2023-09-02 11:14:41 anyway, just saying you can see in those presentations what's possible but that is sometimes constrained by the calculator series. the 48 models sounds like more of what you want. you can definitely drop down into a language close to Forth then return back to the high level stuff when you're ready 2023-09-02 11:26:55 veltas: My Oracle thing timed out its free trial and went down to "free services only." 2023-09-02 11:27:08 Along with this, my previously working ssh connect process stopped working. 2023-09-02 11:27:28 I don't know if that's something that's "paid only" or if I now just have to do it differently somehow. 2023-09-02 11:27:43 Can't find any good documentation, and when I try to chat with them I learn that live chat is a paid only feature. 2023-09-02 11:27:49 So that's kind of frustrating. 2023-09-02 11:28:49 MrMobius: My guess is that I'd still find the HP48 to be not quite what I want - I feel almost certain that it's still too "protected" to suit me. I want AT THE HARDWARe. 2023-09-02 11:29:17 I.e., I want the most direct, high-performance possible access to the resources. 2023-09-02 11:29:18 fair enough 2023-09-02 11:29:29 then youll be doing assembly :P 2023-09-02 11:30:20 I'm guessing here, but the general trend in the world seems to be a compulsion to provide a "safe cocoon" for users. 2023-09-02 11:31:35 As though letting a user stub their toe is going to reflect badly on your product or something. 2023-09-02 11:32:23 I do think measures should be taken to guard against users BRICKING their device, but short of that I'd like to be able to "take my chances." 2023-09-02 11:32:55 not advocating you get into 20+ year old calculators but you would definitely have that on one of the hp48 models. sysrpl is just threaded code like forth. you can do anything that can be done including shooting yourself in the foot and crahsing the whole thing. the OS is written partly in it 2023-09-02 11:34:20 Yeah, 20 years is along time - along with that would come the performance and power consumption of 20 year old technology. :-) 2023-09-02 11:34:50 I expect if I ever do try to make a calculator I'll wind up taking an old unit whose keyboard I like and replacing the internals. 2023-09-02 11:35:09 I suspect that making a keyboard I could accept is just beyond my talents. 2023-09-02 11:35:30 keyboard are actually really, really hard 2023-09-02 11:35:47 I plan to present partly on that this year 2023-09-02 11:35:54 Yes, and most often not very good. 2023-09-02 11:36:10 I think my college years represented a high water mark, at least in terms of what was readily available. 2023-09-02 11:36:16 HP had the keyboard down pat in those days. 2023-09-02 11:36:39 the hp 48 ones are considered just as good. the hp49 was rubbery keys everyone hated 2023-09-02 11:36:52 Oh, I don't think the HP-48 is as good as the HP-41. 2023-09-02 11:37:06 I used the 41 family - my wife used the 48, so I've at least had the thing in my hand. 2023-09-02 11:37:21 then came the HP49g+ which was a 72mhz arm emulating an hp49. some of those had crap keys so discontionued and latest was hp50g which has good keys again but those have been discontinued for a long time now 2023-09-02 11:37:36 I may not be talking here about only the mechanicals - my opinion is sort of a wholistic one that has to do with the tactile response and the size and spacing of the keys. 2023-09-02 11:37:50 hehe, ya the HP users are endlessly obsessed with the keyboard. people get very opinionated 2023-09-02 11:38:04 Whenever I did use my wife's calculator, I felt the need to "thumb" the keys for the most part. 2023-09-02 11:38:10 I *touch typed* on the HP-41. 2023-09-02 11:38:20 Using all of the fingers of my right hand. 2023-09-02 11:40:05 Also, I utterly despise that method of entering text where you first hit a "group" key and then hit an "item from that group" key. 2023-09-02 11:40:17 That just slows things down to a crawl. 2023-09-02 11:43:21 The DM42 offers that mode, but it must offer a different mode thata I haven't figured out how to access yet, because there is a letter label beside practically every key. 2023-09-02 11:44:18 On a slightly related note here (only slightly), I've had on my mind for a while a "calculator application" for my computer that exploits the computer keyboard in what I regard as a proper way. 2023-09-02 11:44:46 When you have a keyboard in front of you, it's just a lot faster to type s i n than to grab the mouse and poke an on-screen sin button. 2023-09-02 11:45:19 Also, I want to create a way to enter digits without having to use the ones spread out across the top row of my keyboard. 