I wonder what they envision for their new text functions like TextSentences and TextWords, it's like they're making more semantic versions of some string patterns.
@Guesswhoitis. Make a login at Glassdoor and read the reviews there to find out. And I'm fully on your side about the fine developers! I know only a bunch of them, but I would love to work with every single one. Talks or chats were always inspiring and if you'd heard a talk from @user21 only once, you would instantly know who much he likes his FEM stuff and how much he cares about the users.
I guess I don't have to even start about Leonid, Daniel Lichtblau, @JohnFultz or all the others who are an integral part of this site now. Giving tips, awesome insider information or algorithmic details..
I don't see any link on the main Mathematica.SE page that leads directly to http://meta.mathematica.stackexchange.com. There is a panel in the top-right corner that links to "Hot Meta Posts", but this requires an additional click from one of those posts to the main meta page. I think there should...
Yes, generalizations to convex polygons. I'm asking because Wachspress used this for his version of FEM, and I had thought you are well-acquainted with the literature.
(I suppose, not really that applicable in the case of Mathematica's FEM, since you do a simplicial decomposition anyway.)
@Szabolcs in my experience, there is never a functional difference apart from the banner which disappears if you enter your actual activation code when you get it.
Despite the ever-growing list of tools I have for communication, email remains one of the most important. I depend on email to find out about all sorts of things: my ultimate Frisbee game is rained out, flights to Denver are only $80, my Dropbox account is almost full, my neighbor’s cat is missing (again). While [...]
@kirma 10.2.1 is unlikely to happen, except for a global emergency (like Mathematica 10.2 not working on the next Microsoft Windows 10 operating system). Note that besides new functions (reference.wolfram.com/language/guide/…), the 10.2 release also includes many bug fixes (not enough in my opinion, but a non-trivial amount, about 700). We're in the process of updating (fixed) issues reported here StackExchange.
2
Additionally, we will use paclet updates wherever possible to update documentation issues and code problems. This is becoming easier and easier and is beginning to take the place of 0.0.x releases. We're not quite there yet, but addressing any issue automatically with paclet-like update is a goal.
@halirutan We can't yet reproduce this, but this issue is now reported internally. If you have any more details, please do let us know (ideally via support@wolfram.com). Thanks.
@halirutan The documentation has now been updated. Should be live on the reference web site with the next update (usually within a week).
@SjoerdC.deVries Did both issues resolve themselves after you restarted?
@ArnoudBuzing Version 4 .mat files have a simple structure: header-data, header-data, ..., sequentially. The header has information about what type the data is, and how much data is there. There's a small mistake in the importer code in the mapping from numerical codes to datatypes. 1 means 32-bit real, but the importer treats it as 64-bit integer.
This means that it'll try to read twice as much from the data block as it should (64 bit vs 32 bit values) so it runs over the end of the file. Hence the "Insufficient data found for MAT format" error.
ok, I've fixed this in the source and made a pull request for it. I don't think this is paclet updatable (yet), but I will look into that next. Thanks for reporting (and the additional sleuthing)