Beiträge (durch)suchen
Forum Verzeichnis
Alle Beiträge mit Daniel Simu werden angezeigt
Daniel Simu - #
Pocket state explorer
It's been a while since I've last posted here, but I got something fun to share!
https://youtu.be/CSxRB_KrN-E
Together with Sylvain Garnavault I made a toy to explore siteswaps. And you can get one!
They will cost about €25 and I can bring some to EJC. Let me know in the comments if you're interested, or send me an email (hello@danielsimu.com) , so we know how many to make!
Thanks for watching 🙂
Little Paul - - Vorredner #
Oh now that’s fun! I like that a lot… I don’t juggle enough any more to justify getting myself one, but if I did I’d be all over that!
Siteswapordle
You can blame Graham Haslehurst for this. Or perhaps I should just let jokes pass by without doing this kind of thing...
https://siteswapexplorer.com/siteswapordle/index.html
1. Damn you Barnesy!
2. Siteswaple Shirley?!
3. Go on then. Siteswapordle https://siteswapexplorer.com/siteswapordle/index.html?utm_source=share
278 5/6
⬜🟨⬜⬜⬜
⬜🟨🟩🟨⬜
⬜⬜⬜⬜🟩
🟩⬜🟩🟩🟩
🟩🟩🟩🟩🟩
https://siteswapexplorer.com/siteswaple/index.html
I'll have to fix the redirect for the old one tomorrow - time for the school run!
This is really cool, could you put a link to or animation of the pattern once you have solved it, like they do with phrasle
Maybe! I’ve not seen phrasle, will have a look. I could also put something in the share text to link to siteswap explorer but that’s probably too mixed a message!
Oh yeah, more than a mixed message, it would spoil the answer! I’ll see some time if I can add a link to the success message.
el_grimley - - Vorredner #
For something based on a flippant comment its surprisingly engaging.
I thought it was just a flippant response but now I’m enjoying the game too!
The work I put into it was comically slight: I just replaced the word lists and gave it a new name. Now that it has users I’ll have to make sure it keeps working!
Daniel Simu - - Vorredner #
Oof, I feel like this would be a lot easier with my state explorer! But I sadly broke it for the time being... :(
Nice work!
I got ...
1 gold, 2 green, 3 green, 4 gray, 5 gray.
That leaves 0, 6, 7, 8, 9 as different digits, if needed. Or also more of 2, 3, 1.
Now, the 1 can only belong, where I guessed 4, so as to not bite with the 3 landing. It will then from there land, where I guessed 5.
So, then is still open where the 2 and 3 come from.
I musn't now let the fifth position land on the 3, 'cos the 2 could not anymore be accessed ( but by a 1 before it, which was wrong, let alone that 1 then not be accessed ). So my fifth position will have to land on the 2, done so by 2 on it ( the former ).
That then leaves 2 or 7 to access the 3 from first position:
7 gray, 2 green, 3 green, 1 green, 2 green.
Only 2 2 3 1 2 left now, and was right, all green °•yipih•° :oD
My strategy is guess 97531 & 86420 which at the very least tells me what all the candidate digits are. So far more often than not I get enough information to work out the correct answer on the 3rd guess. This actually makes a very nice little logic puzzle!
I also use the first two guesses to get all the digits. I think that's a good strategy, but often there is still more than one possible solution after that. (I have used 12345 and 67890 or something similar.)
I have used more guesses than necessary at least twice due to not paying attention (placing a digit in a position where I should know it didn't belong because it was yellow there). But yeah, without mistakes like that it should be possible to get pretty good stats.
Little Paul - - Vorredner #
You should turn on “Challenge Mode”
I like how such different aspects as thinking in orbits and having to close orbits, their loop, then divisibility of the sum by 5 ( leaving only 5, 10, 15, .. multiples of 5 for sum ), then bare guessing to start off, and knowledge about states, in complexity melt to one siteswap.
Very fun, thank you for making it :)
Is there a reason it doesn't like siteswaps that start with 0? E.g. it doesn't accept 00555 or 05550.
Glad people are enjoying it!
There was a good reason for that: I made a mistake! I fixed that a good few days ago, so I hope you just need to force a refresh to pick up the newer javascript.
The code should contain a list of every suitable siteswap. It could of course validate entries instead, but what seems to be working on the site is my half hour hack!
My mistake meant that everything starting with a zero wasn't on that list. I can certainly enter 00555 here, so I think it is a caching thing if you still can't.
