@Kuba You're on Windows, right? Does the updateMaTeX[] function here hang your kernel for good? You can always get rid of the installed paclet afterwards with PacletUninstall["MaTeX"]

On OS X everything is fine, but just tried on Windows (with v11.0.0 not v11.0.1) and it hung. I don't know why. Evaluating the same commands one by one doesn't hang. That temporary progress indicator may mess up something.

Hi everyone! To complete my previous text, I also think that it is missing other type of user defined docked cells. We have a top docked cell, but why not a left and a right, or a bottom, or bellow selection (that is the predictive bar, right?). In this way, other missing interface "objects", like processors and memory usage, symbols, etc., could be docked on these cells.

...and this is so internet style... so, let's just look at the open programming cloud... At least, two docked are needed to build that interface... that we can't reproduce on the notebook.

While writing this, a funny thing happened. One more docked cell appeared on the cloud interface! A "sign-in" bottom docked cell... So, we are at three different locations of docked cells on this alternative interface

So, at the end, why can't I just have the input / output sequence be one more "object", and repeat it where I want, within the interface itself... I mean, I just programmatically cut my notebook screen in multiple regions, and say what should go where, from a simple dynamic a la docked cell, to an entire input / output sequence (that could be repeated on multiple regions), a spreadsheet "object" on another region, dynamically linked to a given variable (list or association), etc.

@OliverRange you ask a question if it is short and wait:) How can one know if that could be specific differential equations question or related to graphs.

... this is absolutely nothing new to the world of interfaces, seems (to me) as a beneficial evolution...

@Szabolcs works well on win 10. Can try later on 7. Maybe it is related to network speed and an asynchronous task?

Pane clearly does not respect Alignment settings for contents larger than its ImageSize:

Framed[Style["L", 100], ImageSize -> 1 {50, 50},
 Alignment -> {Left, Bottom}]
Framed@Pane[Style["L", 100], ImageSize -> 1 {50, 50},
  Alignment -> {Left, Bottom}]

Should I consider this a bug?

I'd love to reopen this question and award a bounty since a full answer is needed and would be usefull for many users

Help me reopen :)

Still using Outer??

In[13]:= points = RandomReal[1, {1000, 2}];

In[14]:= Outer[EuclideanDistance, points, points]; // AbsoluteTiming

Out[14]= {0.990974, Null}

In[15]:= Table[
   EuclideanDistance[p1, p2], {p1, points}, {p2,
    points}]; // AbsoluteTiming

Out[15]= {0.807154, Null}

(I know that this specific task can be done much faster with other methods.)

Must have something to do with auto-compilation and will probably not be like this in all applications.

@Szabolcs If I am not mistaken, it looks like you have made a mistake using Outer, due to the nested structure of points. When Outer is applied correctly, it is faster.

points=RandomReal[1,{1000,2}];
resOutWrong = Outer[EuclideanDistance,points,points];//AbsoluteTiming//First
resOut1=Outer[EuclideanDistance,points,points,1];//AbsoluteTiming//First
resTable=Table[EuclideanDistance[p1,p2],{p1,points},    {p2,points}];//AbsoluteTiming//First
(*1.347639`*)
(*0.628976`*)
(*0.970818`*)
resOut1 == resTable
resOutWrong == resTable
(*True*)
(*False*)

It seems to me that autocompilation cannot play a role, as EuclideanDistance "is already compiled" (it already as fast as it can be / optimized).

@JacobAkkerboom You are right.

@Szabolcs Weren't you claiming, around 100k rep, that it is time to rest from mma.se? :P

@JacobAkkerboom Hi Jacob

I should stick to what I said.

Hmm I suppose that last sentence of mine about autocompilation is not correct. I mean something like probably with or without compilation some of the same code is used, but there may still be a significant speedup if the function is compiled just because the "variables can be put into the function faster" (as the types are known).

@Kuba heya!

Uf I have to go again. Just a small warning @Kuba I will probably downvote you to reach 8888 :P.

@JacobAkkerboom sounds fair :)

I'm wishing someone with way too much free time is going to write either Mma-specific support for Apple Touch Bar, or generic utility which could be adapted for Mma. ;)