Live CEOing Ep 754: Language Design Review of Special Project Features for 14.0 continued

All right. Hello, everyone. Welcome to yet another version 14 design review. We are getting ready for something that doesn't wait, which is the eclipse of the sun in April Release. Right. Of 2024. Yes.

There's actually going to be another one this month, the 14th.

Okay. Where's it visible from?

Well, the. The maximum eclipse is going to be somewhere in Mexico. Here in the US we will see a partial eclipse in Mexico. It will be annular.

Okay.

It's not total this time. That's why it's not that.

It's not such a big deal.

Exactly.

Okay. All right, so where do we start here?

I think we go again to the page of Solar Eclipse. It's Solar Eclipse New. And have you updated the documentation?

Oh, I didn't. Not in the last. Within the last hour. This updated within the last hour, perhaps.

We keep adding things to that page, so I think it would be safer to update.

Okay. Okay. Just coming here.

And which version of prototype is this one?

I don't know. Let's take a look.

Perfect.

Seems modern since it's the same date as today.

Yep.

Okay. Is it like that?

Yep.

Okay. All right. What to look at.

So, okay, so we. We can have a look at various properties, but something that we are wondering right now. If you go to the first examp.

I have an even more basic question. Okay. Things like this, that picture, presumably if I go to the notebook version of documentation, I will find that as vector graphics. True.

If you go to where, sorry, the.

Notebook version, the in product version of documentation, I will find that image as vector graphics. Is that a true statement? You should.

Yes.

Assuming that it's not so big that we're forced to rasterize it. But that graphic should be relatively small.

But here's my question. In general, all this illustration that's going into documentation, you know, I can imagine wanting to use that for, you know, various kinds of educational materials and so on. And the question is, how can you lift it out of there and have it be. I mean, I'm just wondering whether we should have a click to copy for these things. Do you see what I'm saying?

Yep.

I mean, Roger, that would be a general thing for.

Yeah, no, it is. I mean, so I. In the notebook, it's definitely depends on how what the author did. So most of the time is going to be vector graphics.

But even if it's not, a click to copy would be.

Yes. And people are actually doing that, Steven. So we started to see it show up in various kinds of articles. These illustrations that we are now Creating. So for instance, if you go to complex plot, it tells you how you spot a zero and. And a poll and.

Right.

Essential singularity. And I've seen that now show up like in many, many articles and textbooks.

Okay.

Straight copied, straight out.

Be nice if they referenced us. Do they?

Maybe not for the illustration, but for.

Yeah, yeah. Okay.

And I just saw a book this morning on statistics and it copies straight out sort of the. The explanations we did for box whisker.

Okay. That's nice. The particularly. Right. If. No, we should tag those things. That's good. And contact the authors and say thanks or growl at them if they did something unreasonable.

Well, maybe we should have some kind of. Maybe we should say how people cite things same, you know.

Yeah, we probably should for these pictures.

Yeah.

I mean if. If this thing had a mouse, had a. Had a tool tip would make it easier. Yeah. And if it had a click to copy that had something that had.

You know, I found that by the way, so already when we started this many years ago, that when I give presentations, you know, for, you know, introducing a new concept using that illustration is like super good way to introduce the concept.

So, so let. Let's look. Maybe we need a click to copy. And the thing which is copied in the click to copy contains some little signature. Yes. Okay.

I think that's something like that we.

We should figure out. All right. Okay. So that. That's a separate project and that should go to the folks who are doing project management for documentation. Right. In terms of where this actually happens, who is. Who is doing project?

Kristen? I am. Yes. This can pass to the docs team.

Okay. Okay. All right.

So something else that Jeff and I have. Have been testing in this regard is illustrations that move. Because in Astro, time is so important that when you see how things move, you typically get a better idea of what's going on.

Wow, that's getting very fancy. I've never heard of an antumbra, but it's rather easy to figure out what it must mean.

Exactly. Yeah. We want to put the three names in this.

There's going to be callouts and stuff. Right. It tells what to do.

Exactly. Yeah.

And that obviously should be a spelling word.

Yeah. So monitoring the three cons. There is the key thing in all this problem.

Okay.

