« first day (2595 days earlier)      last day (2190 days later) » 

02:06
@HenrikSchumacher Thanks, but I may have to go again in a few. It's really hard to do stuff without a computer of my own.
:(
02:43
Hello. How can we do constrained linear optimization with parameters as coefficients?
\[Pi]TT33 (** Maximize this objective function wrt p1,p2,F **)

(* These are my consraints***)
0 < F < -(-1 + \[Gamma]) (p1 + p2 - \[Theta]h - \[Theta]l)
F < p1 - p1 \[Gamma] - \[Theta]l + \[Gamma] (-p2 +
v + \[Theta]h + \[Theta]l)
F < p1 - p1 \[Gamma] - \[Theta]h + \[Gamma] (-p2 +
v + \[Theta]h + \[Theta]l)
F < p2 - p2 \[Gamma] - \[Theta]l + \[Gamma] (-p1 +
v + \[Theta]h + \[Theta]l)
F < p2 - p2 \[Gamma] - \[Theta]h + \[Gamma] (-p1 +
v + \[Theta]h + \[Theta]l)
F < \[Gamma] (-p1 - p2 + 2 v + \[Theta]h + \[Theta]l)
 
1 hour later…
03:57
@Roman I simply love the capabilities of IntelliJ
 
3 hours later…
07:19
@halirutan Wow that's amazing and precisely what I meant!
07:58
@Szabolcs @halirutan @other people who have tried to set up compilable C++ packages across multiple systems, I'm trying to resurrect my PJLink project but the core issue I originally ran into was trying to get the damn thing to compile on machines other than my own. Is there a way I can precompile all the necessary parts and just supply those to people?
Like lots of projects just supply a precompile binary, I think. Is that likely possible here?
08:36
If anyone has ideas they've wanted to do/thought would be cool to have, I'm creating a listing:
1
Q: Projects you wanted to do / are interested in seeing done

b3m2a1I like to have at least one side project going at all times and often get bored with one and would like to play with another. Sometimes I find myself (as now) unable to come up with an inspiring, quick side-project. As I was thinking about it I thought it might be nice to have a community record...

I've contributed a few of mine. Hopefully others will contribute theirs and we can all take inspiration from eachother.
I'd love to see a Language Server Protocol implementation for WL (I think @halirutan and I had a discussion about this about six months ago)
It would mean that you could (with relative ease) write WL modes for any IDE/text editor, because the underlying functionality lives in the LSP server and all you need to do is write a little wrapper for emacs or sublime text or whatever else you like
I think you could use a lot of the work done for the IntelliJ plugin to start with
08:55
@CarlLange worth adding it as a project idea so at least it's documented in a more searchable format :)
09:05
@halirutan is there something like slot scope highlighting when hovering over a slot symbol? Hovering over one slot would highlight all equivalent slots (but only them) plus their scope.
09:30
@Kuba another useful option to add for embeddable cloud objects. You can specify your notebook width (first noticed here).
The syntax is to add _width=... to your URL query to specify an iframe width
Seems to only make sense within an iframe though as when I try this: sandbox.open.wolframcloud.com/app/view/… it strips the _width arg on render
Wow I feel like the cloud is just finally becoming actually useful. Took like 5 years.
09:52
4
Q: Cool Project Ideas

b3m2a1I like to have at least one side project going at all times and often get bored with one and would like to play with another. Sometimes I find myself (as now) unable to come up with an inspiring, quick side-project. As I was thinking about it I thought it might be nice to have a community record...

10:12
@Roman Anonymous functions are italic in my IntelliJ plugin per default. That means, you always see the scope explicitly. Highlighting equivalent slots is a disaster since there are so many ways to use them nowadays
Function[key1,
   With[{key = "Test"},

    {#[key], #Test, #"Test", #["Test"], #[key1]} &

    ]]["Test"]@<|"Test" -> True|>
It's not really possible to find equivalent elements in such situations with a static analyser.
10:34
@halirutan How to install most recent plugin beta release?
@halirutan p.s. hyperlink to 'please leave a comment and rating at the' from readme is invlaid.
Thanks
@Kuba Nowadays, you must be logged-in to provide a rating.
I see, then they should say so instead of 404
@Kuba No, the link is also wrong. Someone had the clever idea to rename Mathematica to Wolfram Language :)
:)
11:13
@b3m2a1 In general, it is possible to bundle pre-compiled binaries. Making sure that it works on most systems can be some (or a lot of) trouble. I am not familiar with Python and I do not know what is required for binary compatibility with it. You would have to ensure: 1. binary compatibility with multiple versions of Mathematica (not so difficult) 2. multiple OS versions (I still do not have a good understanding of this on Linux, which has been by far the most trouble) and
3. different versions/distributions of Python. I have zero experience with this.
You might want to create binaries for 32/64-bit Windows, 32/64-bit x86 Linux, ARM Linux (Raspberry Pi), and OS X (on which M had a compatibility break between v10.3 and 10.4)
Then there are the rumours that Apple will switch some of their machines to ARM CPUs in less than 2 years.
Personally, I'm close to giving up.
As far as I know, IGraph/M is the only 3rd party package that tries to ship pre-compiled binaries and actually make them work.
There's GitLink, but it uses Wolfram's internal build system, so they're in a good position to do this.
To make all of this work, one needs access to all these platforms and a Mathematica installation on all of them. Licensing will be a problem for many (except those with unlimited academic licenses)
Actually, this might not be true. I believe it should be possible to compile binaries without having access to a full Mathematica. Of course those binaries also need to be tested ...
I have not yet tried to set this up. It's some work.
@Markus Roellig Have you continue your work on RTE using spherical coordinate system? Are you an academic professor. I will like further develop your work as demonstrated in the post wile back mathematica.stackexchange.com/questions/10767/…
So there is actually a chance that some automated build service could be used. Is anyone here familiar with Travis? Can it be used for this purpose in practice? Importantly, is it possible not to have to re-build all dependencies on each change added to the main project?
For Java, there is a standard mechanism to just add the jar file. But for Mathematica, there is no standard way to embed a sub-package into a main package. For LTemplate it wasn't trivial (and I would not expect someone new to packages to be able to do it).