I was testing in an incognito tab as I'd already done today's, so would have assumed it wouldn't depend on the same one as my normal browser? But maybe not, either way it seems to now work in an incognito tab so I assume you're right about it being caching...
Oh, I think it may also have been because I just moved to a new host and the redirect from the version linked at the top wasn't working any more. And I don't think I had applied the fix to that version.
Easiest answer was to delete the old one and do some work! The correct link is https://siteswapexplorer.com/siteswaple
I was asked if Siteswaple could be made more difficult. Right now I'm not entirely sure that I succeeded...
https://siteswapexplorer.com/siteswaple11/
Daniel Simu - #
Siteswap State Explorer
Now that this forum is already into explorers, I present to you my new state exploring toy:
I hope you like it!
Daniel Simu - - Vorredner #
I'm glad the Edge agrees it is valid!
Very nice!
I’m going to have to read about states. I’ve been enjoying playing with siteswap recently, particularly because of the ‘well of course that works!’ feeling I’ve been getting after understanding something. But I don’t yet have that feeling whenever I see state stuff, even though it should clearly eventually give me that feeling!
It took me a while to get to 520401 with Siteswap Explorer, by the way. What with just typing the URL in seeming unfair.
I went: 333
Swap to 423,
Subtract period from the 3 to get 420
Add cycle to get 420420
Swap a 2 and 0 to get 420411
Swap the first 1 with the first 4 for 520401.
That's a neat little gadget. I'm always amazed by the creative juggling related non-juggling things we do!
el_grimley - - Vorredner #
I like this. I think Daniel Shultz did a similar thing with a wheel and markers. (I think it was a clear cd in a case but my memory is hazy)
Daniel Simu - - Vorredner #
Your memory is correct!
I had not seen it when I made this, but he showed it to me as soon as I posted the video.
https://youtu.be/uBL7EIAfEIg
Siteswap Explorer web app preview
About a year ago I released my iOS App 'Siteswap explorer'. My Apple Developer account has just expired and Apple have removed the app from the store. Being a fool, I had some stupid idea that it was going to stay there. Ho hum. I'll probably renew the account but maybe only when I have something new to release.
Meanwhile, I've been working on a cross-platform rewrite. The intention there was to release new iOS and Android versions using the same code base, but then I experimented with reusing the logic in a web app and this happened. I've used a siteswap animator which should be familiar to Edgizens!
Here it is: https://siteswapexplorer.azurewebsites.net/
The design is very rough and ready at the moment. Fixing that's the next phase. What you see here is the 'make anything work' version.
Enjoy!
Dave
Oh yeah, if you have an iOS device and you add the page to your home screen, the navigation buttons will be hidden, making it look more like an app.
I'm planning to make the javascript less demanding, a bit in line with what our great lord and master did with the Edge version (but maybe I'll strip a little less out). The edge version definitely runs more acceptably on my 2017 ipad.
I've added 'excavate ancient devices' to my to do list, but I don't think success is guaranteed there! Success with the code, that is. There are ancient devices to be found! Some may even work.
After making 51 one handed (to a020), there is then an option to Make One Handed. Which seems... interesting.
Hah! There are lots of algorithm tests but that one's a little too concrete for the headspace I must have been in at the time so I didn't think of it!
Fixed.
It still lets you do 'showerify' multiple times which does strike me as strange, but it's a touch more valid than the issue you mentioned.
I've also made slight changes to the animator (draw a few fewer things, change some colours, probably not going to make it burst into life on devices which didn't work before that change).
Thank you for making this available to non-smartphone owners!
This is great work, there are a number of functions in there that I didn't know about (extend up/down, dual) & I can't immediately work out how to calculate them myself. I have less inclination these days to think about these sort of things so it's nice to have some software to do it for me!
I was pretty savage with my optimisation of the animator. From memory I cut out all the considerable amount of different prop related code. Then I found there was a lot of work being needlessly repeated every frame in both the hand & prop update functions that I moved back into the constructor functions eg. the body was being recalculated each frame, I calculated the path once in the constructor & saved it as a variable (I don't think the nodding head is necessary!)
I don't think your siteswap explorer app needs anything that is not in my stripped down version (Tom's site needs a lot more features) so I would've thought the Edge version might have been an easier starting point for you? There's a little bit of code for synch hand movements I think that you could strip out if you want to optimise further!
Thanks, glad you like it!
