« first day (2959 days earlier)      last day (42 days later) » 

2:29 AM
My test case for all this package framework stuff I've been working on is to build a package to help with the process of writing paclets (which also makes paclets that make use of this stuff). Any requests for what might be nice in a single notebook interface that allows you to define the content of the PacletInfo file / list all the packages / front-end content / dependencies for the package?
The plan is that it will also handle to business of bundling the paclet up and maybe adding to a paclet server or something
I think that's in the future, though
 
 
3 hours later…
5:57 AM
@Szabolcs You are right to be suspicious of R^2 values. (How can a unitless statistic describe everything about any and all datasets no matter what their subject matter?) But for the example you give, a quadratic fit does provide a pretty good fit to the exponential data you constructed.
 
 
6 hours later…
11:32 AM
is it just me or are the mappings of modifier keys between macOS and Windows super strange? (talking about e.g. CurrentValue["AltKey"]) - to be fair, there is no perfect mapping, but to me, the mapping Cmd->Ctrl, Option->Alt, Shift->Shift from MacOS->Windows seems more natural than Cmd->Alt, Ctrl->Ctrl, Shift->Shift, Option -> Nothing.
Some specific annoyances: Most shortcuts on macOS of the form Cmd+... translate to Ctrl+... on Windows - this is not easily doable using CurrentValue. Ctrl+Click is always right-click on macOS (again, Cmd+Click seems better suited as analogue).
Or maybe I'm just not used enough to the "philosophy" behind macOS modifiers - maybe a long-time macOS user has some insight into what could have driven the decision?
 
 
6 hours later…
5:42 PM
@LukasLang For many years neither the Lisa nor Mac keyboards had control keys. Ancient history now, though.
The mapping back then was essentially what you suggest. But I suppose at some point ater the ctrl key was introduced they decided to use it for cross-platform consistency (as opposed to Mac-only consistency).
 
 
6 hours later…
11:43 PM
A fun application for the package framework I've been working on: since I had to write it to be able to handle dependency packages without conflicts with some main version of the package, it can also provide a way to distribute all your packages within a single "namespace" if you will. Like it can make sure that every package I distribute lives in the parent context b3m2a1`...
I don't have a direct usage for this, now, but this could be useful with regards to stuff like the paclet repo that WRI is supposedly working on
 

« first day (2959 days earlier)      last day (42 days later) »