3 ms·

I can't help but feel that the Wolfram Language is less of a programming language and more of a DSL for querying a very precisely defined set of data. I think t

by nv-vn 10y ago

I can't help but feel that the Wolfram Language is less of a programming language and more of a DSL for querying a very precisely defined set of data. I think that the whole thing falls apart once you step outside the boundaries of the data they offer. In that sense, it doesn't seem very well suited for teaching computational thinking at all -- there's no logic involved in letting another person solve a problem for you. The examples show how computers can enrich the work of people in the future, but to me that's entirely different from computational thinking. What's going on here is little more than teaching kids how to use Google efficiently.

SolarNet 10y ago
I think the google comparison is fair, but I feel like Wolfram (and let's be clear, this piece is basically a giant ad) is like an order of magnitude ahead in the search engine aspects, if the sub-field is not-web-page datasets.
- nv-vn 10y ago
  Exactly! It's a huge development in terms of technology and usability, I just think it's largely mis-marketted.
ecfxixiuabhyqnm 10y ago
> I can't help but feel that the Wolfram Language is less of a programming language and more of a DSL for querying a very precisely defined set of data. If what you're thinking about is Entity and EntityValue, you can't be more wrong: the built-in data is so slow it's impossible to use, even Wolfram employee steer away from it when giving presentations. You clearly can't give a demo where returning a bunch of missing data takes several seconds. If by 'data' you mean a certain library of algorithmic content, then you might be onto something.
- nv-vn 10y ago
  Well, both here really. But the amounts of data normally showed in demos is a bit over the top. Showing me that you can color states on the map of the US in 3 lines of code is not impressive when you already have the map data set up in order to be colored. The interesting part of the problem is finding the boundaries and creating shapes to color using that. The algorithmic content also detracts from the examples showing programmatic thinking, but not to the same extent.

by nv-vn 10y ago

SolarNet 10y ago

I think the google comparison is fair, but I feel like Wolfram (and let's be clear, this piece is basically a giant ad) is like an order of magnitude ahead in the search engine aspects, if the sub-field is not-web-page datasets.

nv-vn 10y ago
Exactly! It's a huge development in terms of technology and usability, I just think it's largely mis-marketted.

ecfxixiuabhyqnm 10y ago

> I can't help but feel that the Wolfram Language is less of a programming language and more of a DSL for querying a very precisely defined set of data. If what you're thinking about is Entity and EntityValue, you can't be more wrong: the built-in data is so slow it's impossible to use, even Wolfram employee steer away from it when giving presentations. You clearly can't give a demo where returning a bunch of missing data takes several seconds. If by 'data' you mean a certain library of algorithmic content, then you might be onto something.

nv-vn 10y ago
Well, both here really. But the amounts of data normally showed in demos is a bit over the top. Showing me that you can color states on the map of the US in 3 lines of code is not impressive when you already have the map data set up in order to be colored. The interesting part of the problem is finding the boundaries and creating shapes to color using that. The algorithmic content also detracts from the examples showing programmatic thinking, but not to the same extent.