Yeah, I see how moving some of the drawing calculations up will help matters. Some other changes are underway which mean I think I'll need to compare the Edge version with mine when those changes are done and lift whichever optimisations I can. I'm going to have to learn more about benchmarking javascript: I see loads of information in the debugger but I only want to optimise when I can turn that into a clearer objective performance measure. Javascript is not a source of happiness for me. Baby steps.
The other changes I mentioned relate to 'colour by orbit' which is my favourite feature form the mobile app but is still missing on the web version. The problem there is that the current version creates a new ball for each throw. Quite some sleight of hand for a hand which is the end of a line! I did some ugly stuff to make ball IDs work for balls which have been thrown, but I really want to be able to colour each hand's next 'not yet thrown' ball too, which would get even uglier. And of course I've certainly broken non-vanilla siteswaps (which I suspect will never be handled by Siteswap Explorer, but whatever). I asked Tom who I've not met but is clearly awesome about that and he's offered to have a look at restructuring the code so the balls can persist.
I'm irritated that Apple want 100 hard-earneds a year for me to keep a free hobby app in the store. The cross-platform rewrite is working on iOS, Android and Windows. MacOS and Linux are probably also possible if I wanted to maintain even more binary releases. But now that the web app is looking viable I'm experiencing a bit of 'but I'm a mobile developer' existential angst!
I was quite surprised when I learnt that the props were created anew each throw. I thought that was a very clever solution I would not have come up with myself.
Instead of rewriting everything to track props I think it would be easier to precalculate the sequence of colours to be thrown from each hand & just look up the colour & assign it to the prop when it is created & pass the colour to the hand when the prop is destroyed.
In the ValidateSiteswap function it populates the this.left & this.right arrays with the throw each hand makes on each beat.
eg. for 534 it calculates:
this.left = [5,0,4,0,3,0];
this.right = [0,3,0,5,0,4];
At this point in the code you should have all the info required to calculate the full colour sequences:
this.leftcolour = [red,0,blue,0,green,0,blue,0,octarine,0,green,0,octarine,0,red,0,green,0,red,0,blue,0];
this.rightcolour = [0,green,0,octarine,0,red,0,green,0,red,0,blue,0,green,0,blue,0,octarine,0,green,0,octarine];
Or at least that seemed easier when I thought of it but now that I've written it down & read it back I'm not so sure!
If you go this route then you may be better off sticking with the original code because one of the other optimisations I think I did was to draw all the props as one path whereas I believe the original code drew each prop as it's own path. One of the best ways to speed up canvas animation is to reduce the number of changes to the context state, I think of it as pick up pen draw circle, put down pen, pick up pen draw circle, put down pen... vs pick up pen, draw circle, draw circle... put down pen). However if you need to pick up different colour pens anyway this doesn't matter.
I’ve been meaning to reply but I’m still thinking!
Tom has come through with changes to support a list of ball colours.
I think I will still come back and see how much of your optimisation principes can still be applied later but yeah, I’ll be starting with the original code.
The site has gone live with its new URL and a few new features: https://siteswapexplorer.com
Daniel Simu - - Vorredner #
Nice, adding orbits is a great idea!
Nice, although not sure colours are working consistently for base patterns, eg 3, 33, 3333, 33333 shows the balls all in red, but 333, 333333 has 3 different colours
Hah, I’ve had to get creative with my answer: my phone is determined to give me a transparent nothingness where I would hope for the Edge to normally show a keyboard!
The reason for that colouring isn’t immediately clear but I think it’s correct. For the first 3 of 333, the next ‘let’s call it a throw’ is 3 beats later. So it’s always that first 3. If you view 300 and 330 that may become clearer.
In 33, the first 3 is thrown, then the number three beats later is the second three. Three beats after that is the first 3 again. So both are in the same orbit.
I’ve been enjoying others examples of how changing the number of cycles also changes the orbits. A favourite is:
https://siteswapexplorer.com/534534534
Of course, I've only just realised that I could also have said "Yes, it defaults to 'colour by orbit' mode". I need to think about how to communicate stuff like that better on the site!
Hah, perhaps I should better have anticipated the need for a good robots.txt in my 'generate all the links you can think of' site!
Magician Ricky Jay is auctioning off his collection of memorabilia.
https://www.sothebys.com/en/buy/auction/2021/the-ricky-jay-collection
Lots of interesting stuff going under the hammer. I feel a bit sad when collections like this are split up.
