2023-02-17 08:02:01 <veltas      > KipIngram: Having some fortran fun?
2023-02-17 08:08:27 <KipIngram   > Not too much.  I just copied a silly little random number programm from a web page and compiled and ran it, just to make sure it worked.
2023-02-17 08:25:21 <KipIngram   > Oh, interesting.  Turns out none of us have ever seen a Mercator projection of the entire planetary surface.
2023-02-17 08:25:35 <KipIngram   > Because in a correct Mercator projection the poles are at infinity.
2023-02-17 08:25:46 <KipIngram   > Map would have to be infinitely tall.
2023-02-17 08:26:33 <KipIngram   > I knew things were stretched left/right on a Mercator map, but I didn't realize there was vertical stretching too.
2023-02-17 08:26:44 <KipIngram   > I guess now that I think about it it makes sense, though.
2023-02-17 08:27:35 <KipIngram   > Just learned what a "rhumb line" is.  It's a path on Earth along which you hold a constant compass bearing.  Depending on the direction you go, you'll move toward a pole along a spiraling path.
2023-02-17 08:29:32 <KipIngram   > https://upload.wikimedia.org/wikipedia/commons/d/d6/Loxodrome.png
2023-02-17 08:38:35 <KipIngram   > Rhumb lines wrap around the poles an infinite number of times, and yet the length of a rhumb line, pole to pole, is finite.
2023-02-17 08:49:46 <KipIngram   > Before the invention of the marine chronometer, navigators often sailed a special case of a rhumb line, the special case being to stay on a constant line of latitude.
2023-02-17 08:50:22 <KipIngram   > It's really easy to see how achieving quality timekeeping at sea sooner than everyone else gave the Brits an edge in maritime activity.
2023-02-17 08:50:32 <KipIngram   > They could get places faster.
2023-02-17 08:52:05 <KipIngram   > One not always talked about aspect of the voyage of the Bounty was that one of Captain Bligh's duties on that trip (before the mutiny, obviously) was to evaluate the performance of a particular prototype marine chronometer.
2023-02-17 09:05:22 <KipIngram   > I suspect Bligh's performance on that voyage was less than ideal - no doubt he carries some of the blame for how things went.  But that doesn't change the fact that his navigation after they were left adrift saved those men's lives.  That was an amazing achievement - getting them to safety with the resources he was equipped with.
2023-02-17 09:11:47 <KipIngram   > Anyway, as usual for this kind of thing, the Wikipedia article on rhumb line is quite nice.
2023-02-17 09:12:14 <KipIngram   > Long as your topic of interest isn't a controversial one Wikipedia generally works very well.
2023-02-17 10:16:17 <veltas      > Yeah I love clockwork watches, really amazing technology
2023-02-17 10:17:39 <Zarutian_iPad> I wonder if a leaky bucket on a ballance with bell was ever a thing
2023-02-17 10:17:44 <KipIngram   > Yeah - very much.
2023-02-17 10:18:06 <KipIngram   > I've also found the Wikipedia article on the pendulum clock really interesting too.  Fascinating history.
2023-02-17 10:18:24 <KipIngram   > And up until the 1930's it was far and away our best timekeeping technology.
2023-02-17 10:18:36 <Zarutian_iPad> probably such bucket setup was not accurate enough
2023-02-17 10:22:14 <KipIngram   > What was particularly interesting was the stream of improvements made to pendulum clock tech in the years following the initial discovery.
2023-02-17 10:22:27 <KipIngram   > Ways of compensating for thermal expansion and things like that.
2023-02-17 10:22:42 <KipIngram   > At first it made no sense to have a minutes hand - they weren't accurate enough.
2023-02-17 10:22:49 <KipIngram   > But then some decades later that got added.
2023-02-17 10:23:16 <KipIngram   > Then a little mechanism for adjusting your clock for local gravitational g value, etc.
2023-02-17 10:24:21 <KipIngram   > A particularly interesting one used a reservoir of mercury for the pendulum bob, and it had a glass tube that the mercury would move up and down in as temperature changed.  That changed the mass distribution, and they sized it just right to compensate for other thermal changes.
2023-02-17 10:24:48 <KipIngram   > I just thought that was particularly clever.
2023-02-17 10:26:03 <KipIngram   > They also developed a special one that resulted in a kit that maritime folks could carry around with them and use to measure local g.  For a period of time in there is was standard practice to do a g measurement whereever they went, so that a big catalog of such values could be compiled.