2023-09-02 11:46:36 I think something like nm, = 123, jkl=456, uio=789, 0=0, .=. 2023-09-02 11:46:57 The trick would be working out in a really clean way how to get in and out of that mode. 2023-09-02 11:47:14 I'd like that transition to be automated to some extent, with a manual switch also available. 2023-09-02 11:47:36 The driving design imperative would be to do as much as possible without taking my hands away from home position. 2023-09-02 11:48:47 One possibility for mode switching would be to hold down the left control key when hitting the first digit of a number, and then remaining in that digit mode until either a non-digit key was hit or ctrl was held down again. 2023-09-02 11:49:12 The key aspect being that you'd not have to keep crtl down for all of them - only for transitioning. 2023-09-02 11:50:01 Using such a thing would have to be learned, of course. It's not something you'd just immediately be fast at. But I think with some training one could get very fast. 2023-09-02 12:34:28 This is a nice video on roots of cubics: 2023-09-02 12:34:30 https://www.youtube.com/watch?v=ON2IMEC_pTE 2023-09-02 12:40:01 I'm just trying to think now about how one would really "like" to deal with mathematical entities, whether working on a computer or a calculator. I'm thinking that a lot of stuff just needs to happen automatically. 2023-09-02 12:40:55 For example, on most MODERN calculators (good ones at least), if you take the square root of a negative number, the calculator simply produces a complex number, and that's A number as far as the calculator is concerned. As in, the whole complex number can sit in the top stack element. 2023-09-02 12:40:58 That's the right idea. 2023-09-02 12:41:04 But it should go further - much further. 2023-09-02 12:41:21 Say you do some calculation that results in a 3D vector. 2023-09-02 12:41:38 Well, that vector is A thing. And it should be able to sit entirely in one stack item. 2023-09-02 12:42:42 Or perhaps you solve for the state of a vibrating string. Well, that's a whole infinite vector of displacements, but it should somehow be represented as A thing and be able to live in a "single quantity" as far as the computing environment is concerned. 2023-09-02 12:43:15 Same idea all the way up - solve for the electrostatic potential in a 3D reigon - that's a field and it's ONE entitity. 2023-09-02 12:43:39 And you should then be able to do anything to it you can do to a field, like calculate its curl or its divergence or whatever. 2023-09-02 12:44:08 I'm saying these things should NOT show up in the environment as a "user level array." 2023-09-02 12:44:34 They're wholistic entities, and the human interface should treat them that way. 2023-09-02 12:45:30 For some of these fields we know from finite elements that we want to represent them as a grid with nodal values. Well, such grids should just appear more or less automatically as they're required. 2023-09-02 12:45:57 Knowing all those nodal values and how they relate to one another is no different from knowing that a complex number has a real part and a complex part. 2023-09-02 12:46:01 I mean imaginary. 2023-09-02 12:46:27 The user shouldn't have to fuss over the "internal representation" of these things. 2023-09-02 12:46:41 Rather should be able to focus on their mathematical relationships. 2023-09-02 12:48:16 something just like the solvers we were talking about last night should work. Say you enter a differential equation that defines a field on some region. Well, you need boundary conditions - specifying those is like specifying the other variables in a solver problem. Then when you've done that you can solve, and it will just go calculate the entire field numerically. 2023-09-02 12:48:24 And present it to you as a single contained item. 2023-09-02 12:49:24 There should be a way to interact with that required gridding process - you might want to refine the grid manually in some way. But it should at least try to do it itself and be intelligent about it. Seems like that's an area where our fancy dancy AI methods should be helpful. 2023-09-02 12:50:04 You know that you want a fine grid where the field is changing quickly and a coarse grid where it's not - that should be fairly "automatable." 