Jupyter Community, 30 Mar 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Jupyter Community Call - March 30, 2021

Description

Recording from the Jupyter Community Call in March 2019.

The notes from this call can be found here:
https://jupyter.readthedocs.io/en/latest/community/community-call-notes/2021-march.html

Read more about these calls in Discourse:
https://discourse.jupyter.org/t/jupyter-community-calls/668

A

Record, yes, please awesome, um okay, well, hello, everyone and welcome to our march 2021 jupiter community call I'm isabella. I am going to be your host today. I am a designer who likes to hang out in the jupiter community and wants to make sure we all get to connect with each other better so great to have you here. Thank you for taking the time uh to summarize really quickly what the purpose is here.

A

Jupiter community calls are a place to just celebrate what people are doing in the ecosystem, make connections across communities and projects so that we can all continue work better together and remember. You know why we do all this work for all these cool things and people getting excited and doing amazing stuff with all these tools and standards.

A

um The first thing I do want to remind- I know you heard it jupiter community calls are recorded whatever. That means for you. Just keep that in mind. I'd also like to remind people. This is a jupiter event. Like all jupiter interactions, you are held to the code of conduct. That includes me. By the way, you can always find that at jupiter.org conduct and yeah, if you haven't been here before what we normally do, is we have an agenda which I have linked in the chat.

A

Hopefully you are signing in and we go down what people have posted before like hey. I want to demo this or share this. Today, though, we actually have an empty agenda. So after I get through my also a short list of celebrations and quick reports, uh it's open. If you want to present something just let me know, and we can do a quick demo, it's a totally open time for you.

A

If not, I do have some discussion items that we can do since we have especially a nice kind of smaller group today, um some fun community stuff. We can talk about that. Maybe isn't, as emphasized all the time, so I'd find that exciting too, but yeah. Does that sound good to everyone see at least one not.

B

A

Oh I'm recording with here I can go to speaker queue. I always forget to switch that. Thank you. I've seen some thumbs up nods. Thank you carol. Okay, awesome!

A

So, first things. First on that, looking at the short report celebrations and shout outs, the first one is me: it's just me saying. I know that uh cycle four czi open source, essential, open source grant work has been happening and there's been several people in the community that have been talking about grants that they're starting the application process for- and I just want to say, congrats to them. That's a lot of work and we appreciate everyone's efforts of you know trying to work hard to improve this ecosystem.

A

So thanks there oh and we have we have a newcomer yay to the short announcement section, so I know you're still typing nick, but go ahead whenever you're ready.

C

Unmute there we go. That's me um hi y'all, so oh.

B

C

I'm on the camera- I didn't know that uh so the uh I've been doing. Jupiter starters for a couple years now and we've shipped in a bunch of products and just hadn't done a real release in a long time. So I got that out this morning. It's up on uh kind of forged in a tick. um It's basically a technique for adding files to your jupyter client. So you would make a collection of files like I wanna.

C

I want a to-do list or I wanna. I don't know whatever something you press the button. You get the file right or you get the files or you get a workflow that goes around in a circle where it's like. Well, do you want more files? Okay, you have more files, um and so we've used it for a bunch of things I'm going to be using it for shipping some course material.

C

um So you know you, you put your content inside of a package where it belongs, um and then you get a button and that gives you a local copy of it. So you're, not you know you don't have to learn how to.

B

C

To get your your content for the first time or something like that, but.

B

C

It it's not on someone else's server, um so the the one of the things under the hood there is react on schema form, um which is a cool piece of tech and we'd love to get that into the settings uh editor on on lab. But just you know finding the time, um but that is a reusable component and we use that on wxyz and some other places. So if you ever like man, it sure is annoying to actually write forms and pick a react.

C

Library just ignore that write, a json schema and be happy, and then you know you get all that stuff for free. So that's all.

A

Thank you. That sounds wonderful. I hope it goes well getting it out to students soon. um We also have carol on the short reports take.