So. So this is. This in the end is a problem of the geometry of the intersection of an ellipsoid and a cone. And the cone is moving, the ellipsoid is rotating and tilted. So all these angles intervene.

But it's all about that so we're on the ellipsoid.

Exactly.

We also have geo elevation on the. It's not even an ellipsoid. Yeah.

So one can specify the position of the observer. And this observer has the 3 degrees of freedom. Latitude, longitude, elevation.

Okay, but so when you. Yeah, okay. All right, so let's take a look here. So what should we look at here?

So let's go down a bit to the basic examples. And we have a map. And something we are wondering is whether this type of map, we have improved it a bit from. I mean, Jeff has been working with Tim and on this, and we are wondering whether this deserves its own function rather than a property in solar eclipse.

It's a very specific thing, I mean, with a lot of detail. I mean, in.

Well, the idea of a function would be that the user could then specify styles and have control of that, whereas if it's a property, they just get what we decide.

Yeah. And for example, specify a region. I mean, what happens if you just want to see what's going on near Dallas?

Well, so, I mean, what would we call it? Solar Eclipse Map.

It could be, I mean, the.

The other geographic. So I mean, it reminds me of sort of the overlay of. Of geovisible region and things like this. Right.

Because the users can extract all the primitives from Solar eclipse one by one. You provide access to all these elements. But then of course, it's very many. And reconstructing the.

Well, so I mean, so what you're suggesting is that the correct way to do this would be something like geographics. Just like we can say, you know, if I say something like geographics of geovisible region of iss, for example, I should get. Why did we just crash here? Just crashed. Let's try it again. I could try this in the release build, I suppose. It just crashed.

My goodness.

Yeah, well, let's try it in the previous version.

That's new.

That's.

Okay. So I mean, you know, doing something where we're essentially saying, I mean, it's kind of like geomarker, geo, whatever else, you know, geo. Because what you want is all of that stuff as an over overlay. I mean, these different elements as overlays.

Yes. I mean, we could have a single property of Solar Eclipse that returns all the elements.

Why do that? I mean, why not. Why don't just say solar eclip. What is it really? It's markers, it's Solar Eclipse, it's not graphics. Because that would be the whole graphic.

Elements. That sound quite right.

No, it's not Quite right. But I mean, so, so the idea would be. You'll be specifying which solar eclipse, and then you'd be specifying like, I just want the. The region of totality or something. And then that would become something like this that you could style however you want. That make sense?

Yes, but solar eclipse can now return the polygons, for example, all the lines.

Yeah. There's a long, long list of properties up above that lists all the things we can do, and some of those are elements that are. Can be directly fed into geographics. And that's how this graphic was actually constructed.

And what I'm, what I'm saying is each of those elements in the function. Solar eclipse elements or whatever they are, Solar eclipse markers, whatever, whatever we call it, each of those elements you can call for here. And then it basically gives you a bunch of primitives that you then feed into geographics. Am I making sense?

Yes. I guess what I don't understand is the difference between solar eclipse elements and solar eclipse, because the elements are properties of solar eclipse. So we could just put them in the second argument of solar eclipse.

Well, because I think the defaults would be different. I mean, but, but I mean. Okay, so the difficult is.

Yeah, so I think the difficult thing is how to enter, how to refill the styles into the elements.

Yes.

So I mean, do you want a consumer function or a programming function? I think if you get elements, it's becoming kind of a programming function where you assemble it.

Well, are we caching well enough that if you, If I ask for several different things from solar eclipse, that's not going to recompute anything?

It's. It's getting better on that front. Yes.

Okay. Okay, so then. Well, let's say I just wanted to show the path of the region of totality. How do I get that here?

So. Yes. So there is a property, right. It's called. Total face polygon.

Oh, yeah, we had that before. We had that years ago.

Yes, it's. Exactly. It's one of the, One of the things that actually existed before.

Why is this taking so damn long? Okay, so now if I say geographics of red comma, that.

Hope that this.

One didn't have to. It crashed again.

That's, that's. That's completely new today.

Well, then maybe I should use yesterday's build, although I won't have other stuff. Right.

These.

This was there before, right?

Yeah.