2023-02-17 10:26:30 <thrig       > why does it sway northwards at Karachi, one wonders
2023-02-17 10:26:41 <KipIngram   > That kit was made so you could flip the pendulum arm end over end, and you did a measurement with it turned both directions, and somehow the average of the two was the number you wanted.
2023-02-17 10:27:52 <KipIngram   > What's special about Karachi?
2023-02-17 10:29:24 <thrig       > the large mass to the north
2023-02-17 10:29:44 <KipIngram   > Oh, is there like a big mountain up there or something?
2023-02-17 10:30:26 <thrig       > there was a small pile-up between the indian plate something else
2023-02-17 10:30:35 <KipIngram   > Interesting.
2023-02-17 10:31:50 <KipIngram   > So, I'm almost embarassed to tell this, but I will anyway.  One of my daughters just moved to North Carolina to go to graduate school.  And she had heard about this thing: https://mysteryhill.com/attractions/natural-gravitational-anomaly/
2023-02-17 10:32:17 <KipIngram   > So she went to see it and now she swears it's legit.  And I just don't see how it can be - I figure there's got to be some trick going on.
2023-02-17 10:32:29 <KipIngram   > But I don't care enough to go there myself.
2023-02-17 10:33:08 <KipIngram   > Something as pronounced as that photo depicts just seems beyond reasonable to me.
2023-02-17 10:34:41 <KipIngram   > So I feel like I somehow failed to deliver a suitable level of skepticism to that daughter, or something like that.
2023-02-17 10:48:23 <veltas      > When is someone writing TTF Forth https://developer.apple.com/fonts/TrueType-Reference-Manual/RM05/Chap5.html
2023-02-17 10:52:01 <veltas      > DUP, SWAP, ROLL, sound familiar?
2023-02-17 10:53:45 <KipIngram   > :-)
2023-02-17 10:54:24 <KipIngram   > You'd think they could pat at least a bit of homage.
2023-02-17 10:54:30 <KipIngram   > pay
2023-02-17 10:54:59 <KipIngram   > Since they kind of just got their language pre-packaged for them.
2023-02-17 10:55:48 <KipIngram   > That's fairly interesting material, actually.
2023-02-17 10:55:57 <KipIngram   > Nice usage of fixed point.
2023-02-17 10:56:19 <thrig       > STOP DROP ROLL oh wait different fire drill
2023-02-17 10:56:31 <KipIngram   > hahahaha
2023-02-17 11:00:14 <KipIngram   > I expect studying how they've tacked that particular problem in geometry would be instructional.
2023-02-17 11:00:35 <KipIngram   > Chuck always claimed fixed point was enough, if you arranged your setup intelligently.
2023-02-17 11:01:00 <KipIngram   > Sort of implied floating point is laziness.
2023-02-17 11:01:54 <KipIngram   > I don't necessarily disagree - if you keep track of what your exponents would be by virtue of knowing your application, then you've got room in your numbers for considerably more precision.
2023-02-17 11:02:35 <KipIngram   > So it definitely seems like you could write superior code if you do it that way.
2023-02-17 11:03:09 <KipIngram   > But... more for the programmer to keep track of.
2023-02-17 11:04:09 <thrig       > starting forth gives a calculator example but there was a bunch of conversion (bugs?) work at various points to keep it within 16 bit
2023-02-17 11:04:26 <KipIngram   > And in the occasional case where a parameter can vary over a wide enough range to make that approach problematic, you could always add an extra variable and manage its range yourself.
2023-02-17 11:04:59 <KipIngram   > Yeah, I'm sure setting out to manage those aspects manually opens the door to many more error opportunities.
2023-02-17 11:05:59 <thrig       > of course floating point isn't without hilarious problems
2023-02-17 11:06:32 <thrig       > such as negative standard deviation if you math... poorly
2023-02-17 11:06:46 <Zarutian_iPad> dec64 and posits enter the ring
2023-02-17 11:06:51 <KipIngram   > Oh yeah, you can definitely stub your toe.
2023-02-17 11:08:55 <KipIngram   > And when you get some physical system that has chaotic dynamics things get extra fun.
2023-02-17 11:09:08 <KipIngram   > And you just can't get a good answer beyond some certain point in time.