D

It away yeah, I just wanted to give a shout out to darian who's, been doing lots of work behind the scenes to get the jupiter distinguished contributors group uh moving and working on the next cohort and as well as all the work he's done with governance and steve and others on that. So thank you.

A

Yeah, thank you very much from me as well. That's a great thing to call out. Thank you, okay, anyone else on short reports. I will be quiet for about 15 seconds to give you a chance to think about that.

A

Okay, that was about 15 seconds in my arbitrary counting um great well, thank you. Everyone who took the time to share that is there. Anyone who does have a spontaneous share they want to give. This could be a demo, but it doesn't have to be. I remember a couple community calls ago we had someone, ask hey like what is a jovian. I keep hearing this term right, so this is open for questions any kind of discussion. If someone wants to lead that or it has something they want to say.

E

If nobody suggests something in 30 seconds, I can, I can try and do a demo. uh It might be a demolition but we'll see.

A

We we, we say demo so that we can include all demo ending forms. But yes, that's a great plan.

A

C

I guess just a quick question: people are probably familiar with if this, if this than that the ifttt uh has anyone ever stood up, something like that like huge on a on a jupiter hub that can you know if I get an email, then then make a ticket.

A

You muted yourself in the middle of that I think.

C

Oh sad, um I do that a terrible keypad uh if that's than that on jupiter hub has anybody ever done that.

D

Within a notebook.

C

uh Within a hub so having the persistent, if this and that service be in the hub and not be the big assistant that, like an on-prem.

D

I don't know I mean somebody may have done it using github actions or something, but I haven't heard of anything in specific. I don't know.

B

D

C

Yeah, so the the open source, one that I saw is ruby bass, is called hugin um and might be a really really cool thing to give to people. You know like I'm gonna email, myself, a notebook, and then I get a binder back or something like that. Right.

D

How do you spell the ruby one you.

D

B

I think you may just paste a link into other zoom or maybe the notes.

D

Oh, that would be even better still.

B

C

I was talking, I was talking, but it was muted yeah. So huge is some raven from north norse mythology or.

B

B

Okay, I'll see you, I can see what I mean, I'm basically running hogan as a jupiter hub service and then do.

D

Everything yeah, I mean it seems reasonably as long as you get the permissions and the security.

B

Right, yeah and being in ruby shouldn't be a problem anyway, because it uh services incorporated directly into cheaper hub yeah.

C

B

C

Have a very it's a very opinionated deploy, uh I I don't know.

E

C

E

Could you guys talk tell us what if this, then that is for folks who might not know.

C

Sure so, if this than that is a is one of the low code uh approaches to doing complex things that span a number of services.

C

So uh let me find the link to uh if this, then that um so you know, basically you put all your keys up in their big key box uh so that they can pretend to be you on everything. And uh then you create these workflows. Where you know the workflow would be triggered by a release happens and so, okay, the release happens.

C

Now I want to post to twitter and I want to post to facebook, and I want to post to my blog and I want to you know, trigger a security scan or something like that. Yeah zapier yeah, it's just meant to be more. You know you give people easy to use, workflows that uh once the the integration is written, then it just kind of works.

C

There's a lot of workflow engines, it's just that's a one that that people like to it has many integrations. So people like it.

A

Would you want to talk about the use cases you're looking into this, for no pressure by the way.

C

Yeah well so, if you're supporting a team of scientists that are um you, know, kind of all over the uh the spectrum of of computer savviness and you want the interactive computing to be the focus of what you're getting done, you know, so you want stuff to.

C

You, know end up deployed as a voila dashboard, or something like that. uh If you have easy to use entry points into it like email and like your your chat machines, like your slacks and your hip, chats and your uh zulip and stuff like that, if you can trigger interesting things happening like I'm gonna email, I just got this new data set from my offline sensor.

C

I'm just going to email it to myself and to a bot. um When I get back to my lab, I've already got a binder built that has that run with some analysis run next to it, and I can share that link to my team um without having to re-plumb every you know like. I don't want to write an email kernel right, um yeah and there's not a ton of jupiter integration stuff on there, but I think once an organization is large and diverse enough.

