@rantingbob: I have clojure code modifying a polygon mesh using the http://thi.ng library output to an X3D file that I'm uploading to my Shapeways account where they render the file, which I'm simply saving as a jpg file. You can view/manipulate the model using the Shapeways 3D tools here: https://www.shapeways.com/model/3dtools/4108313/1/26?key=693ebbbc85d333d8708ec57a96dff6e2

Thats really cool, thank you

Glad you like it. I'm having a lot of fun making them. You can see a bunch of them on my profile page on Shapeways here: https://www.shapeways.com/designer/pkobrien

Wow, that link came with a pretty spammy description. Wish it didn't do that.

Anyhow, it'll be fun to see some of these actually printed.

All my code is on github here: https://github.com/pkobrien/cad

Might have to change that description since I'm not really using OpenSCAD anymore as I can do more by manipulating a mesh directly in Clojure and outputting to X3D.

Here's one in silver. Don't ask what it would cost to actually print it. https://images3.sw-cdn.net/model/picture/625x465_4093153_13239119_1449113030.jpg

How well do they print out?

@rantingbob: I have no idea. I haven't gotten any of them printed yet. None of them are actually for sale. I'm still working on my process. These are just baby steps.

@meow: What's the end goal, or is it just for fun?

Sell a lot of these, have fun, buy an island.

I just need to sell one of these and I'll be set: https://images4.sw-cdn.net/model/picture/625x465_4088521_13239589_1449116396.jpg

Or a million of these: https://images4.sw-cdn.net/model/picture/625x465_4088521_13160630_1448989545.jpg

That gold one is beautiful

@meow: that's really cool!

Im trying to find out where this error message is coming from : CompilerException java.lang.IllegalArgumentException: Don't know how to create ISeq from: java.lang.Long, of this code : https://www.refheap.com/112295. Anyone who can give me a hint where the problem is

@roelof: apply's last arg should be a sequence.

you mean args ?

@roelof: yes

@andy.fingerhut: thanks, I deleted apply and everything seems to work fine

any remarks : https://www.refheap.com/112297 from the experts

@roelof: A minor comment is that if you wanted to change your code to allow a variable number of args as the name 'args' implies, you could put the 'apply' back in, and change process-user-action args to [action & args]

...or call it like (process-user-action :deposit [200]).

args is only 1 number

it's meant as very simple accounting system

The body of the deposit function could be changed to (update amount :amount + deposit-amount).

@henriklundahl: oke, I Always thought that only swap could change the contents of a stom

stom = atom

@andy.fingerhut: as I meant it args can only be 1 number

Well, the update function doesn't really change anything. The code I proposed does exactly what your function does.

oke, I did wonder if my function is then still pure

and I did not know update as function exist. The clojure programming book makes no mention of it

but then I could also change the body of the withdraw function to do the same

I believe it was added in Clojure 1.7.0. Before that there was update-in.

oke, thanks , I like update more because it more clear what is does then swap

(update-in amount [:amount] + deposit-amount)

thanks, Learned more and more

swap! is something else. It's for changing atoms. update and update-in just returns a new map.

sorry. I was confused. you are right

the swap is in another function

So you need swap! in process-user-action, but not in deposit or withdraw.

as I said earlier. you are right. Think It's here to early

Now I can think of how I make it work that the atom is updated and a description is added

So I can make a sort of report of the account

I think I need to "update" a seperate map for that

or is there a better way ?

Thanks for bringing update to my attention, @henriklundahl. I always hated my solution of (update-in coll [:one-key] f) 😄

NP.

Hmm

@arohner: replying to yesterday's question: yes, clojure core is AOT compiled with direct linking on (since 1.8.0-alpha3)

does anyone have a link handy where I'd be able to read about the work that's going into 1.8?

just curious about what's coming

found this: http://dev.clojure.org/jira/browse/CLJ/fixforversion/10254

there's release notes linked from that page too that might be a better summary

@clark: http://dev.clojure.org/jira/secure/ReleaseNote.jspa?projectId=10010&version=10254

@matt @birdspider awesome, thank you!

Looks like lots of perf improvements. Always welcome.

My first polygon mesh model is now available here: http://shpws.me/L0c3

Or in Platinum: https://images2.sw-cdn.net/model/picture/625x465_4110302_13243544_1449152760.jpg

i'll take a platinum one for review, i'll write something nice I promise 😛

jokes aside your work does look amazing

@marvotron: tyvm

@clark you can always find the latest change log at https://github.com/clojure/clojure/blob/master/changes.md

@alexmiller: perfect, thank you!

