern era. I mean, here's a guy of essentially my own age, embarking with an outwardly brave face to delve into study in order to make sense of the insanity around him, only to be beset by recurring doubts when the insane world continues to get weirder in new and disturbing ways.

Did somebody besides me just now say "pandemic"? Although Roth's diary deals primarily with troubled economics, and it was was published at a time when we were facing a modern financial parallel, I can't read it today without comparing Roth's notes to my COVID experiences during the past four years. I am instantly reminded of my own journal posts from March 2020, not October 2008. And though the causes and events may be different, I can't help but shove a bookmark into the spine at a particularly apt passage. A 45-year-old Roth, writing in mid-September 1939 days after German tanks rolled into Poland, laments: "For 10 years I have longed for normalcy but it does not seem so destined. My generation has already lived thru war, boom and panic but evidently we still have some excitement ahead of us." Us too, my friend; us too.

(no subject)

Wednesday, 8 November 2023 10:43 pm
gregh1983: (Default)
Yes — finally a weekend with consistently nice weather! I tried to make the appropriate use of it.

On Saturday I did quite well, spending almost seven hours continuously outside. At 10:30 I left on my bike to meet Laura for a walk in Frick Park at 11. It was the first time we'd seen each other in about six months, so the outing prolonged itself until the last possible minute. Then I raced on my bike from Squirrel Hill to Garfield to meet Alan for an outdoor lunch at People's before the buffet stopped for the day. We had a quick indoor interlude at the grocery store on the way home, but then it was yardwork time immediately afterwards. The year's first leaf-raking session netted me three full bags (and just a bit on into a fourth) from the deck and the back yard; Alan contributed three buckets of weeds from my long-neglected garden and hydrangea bed.

A guy walking by on the street, who saw my lineup of yard-waste bags on the front porch, helpfully informed me that the city's one and only fall collection date for yard waste had occurred that morning. Argh! I know this event is always uselessly early, which is why I was pretty sure it was next week. Well, this year apparently it's mind-numbingly uselessly early, and I'd stupidly failed to write it on the calendar. Not that it would have mattered too much in practice even if I'd known: the leaves in our yard that have indeed fallen really only did so all in a clump on Thursday and Friday, which means the only way for me to have had anything collectable by 8 a.m. Saturday would have been to take Friday afternoon off work to do the raking. Very likely I wouldn't have done that anyway.

On Sunday Alan and I walked down to Oakland and had a special outdoor lunch at the Union Grill. The line to get into the restaurant was out the door practically the whole time we were there — it was apparently Family Weekend at Pitt — but almost no one wanted to sit at the six outdoor tables. We seemed to enjoy priority service as a result, both in getting seated and in having our food brought out. Usual trip to the public library afterwards, getting home just in time for a nice cup of tea.

That day, of course, also began this awkward bit of the year that follows from setting the clocks back an hour. I've proven to myself before that I support the shift to summer time, but I've also written before that I don't think the problem of daylight saving (let's more properly call it daylight "moving") can have a unified solution. The northness of Seattle compared to the southness of Miami makes too great a spread in the amount of daylight to work with. The eastness of Maine compared to the westness of Indiana within the same time zone makes too divergent a direction to move the daylight to. And it's trivial to find people with wildly different work/sleep/activity patterns even within a single locality.

At any rate, no matter what I think of it the sun is now setting at 5:15, and I've had my usual November transition when I briefly turn into a normal adult for a few days. On Sunday I woke up and went to bed purely according to the old times still. On Monday I appeared at work on my bike before 9:15, rather than scrambling through the morning routine in order to just make my team's 10:00 meeting. During the following two days I've been half shifted, though in a dumb way — going to bed on the new late schedule but still waking up at the old early hour. It shouldn't last long. I'm doing this business for the 41st time in my life, after all.

