Those tweets on Wolfram Technology Conference are close to causing me cerebral hemorrhage... but so do all tweets. Somehow I feel the whole point of being a Twitter user would seem to be being an attention-seeking jerk trying to appear brilliant on every tweet, and failing on it by overdoing it every time.

Sounds like FE to me.

Guys have a look hype or paid article? gamasutra.com/view/news/212709/…

@PlatoManiac There have been demonstrations of Mathematica-Unity integration... but I am deeply suspicious of "Wolfram Language" (I would really rather call it a runtime) being suitable for serious game development. Performance is what it is (especially on symbolic side), it's unpredictable, and Unity tends to concentrate on things where dependable real-time response is essential.

Unity as a modern visualization/interactive environment would be nice addition for Mathematica, but that article once again gives the false image that WRI could just enter a fiercely competitive, high-budget market and trivially succeed at something that others have probably spent tens of thousands of man-years of fine-tuning.

@kirma I suspect similarly too! WRI trying for a kill all solution and spending too much time on PR that actually halted core development and bug fixing of MMA. Slowly everything is becoming API dependent and subscription based. Bad planning in my point of view. They added CUDA support but did not do any further work to support newer releases of CUDA.

So it proved to be just a namesake. Same is true with other new areas like FEM. Meshing capabilities are way too limited CAGL is open source they could simply port to it and add their own innovations.

Before turning into a publicity maniac Mr. Wolfram could have done some solid work so that MMA could solve real life problems and become fast like other programming languages.

But anyways those dreams of mine are not to be fulfilled in this life with an ego maniac defining the development pipeline. It is at the end not much use to be very proficient in MMA you can not really make applications with it that runs on any platform like C/C++...So I sometimes think why did I get so damn addicted to this beautiful peace of software...:(

In theory, one could attempt to build neat "applications" on WebGL + JavaScript (+ asm.js) + cloud back-end. I think it would be overestimating resources of WRI and attention span of Mr. Wolfram on these kinds of features that to actually happen.

Keep symbolic stuff in the cloud, run compilable stuff on the browser. From Unity, one can create WebGL applications, I mean...

Sure that is quite plausible. For example 4-5 years before a web developer could only dream for graphics that MMA could produce but now with WebGl, D3.js Sigma.js and countless many they can make visualizations that are dynamic and fluid and definitely much more compelling and flashy than current wolfram language and they need no cdf and no damn API call that has to be slow given the paid wolfram infrastructure...

I wish advancing the CAS and fixing known issues were at the top of the priority list. I still don't really know what to make of the Wolfram Language marketing speak.

jk, of course :)

My five cents is that WRI has no intention of making the language fast because it's too much work. So they expect to be overrun by Julia, Python etc. in the scientific computing sphere. Therefore they are trying to find new markets especially in education and in knowledge curation. I think that justifies many of their decisions and the advertisement, which is clearly not meant for us.

Also I think they predict a future where a substantial part of the population use programming occasionally. They want to reach this kind of users. WL is actually very well suited for such hobby usage.

I'm playing with Hough transform for parabolas in relation to this question:

My code works wonderfully with my own generated test data, but provides only non-results with data points extracted from that image...

Hi all, I want to take a graph from a paper and have mathematica give me a list of the points that make up the graph (obviously I'll have to supply it with the tic mark values from the graph)... how can I do this?

@FdotFloss Do I get it right, you would feed the graph into Mathematica as an image?

@kirma, yep, that's all I have from this paper

That's probably not too trivial... but I'd consider correcting the coordinates with FindGeometricTransform and Dynamic like here: mathematica.stackexchange.com/a/63897/3056 and use image processing and binarization tools with Position[ImageData[..., 0]] to extract points in the plot, which would then be transformed with the above transform.

Cleaning up tick lines and such might be tedious.

kirma, I think I've found an easier way

I hope so... :)

Use this reference.wolfram.com/language/howto/… to get the coords of several pts on the curve, provided it's fairly well behaved, then interpolate them

then scale them, provided 2 reference points for each axis

Yep, my second suggestion (depending on the plot!) would have been measuring more or less manually some keypoints and attempting interpolation with a suitable function.

@kirma, yep, it works now

I have another quick question: right now my function to do this involves clicking around a bunch and then having to copy and paste the points and such. How hard would it be to make a program that automatically enters the points I click?

@Pickett wow I think you are right about how they have been positioning themselves re scientific computing ... never thought about it this way -- have just felt more and more enraged that they are not focusing on their science stack more (which is what I care about personally)

I totally agree. I feel I have been forced to use other platforms increasingly due to this lack of interest in scientific programming. But I always miss just using MMA on its own as I really like the platform

@FdotFloss It should be easy, but I don't remember the details now. :o

alright, thanks

Something is driving me nuts: I'm trying to return a really simply linear function from within a module

@FdotFloss

list = {}; Row@{ClickPane[Graphics@Circle[], AppendTo[list, #] &],
  Dynamic@Column@list}

so for example testFn[d_] := (c = d + 2; Return[Function[c + #]])

The problem is, if I do g=testFn[6]; h=testFn[8], the c becomes the one from the most recent run, from testFn[8], so they return the same thing

if I do g[5]; h[5], for example

How can I get around this? I know I could do testFn[d_] := (c = d + 2; Return[Function[c + #][x]]), and then g/.x->5 or something, but I'd like to be able to just do g[5]

@FdotFloss The Return in Mathematica is only used to force the function to end. It doesn't serve any purpose at the end

@FdotFloss Function will hold the c so it will end up having the value at the time where you USE the fucntion, and not the one where you created it

You can "inject" the value of "c", fore xample, using With

With[{c=d+2}, Function[c+#]]

(but why not just Function[d+2+#])

@Rojo, because that was a simple contrived example of a more complicated problem I'm having

@PlatoManiac "now with WebGl, D3.js Sigma.js and countless many they can make visualizations that are dynamic and fluid and definitely much more compelling and flashy than current wolfram language and they need no cdf" ...couldn't agree more. I'm about to start playing around with D3.js, also RAW project. This seems to be the future ...in 2D at least.