« first day (2680 days earlier)      last day (1794 days later) » 

4:00 AM
@ArnoudBuzing OK. The wording is a little more understandable than on many WRI licenses I have seen and never got convinced that I can confidently interpret that a use case isn't a breach of license terms. But still, what on earth are "open-source projects approved by Wolfram?" Does this mean that if I publish my code which can be set up to work with Wolfram Engine, publication (or use) of this code for other purposes listed as accepted use is a breach of license terms?
Being pedantic, software which requires publication permission from a third party is often not considered open source by definition.
I mean... if I publish such code without explicit permission acquired from Wolfram...
 
4:59 AM
@kirma on the other hand how many OSS projects bundle a >3GB dependency?
Reading here it sounds like the issue is only if you want to directly include it in a project
 
 
2 hours later…
6:32 AM
@b3m2a1 Yeah, but they could for once write license terms in a way that wouldn't induce uncertainty. ;)
 
Amen to that
 
 
7 hours later…
1:21 PM
Mm, not totally delighted with those terms (audits, hmm), but I haven't had my third coffee yet.
no definition of commercial, or "production software application" - aren't they all, really
> "using a password, activation key or other means of accessing the Software other than as specifically authorized by Wolfram."
so I can't use an ssh key to access the software on a remote server?
(I'm being maybe more nitpicky than usual, but licensing is my favourite thing to grumble about)
 
Quick update. Stephen is at Collision Conference this week, so livestreams will be minimal. However, we do have one today at noon (Central Time) discussing the new features of the Notebook interface in V12.
 
@JKlug I take it this is by the developers on twitch, rather than a meeting?
oh, my b, I'm actually OK with this regarding "production":
> Development but not production of software applications. While determinative classification of use shall be solely made by Wolfram, allowed uses include but are not limited to:
pre-production/pre-distribution development of a product for Yourself or Your organization;
undertaking non-commercial projects for personal use; and
exploring the Wolfram Language™ for consideration of use in subsequent projects.
@JKlug And thanks as always for the forewarning!
 
@CarlLange yes, this will be a developer livestream on Twitch.
 
@JKlug 👍Thanks!
 
2:29 PM
posted on May 21, 2019 by Stephen Wolfram

Why Aren’t You Using Our Technology? It happens far too often. I’ll be talking to a software developer, and they’ll be saying how great they think our technology is, and how it helped them so much in school, or in doing R&D. But then I’ll ask them, “So, are you using Wolfram Language and its [...]

 
It's odd that the blog post begins with the question "So, are you using Wolfram Language and its computational intelligence in your production software system?" considering the free engine explicitly does not allow that
4
 
2:52 PM
Does anyone have a good example of Polygon unwrapping/rewrapping for this user? I'm having a hard time finding one. mathematica.stackexchange.com/a/198786/57593
 
3:32 PM
The wolfram engine blog post has the following sentence: " Soon it’ll also plug directly into web servers (J2EE, aiohttp, Django, …)." My question is what design principle they will use? How concurrency and async behavior of common servers will be mix withwolfram engine. How the deployment will be? Will it be container friendly and so on. Any guess about this things anybody or the insiders ;)
 
4:12 PM
So, the Wolfram Engine for Developers is a free Mathematica without notebook interface?
3
 
4:44 PM
@PlatoManiac I think the approach will be similar to how webMathematica used to work (although upgraded to support newer primitives like APIFunction, and co)
For tomcat (webMathrmatica) that means having a servlet that can serve "active cloud objects" (the ones that work through GenerateHTTPResponse) that can be configured like any other webapp in tomcat
the idea is to have server bindings for all common server environments
those will communicate by sending the kernel a HTTPRequest and an object (APIFunction, URLDispatcher, ...) and by returning an HTTPResponse.
Most of the work is translating a java/python whatever HTTPRequest into a WL HTTPRequest and viceversa for the HTTPResponse
but there is nothing stopping you from doing this now in the language of your choice
(apart from maybe licensing)
I think it would actually be very good if some of these translation layers were owned and contributed by the community
rather than by WRI
 
I wonder : Since there is the Wolfram Engine, does it make sense now to have a limited number of subkernels for 1) those who have a hobbyist licence ? 2) those who have a classical desktop commercial licence (allowing 2 frontends, 2 main kernels, 4 (or 8) subkernels running simultaneously) ?
 
 
1 hour later…
6:11 PM
It would be cool if you could use WolframEngine to set up something similar to a private cloud, but just your own $5 server somewhere, just to serve a few web pages or whatever. Not gonna happen though, surely.
 
@CarlLange if you don't really need cloud notebooks, but just want APIs that can be done with the technology I was talking about above
 
6:44 PM
@Carlo Yeah - I guess I want it built-in to the language easily ;) I think it's a wait-and-see for me for now, but I'm excited what people will do!
 
@CarlLange I'm guessing a Django/python/WL wouldn't be a crazy amount of work. A little bit of glue code here to get from Django's HTTP response and serve it in WL and vice versa shouldn't be brutal to write, esp. with the existing client library for data structure transfwer.
 
@b3m2a1 exactly! and the only WL code you'd need would be: GenerateHTTPResponse[active-cloud-object, $HTTPRequest]
 
@Carlo on the other hand, you know as well as anyone that WRI has set things up to make the average Mathematica user suspicious of any code that doesn't come down like manna from heaven from WRI
We can write it, but people probably won't use it
 
@b3m2a1 I'm not sure you're right. The real power users that would actually build their own mini-cloud probably rate some of the stuff from this community higher than our manna
I for one use the intellij plugin instead of the eclipse one :)
2
 
I'm sure @halirutan will be happy to hear that :)
 
6:59 PM
Thanks @halirutan!!
Anyway I think the blog hints at some upcoming releases in this regard
so maybe you won't have to wait too long
 
 
1 hour later…
8:28 PM
@xzczd Yes. But you can use it with Jupyter notebooks. Hard to believe, right?
Just make sure you are using jupyter notebook and not jupyter lab (it's wonky with the latter)
 
8:50 PM
posted on May 21, 2019

Wolfram Research developers demonstrate the new features of Version 12 of the Wolfram Language that they were responsible for creating. Previously broadcast live on May 21, 2019 at twitch.tv/wolfram. For more information, visit: http://www.wolfram.com/language/12/notebook-interface/?

 
 
2 hours later…
 
1 hour later…
11:48 PM
It'd be nice if there were some more sophisticated support for Typed in patterns where things like:
f[obj:Typed[_Integer]]:=...
Would unwrap to:
f[obj:_Integer?IntegerQ]:=...
And in places where there's no Q function it'd just ignore it
And this could get gradually smarter with a RegisterTypeRule function that would introduce more mappings for things like QuantityArray too
Since there's no QuantityArray head it'd go to something like:
_StructuredArray?(Developer`PackedArrayQ@*QuantityMagnitude)
or something
Honestly this might be worth writing a small paclet for... And then update the rules whenever I need a new pattern.
I'll probably want a shorthand for these things too I think.
 

« first day (2680 days earlier)      last day (1794 days later) »