(no subject)

Tuesday, 24 October 2023 09:59 pm
gregh1983: (Default)
I had forgotten it was Statistics Day until I woke up this morning, saw the date on my computer, and thought "There's something about October 24..." Yes, 8:00 brain, there is. Today my online journal is old enough to legally drink in Ontario. For better or for worse, it seems like it's not just this morning that I was scatter-brained: I just spent 30 or 40 minutes copy/pasting posts into my OpenOffice document and counting them up, because apparently I hadn't been doing that for the past six months!

Anyway, with my sprawling 19-year text now onto the 1889th page, I have the honor to present to you my annual recap of just how sprawling this journaling effort has become. I'm sticking with the new approach to counting words that I settled on last year, though I am not going back to verify or re-compute any prior year's data beyond what I did last October. New numbers are therefore confined to the bottom two rows of the table. One thing that will never change: statistics don't include this post!

Year Period Days Entries Words Days/entry Words/entry Words/day
1 2004-05 366 188 102,436 1.95 545 280
2 2005-06 365 220 114,814 1.66 522 315
3 2006-07 364 179 82,638 2.03 462 227
4 2007-08 366 128 56,181 2.86 439 154
5 2008-09 365 114 74,605 3.20 654 204
6 2009-10 365 136 92,012 2.68 677 252
7 2010-11 365 118 73,348 3.09 622 201
8 2011-12 366 100 66,408 3.66 664 181
9 2012-13 365 93 60,366 3.92 649 165
10 2013-14 365 88 67,667 4.15 769 185
11 2014-15 365 85 74,322 4.29 874 204
12 2015-16 366 87 69,019 4.21 793 189
13 2016-17 365 95 75,975 3.84 800 208
14 2017-18 365 73 65,013 5.00 891 178
15 2018-19 365 54 49,650 6.76 919 136
16 2019-20 366 118 91,456 3.10 775 250
17 2020-21 365 114 82,625 3.20 725 226
18 2021-22 365 80 62,973 4.56 787 173
19 2022-23 365 101 79,843 3.61 791 219
All 2004-23 6939 2171 1,441,351 3.20 664 208

Kind of a mixed year, right? My own private guess — based on the limited information I allow myself to access in order to make it — was that I'd written 85,000 words in 100 posts. I was thinking that this year was roughly on par with either 2016-17 or 2020-21 in terms of total volume. It absolutely is, in terms of both post count and word count, so I'm not sure why I tricked myself into predicting a larger number of words than either of those other two had. Anyway, it's nice to be posting a little more often again compared to last year; on the other hand, my habitual inability to control the post length has run unchecked again. That's obviously the fault of those massive travel logs from August and September. Perhaps that makes it even more important to thank my Dreamwidth audience (currently 21 active accounts) for following along with my exploits during another trip around the sun!

I have no particular plans for this journal's upcoming year, other than to keep writing it. It's been ages since I've run playful textual analysis or explicitly asked for content suggestions as part of my annual checkpoint. (Of course, feel free to send in any ideas if you have a particular request!) Some kind of grand jubilee probably ought to take place a year from now, when we hit the big 2-0. Until then, stay tuned for the usual mix of middle-aged gripes, activity reports, vague work updates, yardwork concerns, husband dates, and the odd social outing. After all, as it said on side of the semi truck that was delivering windows across the street this morning, "Life is for living!"

(no subject)

Wednesday, 28 June 2023 08:56 pm
gregh1983: (Default)
Call it officially the Summer of Smoke. Those Canadian wildfires — now about 275 in number nationwide, according to The New York Times — are sending their by-products down this way again. Pittsburgh is having a more intensive exposure this time around compared to three weeks ago. The fun centered yesterday on Chicago; our local air-quality index passed the 100 mark around 9 p.m. last night, and this morning I put the trash out this morning under the worst AQI reading I've ever experienced at 196. Our 9 a.m. report was 210, which seems to have been the peak so far. The rest of the day's been largely spent in the 160s and 170s. Not exactly prime condition, and still worse than anything I saw here earlier in the month, but at least this morning's pervasive smell of campfire has dissipated.