2023-02-17 11:09:45 <KipIngram   > Where that point in time is depends on your precision, but the important point is that the precision required is exponential in the time forward you try to look.
2023-02-17 11:09:49 <thrig       > I hear the three body problem is good for book sales
2023-02-17 11:09:55 <KipIngram   > :-)
2023-02-17 11:10:01 <KipIngram   > The three body problem is fascinating.
2023-02-17 11:10:30 <Zarutian_iPad> thrig: yeah but the next two books suffer from the Second Album Syndrom
2023-02-17 11:10:50 <KipIngram   > Sequels always struggle to live up.
2023-02-17 11:11:02 <KipIngram   > Unless your the Alien or Terminator franchise.
2023-02-17 11:11:14 <thrig       > good thing they only made that one matrix movie
2023-02-17 11:11:24 <KipIngram   > In both of those cases I think the sequel at least matched and maybe surpassed the original.
2023-02-17 11:11:34 <KipIngram   > Exactly - thank you, thrig.
2023-02-17 11:11:48 <thrig       > those were horror -> blockbuster shifts
2023-02-17 11:11:55 <KipIngram   > Though... Monica Bellucci...
2023-02-17 11:12:06 <Zarutian_iPad> well there is a band that only releases a new album every two or three years. Basically the same amount they spent on the first
2023-02-17 11:12:09 <KipIngram   > I'll watch those sequels just to look at her.
2023-02-17 11:12:37 <thrig       > xkcd://566
2023-02-17 11:12:42 <KipIngram   > Movie-wise they're pretty bad, though.
2023-02-17 11:14:13 <KipIngram   > The Matrix is a movie I make a point to watch every couple of years.
2023-02-17 11:14:21 <Zarutian_iPad> I rather liked the line of the Merwincian: cursing in French is like wiping your arse with silk.
2023-02-17 11:14:25 <KipIngram   > I think it's more pertinent to the real world than some folks realize.
2023-02-17 11:14:36 <KipIngram   > In a metaphorical way.
2023-02-17 11:15:16 <KipIngram   > We are in a matrix - a web of customs and protocols that we can't get out of without "disconnecting" ourselves from the world.
2023-02-17 11:16:25 <Zarutian_iPad> ”so hopelesly innerved and dependent on the system that they cannot survive outside it” comes to mind
2023-02-17 11:16:47 <KipIngram   > Yes, and that bit about you see the matrix when you pay your taxes, go to church, etc.
2023-02-17 11:17:00 <KipIngram   > That's the point - some of the dialog from the movie is just so spot on it's scary.
2023-02-17 11:17:50 <KipIngram   > Ok, maybe I need to watch that this weekend.  It's been a while.
2023-02-17 11:18:40 <thrig       > if it's really the matrix then why did they not cover eigenvalues
2023-02-17 11:21:20 <KipIngram   > Would have gone over everyone's head.
2023-02-17 11:27:09 <KipIngram   > It's really not a very complicated concept, but I rarely see it explained in terms that expose that simplicity.
2023-02-17 11:28:36 <KipIngram   > It's usually presented in linear algebra terms.
2023-02-17 12:48:10 <veltas      > thrig: That's precisely the point, floating point is very complicated and expensive, people have no idea how much so
2023-02-17 12:51:14 <veltas      > Go look at the formula for sin and cos in glibc, there are like 5 different cases, and no normal programmer can even understand what's happening without some long winded article
2023-02-17 12:52:24 <veltas      > The work is so much simpler in floating point but it requires slightly more thought on the part of the programmer
2023-02-17 12:52:59 <thrig       > yeah but most programmers ain't thinking and have time pressures
2023-02-17 12:53:28 <veltas      > Well in real life I have seen programmers opt for fixed point and fuck it up endlessly, when floating point would have worked
2023-02-17 12:53:46 <veltas      > And the reason they did this is because they had an inherent distrust of floating point and didn't understand it
2023-02-17 12:54:40 <veltas      > And the truth is I'm quite possibly one of the only engineers in a floor of like 20-30 programmers who understands it at work, although I have been training people in it
2023-02-17 12:55:22 <veltas      > Floating point really matches calculations that you do with "significant figures", and fixed point for "decimal places".
2023-02-17 12:55:48 <veltas      > Most scientific stuff is SF, engineering often DP is sufficient
2023-02-17 12:55:59 <veltas      > And maths can be all of the above
2023-02-17 12:59:11 <veltas      > Pretty much none of the calculations I find interesting need floating point, personally.