C

More tools like that are they start to look pretty attractive.

A

I know that sounds great. um Thank you.

A

Anyone else have questions or thoughts on that.

A

I'll take that as a no um thank you all tony did you have a possible demolition? You wanted to share no move.

E

On I can do it, we can go for it um if uh people can ignore the banging. In the background that's going on because of housework. uh Hopefully that won't be too big of a problem.

A

I don't hear it right now if you're worried.

F

Well, it's not going on right now, but okay, okay, let's party.

E

um So this is like intermediate work. uh I I still have some tests. I need to fix, and um thanks- and uh I gotta thank nick a whole bunch for all the code. I'm gonna use of his, but also a destiny, wrote uh testing infrastructure. He wrote for this, um so what I want to talk to you all about is a library called schemata.

E

um It's not imported import, schemata um and schemata is uh a vision of a future traits library um that complies more with json standards than say traitlets does, um and you know this is inspired by tools like pydantic which allow you to write, json schema, but all of them do it as objects, and uh I wanted these things that these sort of types that can describe themselves, because when types can describe themselves like saying that a string is marked down, um I can go and print it in a rich terminal format with a tool like rich or if I just tell it that it's a plain string, it shows up a certain way, but the nice thing is: is that all of these objects?

E

um All these things are actually types, so this is only for ipython people. This is not generic, um but I believe the specification can be used a little bit further, um so everything that is instance type. um So all of the things in this library are actually types like even this thing here. So let's go and take a look at this crazy little situation. So what this actually gives us is some symbolic stuff, on top of um sympolic type composition, sort of similar to what um you get in typescript.

E

So what this says is if I have an integer um with that, has the properties bar and that must be an integer. If I have a dictionary with a property bar that must be an integer. I can then pass it into a ginger string. uh Then I can tell it to render as markdown um and then I can go and print it. So when we run something like this, we get this nice display.

E

um But then, if we go and look at this here, we get our pigments display. um I think I have another marked yeah that one's not working yet part demolition, but nevertheless we can keep sort of pulling these things away and eventually get back to the original types. But what's super nice about this is that we also get all of this validation that's going on, um so let me go a little bit further down.

E

um Let me take this notebook that I'm working in, for example, um I'm gonna load this thing in as a dictionary, um so let me go and do this and yes, we get a long string. You expect that. But now, if we say that it's a content, media type of application json, uh we use the ipython rich display system to show us a pretty view, and now I can actually go and even search.

E

My oh, this is that might be slow, but I can go and search for things now and query my json content or my json blobs. So if we go and look if we go to this object here- and we just pop schema on here- it tells us what the schema of the thing is. So this is complying with json schema and my hope is to kind of make a a system where you can comply with json schema, but still use like a native type interface.

E

So, let's go and take a look at my dictionary here um and if we look at this schema, um the uh what we wind up seeing is that uh the annotations here become our properties in the schema for those that are familiar. This is going to check whether these types are correct and since um the this should not be required, so I got a little bug in here, but it'll go and figure it out what the required features are too.

E

So if I have my dick um and then I go and show it to you all, you know nothing fancy we just get dictionary, but then, since I'm carrying along the schema with this thing, I can take react. Json's schema form which nick talked about earlier and get a widget right out of it, which is super, duper pretty, and I'm still working on that. But then you know I can take the same type and append a rich display to it. And now I have a nice view for my terminal.

E

uh So one of the features of uh traitless, that's nice is the observable pattern. So if we say take something like t- um and I give it a and b it goes and computes c and the reason that it does- that is because, um in the self definition of the class we've defined the things that t depends on and now, if we go and look at schema, um it's actually adding the dependencies property, which is a part of the schema convention here.

E

So this dependencies property is really nice for observable patterns, because it allows you to like a priori figure out what order things need to be executed in um so with something like this yeah. Let me go, let me grab this one here. So here's another example.

E

um So this one is saying: uh let's go and look at the schema and it'll tell us exactly what it says. um So it is saying that z is a string. W is a number y is an integer x is a string and v is a list, and I have these dependencies here and again. The dependencies are defined by the annotations um on the self object. So y requires w and it turns w, which is a float into an integer.