(P.S. Do we not agree that this "air-quality index" really ought to be re-named something like the air-degradation index? The higher the number, the lower the quality of the air. This is as backwards as saying a UV index of 1 is the strongest sun or that a Category 5 hurricane is the weakest wind.)

As luck would have it, I was requested to go into the office today for a big team brainstorming meeting. If this is the way our climate is going, on top of the COVID pandemic, I guess I now need to learn how to make other kinds of major health decisions at an individual level with no data. Our house is quite old and absolutely not air-tight... but is my super-sealed office building with the constantly running HVAC system better enough that I can still come out ahead after spending 40 minutes outside and 50 minutes on the bus to get there and back? How about if I wear a mask outside, in addition to as per usual in the office?

In the end I went for the trip and the outdoor mask. It was certainly strange kicking around with the whole optical world fading to an intermediate grey. Buildings more than a block away were visually obscured, and downtown was hardly visible at all from the South Side Works. Pittsburgh was well represented in the lead story on The New York Times all day, under the headline "Smoke from Canadian fires blankets Great Lakes region." I tried grabbing a few photos on my way down to the bus, from the bus window, and when I stepped outside to grab some lunch. The Times did a better job, though: for a while the first photo on the entire website was a shot of the downtown skyline from the West End Overlook.

There was a proposal among the five of us teammates in the office to go out for drinks after our three-hour meeting. On a "normal" day in June, I would have of course advocated for this activity to take place on an outdoor patio or something, such that I might happily join in. With the air like it was today, it made a fun little choice between risk of COVID in a crowded bar and risk of pollution at an outdoor table. I didn't see a good way through it and so just excused myself from the event. Sometimes nowhere is safe! Or perhaps I've gone off the deep end into tinfoil-hat territory. On this day when major authorities are warning people to stay inside with the windows shut and to avoid outdoor exercise, I counted on my way home from the bus stop and found only three mask-wearers against 17 without. This morning there was a wonderfully fit-looking guy running with perfect smoothness past my office building. I hope this pursuit didn't cost him more health than he gained by it...!

(no subject)

Saturday, 29 April 2023 12:14 pm
gregh1983: (Default)
I just counted that I have 77 open browser tabs, so I really need to process some of the ones whose sole purpose is to remind me to write about them.

Several weeks ago, Owen sent me an absolutely incredible song-and-dance video advertisement for the Montreal metro system. They definitely don't make public transit commercials anymore the way they used to in 1976! I don't know how many total viewings I spent laughing over this thing — first over the visuals, then over as many of the lyrics as I could understand, and finaly over the rest of the words as revealed by the YouTube closed captions. The close translation of "Il fait beau dans le métro" would be "The weather's nice in the subway"; in this instance I prefer the more general "It's nice in the subway," as the song ultimately turns into a broader set of pro-transit gushings.

As befits such a rare piece of silliness, the '70s advertisement has taken on a life of its own on the modern Internet. Among many others, there's a straightforward imitation from 2012 and a 2010 parody ("It's hot in the subway") protesting the lack of air conditioning. Some inventive rhyming and good imagery there: the original second line "Everyone's cheerful; everyone's heart is in the sun" turns into "Everyone's hot; everyone's sweating down to their toes."

Previous users of the Montreal system may recognize that the start and the end of the song refer musically to the three-note progression of sounds you hear on the trains as the engine gets moving. I remember this property well from my first visit to the city in 2002. It's a great little Easter egg, though maybe not as subtle as it would have been when I was riding: French Wikipedia says that they're now officially using those tones inside the train as the "door is closing" signal. There's a cool audio clip of both side by side, along with some explanation if you read the accompanying paragraph.

(no subject)

Monday, 24 October 2022 09:52 pm
gregh1983: (Default)
I successfully had my Montour Trail hiking adventure on Friday and Saturday... but the report of that is going to have to wait. Instead it's Statistics Day — the closest thing we have to a national holiday in Greg's Journal Land, when regular posts are closed and we spend the time numerically examining everything that's gone before. This is the date it all began, 18 years ago tonight. So while the journal fills in its voter registration card and starts considering its options for the 2022 midterm elections, let's have the old giant table out and see what new row we can add to it. Year 18, down at the bottom there, covers up to but not including this current post.