2023-02-17 14:57:53 <KipIngram   > I agree with your "Well in real life" comment.  For *most* applications floating point will just take care of you, and certainly will do so better than fixed point.
2023-02-17 14:58:04 <KipIngram   > Yes, you can still get in trouble with floating point.
2023-02-17 14:58:11 <KipIngram   > But it works just fine in a lot of cases.
2023-02-17 15:00:06 <KipIngram   > I didn't mean to be advocating for or against either one (floating or fixed) - just noting that there is a tradeoff.  Taking some of the thinking onto yourself can let you achieve greater precision with a given number of bits.
2023-02-17 15:00:24 <KipIngram   > Because you're mentally tracking something instead of using part of your number to do so.
2023-02-17 15:02:35 <KipIngram   > By the way, I certainly would not claim any particular level of "expertise" in floating point nuances.  I've read a couple of the well-thought-of papers on the subject, so maybe I'm generally aware of some of the potential issues, but I've mostly just used floating point for rather friendly applications.  I've never really had to bring out heavy artillery on that front.
2023-02-17 15:02:48 <KipIngram   > Maybe if I needed to I'd know how to go find the information I needed.
2023-02-17 15:02:58 <KipIngram   > But it's not in short term storage.
2023-02-17 15:03:59 <KipIngram   > Though, in fairness, I probably know more about it than everyone else at my office too, but that's a comment on the nature of the work done there, which doesn't even fall remotely  near that arena.
2023-02-17 15:04:59 <KipIngram   > I had a past job that involved modeling physical systems, but that work is way back in my past (as in last century back).
2023-02-17 17:01:00 <veltas      > Well my point is there are plenty of good programmers at work, but very few people generally seem to learn "what every programmer should know" about floating point
2023-02-17 17:06:49 <MrMobius    > at a job I had before they had taken ID numbers in an Excel sheet and imported it into our SQL database. the modeling team head guy called over the big boss and they were puzzled why it kept matching multiple people to the same number. the big boss finally said he found out about something called a float on google. i was just dying inside the whole time but knew not to butt in
2023-02-17 17:12:56 <thrig       > ... eventually they renamed some human gene because Excel kept corrupting them
2023-02-17 17:27:10 <MrMobius    > lolol
2023-02-17 17:30:05 <thrig       > fix excel? oh no, that's a sacred (cash) cow
2023-02-17 18:00:15 <KipIngram   > There's a prime example of how Microsoft bought it's way to the front row.  It's not as though they invented the spreadsheet or anything, but nonetheless, when most people think "spreadsheet" they think "Excel."
2023-02-17 18:00:34 <KipIngram   > Just because of sheer financial muscle - not because of any brilliance.
2023-02-17 18:01:03 <KipIngram   > I remember the first one.
2023-02-17 18:01:56 <KipIngram   > I was doing coop work at IBM in either 1983; I worked for a group that quality tested software that had been submitted by 3rd parties for an "IBM approved" logo to put on their boxes.
2023-02-17 18:02:05 <KipIngram   > One that came through for approval was Visicalc.
2023-02-17 18:02:11 <KipIngram   > That was the first spreadsheet program.
2023-02-17 18:02:51 <KipIngram   > It was released first for the Apple II.
2023-02-17 18:03:18 <KipIngram   > Some people say IBM's decision to make a PC in the first place was influenced by Visicalc.
2023-02-17 19:17:29 <veltas      > Apparently Excel is actually better than Lotus 1-2-3 and Visi Calc though.
2023-02-17 19:17:53 <veltas      > I've not used Lotus but Visi Calc is certainly harder
2023-02-17 19:18:27 <veltas      > Excel has the advantage of coming with Office and therefore being everywhere already before anyone can question which spreadsheet they'll use
2023-02-17 19:19:46 <thrig       > clarisworks was pretty good before Apple killed it, but then I stopped using spreadsheets
2023-02-17 19:20:27 <thrig       > Office has never left the "ugh, no" zone
2023-02-17 19:27:03 <KipIngram   > Oh, I'm not slamming Excel - it may well be the best one around now.
2023-02-17 19:27:14 <KipIngram   > I'm just harping on Microsoft's lack of innovation.
2023-02-17 19:27:26 <KipIngram   > In general - not that they've never innovated at all.