E

It turns and then it goes and takes y and turns that to a string and then it goes and takes x and turns it into x. So now, if I go and just give it w, um I get back the whole thing um and oops. Let's do that's not what I wanted to do. This is the last thing I'm going to type for you all all right there. We go uh all right so now I've got t here and then I can go and say w um equals 999.

E

And now we look at t and it goes and gets updated again. So these dictionaries have observable patterns and again since there's this dependency feature, um we can go and do some super smart stuff um and then uh one of the last things that's kind of uh actually no I'll stop here. Yeah, there's a lot of other things that this can do.

E

uh It can handle conventional python types and things along those lines that can go and make its own uh applications using typer, um but anything that uses schema we're using um and uh even something like this dictionary here has something on it called a strategy um which is a hypothesis testing strategy and um these strategies uh I'll actually allow these type systems to sort of test themselves.

E

There's a lot of nice properties that fall out when you know a lot about your types, so I'm gonna stop there um and yeah. My hope is sort of a more descriptive uh trait type system that we can use to be a lot more expressive about the applications and things that we build.

A

That was amazing. Thank you so much especially for showing that unexpectedly, can I ask the ignorant question and say you were talking about this approach applying more with json schema. I guess what are the benefits of that or what is the? Why, like? Is there a specific reason or just because following schema is generally good.

E

Well so json has a standard, has a bunch of standards, organizations and there are specifications and implementations of a lot of json things and when we're working in notebooks most of the time we're exchanging jsonish objects and json structures. So right now the trait and type systems that we have are either closer to python or typescript.

E

But if they can be closer to json, then things on either side of the pro program can understand it, because we can deal with all of these json standards in python and these json standards in javascript and ruby and go because all of the there are tools built around all of those specifications.

E

So as a reference implementation, it can be used in most languages.

A

Yeah that sounds critical or very helpful, at least for sharing. um Can I also ask one more thing: does anyone else have thoughts? First, actually,.

A

This is your seconds of silence to jump in.

A

No okay, uh oh no go.

D

Ahead carol, I was just gonna say there was a question in the chat that um can you use it with something like fast api, which is actually a question I was gonna ask so the answer that nick gave is a very happy answer.

E

Yeah yeah so right now I don't have that going yet. um So one of the really cool things is yes, it does work with fast api, um so I can generate pedantic types from this system. It's super duper easy to do that. That's awesome! So, by being able to write the pedantic types, it becomes really easy to write um fast api and, in fact, fast api. The way so fast api uses format, python format strings, but the approach we're going to take is to use uri templates instead because that's again another you know cross-language standard.

E

That applies a little bit more generally so yeah. I I think it's going to be pretty easy to get these things consistent. So right now all of the schemata types convert into typer types and they convert into pedantic types, which means any of those tools in that ecosystem or space become accessible from schemata and uh and we can also turn schemata into traitlets. We can turn scamada into params.

E

We can turn scamada into ipi widgets, so uh yeah, it's really thinking about, like being a general lego sort of interface for types that work with all the different back ends.

A

Yeah, thank you for asking that christian too. That was great question.

A

So what I was thinking, let me share yeah. That's the idea, I'm so glad that you did. Thank you for sharing um I or do you also have. I know that we talk about accessibility, stuff. Some are there any implications for that with this, because you were talking a little bit sounded a little bit like labeling to me in some of these cases like? Are you labeling outputs? Is there any kind of.

E

Well, so, what's nice is that for accessibility again there are standards and specifications for them, so what it really would be so a lot of times. What actually one of the things that has helped me learn under learning about a program's type system really helps. You understand the program sometimes.

F

uh Where oh crap, where was the? What did you ask me again sorry.

A

I was asking just about accessibility implications.

F

E

A

F

E

Yeah, so one of my favorite things about schema is that if you spend time understanding the program and understanding its type system and writing down at schema, you often understand things.