That looks different, which is a bit odd. But anyway. Okay, lovely. I don't know why it has opacity, why it has low Opacity. Like that.

That's the default in Geographics for all polygons.

Okay, so If I said opacity 1.

In geostalion, rapid in Geostalion.

Why?

Because that's the convention in Geographics that the geopolygons need. Data styling wrapped in geostelling. Let's put geostalign override, for example.

Why did it download the polygons again?

No, no, it's your styling of red. So your styling of red is the styling comma.

Yeah.

You know, we really shouldn't. Okay, fine. We really shouldn't have to download the default world tiles. We should surely have downloaded those already.

We currently download all the tiles. I think it's.

Yeah, we need to optimize this. I assume in this version we are continuing to optimize vector Geographics.

We continue, yes. But it doesn't go as far as I wanted because the time was needed in the Astropart. But the new underlying functionality for Astrographics and Geographics is getting much better. So it will get to Astrographics this time, but. But not to Geographics.

Okay, Geographics. I mean, the surface of the Earth is more important than the sky.

Yes, but it's more difficult to handle, too. So Astrographics is much simpler.

We could push back the eclipse.

Okay, all right. So the question here is there are all these elements, where are the things that are sort of visual elements? Umbra, envelope, polygon, all that kind of stuff?

Yes. And all the positions, for example, all the contact positions for the points and.

Yep.

Okay. But the labeling of those and so on is in this. There's some default labeling here, Is that correct?

Yes, that Jeff has added that. So a function that does much of these for you would be interesting.

Well, but I mean, it's added in this eclipse map case. And, I mean, why couldn't we give, like, an association with all kinds of crazy things you can set.

Yes, but we can certainly have a property graphics data that returns everything.

Then I think you're suggesting, Stephen, that this is almost like its own new object that you can feed into Geographics, and Geographics would just ingest it. Is that right?

No, no, I'm saying that when you say eclipse map here, there can be additional options to that, which is an association of all the different settings of different styling features of this.

Right?

Yep. Yep, we can do that.

I mean, that seems like the right way to handle. I mean, I. I'm not convinced that a separate function that makes this is worth having. This is pretty simple.

And it can get messy. Really quick because you have which elements do you want to turn on? And then if you turn them on, what style do you want them to have?

Well, that's why I'm saying what are, what are those elements called? They have names up here, right? Every element has a name, correct?

Yes. So the contact points. Have a look, guys.

Okay, so then what we want to do is we want to give essentially an association where it says, you know, I don't know exactly how this should work. I mean something like this isn't quite right. Eclipse map and then an association that says contact contact point one goes to. And then the styling. Well, actually what we really want, I mean that could be a style or you could have even more pedantically contact point one function that gives, you know, the geoposition of the contact point and then you can wrap it in whatever you want, a marker or whatever you want because that's what you're ultimately going to need to do if you want to make a custom map. But the alternative is if you really want to make that level of custom map, just extract the contact point and build the map yourself.

Yep.

Yeah.

I mean, I gotta say I, I just think people are just gonna use our default map and the only thing they might say is, you know, if there's a plot theme, for example, that might be useful. Mm.

Or idea range. We could just have eclipse map. GEO range goes to this area and we just show a nice map of that area.

Sure, that would be good. I mean, why not just have that be a top level option to solar eclipse that only applies to, you know, I mean, I think the GEO range thing be super useful if it could tell me what eclipses are going to be visible for this from this area. I mean that's a slightly different meaning to GEO range. I mean in other words, if I say an eclipse and I ask for this eclipse but in a place where it isn't visible, we don't want it. Well, I mean, I guess we asked for this eclipse. Yeah, I mean if we say GEO range and we don't specify a date, then giving us the next eclipse visible from that GEO range seems like the right thing to do.

I see. It's a bit like the option eclipse type which looks for the eclipse, the next eclipse of that type.

Yeah, I mean the GEO range thing is exactly the thing that I've wanted to do a zillion times. It's like, when is there next going to be an eclipse from wherever? Mm. Now another thing, I presumably I can get the GEO region and I Can say, is this town in this region? Right. So if I. If I get that for the next solar eclipse, how do I get. Okay, total phase polygon. So I can say something like, how do I find out if a particular town is in this region? Geo within Q. Is that right?