2023-09-02 12:50:31 to handle things like this you'd need a lot of RAM, but that's really ok - a lot of RAM can be available these days. 2023-09-02 12:53:42 especially after you turn off firefox 2023-09-02 13:05:42 I'm kind of all over the map this morning. Say I'm at a nice computer and am trying to code up something like this. Say I want an expression that is an integral. I can imagine wanting to just type the word "integral" and then, say, hit ctrl-Enter or something, and have that word disappear and a lovely integral symbol appear. And then tab will cycle me around the lower limit, upper limit, and integrand 2023-09-02 13:05:44 fields. Depending on what I did in those areas new fields might appear. For example, opening a fraction would give me numerator and denominator fields, etc. Typing sin ctrl-Enter would put in a sin() and there would be a new field in there inside the (). 2023-09-02 13:06:00 I'm harkening back to that old Equation Writer that was part of Word many many moons ago. 2023-09-02 13:06:11 I remember it as being quite friendly to work with. 2023-09-02 13:06:58 But I'm thinking of more than just creating an image - these would be things you could do math on after you got them built. 2023-09-02 15:48:42 KipIngram: Yeah mine's timed out too 2023-09-02 15:49:49 I've just been using ssh with my own details, I removed the default account it made 2023-09-02 15:50:16 I don't really know what "ssh connect process" means 2023-09-02 15:52:00 By "timed out too" I mean my free trial is over and I'm on free-only services 2023-09-02 15:53:37 KipIngram: agreed. the menus are nice to scroll through when you dont know what you want but good to be able to type it in letter by letter like all the graphing models 2023-09-02 15:55:51 a guy was working on custom firmware for the HP-50G starting over from scratch using the intel decimal libraries with 34 digits of precision. he had a small box in the bottom corner to suggest auto complete stuff while you type 2023-09-02 15:55:54 which is a neat way 2023-09-02 17:31:27 veltas: I just meant I can no longer ssh to the account I could before, nor can I ping the IP. 2023-09-02 17:31:59 So you had to do something through their web interface to set up a new account to ssh to? 2023-09-02 18:47:41 KipIngram: No I ssh'd using the details on web interface, set up an account, and then deleted their account 2023-09-02 18:48:52 I haven't really touched their web interface since setting up 2023-09-02 18:49:29 Maybe I somehow used the initial account instead of setting up another one, because mine is just like it's not there now. 2023-09-02 18:50:08 Oracle vms evaporating? 2023-09-02 19:02:16 KipIngram: Well is your VM still on the web interface? 2023-09-02 19:03:36 Check your emails if you haven't alreayd 2023-09-02 22:13:17 I haven't rolled up my sleeves to dig into it yet. 2023-09-02 22:13:24 Kind of a busy day. 2023-09-02 22:13:44 BTW, this looks like a quite good coverage of spinors: 2023-09-02 22:13:46 https://www.youtube.com/watch?v=j5soqexrwqY&list=PLJHszsWbB6hoOo_wMb0b6T44KM_ABZtBs&index=1 2023-09-02 22:13:56 I've had a hard time finding good material on that topic. 2023-09-02 22:29:17 fidget spinors? 2023-09-02 23:02:10 ;-) 2023-09-02 23:02:24 No, spinors as in quantum spin. Also as in EM wave polarization. 2023-09-02 23:02:54 I'd never seen them connected to polarization before - that was very helpful to me because I have a good mental picture of that. 2023-09-02 23:03:33 The really weird thing about them is that if you subject them to a 360 degree rotation it doesn't leave them in an unchanged state. 2023-09-02 23:03:50 It NEGATES their state. You have to do 720 degrees to get them back to their original state. 2023-09-02 23:04:20 360 = half of 720 - that's the "1/2" in "spin 1/2." 2023-09-02 23:05:07 That other stuff I linked the other day - the "cultural / social history of science" lectures - they did a good history of the gradual understanding of spin. 2023-09-02 23:05:46 When they first began to notice that something like that was going on, they tried to make it be normal physical rotation. 2023-09-02 23:06:23 But Pauli ridiculed the guys who suggested that - pointing out that given what they already knew about the electron, in order for it to spin fast enough to have the necessary angular momentum the surface of the electron would be moving faster than light. 2023-09-02 23:07:03 Those guys were actually on the right track, but the desire to kep it all "physically visulizable" hung them all up for a few years. 2023-09-02 23:07:20 Finally they just chucked that hope and started thinking of it as more "abstract." 2023-09-02 23:08:02 Now they say that it's a rotation in some "internal state space." 2023-09-02 23:09:02 But they wound up keeping the name "spin," which I think has caused endless confusion for us non-physicists.