Fork me on GitHub
#off-topic
<
2019-04-03
>
tagore01:04:50

@richiardiandrea I think you really do have to be a bit careful about introducing new "weird" things, even if they seem objectively better to you (and even if they are pretty much better.) Adoption requires buy-in from a lot of people in most cases. And most people have an aversion to the new and weird. The ground must be laid very carefully.

richiardiandrea02:04:27

Yeah I think I have definitely noticed this šŸ˜„

jaide16:04:43

Anyone know how to convert a bezier curve like cubic-bezier(0.42, 0, 0.58, 1.0)into a function of time?

jaide16:04:13

Using this I can see the relationship but not sure how to turn that into an equation http://cubic-bezier.com/#.42,0,.58,1

polymeris19:04:28

The CSS form is cubic-bezier(P1x, P1y, P2x, P2y), with implicit P0 = {0,0} & P3 = {1,1}

jaide19:04:38

Thanks, I usually find wikipedia too dense to learn math concepts from but this is really approachable. I shouldnā€™t have ruled it out so early on. Thanks again for pointing me to it.

jaide20:04:05

Though admittedly, Iā€™m not clear on how to plugin the P coords into it. Where it asks for (1-t)^3P0 for instance, is P0 either P0x or P0y or some combination of the two?

polymeris20:04:29

You apply that fn twice, I think, plugging in P0x, P1x, etc, to get B(t)x

polymeris20:04:46

Then again with P0y, P1y for B(t)y

polymeris20:04:10

The coordinates should not depend on each other, I think

jaide20:04:43

Ok, I can work with that!

polymeris20:04:35

Also, P0x & P0y are 0 anyways, so that simplifies the equation quite a bit

jaide22:04:14

So by plugging the coords I get an x and y where 0 <= t <= 1

jaide22:04:37

But what I need is another percentage between 0 and 1

polymeris22:04:08

not sure what you mean by percentage?

polymeris22:04:35

does that ā˜ļø look like the curve you expect?

jaide16:04:23

Thatā€™s not quite what Iā€™m looking for, however I did find exactly what I was trying to do: http://greweb.me/2012/02/bezier-curve-based-easing-functions-from-concept-to-implementation/

dpsutton18:04:15

and related job posting: https://redplanetlabs.com/careers

šŸ‘ 4
drone19:04:02

cool, I wonder what is referred to by mentioning compilers in the closing sentence: excited to push the boundaries of whatā€™s possible with compilers

Alex Miller (Clojure team)19:04:22

it compiles AND its a flight simulator

šŸ˜† 4
dpsutton19:04:02

so basically excel

Alex Miller (Clojure team)19:04:15

compiled excel

šŸ‘ 4
Daniel Hines19:04:03

Is GraalVM+Clojure a thing people are doing in production?

Daniel Hines19:04:30

Err, people do all kinds of things, but is it a relatively common and/or not-insanely-difficult thing?

souenzzo20:04:32

@d4hines once after "warm up", JVM is faster then native-imageĀ¹ I think that makes no sense to move from JVM to a native-image But i see people talking about it on "lambda" scenario https://gist.github.com/mfikes/07be413e34750a9f451bd76d937ff545

dominicm20:04:00

Graalvm isn't the same as native image.

šŸ‘ 8
dominicm20:04:32

We are evaluating it for an internal system due to the interop capabilities.

Daniel Hines20:04:45

ā˜ļø Thatā€™s what interests me - the interop.

dominicm20:04:42

I don't get the impression it is particularly harder than graalvm+java. There's a gotcha or two in clojure core iirc.

Daniel Hines20:04:42

If I worked in a Java shop, I would try the whole ā€œClojure is just a library for Javaā€ trick, but we do native C++ stuff, where that isnā€™t an option.

Daniel Hines20:04:14

So Iā€™m interested in Graal as a way of making our stuff polyglot.

dominicm20:04:31

Llvm bytecode?

Daniel Hines20:04:00

I donā€™t know much about LLVM (or native stuff in general actually - this a new world for me). What would it look like? Clojure => Java Byte Code => LLVM => consume from C++?

dominicm20:04:02

I think graalvm can consume llvm bytecode for interop, so you can call existing code.

dominicm20:04:10

Although you can already do that with JNI

Daniel Hines20:04:05

Oh, so go the other way C++ => JVM => Clojure.

dominicm20:04:17

I think I remember a demo of going nodejs=>java, but haven't seen it actually released

Lennart Buit20:04:37

yeah, the polyglot nature of Graal sounds very interesting to me