Yep.

Get that polygon again. Geo within Q of the polygon comma and let's say New York City. Okay, so now how do I find the. The cities that are within this? So is that geo. What's that called? Geo. What is that called?

Yes, that would be geo. Entities of city comma, polygon.

I think I tried this the other day and ran into some timeout problems. I asked the GEO team about that, and I think they're looking. They're working on it. I think it's a timeout. I think because it works for administrative divisions. It just doesn't.

Perhaps it's the other way. I. It may be the other region. Yeah, sorry, it's the other. Is the reverse order.

Well, let's say. Oh, God.

Yeah.

This is what jeff was mentioning.

That's not very impressive. Yes.

Yeah, those worked for me the other.

Day, but they don't work today. Yeah. At least on this prototype build and so on. Okay, so my, My statement is, if you want to build your own picture, you got to get the primitives out. And, you know, the one thing we clearly need is a GEO range. You know, it will be good to try to cash, you know, like. Like looking for eclipses in the next hundred years in different places in the last hundred years. Surely we can cache that. And we should. Okay.

Yep.

Okay. So now what else should we look at?

So another question we have is whether to return dates. So many of the properties here are dates. Of course, if you are interested in the eclipse you are going to be seeing, you want all those dates in your time zone. But if you are studying eclipses globally, the history of eclipses all over the world, you want the results in the GMT zone.

Why aren't we just giving a time zone option for the whole thing?

It has the question, is the default?

Well, I think it should be the same as for now. Which is your local time zone.

Your local time zone.

Okay. I would think so. I mean, isn't that what all date object date functions default to your local time zone?

Most do.

Most, yes.

What doesn't I mean local time. If you specify a different place, obviously.

Yeah.

Okay. I guess because we are changing from. To have so many, many eclipses in this function, it's like 72,000 comparing them fields nicer in UT1, but I understand people will use them individually for the.

Comparison is just going to work because we know we're keeping the time zone. So if you say what's the time between these things to the second, it's going to give the right answer.

Yes, of course, of course, yes.

But I mean, this is in your personal time zone that you. The one that you're in. So if the eclipse is off on the other side of the earth, it's still going to give when that eclipse occurs in your local time zone, not in the local time zone of the eclipse. Right.

For example, when you compare to other references or books, they are always reported in universal time.

Yeah, I understand. Well.

I understand it would be an exception with respect to the other functions. So users can always specify timesong. Goes to tmt.

Yeah, I think they should. I think they should. And I think people, you know, if they're confused and they say that eclipse is happening in the middle of the night, for me, how can that be? Or that eclipse appears to be happening in the middle of the night, they're going to figure out it's a different time zone. Okay, what's next to look at here? Okay, yeah, you've got time zone here. So.

Yep, exactly. Yep. And eclipse type to look for.

Yep.

For a given particular eclipse.

So GEO range should just be another option here.

Okay.

That can look for eclipses that are visible within that GEO range. Now, we got to be a little bit careful. If you say georange arrow or city, I don't know whether we have, you know what I'm saying? Whether. Whether we don't have polygons for all cities.

No, it would be just mostly a point.

Right. But so actually that's still okay because it's a question of is this eclipse visible from that point? Which is fine. That's perfectly well defined.

And presumably, Stephen, every time you say when an eclipse is visible, you always mean totality. But if it's a partial eclipse, you don't want to ever include those.

No, no, no, no, no. I think an eclipse being visible, I think it's any eclipse. And if you say eclipse type, arrow, total.

Right, right.

Yeah. So I mean a partial eclipse, now a lunar eclipse, I mean, it's clearly visible over half the Earth. Right, right.

Something like that.

So. But that's okay. I mean, you'll just see a bunch of lunar eclipses and then you'll say, okay, well, fine, I only want total solar eclipses. Set the eclipse type.

Okay.

All right, what's next?

Here I'm thinking it might be an expensive operation to, to look through so many eclipses, so many polygons.

Absolutely. Yes it is. So let's make sure there's a progress indicator. But I mean that's. Yes, it's non trivial but I mean, how else are you going to do it? And again, I think it's important to cache what we can, Right. That like we should be able to have a bunch of polygons for a bunch of, you know, eclipses in the last, I don't know, 100 years or whatever it is and then, and next hundred years and then, you know, we can easily do the test against those. Am I making sense?

