@kirma Ah but I already have a bunch of keys on the keyboard that I hardly ever press. I would much rather work on making the computer do more interesting and useful things (automatic game design) when I press the keys.
For all the non-native speakers here I need to share an excellent tool. It's called Grammarly, and it is just unbelievable what mistakes and grammar mistakes they find and highlight. It works in almost all website input fields. The basic version is free and can be installed as browser plugin.
(And I always thought I'm not that bad at English)
I was very skeptical when I first saw the new region functionality ... It is so insanely general. It can express just about any possible region operation on any possible combination of region representations. There is no way that all this can work seamlessly together. It is expected that many operations that are easy to express will fail.
I am actually quite impressed with what it can do in version 11.
The disadvantage of the generality is that people will get annoyed when stuff doesn't work, and many won't realize how much functionality there is already ...
Another disadvantage is that it isn't 100% clear what I should expect to work and what I shouldn't, what I should expect to be fast and what I shouldn't. In some specific situations I would rather have a much more limited function with a very clearly defined scope.
But I see the advantages of the generality too.
I wish they exposed the specific implementations in addition to the general interface ... and not just for regions.
I really like how Oliver put together the FEM stuff and how he documented parts of it. The completely general approach would give us one function that can solve PDEs fully automatically with minimal user control. We do have something like that in NDSolve. But we can also access the details, and there are tutorials which show us how the solution works step by step. We can build our own solver, and we have access the toolbox that the high level functions use internally (NDSolve`FEM` ).
Hey @Szabolcs by any chance you have any clue of what is going wrong with SQLInsert for me? When I try to insert a mma integer expression in a SQL table on a particular column that hasint8 type assigned, my mma integer is converted to something else. For example an integer 20161102135641 in mma becomes 545533010 when inserted. Any idea what might be going wrong?
Packages can be made into paclets, which provides easy distribution and versioning. The paclet metadata is in the PacletInfo.m file. The settings here also determine how the paclet can extend Mathematica.
What settings and extensions can be used in a PacletInfo.m file and what are their effects...
Let's try to figure out what we can do with PacletInfo.m
There are many guesses about what might work, which I don't have time to personally verify. If you feel like trying out any of them, please edit the community wiki answer.
If you feel that it's appropriate, also feel free to post a new answer.
@Searke Do you know if there is anyone from WRI who is familiar with this stuff who we could ask nicely to contribute? I understand that many of these are probably irrelevant for users and are strictly internal. But some are clearly extremely useful, and even essential for cerating our own applications. I am interested in these useful ones.
@halirutan and everyone else: I just added the following note to the post. Spelunking help will be most welcome!
> Look at the contents of First[PacletFind["PacletManager"]]["Location"]. This contains at least a partial implemention of the PacletManager. The code is well commented.
Extensions.m has most of the stuff I wrote up in the post ... if only I found it earlier!!! (@halirutan)