E

So the exercise for schemata would be to take the uh arya standards and wright do a little bit of research and convert the aria specification into schemata types, and then we could validate and verify those things and basically say: hey we're not going to let you put an html element on the page or an image object on the page if there is an alt text or if there isn't aria so, like I kind of like the idea like that's, I look.

E

I've been thinking about this idea of an html, accessible type system that actually does validation on whether or not it includes certain standards.

A

Yeah, no, that sounds oh, go ahead. Nick.

C

The other dimension of that is that um there are also standards for internationalization that are encoded as part of a json standard. So.

E

Yeah, oh yeah, this is going to do translation too. uh It'll do translation closer to json, rather than like the i18 or the program language stuff. Sorry to interrupt! You forgot about that thanks nick for reminding me no.

C

Big, so I mean it's. The the only unfortunate thing- and I raise this in the chat- is that the json schema itself is not actually a standard. It is a draft and there is some challenges in the community of the maintainers of those libraries not wanting to support things that different stakeholders have gotten in there. So we can hope for a spec at some point, but it's the best thing we got today. You know.

A

Yeah that makes sense.

A

That sounds very exciting, though, as it does, I know I would be excited by things that tell people hey this needs to be accessible. It's not okay, so that's amazing or nationalizable is that the yeah.

E

Well, you know I mean so it's tough right because, like you want to be able to tell somebody also why it's not accessible right so like with this idea of using schema, you can put little pieces in there that say: hey yeah, when you see this error, it means this kind of thing too. Right, like. I think that one thing that's going to help us to accessibility is like really good error messages to help people along.

A

Yeah, because these aren't things, people tend to be learning, you run into it. When you have a problem and that's kind of it, I find um yeah, that's awesome. Does anyone else have comments? There, gonna have your moment of silence again to.

A

Think, okay yeah. Thank you once again for that sudden share that was really great to see especially work in progress. I know not everyone is has enough courage to share that. I know I don't always so thank you.

A

Does anyone else have sudden inspiration, sudden questions anything like that? All.

A

Welcome this could be, I don't know. I have some discussion topics I can bring up if we're ready for that, but yeah, okay, I'm seeing at least two nods, so I will take that for the whole group here. Thank you um so one that might be harder, but I think, is a good question, since we kind of started it before, recording the call anyway like what our favorite notebooks people have had, they could be something you've authored. It could be something that you've seen.

A

I know that's why we were looking for the notebook hall of fame notebook, which is kind of funny meta, notebook being a favorite notebook.

A

I can also, I know we put the link in the chat, but I can probably screen share that. If we want to look at it closer to people of other notebooks, they remember something they loved, that they wrote.

E

I have a hard time, loving myself, maybe for some people- that's uncomfortable to hear, um but.

A

They can be someone else's notebook, you're.

E

Good, no one of the earliest um one of the earlier deathbeds blogs. uh I realized that this is still one of the least sucky notebooks I've ever written um as far as being useful. I know we talked about it this week too isabella, um but this uh notebook talks about at least in python, land writing notebooks that are testable um and or formally testable and stuff like that. um So, oh!

E

No for me and for the people that I've shared it with folks have found it kind of useful and understanding a little bit about how they can understand running tests in the notebook um so yeah, that's one of my favorite personal notebooks.

B

Can I share plugins instead of a notebook yeah, so this.

G

One's not working.

B

Ages ago and then uv panda from jupiter have joined in so that basically lets you get all food in its desktop uh inside juvederm. On my binder, I guess some people have probably seen that before I see if it's quite cool, this is awesome.

A

Should I screen share these? Does anyone want to see these? Are you all just clicking and looking on your own.

B

You can always click on it.

A

Yeah nope, that's true.

E

Maybe we can add favorite binders to this, like I want to know. I'm personally curious is jeremy still on the call, I'm curious if he has a favorite binder of his because he's made so.

A

Many don't know.

A

E

Why are these people's.

A

Favorites, maybe I should ask that too, like we've got a few shared, but what is it novelty? Is it something about how they're structured like? What's the important part to you,.