Yep.

What's the average time? I worked that out for that thing I wrote about eclipses. The average time between eclipses for a given point on the Earth. Like a few thousand years, I think.

I think so. It's quite a lot in, in general.

Right. But that means within a few thousand years every place on earth will have an eclipse. So if you say solar eclipse of you know, my little tiny village and wherever, if we had cached, you know, how many thousands of years, there's a good chance that you will hit something, but not enough in a cent.

Perhaps we can create some sort of reverse index of some grid of the, of the earth.

I see. Yeah, that's a sensible idea. Yeah, we just grid the earth and we just give a list for every grid square on the Earth. We just give a list of what eclipses are candidates.

Right. And we just give results with some resolution.

Yeah, well, but I mean internally that's what we would do. So we would pick, take the GEO feature, ask what grid squares it covers, you know, or then just get the list of eclipses because for every grid square the, you know. Okay, so claim, okay, if, if the grid squares are, let's say 100 miles on a side, okay. That's a modest number of grid squares. You know, what is it? 240 are going around the earth squared and then each one of those will have, I don't know, I would guess, you know, from our full Data set of 72,000 eclipses, I would guess each one of those, we can just work it out. It's one over 24, 240 squared, isn't it? I mean it's just, you just break that number of eclipses into roughly those grid squares because. Well, it's not quite because each of cups goes over a bunch of different grid squares, but this is a good approach. We should create that cache and then from a given point, we can easily pull up.

Yep.

Because when we've got a small set of comparisons to do, and particularly if we're looking for the next eclipse, it's a very straightforward piece of comparison. We just keep going until we find one that covers that particular point. I like it.

Yeah, I think that's certainly doable. So another property I would like to have a look at is about what we call the panorama, and it's further down in the scope section.

Why is it in the scope section? If it's not documented, then it's not real.

Well, it is documented. The example I want you to have a look is the one called eclipse cycles. So the property is documented. It's a particular example. The output of this one. So there are these two numbers that identify the eclipses.

I remember. Yes.

The salus and the inex. And so I was thinking of having this property, the Saros INEX panorama, such that we show kind of the local area around a particular eclipse so that you have an idea of where in the Saros. Serious. Which are the vertical lines. It is. Because if it is closer to the borders, you know, it's nearly so like.

Like a chromaticity plot. There's a lot of background detail here.

Yep.

And then you've got a point on that background detail. By the way, I. I hope somebody paid attention to the thing I reported about plotting colors on the chromaticity plot and the. The. The horrible features of. Of, You know, being being able to. Not being able to see those colors on the chromaticity plot because they're plotted in the same color as the chromaticity plot. And I couldn't figure out how to plot them in a different color. That would be a thing for Shadi. Somebody could bring that to her. Please. I did send mail, but I don't know what I got an answer.

Yeah, I'll check on that.

Okay. All right. So.

Yeah, so. So, yeah, I wanted to check on this idea of highlighting the. The position of a particular eclipse on the background. And we could color. For example, I think we should color the total eclipses of the background should be sort of have like, layers of different types.

Why are there. Why is there that grid? Is that. Are those eclipses or what was. What is the grid?

So Those are like 5,000 points of our 72,000. Those are around 5,000 eclipses. Each one is a. Is a point, and it's a grid in families, in vertical families. And the distance vertical is always 18 years. The distance horizontally is about 38 years.

Looks like the kind of thing that should be in my description of how eclipses work.

Yeah, this is a very interesting thing. So eclipses are uniquely there. So each unique is a single point. And then.

So those are the eclipses.

Exactly.

Anti aliased into something weirdly different and.

Yep. And so when an eclipse is reported, you always say it belongs to Saros family. For example, this is the one in two weeks. This is 134, which is the horizontal from when. It's more or less arbitrary.

Okay. I mean, they knew about Soros families and antiquity.

Yes. And those corresponds to, I don't know, minus 200.

Okay.

Okay, so you see that the total plot we have below in blue. So we have families from minus 400 or so to 700.