Year Period Days Entries Words Days/entry Words/entry Words/day
1 2004-05 366 188 102,621 1.95 546 280
2 2005-06 365 220 116,725 1.66 531 320
3 2006-07 364 179 83,463 2.03 466 229
4 2007-08 366 128 57,380 2.86 448 157
5 2008-09 365 114 74,612 3.20 654 204
6 2009-10 365 136 90,289 2.68 664 247
7 2010-11 365 118 73,629 3.09 624 202
8 2011-12 366 100 65,970 3.66 660 180
9 2012-13 365 93 60,629 3.92 652 166
10 2013-14 365 88 67,900 4.15 772 186
11 2014-15 365 85 74,566 4.29 877 204
12 2015-16 366 87 69,259 4.21 796 189
13 2016-17 365 95 72,623 3.84 764 199
14 2017-18 365 73 65,013 5.00 891 178
15 2018-19 365 54 49,650 6.76 919 136
16 2019-20 366 118 91,456 3.10 775 250
17 2020-21 365 114 82,625 3.20 725 226
18 2021-22 365 80 63,659 4.56 796 174
All 2004-22 6574 2070 1,361,700 3.18 658 207

The first thing I noticed when filling in the new data is that this has been a predictably bad writing year. I say "predictably" because last week I made a private guess that the numbers would turn out around 64,000 words in 75 posts. Not bad, if I do say so myself! I knew that I had a pretty significant dearth of entries during the late summer, which seems to often happen, and that the kind of pandemic-fueled unmooredness that was causing the past two years to be so well documented was notably wearing off. Still, I'm not exactly pleased to find I just completed my third- or fourth-worst year ever, depending on how you count, with only the likes of major depression and Amazon Translate to keep it company.

The second thing I noticed in doing some basic sanity checks on the arithmetic is that something fishy is going on with the word counts. The sum of those 18 numbers comes to 1,362,069, not the nice and round 1,361,700 that I get from Tools > Total Words in LibreOffice. Well, that's interpretable enough, as I already discovered in 2019: with all the different operating systems and software tools this journal's been through over 18 years, naturally the word-count definitions used in my "current" setup will have evolved as well. The exact same text for (say) 2007-08 that registered as 57,380 words back then in OpenOffice 2 might easily ring up as 56,181 today in LibreOffice 6.

Now I never actually fixed the discrepancy for Statistics Day in 2019, 2020, or 2021; I just commented on it in the first two years. So this evening I went through my giant LibreOffice document, selected all the posts for Year 1, ran the word count on that, selected all the posts for Years 1 and 2 combined, ran the word count on that, etc. — thinking to get a clean set of statistics that would at least be consistent with each other. For a while this was making so much sense. The counts were fluctuating by moderate amounts for every year through 2016-17, followed by a string of no changes. ("Aha, that must be when LibreOffice's current definition came in.") And then this most recent year — the one whose word count I computed this morning in the exact same version of everything that I have tonight — was off by 686. This is how I discovered that I can hit Control-A to select all the text in a document, run the word counter, and see the mystifying message "Selection 1,361,508; Document 1,361,700"! So what part of "Select All" do I not understand? On Alan's suggestion, I tried pasting that full selection into a new document... which after some number crunching told me it now contained only 1,280,182 words, or magically 1,361,508 again if I did the Select All over there!

