Since CDF's no longer work on browsers, I am thinking of removing all the javascript code to embed them in the web page and just have a link to download the CDF. Sad state of affairs that one can no longer run CDF on a web page :(
Frankly, everybody has known for years that browser plugins, even Flash, are going to go away through browser vendor deprecation. I just hope WRI (or whatever it's called nowadays, Wolfram Group?) has actually prepared for this in sufficient detail...
@kirma It is not so smooth. It won't look the same (100% you will have to make edits to fix misplaced labels etc.). People using it will eat your CloudCredits.
@kirma so it is not worth converting that in this case probably
@Nasser It will continue to work in Firefox until the end of this year, and it will continue to work in Internet Explorer and Safari after that.
It never was a natural part of the browsing experience, since people had to install the CDF player. If you have to require the user to install the player, then I think you can also require the user to open a specific browser.
Is this really off topic? mathematica.stackexchange.com/q/120024/12 It's not unlikely that there's someone here who has M7 installed in W10 and could at least say whether it starts up and whether basic commands work.
Something I would like is the option to bundle the CDF player with the CDF. When I was young I was into programming games with Game Maker. It uses an interpreted language, and in order to distribute the game it would simply bundle the interpreter with game code into an .exe file. Users just saw the .exe file, they never knew what the game was made with or had to download anything.
Nowadays you can write games in Game Maker and deploy them to iOS, Android, OS X, Windows... the interpreter is bundled with the game so again the user doesn't have to do anything.
The only downside is that it makes the file size larger but I could definitely live with that.
@Pickett How much larger though? Mathematica s unfortunately not very modular, so it seems hard to strip out unused components. Is it at all possible to keep the runtime below 100 MB (which is already very large for something that does as little as a typical CDF)?
I really think WRI has dropped the ball on the whole CDF thing. Instead of CDF taking over, I see very few using it. At school I tell the teachers about it and show them some demos and they can't believe it, and they never heard of it, yet no body uses it there. They spend so much time on Matlab GUI, when the same thing can be done in M much easier.
There is a lot of new stuff starting to compete with CDF's out there now,. Python Jupyter notebooks and few other things based on Javascript. Still nothing as advanced as Manipulate and Mathematica Graphics.
The is nothing inherent in Jupyter that makes it good for CDF-like things.
I haven't seen any interactive examples with Jupyter.
And each language plugin is responsible for its graphics, so the abilities there vary according to language. There is no unified effort on behalf of Jupyter as I understand it.
@Pickett I was going to take a course at school in optimization. I found the teacher insists on us using Python and Jupyter notebooks. After trying it for few days, I dropped the course :) It is still work in progress. But there is a lot of interest in it out there.
At school, students can not download and install anything on the PC's. So in libraries, labs, study places, one can't use CDF's since one can't install player on them or any software. But many use laptop pc's these days and they can install it on the laptop. The problem is that they do not even know about CDF's. WRI need to do more to make it more known.
> Actually, there's one very big Wolfram Alpha installation which will be coming out - it probably won't say it's a Wolfram Alpha installation but it's a big, consumer-facing retail-type thing that's coming out in a few months that's based on Wolfram Alpha.
Anyone know what Wolfram is talking about?
Would this be an on-topic question for the main site?
@Kuba it seems that Dynamic[x] is added automatically when x is there. If you try Dynamic[y]; y = 5; {Slider[Dynamic@x, {0, Dynamic[y]}], Dynamic[x]} and then change y in new cell, you'll see that now you can move slider more, up to the new y dynamic value. But if you do this Dynamic[y]; y = 5; {Slider[Dynamic@x, {0, y}], Dynamic[x]} then it will not work.
How do you write a function f, that 1) enters the function's RHS to perform an argument-test and 2) returns f[arg...] unevaluated (with error message) if the test does not pass? Like Solve[x + Log[5 + x^2] == 0., x] does. Clearly, the test is not simply a f[arg]/;argTest:=body. Anyone has an idea about the standard (Wolfram) way to do this?
@IstvánZachar I'd have to look up the details as I don't often do this, but it's something like f[arg] := Module[{}, ...; ...; ... /; argTest]. Check Module, Block, etc. docs.
How do I get Mathematica to return a function call (conditionally) unevaluated? I expect this may use the slightly-mysterious Hold function.
As a toy example, suppose I want to define AlgebraicQ such that AlgebraicQ[x] returns True or False when Element[x, Algebraics] evaluates to True or False,...
@Szabolcs Perhaps we could set up a fund where anyone can contribute some credits. I certainly not using mine as anything that I plan to put in the cloud fails due to heavy package-dependencies.
Cloud CDFs are not scoped, if you forget about DynamicModule they will share variables! Which may be unexpected since desktop CDFs were using Notebook context by default.
@Kuba FrontEndObject::notavail is a very generic error. It just means you're doing something that involves image or graphics or something that normally has code in the front end which isn't available.
Sometimes you can get around it with Developer`UseFrontEnd, but I'm not sure how that would work in the cloud.
@Searke not really, I was testing a lot of things and yes, when one particularly heavy Manipulate, with couple of simple bitmaps was deployed, first there was a timeout and then each deployment was givin me that messages.
@Szabolcs It's strange that "DistanceFromSun" doesn't work. If I had to guess it'd be that Entity["Star", "Sun"][EntityProperty["Star", "DistanceFromSun"]] returns Missing["NotAvailable"]. Which is a strange fluke.
We can consider it an experiment, I had 1250 cloud credits when it started and I'll tell you how many I have in a week. That way we'll know approximately how expensive it is to run this type of ad in the Wolfram Cloud.
@Szabolcs "QueryTooSlow" is just a heuristic that stops queries that aren't going to work for some reason. There was a false positive for this with "Star" in the past. I filed a bug report about it.
@kirma The documentation says: "If no explicit Shortest or Longest is given, ordinary expression patterns are normally effectively assumed to be Shortest[p]", so most patterns would only give one result if ReplaceList didn't ignore Longest and Shortest
I think it's just how it's meant to work. If you look at the "basic example" under ReplaceList you see an example that would be exactly identical to just normal Replace if it didn't ignore Longest and Shortest.
But I don't know of any explanation in the docs about this either.
I could see situations where not ignoring them would make sense. Like that "longest repeating pattern" question today.
That behaviour ignoring greediness preference caused amount of intermediate results to grow too much, and to cause an answer candidate I considered to become quite inelegant.