And that corresponds to what range of years?

Minus 13,000 to plus 17,000.

Okay, wow. Okay, so when that started, nobody was recording eclipses. Right.

So this is because we can compute.

Obviously they might have been studying eclipses, but they were probably throwing rocks at it.

Yeah. Or something. Okay, okay. So the issue is. So for example, computing. So we've got this whole. How are you specifying which region to plot in this default Saros in expert panorama?

So right now what we are doing is that for the eclipse at the center, we go around 1,000 years back and a thousand years forward.

Yeah, but I mean, that's totally arbitrary. It could be 100 years as well.

Exactly. But we need some margins so that we have the full family. A Saros family will last for many hundred years, potentially a thousand years. And so in order to make sure that the vertical line is complete.

Well, I mean, we could just decide that it's. I mean, but this cutoff, this. I wouldn't call it space, like hypersurface, but it's some kind of time, some kind of weird. You know, I don't even know what it is exactly, but it's some. What is it actually? I mean, that's obviously the time. I don't understand well enough what these saurus cycles actually are. To understand why. I mean, that's kind of like where.

Well, okay, so these are families of eclipses that are, for example, vertically always separated by 18 years. It's very precise. 18 years and a few days. And yes, this affront and the inclination is just this quotient of 18 and 38. 38. So horizontally the two eclipses are separated by an annex, which is about 38 years. And vertically it's a saros, which is about 18.

Okay.

And so then eclipses jump around. So the Next one will not be next to it. They are jumping all around this. This plot. Yeah, more or less.

It'll be nice to have a picture. I would like to see a picture that shows the, the, the trajectory in Saros in X space.

Yeah, yeah, that's a good idea. We can do that.

Okay. But the issue is you want a. I mean, clearly, this thing here, I mean, we could clearly have a property Saros in X series or Saros NX coordinate or something. Saros inex value, which gives a pair. So that gives you the position of that center dot.

Right.

I have to say that my tendency would be to not make an arbitrary plot here, but just to have an easy way for people to get a big bucket of Soros and X coordinates.

I see.

Because I just. So many different kinds of styling you might want here. And you might want. You know, it's just, it's very unclear. And if what you did was to say you had a thing which said, you know, either within the Soros range or.

We have a notation.

Yeah, In.

In the first argument, we have a notation for all the eclipses between two dates. So I understand your point. So we could do solar data. So solar. Sorry, solar eclipse of date one, date two, all. That's the notation for. Give me all the eclipses between one and two. And then in the second argument, we have this property you are suggesting called satosinx, for example.

Yeah.

And it returns. Yeah.

Points. And then we can just plot them. I mean, it should be very straightforward. Honestly, it's easier than this.

That's essentially what this is doing. Just a graphics of point.

I think it's the better way to do it.

Yep.

Okay.

And then if you want to highlight the current eclipse, you just, you know, do that with this. So the only question is the Saros NX result for single eclipse is going to be a singleton pair, whereas the other case is. What am I saying? Is A list of pairs. But is that consistent?

Yes, actually it will be an event series. So our rule is that when you have a list of eclipses in the first argument and a property in the second will return an event series. That's what was done before. So we'll keep it that road.

Okay, so wait a second. An event series. But when you're giving Soros in X indices, then you're still giving an event series which has the actual dates of those things.

Exactly.

And then the value is the Saros NX index. So this particular function would say values of solar eclipse, blah, blah, blah. That seems pretty nice. And that. That way my plot that I was asking for is trivial to make because you're just threading. You've got the things in order in their event series order. And so you literally just take that sequence and that gives you the, you know, the path.

Yep. I like that. Users might also want to keep the keys to do tooltips, to see the various dates of the points.

Yeah. Or other things.

All right, Jose, are there any other pairs that. Other than Soros and nx, are there other indices that people might want to pair up or are these the only two that are relevant?

These are the fundamental ones.

Okay.

There are other cycles, but these are the ones we are exposing because these are the fundamental ones.

Okay, great. All right. Anything else? What else?

No, I think this is the main questions. Any other one from you, Jeff?

Not offhand.

Sorry.

Go ahead.