Clearly, this way madness lies, and I don't suppose I could make sense of this muck short of picking apart the source code. Let's just standardize on the method of highlighting document prefixes, up to and including the whole thing. Here's a newly converted table:

Year Period Days Entries Words Days/entry Words/entry Words/day
1 2004-05 366 188 102,436 1.95 545 280
2 2005-06 365 220 114,814 1.66 522 315
3 2006-07 364 179 82,638 2.03 462 227
4 2007-08 366 128 56,181 2.86 439 154
5 2008-09 365 114 74,605 3.20 654 204
6 2009-10 365 136 92,012 2.68 677 252
7 2010-11 365 118 73,348 3.09 622 201
8 2011-12 366 100 66,408 3.66 664 181
9 2012-13 365 93 60,366 3.92 649 165
10 2013-14 365 88 67,667 4.15 769 185
11 2014-15 365 85 74,322 4.29 874 204
12 2015-16 366 87 69,019 4.21 793 189
13 2016-17 365 95 75,975 3.84 800 208
14 2017-18 365 73 65,013 5.00 891 178
15 2018-19 365 54 49,650 6.76 919 136
16 2019-20 366 118 91,456 3.10 775 250
17 2020-21 365 114 82,625 3.20 725 226
18 2021-22 365 80 62,973 4.56 787 173
All 2004-22 6574 2070 1,361,508 3.18 658 207

Teenagers, right!? This will be the new version of record, at least for now, as my journal journeys on into adulthood.

(no subject)

Monday, 21 March 2022 11:00 pm
gregh1983: (Default)
Daylight Savings Time is lingering in the national news this week, quite beyond its usual temporary visit from the act of actually setting our clocks forward eight days ago. The Senate passed a bill by an amazingly bipartisan margin to keep us on this accelerated time zone year round. While both the House and the president seemed supremely non-committal, everyone else has felt free to chime in and keep the fight going over whether year-round DST is the right answer, or whether it should instead be year-round standard time or some kind of individual or business-level free-for-all.

Someone called Peter Coy capped off this madness in The New York Times today, with a blithely carefree column called "I Think I Just Solved Daylight Saving Time" — which is just asking for your over-confident opinion to be shot down instantly, isn't it? Personally, my thinking had led me to the exact opposite conclusion: I don't believe this "problem," such as it is, is solveable in any kind of a consistent way.

Communities on the eastern side of a time zone, for example, may be biased towards wanting the daylight later, so the sun doesn't set at 4:11 p.m. like it does in Boston. People in Grand Rapids, meanwhile, probably prefer to shift the sun earlier because they have to wait for it as late as 8:22 a.m. In Miami the length of daylight only varies by a little more than three hours over the course of the year, and there's never less than 10½ hours of it: there's not a real problem for Daylight Savings Time to fix in tropical locales. A Seattle December gives a scant ration of 8½ hours of light while June floods you with 16: at these limits, the DST question is again pretty irrelevant because it won't change what you see out the window at the beginning and end of your day anyway. Besides, beyond any regional geography, there are always going to be individually early-shifted and late-shifted people everywhere, each clamoring for the permanent solution the other group hates.

I guess this is where Peter Coy comes in, with his idea that we should each decide how, when, and if to change our own schedules across the year. Employees should just get their companies, in this nirvana, to kick operating hours around as the seasons change. "For example, if school starts later in the winter, that would prevent working parents from getting to the office at the usual hour," he claims, with momentary sanity. "In the past that would have been a career killer. Now, for many, it's business as usual." This is clearly the view of someone who's never had to work hourly. Does he really expect that anyone with a service job could get, say, Walmart to play made-to-order calendar Tetris? And if it did happen, and your in-person job started at 8 a.m. in September but 9 a.m. in February — with kids' school schedules that may or may not match depending on the school system's own ideas, not to mention your co-worker's family who lives in the next school district over — is that not just an even worse version of the current "disrupt your routine twice a year" approach that we all profess to hate?