B

Well, the reason I kind of like without plugging a michelin bias, but I've worked on it, but it's just because it's just something so usually I'm getting using showing off something that you'd never think binder could do.

A

Yeah unexpected.

C

Yeah, I got a, I got a silly one here um and it does have a binder, but it's boring binder. uh Basically, how do you take something that doesn't have a specification and build a specification out of it and the reference implementation?

C

So we hit that in jupiter a fair amount? Where there's you know it works in ipython. So that's the spec now uh and it's hard actually to get. You know these these schemas and and grammars and stuff like that extracted. So this takes the the uh language server protocol markdown source for their specification and the reference implementation in node.js and builds a json schema out of it. uh It's terribly broken, but it's an like as a thing that you can do in a notebook.

C

You can lay out what would be a very complicated piece of software in a way that that you know incrementally builds it up. So it's possible to follow it.

A

E

Nick, I was I shared with pierre, if he's still here, yeah pierre's still here I shared with pierre uh the day I deleted 350 notebooks, with rmrf in front of everybody and ruined everyone's spirits.

A

Was that intentional.

E

Yeah, I actually told people I was going to do a performance, art piece. It was uh yeah, it was a little. It was a little much. It hurt people's feelings.

D

I think I just shared one of my favorite sets of notebooks. It was taught at scipy uh a number of years ago, it's probably five years ago, at this point and it uses pie, die senpai and some other uh scientific python language uh packages to simulate the physics of somebody standing up and it it's really well done. It kind of goes from equations and and each notebook kind of builds on the one prior and there's about eight or nine of them.

E

Does it make these visualizations uh in a crude.

D

Sort of way yeah crazy- I mean because remember it was it was done so long ago, but it would be one that would be worth updating with more model individualizations. I think jason moore um he's done great work. He also has a textbook where they took. It was like classical mechanical engineering textbook, and each chapter has a animation built on a notebook that simulates a physics concept like a cart driving along a road with increased friction.

E

Very cool and it's and they gave it like four years in a row. um I don't know how I've never stuck to a presentation for that long, but that's super impressive that they got that much life out of this.

C

D

Really about that.

C

Yeah, so j uh just moore's been doing um senpai for a really long time um and one of the one of the crazier things that we did was get in the latex parsing into that. uh So this is a. This is a silly one of building up hypotheses that does it work in real lat tech and if it works in real life tech, can we parse it, and if we can parse it, can we round trip it back through senpai to get the real math back out?

C

uh It falls down on like left brackets or something like that, but um there's another one where it's like you'd, never write a program that way but being able to expose an aspect of testing you know as a narrative is, is very cool. I think I think I actually presented that the same day that you deleted all your notebooks.

A

I'm a momentous day for sure.

E

Oh no, the audience. The audience has a rough time uh when nick and I are back to back.

A

I'm sure is lovely. Thank you all for sharing these notebooks. These are really cool. I I don't know I love seeing these.

A

Do we have more there? I can also.

H

So another one I'm curious about is: if people will pull up old, notebooks, oh and see if they.

A

Are are we talking like re-study, your work, see if they still run kind of both yeah.

E

Kind of thing: uh no, no, they don't. I don't think they still have to run uh but yeah. I just I'm just thinking about when it used to be possible to use percent percent javascript in the notebook you could just do the craziest.

E

A

Wasn't there I don't know, isn't there some kind of oh there, it is the like 10-year reproducibility challenge. Do you know what I'm talking about.

E

I have seen that yeah.

D

Somebody was doing research on how reproducible notebooks were over time, but I can't.

B

D

A

Well, I don't want to keep people we're good to go, so I can give you your time back and let you go early if you want, but thank you so much for.

D

Oh go on carol. No, I was going to say thank you for organizing and and thank you for the time back to.

A

Yeah, thank you so much all for being here. It was a quieter one today, but I think I have a lot of resources, so I hope you I hope you enjoyed it anyway, even with our smaller group. So thank you. um What was I gonna say.

H

E

Just shared a four year old notebook which I think is like I just feel really uncomfortable doing it, but it goes to show how comfortable I feel in this meeting. But yes, god there were some gross things that we used to do well. Thank you.