In this thing, I'm planning to put out this eclipse piece. I was thinking of having some kind of appendix which shows how to compute things in our technology stack in a bit more detail. And so, you know, I'll be playing with these things.

I was thinking you're probably going to want to expand. You mentioned the selling elements in your blog, but you don't really go into detail. And we now have stuff you could do that with.

Okay.

Yep. And I think the fact.

That suggestion.

Yeah. Next year, next thing, I don't think I talked about unless it was implicit. And it seems like a nice way to think about eclipses. I don't really understand the meaning of an annex, though, other than it's an index with a D missing within the Saros cycle.

It's a property that was recently discovered, I think, in the 1950s, which is very impressive given that the Saros was known by the Greeks.

Yes, but what does an annex. What is the interpretation of an annex?

So these are numbers that come from periodicities in the orbits of the moon and the Earth. So they reflect something, some physical periodicity.

I see.

So the key fact is waiting long enough so that the geometrical configuration of sun, Earth and moon is very similar so that you get nearly the same eclipse again.

Right.

And in fact, when you compare the different eclipses of the same Saros family, they look very similar, only a little bit transposed.

Well, so you're saying that the time you have to wait to see this is of order 38 times 18, is that right?

That's of the order of a family.

Yeah.

Of a full service family. It's hundreds of years, many hundreds of years.

Yes, yes, yes. No, I understand that, but I'M just saying. Excuse me, this, the second coordinate. The fact that there is periodicity with respect to, you know, in the end it's dotting into different index values. It's not sequentially going through in X values. No, just like you know, in the. And. And so your claim is that it wasn't obvious that there was additional regularity beyond the fact that. I mean everything is on a Saros cycle, but most 18 year period things on the Saros cycle aren't eclipses. Right.

Between two elements of the same Saros cycle there are many other eclipses.

Yes.

So you have to wait 18 years to get the next element in a psychos family.

No, I understand.

So there are many, many sort of families playing at the same time.

I understand, but I don't quite understand why the annexes. Explain again why the annexes were not discovered until 1950 or 1950s.

Because the, their, their similarities between eclipses horizontally is much smaller. So people didn't realize that these two eclipses be. Were together in some sense. Vertically they look very similar. The geometric configuration vertically is very similar. Because this period of 18 years corresponds to precession of the orbit of the moon. It's a fundamental geometric effect.

Okay, so where's the 38 coming from?

Some other combination of numbers in the, in the motion of, of the moon around the earth. But it doesn't have a clear physical interpretation.

Weird. I mean it's very clear in this picture.

It's very clear. Yeah.

Yeah.

That's absolutely weird. There's got to be some way of understanding that. Even, even if it's some, you know, number theoretic combination of a period or something. I mean especially with a number as small as 38 years.

Yeah. For example, I noticed that we have many cycles in our quantity system, but index is not there yet. So just an example of how weird this number is and recent. I'm going to ask for it to be added.

To find the value of a particular. Let's see here. You just look up Sauros. Oh, they see Saros index panorama. Let's see what does. At least in here. Blah, blah, blah. Yep.

So that's, that's.

Well, somebody seemed to have a.

Sorry, I think I, I read here. I, I've been, I've been saying 38, but it's 29. Oh, I said it's.

But so this dude. Might or might not whatever this is. I don't know. Okay. The Saros reproduces distances, diameters and rates of motion of sun and moon with considerable accuracy. Total eclipse is nearly always followed by a total 1, annular 1 by an annular one. Okay. Moreover, tracks on the Earth's surface are very similar. Okay. Okay, let's track some of the Earth's surface. The cycle for which I suggest the name megalosaurus. I guess dinosaurs were known at that time, so maybe there wasn't yet a megalosaurus. But where is the thing here, Huh? Well, who knows? As usual, the history is a bit complicated. I mean, this appears to be a certain John and Stockwell. Not this. This guy.

Yeah, I didn't know about this one. I had read about the Belgian author in the 50s who was the one suggesting that eclipses must be presented in this diagram.

Yeah. Okay, well, interesting. All right, sounds good. Anything else we need to cover here?

No, not today. I think we have all the information we need to complete this function.

Great. Excellent. All right, see you all soon. Bye for now.

Bye.