So yeah, any of these "better" plans being proposed really just feel to me like a personal-preference question of "better for who?" In my ideal version of the world, I would probably take year-round DST pretty easily. It'd put our latest Pittsburgh sunrise at 8:43 a.m., which is a bit borderline but still only around the time I'd be having breakfast. No sunset earlier than 5:53 p.m. would be a nice boost to my late-shifted self to make it feel less like I was at work hours and hours into the winter night. I also don't have a big problem with the current system, which I've lived with for 38 years and after all was probably designed for the population centers of the Northeast such as I now inhabit. But if you have even the ghost of an idea of taking away my summer evenings via year-round standard time, making the sun rise at dang 4:49 in the morning — which is straight-up the middle of the night — I'll pitch incessant ruddy fits and have to start sleeping in a cave. That is not the kind of prospect that leads to rational discourse.

(no subject)

Monday, 14 February 2022 08:56 pm
gregh1983: (Default)
Fads, right? I'm still amazed at how much we all suddenly care about five-letter words, thanks to the Wordle craze. I've seen now an IPA variant, a semantic (word embedding) variant, a version where you play four words at once — and "Worldle," a game with no relation to anything where you're trying to guess a country based on its shape, and the feedback tells you the distance and direction of your error. None of these oddities, though fun, has seriously made it into my daily routine; I'm still just playing regular Wordle in English and French either before, during, or after breakfast.

In English I got an enjoyable surprise last Thursday in the form of my first 2. I don't remember the word now, but my initial guess came back with so much information that I just went straight for a reasonable answer without trying to explore any new letters. Conversely, in Saturday's game I had all the letters (but none of the positions) after three tries but couldn't wrap my brain at all around the resulting anagram problem. I had to burn guesses just to narrow down placement before I finally and sadly logicked my way to ULTRA on the sixth row. My overall streak now stands at 42, with that one 2, eight 3s, 18 4s, 11 5s, and four 6s. The overall average is therefore a slight improvement, from 4.4 to 4.2, since the last time I checked.

In French my main motivation has been to see if I even know enough words to play the game in a foreign language. My strategy's been perforce a little different — interestingly a little more like "hard mode" where you have to use the information gained from one guess in all subsequent guesses. It's supposed to prevent you from being able to spam letters, but in French I find that it's actually kind of a helpful guide towards the portion of my vocabulary that might match the answer. After 19 games, my stats are a little worse than for English: they would be a perfectly symmetric distribution around 5 except for a fluke 3 that I got on the same Thursday as my crowning English success.

The biggest surprise to me in French is that I haven't missed a word yet, including three that I didn't recognize. Trying pronounceable things and seeing what's accepted is a huge help there. Thank goodness that entering a non-word doesn't count against the number of guesses! In this way I've now at least temporarily learned that bramer means "to bellow" or "to cry out" ("particularly of deer during the time of love," notes my favorite French–French dictionary), that broyer means "to crush" or maybe "to mash," and that the English word nurse exists in French specifically in its meaning of "nursemaid" or "governess."

Another amusing result is just how many five-letter French Wordle words have potential five-letter translations for use in English Wordle too. CREDO, INDEX, FORGE, and NURSE work straight off. VENTS, FOLLE, and ARENE are clearly "winds," "crazy," and "arena." Then TIEDE, ENTRE, and TIERS have reasonable in-context translations to "tepid," "enter," and "third." It's not even such a terrible stretch to have CROIT as "think," SAINE as "sound," TARIF as "price," or our old friend BRAME as "cries." If we're allowed to ignore singular/plural distinctions, I could add plural JUPES as singular "skirt" and then singular TANTE as plural "aunts." And that is practically the lot. LINGE means "laundry," which I can't see any way to cram into five English letters. BROYE can only be the composed-past broyé, which sadly requires the six-letter "mashed" or the seven-letter "crushed" instead of a neat "crush." But maybe Scrabble players can rescue our final word LOUIS, because it's the name of an old currency in France, and therefore the English name would be the same and lowercased.