2023-02-17 19:27:40 <KipIngram   > Throw that many people at something and you're going to do something good at least occasionally.
2023-02-17 19:27:53 <KipIngram   > Now I see what's going on with this  TrueType stuff you linked earlier.
2023-02-17 19:28:02 <KipIngram   > I'm not sure I FULLY realized what truetype was.
2023-02-17 19:28:05 <thrig       > I did have customers coming to me at one point saying "please do not install Word 6"
2023-02-17 19:28:05 <KipIngram   > but I get it now.
2023-02-17 19:28:29 <KipIngram   > Yeah, back when I used to use Word I definitely thought it reached a "best point" and then declined.
2023-02-17 19:28:34 <KipIngram   > they just couldn't stop tinkering with it.
2023-02-17 19:29:31 <KipIngram   > Anyway, looks like a truetype font is an extremely high resolution "fixed description," that you regard almost as "analog."  Then you scale it to a desired size, and those Forth like instructions are then used to "adjust" the results of that scaling for a graceful result.
2023-02-17 19:34:34 <KipIngram   > Anyway, it's a nice well defined geometric environment, and seems almost idea for a fixed point approach.
2023-02-17 20:19:31 <KipIngram   > Wow, this truetype reference manual is exceedingly detailed.
2023-02-17 20:19:37 <KipIngram   > Someone did a VERY thorough job here.
2023-02-17 20:20:59 <KipIngram   > it looks like there's plenty of room for art here, though - I think I'm seeing that there is PER CHARACTER code executed as part of this.
2023-02-17 20:21:24 <KipIngram   > Who was it who was saying in the last day or two that we all ought to be using vector font specificatioNS?
2023-02-17 20:22:10 <thrig       > I'll use bitmap when I can, faster and less memory
2023-02-17 20:23:15 <KipIngram   > This whole business looks like how to convert a vector spec to a bitmap device in an algorithmic way, leaving some wiggle room for tailoring.
2023-02-17 20:38:43 <KipIngram   > One of the things this lets you do is attach "cut-in values" to various aspects of the adjustments done on the font - depending on the size you are scaling it to, you either include the feature or don't include it.
2023-02-17 20:39:02 <KipIngram   > I'm not sure whether that applied to the data or the code - I'm suspecting probably the code.
2023-02-17 20:51:35 <KipIngram   > So - I've got a rather technical Linux question; maybe one of you guys will know something here.
2023-02-17 20:51:59 <KipIngram   > At work, I'm working on being able to run several concurrent drive tests in parallel.
2023-02-17 20:52:43 <KipIngram   > Each one consists of a thread that runs the fio performance test program, and a second thread that runs a "logging" process.  The log process sends commands into the drive to get internal metric info and gathers the results and saves them.
2023-02-17 20:52:53 <KipIngram   > It's much lower bandwidth than the test traffic.
2023-02-17 20:53:29 <KipIngram   > Anyway, I've found that if I start both threads from the same user session, then I get a major effect on drive performance each time I gather log info (which happens about twice a minute).
2023-02-17 20:53:53 <KipIngram   > But if I log in independently using a second ssh session to start the logger, I get only a tiny little ripple - it's almost imperceptible.
2023-02-17 20:54:08 <KipIngram   > Both of these threads are launched with nohup <command> &
2023-02-17 20:54:31 <KipIngram   > So, the question is, why would the session association of the two processes affect the impact one might have on the other?
2023-02-17 20:54:42 <thrig       > in theory it shouldn't
2023-02-17 20:54:52 <KipIngram   > Well, that was my thinking too.
2023-02-17 20:54:57 <KipIngram   > But this difference is DRAMATIC.
2023-02-17 20:55:33 <KipIngram   > When they're started from the same session, I can get 75% dips in performance for a small length of time every time I run those logging commands.
2023-02-17 20:55:42 <KipIngram   > When I use  separate login, it's like 1%, or less.
2023-02-17 20:56:52 <KipIngram   > What I'm thinking of doing now is something like t his:
2023-02-17 20:57:16 <thrig       > does it happen if you setsid(1) on one or both commands?
2023-02-17 20:57:37 <KipIngram   > ssh -f root@server "cd <drive directory> ; nohup ../appliance >session.log &"
2023-02-17 20:57:51 <KipIngram   > reasoning that each separate ssh is a completely different ssh session.