If I use atom for a very simple accounting app. Schould I store the transitions into a map

later I will try to put everything in a database

@ajmagnifico: no problem. Can happen

@ajmagnifico: You can delete messages.

lol, thanks @exupero

such amazing technology!

nobody who wants and can think with me if my solution is a good one ?

@roelof I'd use a vector. Unless a time stamp is your map key.

later on I wil try to put a date into the transaction

I have this in my mind at the moment { :date 01-02-2015 :description "taken from the bank" :amount 100 :transaction-sort: "withdraw}

Seems fine. I'd probably do -100. Then you could sum the amounts in order to see a balance at a given point.

oke, but a I can image that a user finds -100 confusing

So then I have to find out how I can put a date into the map

Otherwise it will be a string

@jeff.engebretsen: thanks

np.

@roelof: https://github.com/clj-time/clj-time is a good library for dates

@meow: Are they Catmull-Clark subdivision surfaces?

@cfleming: that one starts as a hexahedron, then conway kis'd, then catmull-clark three times, then another conway kis, then colorized by the face normal

maths is pretty

and fun

Nice, subdivs are like magic

the fun for me is to take these known algorithms, expose all of their parameters (or take a callback function that can do whatever the heck it wants) and then run through strange combinations of values and out come unexpected results - many are ugly, over time I can usually get it to settle on something that I like

catmull-clark is really nice, but also a bit heavy-handed - 3 seems to be the magic number before it just produces something indistinguishably smoother or a sphere

but I'm also intrigued with low-poly objects and using color to play tricks

yeah. i remember being utterly floored by the demo scene back before we had GUI OSes

the amazing amount of things they could pack into a meg or two, considering there were ZERO frameworks or anything. a great example of embracing constraints for creativity

so I recently came up with a different algorithm that maintains the bounding box of the original shape and doesn't move so much towards a sphere and it creates some really intriguing results - plus the parameters make a big difference so I can, for example, run it three times but each time with different parameters to get really controlled results within its basic mode of operation

I've rewritten the catmull-clark one time already from the version that comes with http://thi.ng and I'm about to rewrite it again and make it parameterized so I can mess with how it works

Yeah, just yesterday I rendered all my designs on shapeways in different materials so that you can see how the shapes look without color and how the color is really adding more depth and interest than actually exists in the basic topology.

i love the normals-decides-colour trick

there are so many directions I want to go with coloring and I just have to keep putting that off to work on the mesh operators for now

the color is just a function that gets passed the mesh and the face, so the sky is the limit on how the color is determined

I have one version the blends the color based on the face's neighbors -like cellular automata life rules

distance from camera, distance from mouse, combination of the two, etc

but it could use the distance from the centroid of the mesh, or distance from anything or averages or whatever

I do nothing with camera or mouse

I'm applying the color to the face of the polygon - how it gets rendered is done on Shapeway's servers.

I'm doing as little as possible to get the 2D rendering

ohh i see

Their rendering is supposed to simulate the materials they have available via their 3D printing service.

These are meant to be printed objects. I'm conforming to all their printing restrictions.

i’d love to see a material that changes colours based on ‘normals’

that’s awesome, though. you planning to have some printed?

I'm planning to have other people buy them. 😉

bazinga

made with maths and lisp and love

By the time I upload them to Shapeways I'm kinda tired of looking at them... 😉

and the dozens of variations that get discarded along the way

Actually, its kinda hard to stop spinning them around and zooming in and out. It's addicting.

I want to make virtual reality worlds out of this stuff as well.

in (def columns# 9) what does # mean?

nothing. think of it as the 27th character of the alphabet

just in the name then?

so if i need to use that, i would just put columns# normally..

yes

pvinis: in the context of a macro unsplice

`(let [foo#  ....])

aha

cool. thanks

I was kind of expecting flatten to get me a transducer

What do I do instead? (mapcat identity)?

maybe cat

its a transducer

oh, right; I guess cat is what I want

(well, cat doesn’t flatten all the way, but I don’t need that)

new in 1.7

@jaen: thanks, I will look and play with it

ok so im trying go model a card game, and wanna generate the deck. i have this (map (fn [n] (suit-card :hearts n)) (range 1 14)) which generates all hearts i tried to do this (map (fn [s] (map (fn [n] (suit-card s n)) (range 1 14)))) but i get a fn back so i guess map in map is a bit weird?

