"Letter holder design with Mathematica."

Surely, like my friend put it, it looks like a woodworking exercise from school... but straight edges happen just to be much easier on custom orders.

... but placement and size of components is literally decided by a Minimize expression that attempts to optimize various ratios in order to fit DL, C5 and C4 envelope sizes on the "trays" aesthetically.

@kirma my woodshop-in-his-garage-and-Mathematica-loving friend will LOVE THAT!

@CATrevillian :)

Say, who knows about Splines? Anyone know if or how I can have Mathematica do a Not a knot spline? Or a natural spline? Cubic, etc? Obviously b splines are no issue, but for something one finds in a numerical course, is this possible?

@CATrevillian The "Not" part got me completely lost. :o

@kirma right? It’s apparently what matlab uses, but it is a technique to, for whatever reason, make a tridiagonal matrix in the end for a linear solve. You remove x_1 and x_n-1 from your x_0 to x_n set of values to interpolate, then reimplement them at the end as part of the boundary conditions. However, all I have are pseudocodes for every other type of spline—not a not a knot! Hah!

@CATrevillian Ah, okay. :)

@kirma This is very nice! I've got a 3D printer on the way and I'm really excited to check out Printout3D for this kind of thing!

@CarlLange This is a little more manual work. I went to the downstairs furniture shop, drew my plans on basis of measurements on paper, and they sent the order to their factory where CNC machines machined the pieces... some assembly required at the factory, too. ;)

@kirma Ah yes, who doesn't have a downstairs furniture shop to do this kind of thing ;)

The Wolfram|One personal license price has changed from €35/mo to €26/mo, such great news!

@CarlLange If I wouldn't have had one there, I would have probably given up with the idea, too.

@kirma I think that's why I'm so excited about the 3D printer, since I don't have a downstairs shop of any kind

@kirma In any case, it's very pretty, I like it a lot!

@CarlLange I wish 3D printing turns out equally interesting. :)

@kirma So this is an example of being over-engineered? I still like it.

@halirutan I bet it's engineering over everything else. Like, the minimization problem involves consistency of envelope angles and their relative overhangs over the backing plate, relative distances of plates from each other, and so on. :)

@Szabolcs I usually use Query[Transpose]@<|"ID"->Range@5|> instead of Transpose[<|"ID"->Range@5|>,AllowedHeads->All], for a few reasons: It works in versions <12.0, doesn't need brackets, and has everything in one place (& I'm more used to it ;) ) - essentially Query[Transpose] does the same thing as Transpose for Dataset (it's actually GeneralUtilities`AssociationTranspose under the hood, see Normal@Query[Transpose])

@Szabolcs Only /* has special handling for mixing of descending/ascending operators, see e.g. this answer

@Szabolcs Ideally, the following would work: dataset[SubsetMap[Normalize,{All,"a"}]] - but for some reason, SubsetMap does not yet support associations... In the meantime, I often find myself doing dataset[Transpose][{"a"->Normalize}][Transpose], which, while a bit messy (especially at deeper levels) allows one to easily operate on whole columns (also here, it appears to mess up the display of the dataset until you force it to reevaluate the types...)

@Szabolcs This is not an answer, but you can use something like this to force a rebuild of the types to fix display issues with datasets:

 rebuildTypes[ds_]:=ds
 ds1 = Transpose@Dataset[<|"a" -> Range[5]|>]
 ds2 = ds1[rebuildTypes]

(A pure function like #& will not work since the type system understands what the structure of the returned value will be in that case - a custom symbol with down-values on the other hand is opaque to the type system, forcing a rebuild of the internal structure)

I don't really feel as though the Hot Network Posts thing is very useful and I kind of get the feeling that it leads to brigading of potentially controversial questions (such as the one currently on the front page regarding Fortran vs MMA)

@CarlLange Agreed regarding that post, at least.

@CarlLange Wow this is a 1000000% useless comment thread (on the answer): mathematica.stackexchange.com/questions/208045/…

@b3m2a1 Yes, indeed. And 99% non-mathematica users (at least, non mma.SE users) which makes it almost completely irrelevant.

@LukasLang My internet is breaking up, and Twitch's chat is really bad. I don't even know which message goes through and which does not.

Video stream stopped for me completely.

that's a shame... :/ for me, the chat looks like this at the moment: imgur.com/CCbJnGe (note that I have just joined the live stream a short while ago)