With Java, applications are isolated and do not interfere, even if they run on the same system. With Mathematica, "the kernel is the OS". All loaded packages share the same system. There is a much higher chance of package conflict / version conflict.
If a sub-package is trivial and lives in one file, its code can maybe just be copied. But that's a primitive solution. It makes updating the sub-package difficult.
If the sub-package is not trivial, has multiple files, maybe expects a context, then it is not easy to embed it while avoiding conflicts with other versions.
Of course it can be done (see LTemplate), but that's not a long-term solution either. Imagine if this approach becomes widely used. Then it will be typical to have many copies of the same version of the same package loaded at the same time.
Some sort of versioned context would be a better solution then ...
11:32
@Szabolcs @b3m2a1 On Travis, you can use Linux, macOS and Windows as environment. If you are not depending on a kernel and if you are using something like CMake which allows for easy building on all platforms, it might be worth a shot. Travis can deploy the things it built to several providers like GitHub releases. So if possible, you need to build your C-lib for all systems and store the result on Travis so that you can deploy all products in
one release. If this is indeed possible, I don't know.
12:31
@halirutan I use a patched version of igraph to build IGraph/M. Is it possible to not re-build igraph when I push a change that affects IGraph/M only, but do rebuild it when it's changed? They are in two different repos if it matters.

I'm not asking how to do it but if it's at all possible (if you know the answer). Last time I checked I got the impression that it might not be possible. Today I found "caching dependencies" so it looks like it should be possible after all.
13:06
Mathematica Online terminates the kernel quite quickly after inactivity, at least with my home Mma license. I wonder how hard it would be to serialize/deserialize the kernel state to persistent storage at least while the kernel versions are identical...
 
1 hour later…
14:10
Looks like a cute question by a nice person:
Generate sound from plot curves?
https://community.wolfram.com/groups/-/m/t/1620017
...if you got ideas and visit Community sometimes...
15:02
Greetings, all. There will be a livestream today at 1:30pm (CST). This will be a language completeness and consistency meeting. As of now, that is the only livestream scheduled for this week, but that could certainly change!
@JKlug Is there any news about a new graphs and networks meeting?
15:38
@b3m2a1 @Kuba @C.E. Have you investigated what is needed to make a .paclet file? It looks like it's "just a zip file". I assume I could just take the files, zip them up with an arbitrary tool, then rename the .zip to .paclet. But not every zip is created equation. For example, macOS's default decompressor can't decompress paclet files. They must be using a type of compression that is different from the most common one.
Has anyone investigated this?
Motivation: create a release in an environment where Mathematica is not installed. I am investigating if I can use Travis to compile the C++ part of the project then wrap the result up into a paclet.
2
16:32
Sorry, guys. The livestream has been rescheduled for Friday at 2:30pm CST.
I miss the livestreams :(
@Szabolcs Sorry, but I don't have any news on that front. Streams are really less frequent while the team finishes up the v12 testing. I'll be sure to let you know if/when we can get something on the schedule for that topic.
16:56
@Szabolcs this is what I do and it works on mac and win. Didn't test heavily, paclet only contains .wl and .jar files:
Feb 2 '17 at 11:42, by Kuba
 git archive --format=zip --prefix=Package-2.1.2/ 2.1.2:Package/ > Package-2.1.2.paclet
@Szabolcs the library used to create a paclet is exposed in the PacletManager. Unfortunately I think it might come precompiled but I can look into it when I have the chance. Using Travis as a build system would also allow for more sophisticated package distribution utilities I think.
 
1 hour later…
18:05
https://mathematica.meta.stackexchange.com/a/2445/13
Hey guys, if you think this is cool - please upvote ;-) We opened admissions for the Wolfram Summer School 2019. It is very fun and education is free (the cost is only travel and campus-living). Skill-building, career-boosting academic program for 3 weeks in Boston, around 100 top students and professors from around the world. Networking is amazing. You can be a part of it. Please vote this ad up, apply for the program and spread the word encouraging your friends and colleagues. Feel free to ask me any questions. Thank you & cheers!
18:32
@VitaliyKaurov I kind of wish there was a European counterpart!
I have no idea what it would look like, but I would love GeoGraphics3D... I've been playing with volumetric weather data and it would be cool to see them with all the useful stuff that GeoGraphics has
18:58
@CarlLange ha, we are actually considering it ! Shoot me an email if you'd like with any thoughts of the school format, we''ll take it in consideration.
 
3 hours later…
22:25
@Szabolcs might also be worth asking Todd Gayley/support what protocol WRIunzip.dylib implements
That's the library the paclet manager uses
22:44
I'm not having any problems when using 7z, e.g. 7z e something.paclet will extract all the files into the working directory. If I run 7z l -slt something.paclet then it will me that the method used is deflate.
Don't have time to try and compress files now.

« first day (2595 days earlier)      last day (2190 days later) »