Little Paul - - Vorredner #
Technically, he isn’t auctioning off his collection… he hasn’t done much of anything since he died in 2018! (Lazy sod)
Some very interesting stuff in there though. I wonder how much of it will get slurped up by copperfield and stashed away in his enormous warehouse
erik aberg - - Vorredner #
I managed to buy a very large Cinquevalli poster from this auction. I had heard of its existence for a few years, but never knew if it was actually real. Supposedly it hung in Ricky's living room.
erik aberg - - Vorredner #
It will hang in my living room, I have cleared one wall for it!
Daniel Simu - - Vorredner #
Whoa that thing looks epic! Congrats on the purchase!
Little Paul - #
I’ve just watched this, and thought a couple of people here might enjoy it (and that a couple of other people might grumble that it’s “not juggling” - but we don’t listen to them...)
https://youtu.be/vNWWSf_NR4E
Daniel Simu - - Vorredner #
Did anyone here follow the discussion regarding if Yoann's work is original or not?
https://vimeo.com/509288862
While not all of that could be considered plagiarism (how many jugglers have used Bach's fugues, and tried to implement fugal form, for example, and it's not exactly uncommon to see jugglers performing a box, or finishing with a combined catch / bow) it does seem that M. Bourgeois might have been a bit "literal" with the works influencing his own ...
I'm shutting down my YouTube channel
...in a couple of weeks, so if you want any of the vids, now's the time to grab them.
https://www.youtube.com/user/bad1dobby
I think the juggling videos are pretty much all duplicated on my JTV account (same name), so they'll be there as long as JTV still exists.
Daniel Simu - - Vorredner #
Oh no! One of my favorite channels!
Any particular reason you're shutting it down?
Hello everyone!
I am a 32yo male from the Netherlands, currently residing in Dhaka, Bangladesh. I learnt a 3b cascade and a few tricks in school, but really discovered juggling as a hobby in 2018 in Sudan. Since then, Ive learnt siteswap and a bunch of new patterns.
Im currently working on the following goals:
3b:
working on inverted, box-type, (2x,2x) types of patterns. On my to do list for the upcoming period: broken box, burst box, switched box, threaded box. Oh and of course 60 for the fountain.
4b:
- solidifying (6x,2x)(2x,6x), 633, 71, 651, 64514, 61661
5b:
- learning 645
6b
- aaarrggh the fountain is progressing slowly
7b
- not a focus at the moment, but maintaining current level with the goal of working on it after 6b solidifies.
In Bangladesh the air pollution gets really bad in the winter, which locks me indoors where I can only do patterns with up to 6x throws. Outdoors I work mainly on numbers, but Ill have to find a better spot once lockdown measures due to corona shut down a bit.
Since Ive only juggled somewhat seriously in Sudan and Bangladesh, I have very little experience with other jugglers. I managed to find a few poi spinners here who Im trying to teach a few juggling patterns, but there don’t seem to be dedicated juggling places here.
Anyways, long story short; looking forward to logging some patterns here and being part of the virtual community. And hopefully Ill be able to meet some of you sometime on a convention!
Daniel Simu - - Vorredner #
Welkom!
For my part, I think, it's normal ..
with some work / time put into it, I would get to about ten rounds of a trick - I like towers - which is then already 50 throws of it, being a respectable fraction of even hundreds of catches of the cascade endured.
Siteswapping truly IS genuinely different from doing a same kind of throw on and on, I say; there's more vivid 'music' in them, melody, rhythm, beat, funk, anything, and the chance to miss a certain throw is unlikely greater, and also you have to get .. no, wait, I'll start this sentence anew .. also heights, spacing, thrust thrown, impact of landing props, and surely more aspects and properties of a pattern, are all slightly or notably different, thus all in all more difficult to master.
Daniel Simu - - Vorredner #
Definitely normal.
I can run 5 balls for 10 minutes, but the only siteswaps I have ever managed to run for 30+ rounds are 744 and some synch patterns I think.
James Hennigan - - Vorredner #
What siteswaps are you working on? Also, have you learned any with 4 balls?
No, I think it's just that a lot of siteswap patterns are intrinsically bloody hard to do, and the slightest variation in trajectory can blow things up a way down the line. Juggling 5 in a standard cascade, if you get one throw a bit "off", maybe throw a 5-and-a-half or off to one side, you can mostly recover. Throw a seven-and-a-half in 753, the problem shows up further down the line and in a place where it's difficult to recover.