E

This was when uh this was when I guess adam was written in. Is this a seven-year-old notebook nick.

A

C

Yeah short just id and everything wow wow ipython.html.widgets.

E

And and widgets still works basically exactly the same way.

E

That's good or not. I don't know this reminds me of that guy who gave the talk in chicago's jupiter day where he was like running his laboratory with widgets.

E

Do you recall that nick.

C

Yeah yeah, no, he had for every piece of equipment, he had a widget and and he he he ported like 10. You know: megabytes of labview to 20 classes and widgets.

F

It was very cool.

A

I can only imagine all these great demos I haven't seen over the years good to hear summaries of them, um but yeah. The main thing I want to say before people go just so you know next community call. Hopefully I won't be smothered by deadlines, and I won't forget to post about it earlier than I did. This time will be april, 27th um same time same place and yeah. Thank you. So much all for your time. I really appreciate it and I hope you have a lovely day.

E

But I'm not going anywhere just yet because I know carol has an old notebook somewhere.

A

That sounds like a challenge. I don't know this.

E

Is a challenge.

D

Yeah they're all.

E

D

My repos somewhere they're.

E

All all in the repo somewhere, so just like, dig far back in the history.

D

Yeah like if you search on music 21, you might find something quickly.

E

Okay, okay, okay, uh all those music ones were so good.

D

Thank you. They were fun to do. I still use them every once in a while, but I need to update it because music 21 changed their um yeah data structure type stuff. So oh no.

C

They're the is that is that does that do scoring, or just the music like just the music.

D

It's just it's a music theory, analysis library, so you can do some of the sheet music, rendering and stuff. But it's really for like comparing like how close different pieces of music are um in terms of their sheet music and like comparing all of box concertos or something like that.

C

Oh man, we had a. We had somebody do that at pi data uh from uh one of the colleges up the road here in in atlanta, where they had done some insane visualizations of like everything, everything coltrane ever wrote. You know all together.

E

Yeah but but they we actually, there was actually like a throat singer there too, or I forget it- was they studied um atonal languages and uh yeah? It was really cool like there was a band and stuff wow.

D

Neat yeah, I think that's one. That's probably that's the library for sure the music 21.

E

So santa you know the the pandemics killing me because, like I can't figure out what new things you're preaching about carol, what what are your most recent presentations about.

D

You know, there's nothing really new and exciting. um You know I've been doing some well. I did a p5js one for um play, fest, which was fun. um It's like an intro. It's a conference all about play um which yeah emery thomas who's, the professor up in minnesota.

D

It's her conference and she's like super out of the box, um and so she asked me if I would do something- and I said sure- and then I spent this weekend- uh two hours talking with uh uh middle schoolers about 70 of them about python and jupiter and community. So that was pretty fun. Everything I talk about is on speaker deck, but nothing really new and exciting. I'm sort of missing the traveling around the world and seeing all the cool things up close and personal.

C

I just dropped.

D

C

Has anybody seen the pocket ai thing, so it's a modular 3d printed computer and I don't know how bespoke I mean it's pretty pretty darn bespoke, but um their demos are totally nuts like carol. We've been talking about this for years of a literal, jupiter notebook. That's like a trapper keeper that you put all your computational. You know bits and pieces in it and yeah.

D

C

D

I need to do more education, stuff and less startup stuff.

C

Education startups.

D

C

I don't even I wonder what nicole's up to these days on her on uh go pi, go.

D

Yeah, I think she's.

C

D

Doing this same old same old, she.

C

D

In my feed every once in a while, but.

C

Yeah, if anybody hasn't seen that before uh so this is a robot that runs jupiter um and you you log into it via wi-fi and use your robot from the jupiter running on it uh and you can write in ros, you can write in python. You can write in scratch. It's uh nuts.

D

Yeah all right, somebody is pinging me, so I am going to go, but thank you all lovely folks, and I will see some of you in the next meeting bye. Thank you.

A

Bye, I'm gonna stop.