(no subject)

Monday, 22 November 2021 10:05 pm
gregh1983: (Default)
There seems to be no end to the number of small daily life experiences that have been messed up by the pandemic. I'm thinking especially of the so-called staffing shortage this evening. When it comes to face-to-face service jobs, my read on the situation is that companies laid off huge numbers of their employees during the shutdowns and periods of reduced demand last year, and now they have to get a whole new crew of people back again. Meanwhile, the people who are at a loose end from being let go — and even those who have been continuously employed — are of course re-evaluating along with everyone else what they really want their lives to consist of these days, according to various dimensions. That all adds up to our continued state of scrambling and shuffling.

Sometimes it feels like stepping into an alternate universe. Most of the workers at my gym were different, when I came back again in May, from who I remembered seeing pre-pandemic. At my dentist's office in June, the receptionist and the hygenist who'd taken care of me for years were both gone. In this version of the world, our late-night grocery option is Whole Foods — which just proves what a topsy-turvy reality we're in right now. This setup can't be meant to last the ages: we'll all go bankrupt in short order. But Aldi's is still closing at 8 p.m. (an hour early) and Giant Eagle at 9:00 (three hours off), leaving the local bastion of upper-crust professional hipster food with a solid monopoly over the "Dang, we're out of yogurt" crowd until the now-generous hour of 10 p.m.

What worries me about this state of affairs — aside from fancy cardboard-wrapped imported organic non-BST grass-fed yogurt with psyllium husk or avocado toast on the side, at $1.89 per newly reduced serving size, I mean — is that it seems rather difficult to reverse. Pittsburgh has already been shockingly bad at supporting all-night anything. It was already close to impossible to get proper dinner after 10 p.m. even pre-pandemic. If the grocery stores and my gym keep enjoying how much money they save on wages and benefits by locking the doors at 9, I'm not sure if we'll ever have the momentum to open them back up again. One of the reasons why we live in cities is to have amenities close and convenient when we need them; I'm correspondingly a little extra put out whenever I discover an area where that isn't true.

(no subject)

Friday, 5 November 2021 06:54 pm
gregh1983: (Default)
I haven't yet internalized the exact cadence on which it appears, but recently I discovered the fishing report in the Post-Gazette, and it is a laugh and a half! What an alluring mix of unexpected poetry, inscrutible jargon, and lines just ripe for misconstruction. I read through the column two Saturdays ago (October 23) and couldn't stop shaking my head and chuckling. Of course I have to now present a few choice excerpts...
Lake Arthur (Butler County): Shore anglers reported hookups with perch, crappies and other panfish using live baits. Catfish took shrimp. Walleye went for crawler harnesses pitched from shore over submerged road beds.

Ohio River (Allegheny County): Flatheads took fish heads below Chartiers Creek. A 22-inch channel catfish was caught and released below the Highland Park Dam.

Pine Creek (Allegheny County): With the stream almost to himself last weekend, a veteran spin angler from Wexford released a 30-inch common carp caught while drifting weightless crawlers through grass. The Delayed Harvest Artificial Lures Only stretch received a delivery of state-stocked rainbow trout Tuesday.


This is English, right? I suppose "took" in this case means that such-and-such a fish was observed successfully eating such-and-such other object as bait. But I kind of want to ignore the informational content and just enjoy the lovely meter in bits like "Catfish took shrimp" — this is the "Jesus wept" sentence — and "Flatheads took fish heads." I'm enthralled by drifting weightless crawlers as mental companions to Chomsky's colorless green ideas. And what pertinent news regarding whatever bit of river that is where organic or natural lures are prohibited and where, apparently, you have to wait a while before collecting your bounty! I guess if I ever wanted to have a hook-up with a crappie, and my toilet at home is clogged, I know where to go. It took me so long to chug my way through all this delightful madness that it was only on my second reading that I realized the Highland Park Dam isn't even on the Ohio River.

There's another report out today, I find, which I think reaches its pinnacle with the statement "Crappies were at 18-20 feet and took tipped jigs over structure." If you can understand what that means, you're better than both me and any NLP tool I've ever seen. Are tipped jigs better foodstuffs than structure? Do we have perhaps a novel cooking technique to replace "over easy"? Maybe "tipped jigs over structure" is as common on the bait menu as "marinated pork over rice" is for people? Do the crappies only bite when there's a bit of submerged scaffolding underneath them? Or is that just where the fish carried their jigs off to? The possibilities are endless.