Are you passing the suits to the outer map?

oh shit...

@jeff.engebretsen: thanks… duh

Did that fix it?

yep

cool.

i am not used to clojure yet, so i just closed all parens, without putting the suits there

If you want all the cards, you could also try a for comprehension.

I'm confused why it returned a fn instead of throwing an Arity exception.

(for [rank (concat (range 1 11) [:jack :queen :king]) suit [:spade :club :diamond :heart]] [rank suit])

The function is because it is returning a transducer.

Ah.

You can do nested for's that way?!

hi all, is there anyway on the CLI to get a list of all possible flags for new project? (luminus in my case)

mdaines: i thought of that, but that gives me a vector

right?

vector of vectors. and for me, each card is a map

can i do that with for?

Yeah. The very last part [rank suit] determines what will be returned.

{:rank rank, :suit suit} would return a sequence of maps

right

ok it worked now.. something before didnt work

so i thought for coould only return vectors inside

alright, for is better. ill do it like that

out of curiosity, is there a specific order for works?

does it go always likg for first { for second { for third { bla }}}?

Collections are iterated in a nested fashion, rightmost fastest,
   and nested coll-exprs can refer to bindings created in prior
   binding-forms.

ah

So, yeah, kinda like you said.

alrigt

thank you

np. And if you are just getting started, make sure you know about shuffle for handling a card game.

I guess I can just look at the source https://github.com/luminus-framework/luminus-template/blob/master/src/leiningen/new/luminus.clj

ugh or the docs http://www.luminusweb.net/docs/profiles.md

ok and now for a bit of testing. i have a vec of hands i wanna test. how can i test all of them together? so instead of 10 (is (single? hand1)), can i have something like a map to that?

pvinis, you could maybe use (doseq) for that, which is similar to (for) but just executes its body for side effects (like is)

i was just looking at that

thanks

So like (doseq [hand all-hands] (is (single? hand)))

im also using the -> now..

bbut im sure my code is ugly

`

Another approach would be (map (fn [h] (is (single? h)) all-hands)

Or (map #(is (single? %)) all-hands)

Personally I like doseq for that kind of thing because it gives me a heads-up to look for side effects

oh, can i do map for testing?

ok

what do you think about this code..

One thing is you can simplify (map (fn [card] (:rank card)) trick) to (map :rank trick) which is a little easier to read

a nice

forgot that keys are also funcs, even though i actually used it..

better this way, right?

I'm not sure that function does what you intend it to do either, if I'm reading right after (set) you'd have #{ 4 }

I like that ok

yes

i want to check that we have 3 card of the same rank

so i thought, check the number of cards, and take all the ranks and see if its the same for all

so i thought, make it a set. if its just one number in the set, its fine

I think that works if you only ever have 3 cards, but it would also return true for 7 cards

no, because of the (= (count trick) 3)

right?

Oh, I see, right

Sorry, missed that bit

thats the whole thing again

oh ok

you scared me 😛

is there some kind of counted set in clojure?

or should i just use a sorted vector or something?

ah. frequencies

i want to chck if numbers in a seq are continious, is that close? (reduce (fn [a b] (= (inc a) b)) (list 4 5 6))

so 4 5 6 would give true, and 4 5 7 would give false

pvinis: reduce won't do

how could i do that?

Is there a generalization of this pattern I find myself repeating a lot:

(into {} (for [item coll] [item (somefunc item)]))

@pvinis: lemme see, I'm a bit new to the particular function names

pkobrien: map won't do? edit: oh wait I see

I need the result to be a map

was going to code something like (defn into-map [coll f] (into {} (for [item coll] [item (f item)])))

maybe (zipmap coll (map coll somefunc))

http://conj.io/store/v1/org.clojure/clojure/1.7.0/clj/clojure.core/zipmap/

yeah, zipmap might be the ticket

rather (zipmap coll (map somefunc coll))

yeah, I knew what you meant

still more verbose than if I write the func I wrote above

could i maybe use doseq for my check?

I'm using this pattern a lot and want to reduce bloat.

pvinis: are you looking for something like (= coll (sort coll))

@pvinis: I'm thinking something like (every? (fn [[a b]] (= (inc a) b)) (partition 2 1 [1 2 3]))

ajmorgan: no, i need them to be +1 for every next number

jaredly: didnt know every? let me try