2023-02-17 20:58:03 <KipIngram   > And a similar command for starting the logger.
2023-02-17 20:58:11 <thrig       > or what happens if you run them under tmux or screen
2023-02-17 20:59:00 <KipIngram   > Well, on my old system I used to use screen and ran both from the same bash script - fio in background and logging in foreground, and I never noticed any issue.
2023-02-17 20:59:12 <KipIngram   > The new stand, though, uses Rocky Linux 9, which has abandoned screen.
2023-02-17 20:59:21 <KipIngram   > So my first step was to convert to tmux, but this issue showed up.
2023-02-17 20:59:35 <KipIngram   > I thought perhaps it was tmux-related, so I tried nohup, but still got the issue.
2023-02-17 20:59:44 <KipIngram   > Then nohup from separate logins made it much better.
2023-02-17 20:59:50 <KipIngram   > So I've been all over the map.
2023-02-17 21:00:04 <KipIngram   > In my hip pocket is building screen from source.
2023-02-17 21:00:32 <KipIngram   > I just thought maybe there's something about the structure of the process tree that might be involved somehow.
2023-02-17 21:01:01 <KipIngram   > BTW, these processes run on separate cpu cores.
2023-02-17 21:01:07 <KipIngram   > So I'm sure it's not a core conflict.
2023-02-17 21:04:40 <KipIngram   > Anyway, I've got a working methodology; I just am curious about root cause.
2023-02-17 21:06:32 <KipIngram   > At any rate, the nohup method seems more "pure" to me anyway, because this is a non-interactive script that runs more or less like a daemon.
2023-02-17 21:06:46 <KipIngram   > I start it, stop it, etc. by remote ssh from another machine.
2023-02-17 21:07:13 <KipIngram   > And it uses remote ssh to update status information on that other machine too.
2023-02-17 21:07:29 <KipIngram   > The idea is that I may have a flock of these systems in the future, all controlled from the one control box.
2023-02-17 21:08:03 <thrig       > does rocky linux have systemd? that can run user daemons
2023-02-17 21:08:49 <KipIngram   > Yeah, I just ran into that browsing around earlier.  I am guessing it probably does - if you know offhand the fast way to check I can do that.
2023-02-17 21:08:56 <KipIngram   > My work computer is right here.
2023-02-17 21:09:18 <KipIngram   > I'm going to investigate that for sure - looked like it also lets you get status and log output and all sorts of things.
2023-02-17 21:09:21 <KipIngram   > Looked nice.
2023-02-17 21:09:44 <KipIngram   > systemctl is there.
2023-02-17 21:09:49 <KipIngram   > That means it has it, right?
2023-02-17 21:10:30 <thrig       > systemd may also do wacky things with user sessions
2023-02-17 21:10:40 <KipIngram   > The fio script has to have a particular current working directory, so as long as systemmd lets me do that, it should work well.
2023-02-17 21:11:00 <KipIngram   > With nohup I just cd there before I run it; easy peasy.
2023-02-17 21:21:11 <KipIngram   > Oh, pstree -a is awfully nice.
2023-02-17 21:50:47 <KipIngram   > Linking this here out of general interest - not implying any kind of "endorsement."
2023-02-17 21:50:49 <KipIngram   > https://github.com/seanpringle/reforth
2023-02-17 22:18:14 <KipIngram   > My main misgiving is that when I look through the fairly large amount of sample code he's got there on the github page, the thought that comes to mind is "Looks like C code."
2023-02-17 22:18:33 <KipIngram   > I don't WANT my Forth to look like C (C-style indentation,etc.)
2023-02-17 22:18:42 <KipIngram   > If a definition needs indentation, it's TOO LONG.
2023-02-17 22:19:29 <KipIngram   > My code tends to look more like... sonnets.  As in this type of visual impact:
2023-02-17 22:19:31 <KipIngram   > http://www.poemsearcher.com/images/poemsearcher/14/14eb42476f61983a9df2235c2cf1bf91.jpeg
2023-02-17 22:20:10 <KipIngram   > A small group of related definitions, one per line, of similar length.
2023-02-17 22:20:17 <KipIngram   > Generally similar.
2023-02-17 22:26:37 <thrig       > : when forthy winters shall besiege thy brow ;
2023-02-17 22:57:17 <KipIngram   > Well, I'll count myself lucky if I get 40 more, and looked at from that point of view 40 doesn't seem like too many.