(no subject)

Sunday, 24 October 2021 04:40 pm
gregh1983: (Default)
"Statistics Day, Statistics Day. You shout 'No way! It's Statistics Day?'" It is. Greg's Online Journal turns 17 years old tonight — on par now with the lifespan of your high-school seniors — so for the 17th time I'm taking a moment to add up the total number of posts and words I've contributed to this old thing year by year. The new numbers for the past 365 days appear down at the bottom of the table below. As always, they cover up through and including yesterday's post, but not today's.

Year Period Days Entries Words Days/entry Words/entry Words/day
1 2004-05 366 188 102,621 1.95 546 280
2 2005-06 365 220 116,725 1.66 531 320
3 2006-07 364 179 83,463 2.03 466 229
4 2007-08 366 128 57,380 2.86 448 157
5 2008-09 365 114 74,612 3.20 654 204
6 2009-10 365 136 90,289 2.68 664 247
7 2010-11 365 118 73,629 3.09 624 202
8 2011-12 366 100 65,970 3.66 660 180
9 2012-13 365 93 60,629 3.92 652 166
10 2013-14 365 88 67,900 4.15 772 186
11 2014-15 365 85 74,566 4.29 877 204
12 2015-16 366 87 69,259 4.21 796 189
13 2016-17 365 95 72,623 3.84 764 199
14 2017-18 365 73 65,013 5.00 891 178
15 2018-19 365 54 49,650 6.76 919 136
16 2019-20 366 118 91,456 3.10 775 250
17 2020-21 365 114 82,625 3.20 725 226
All 2004-21 6209 1990 1,298,041 3.12 652 209

Well, there it is! The 2020-21 writing year turned out to be rather better than I thought. Before counting up, my private guess was that I'd managed about 72,000 words in 90 entries. (This was by taking 2016-17 as a baseline, because it seemed to have generated HTML at very much the same pace as I saw for this year, with some adjustment to the post count because I really didn't expect that I had succeeded in reducing the length of my entries at all.) Instead we have favorable results in all categories: actually 82,625 words in 114 entries, with a nicely reduced average length.

In fact it's essentially the same number of posts as the year before, which is probably the biggest surprise to me. My feeling is that the pandemic itself and my resulting altered lifestyle is still leading me to write more than the 50,000- or 60,000-word lean years... but certainly not as much as I put down during the spring and summer of 2020, when the rapidly changing situation and the mental shock of it all generated a lot of text. This year, July and August contained between them only 10 posts, for example. But we are still not looking quite like Year 5 either, which equals the latest Year 17 in post count but came with commendable — though certainly depression-influenced — brevity.

Seeing the running post totals of 1876 (last year) and 1990 (this year), I can't help but instinctively interpret these figures as years. This has happened to me many times before with SVN commit IDs, our team's issue tracker at work, etc. It seems to take hold somewhere around the mid-1700s and fall away again during the mid-2000s. So today I'm mentally mixing my journal stats up with Ulysses S. Grant and not-quite my brother. I see I've also got to stay on high alert about a month from now, when I'm expected to write this journal's 2000th post; it would be a shame to miss that! Otherwise, we'll check in on this stuff again in another year, when my journal will be old enough to vote.

(no subject)

Saturday, 29 May 2021 05:59 pm
gregh1983: (Default)
Partial book review: "Same-Sex Unions in Premodern Europe" by John Boswell.

Very little in terms of spoilers, but still long details )

Profile

gregh1983: (Default)
gregh1983

May 2026

S M T W T F S
     12
34567 89
10111213141516
17181920212223
24252627282930
31      

Syndicate

RSS Atom

Style Credit

Expand Cut Tags

No cut tags
Page generated Thursday, 9 July 2026 03:51 am
Powered by Dreamwidth Studios