Doing several siteswaps in one session worked for me, "grinding" patterns isn't my thing. Maybe that's why I'm rubbish.
James Hennigan - - Vorredner #
Different jugglers practice in different ways, but I can tell you what worked for me.
I think the some of the most important 4 ball siteswaps for 5 ball juggling are 534, 633, 7333, 7441, 7531, (6x,4)(2,4x) and (6x,4)(2,4x)*. I think it would be a good idea for you to work on these if you haven't already.
In your original post, you said that you struggle to do 4 rounds of any of the 5 ball siteswaps. You might be wasting your time working on 97531 if you struggle with 4 rounds of 744. 744 is easier than 753, which is easier than 97531. Therefore, it makes sense to work on them (or at least, prioritise them) in that order. (Some jugglers may disagree with this, saying that working on things way outside your current skill level is a good idea. I don't think spending loads of time on these things is productive.)
In my opinion, 744 and (6x,4)* are the easiest 5 ball siteswaps for beginners. When I was at your level, I spent most of my time working on them. You will probably make faster progress if you focus on a smaller number of tricks. Your progress with 744 and (6x,4)* will, in turn, benefit your other siteswaps.
Maybe you think they're trivial, but I think (6x,4x) and 73 are easier than those. One sided, admittedly, so learn them both ways, like I didn't. (Hmmm, wonder if that has anything to do with why I'm rubbish at... nah, nothing, nevermind...)
It seems pretty normal to me, running a 5 balls and running 5 balls siteswaps is pretty different. When you do a 5 balls cascade, you always throw at the same height, same movement etc ... So you can just "lock" your arms into doing the same thing. It does not need to be "perfect". So when the dreaded time to throw a 7 or else comes, you need to relearn a brand new way of juggling 5 balls.
There's really a gap between juggling 5 balls and doing 5 balls siteswaps. Some are harder than the others.
What really helped me was to learn more complicated and advanced 4 balls siteswaps. It's easier compared to the 5 balls ones, and it can train you to do some high and funky throw combinations! It's also good for the mind because you can learn those patterns in a few days / weeks compared to weeks / months for the 5 balls one.
Introducing: JuggleGym
I have created a new juggling simulator. Inspired by JugglingLab, but written from the grounds up to run within your browser, with mobile support of course. You can also add it to the home screen of your phone. You can find it here:
https://jug.gl/gym/
Warning! Beta -- it can still be a little bit rough around the edges. Feedback is welcome!
Daniel Simu - - Vorredner #
Nice nice! I like how you've made it look!
Some things I noticed
The text in the input field is not selectable, I was confused when I couldn't do that. - correction: On chrome I can select, on Firefox(win10) I can't...?
In some cases when entering an invalid siteswap I get red text, in some cases I don't
What is beat shift? I noticed it only works on synch patterns, but couldn't figure out what happens if I change the value
All in all, works well, will definitely try on mobile too!
Dankjewel! :-) The input field selection issue has now been fixed. There is a random siteswap generator available now as well. The siteswap validation (red text) still needs work though.
As for the beat shift, I think you meant to say that it only works for async patterns right? That is correct. Suppose the time between two throws of the same hand is one second. Then, typically, the left hand throws at T=0s, the right at T=0.5s, and the left at T=1s. So, each beat is spread half a second apart. But, you can shift the beat to throw more closely together or further apart. To see how that works, simply use "3" as a siteswap and drag the slider.
Daniel Simu - - Vorredner #
Hoi Raymond :-) No, I meant synch patterns. For some reason the slider seemed disabled on most patterns, except for synch patterns. Right now it does seem enabled for both, but only show effect on the asynch one, and the effect is rather obvious now. Thanks :)
The generator is a nice addition, but to me a long list of siteswaps is not very interesting. It becomes much more useful once I can filter, for example excluding 2's or finding only 3 beat patterns. Also you could add excited state patterns :)
Very nice. I especially like the ability to rotate my point of view. And, of course, passing.
With Safari on an iPad I couldn’t access the text entry box. I was a little confused with the passing patterns until I realized there were two labelled one- count. After a little while I saw that the text description changed appropriately between them. They were both one-count, I guess the real problem is we don’t have widely accepted names to distinguish them. Perhaps a * or #, etc for the less well known varieties.
I also got lost poking random spots and had to reload the page. No biggie. Thanks for sharing it early, I can use it as is to explain patterns to people.
Subscribe to this forum via RSS
1 article per branch
1 article per post