« first day (1745 days earlier)      last day (2749 days later) » 

Szabolcs
Szabolcs
9:46 AM
@JohnFultz I was trying to compile GitLink on OS X. I get error: unknown type name 'git_merge_tree_flag_t'; did you mean 'git_merge_file_flag_t'? and a few other similar errors. Is this because GitLink requires libgit2 0.23 while I have libgit2 0.24? I checked the libgit2 reference and can't find git_merge_tree_flag_t.
 
Szabolcs
Szabolcs
9:58 AM
Yes, it's because of that, I can see these API changes in the 0.24 notes.
 
Szabolcs
Szabolcs
10:15 AM
@JohnFultz I managed to compile it with a few minor changes. The changes were so small I don't suppose you are interested in them. The things I tried seem to work fine.
Thanks for sharing this, there's a lot of stuff I can learn from how the package is written.
@JohnFultz Are there any plans to release a C++ interface to MathLink that's more convenient than the current C one?
 
Szabolcs
Szabolcs
10:51 AM
353
A: Let's Plan the Second Iteration of the Stack Exchange Quality Project!

Robert HarveyNew users need better awareness of what is expected of them when they ask their first question. This is the single most important thing that needs to happen on the Stack Exchange network, in my opinion. If the new Question Ask page will facilitate that, then I am enthusiastically supportive. H...

 
kirma
kirma
11:26 AM
Time to pay the Apple tax. Now that GPUs are more and more compute-friendly, I wonder how easy it is to use those AMD GPUs on new MacBook Pros for compute in Mathematica. Like, without dwelling into the wonderful world of OpenCL and whatnot.
 
 
2 hours later…
Szabolcs
Szabolcs
1:39 PM
@kirma I thought currently only OpenCLLink and the neural net stuff can make use of GPUs in Mathematica.
I was a bit disappointed with the new MBP. I value my ESC key, it's critical in Vim and very useful in Mathematica. I use the SD card slot frequently. And I have absolutely nothing with USB type C connectors but I use several things with type A daily. Thinner? Who cares? If we are talking size reduction, reduce the footprint instead! There's plenty of bezel to remove around the screen.
 
 
1 hour later…
kirma
kirma
2:50 PM
@Szabolcs In my keyboard, the key next to 1 is completely unused, so I can remap it. Also, I haven't had actual use for function keys for over two decades. Also, my use of any wired ports or slots on my laptop is really marginal... so, I'm pretty fine with it. Price is a bit high, I must admit.
At least 15" version is actually smaller footprint than the old version, but not much. 1 cm narrower I think.
My primary reason is that my current model is really over four years old, and I want to upgrade before it begins to misbehave. Also, I can still get around quarter of the original price back...
 
 
1 hour later…
Szabolcs
Szabolcs
3:58 PM
@JohnFultz This is really my last question :-) By studying GitLink, it seems to be possible to figure out how to package up my own applications as a .paclet file. I have the impression that this gives easier application version management and simplifies installation for users. But paclets are undocumented ... How bad an idea is it for us (non-Wolfram people) to start using paclets this way?
I really wish Mathematica had an official and documented package management solution. Trying ?PacletAnything usually gives PacletAnything is an internal symbol.
6
I can see potential dangers too ... not sure if it would be easy to reverse engineer enough to set up one's own paclet server. If people do that, it will be Pandora's box of security issues.
To be clear, I'm not interested in setting up paclet servers, but distributing my packages as .paclet files seems attractive.
 
user129412
user129412
4:17 PM
I have a question about a piece of code, but I feel as if I'm just making some small stupid mistake so it wouldn't be worth it to put it onthe main site. What I have is a function
GamFun[t_, ub_, A_] := Module[{s, g},
g = NIntegrate[A*SinIntegral[ub*s], {s, 0, t}];
Return[g]]

Now I want to solve for GamFun[t,ub,A] = 1/t, numerically, for some input parameters ub and A. But somehow NSolve doesn't work; it complains about the limits of integration not being valid. I'm not sure why exactly. Perhaps because it is trying values of t that are not physical?
 
Vitaliy Kaurov
Vitaliy Kaurov
5:12 PM
Mathematica on the Pi Zero
 
 
2 hours later…
JasonB
JasonB
7:36 PM
@user129412 Try this:
In[19]:= GamFun[t_?NumericQ, ub_?NumericQ, A_?NumericQ] :=
Module[{s, g}, NIntegrate[A*SinIntegral[ub*s], {s, 0, t}]];
FindRoot[GamFun[t, .1, 5.0] == 1/t, {t, 1.0}]

Out[20]= {t -> 1.58777}
 
 
3 hours later…
William
William
10:21 PM
is there a symbolic c# plugin?
stupid question yes it appears that is exactly what .net is
I have a script that appears to be working in ubuntu but not ubuntu 
SetOptions[$FrontEndSession,
 InitializationCellWarning ->
  Not@Options[$FrontEndSession, InitializationCellWarning][[1, 2]]
 ]
 
William
William
10:45 PM
I have a script that appears to be working in ubuntu but not windows
 

« first day (1745 days earlier)      last day (2749 days later) »