@OleksandrR. Part of the context is also that his colleagues released a great introductory book on Mathematica just a few months ago. They have a lot of experience teaching Mathematica, and they offer a complementary online course for free.
But completely agree that there is no big downside to writing a book of course. Kudos to him for doing it.
@Pickett oh, I see. I didn't know about this other book. Although I suspect that Wolfram had already made considerable progress with his book by the time the other one was released, unless he writes very quickly. Plus Wolfram's book is free and available online, which may help. I was actually surprised that he decided to write an introductory book since to my mind it doesn't seem the most rewarding thing to do.
@ShutaoTANG The wiki page serves as an introduction, there are more complete books on NURBS, like this one (I'm pretty sure there are many books in Chinese on this topic, too).
@Silvia I have been reading the classic textbook "The NURBS Book" this year, however, this book just introduces the open curve, doesn't discuss the closed curve. In addition, I also searched the papers that discussed the closed curve, unfortunately, few paper do that work in detal:)
@ShutaoTANG Did a quick check, I think chap 7 discussed circles and conics, which are closed curve (though not general type). I'll look into your question today later. :)
@Silvia In addition, jojja give me this suggestion: Its the same linear problem you just add more points to the equation that wraps around the perimeter, same way as contol points do. The idea is described in one of the links in the aswers allready provided by you: It reads plain and clear at the end of this page. Your problem seems more fundamental than that, your most probably having the message lost in translation
Here, I will give a Global B-spline Curve Fitting according to the least-squares method.
LEAST SQUARES CURVE APPROXIMATION
Assume that $p\geq 1, n \geq p$, and $Q_0 , ... , Q_m (m > n)$ are given.
We seek a $p$-th degree non-rational curve
$$C(u)=\sum_{i=0}^{n}N_{i,p}(u)P_i \quad u \in[0,1]
$$...
@Kuba Never tried that. But I wouldn't be surprised if it weren't possible. At least on Mac, there's some control of password fields: when you type into a password field, other applications temporarily lose keyboard control (read or write). I have no idea how this might or might not apply to terminals ...
Well, probably it doesn't apply to terminals at all ...
Sorry, my only experience with RunProcess was with MaTeX. I remember finding some weird bugs and sometimes not being able to read stdout properly. I don't remember the details ... that was with 10.1, I think.
Title says it all so let me show it on Git example.
Let's say you have your git executable on PATH:
KillProcess @ process;
process = StartProcess[$SystemShell];
WriteLine[
process,
"cd path\\to\\repository\\directory"];
WriteLine[
process,
"git push origin master"];
Pause[.1];...
@halirutan Just agreeing with J. M. Your second and fourth examples would sound wrong to a native speaker, but they would understand you.
Although actually I have heard native speakers use the fourth construction.
But that would typically be if you interrupt them to ask a quick question they might say "wrote" instead of "written" there. They wouldn't use it in something published, for example.
An example of English evolution I find interesting is dropping the "ly" from adverbs sometimes. If you are leaving someone's house I think it is more common to hear "Drive safe!" instead of "Drive safely!" A librarian corrected me on that when leaving a Wikipedia event a few years ago, so it stands out to me.
Reading S. Wolfram's preface to his new tutorial I increasingly feel he is the Donald Trump of scientific programming. Made his mark in the 80's and can no longer help himself from putting giant, golden W's on everything.
@MichaelHale We have this a lot in Germany. One very common mistake is to use where when it should be when.
When you have a relative clause which indicates a time, they use where. For instance: "Do you remember the time last year, WHERE we were playing football the whole night?".
@halirutan Interesting. I agree that "when" is more correct, but English speakers will often say "where" as well. I think I remember Steven Pinker discussing this in one of his books. About how it indicates the language speakers view time as a physical line, so they can refer to locations on that line using "where" in some situations.
@MichaelHale There are cases where this is fine too in German, but people tend to misuse it and now we have sentences like "This is the best burger, where exists".
@halirutan I just skimmed back through some of "The Stuff of Thought". I couldn't find exact discussion of "when" vs "where" quickly, but I found a few other interesting quotes about time in language.
@MichaelHale On a related note, Pinker doesn't think there is anything wrong with "drive safe," as he explains in The Sense of Style. The main reason is that today's English is a crossover between dialects where adverbs end in -ly, and dialects where some adverbs don't.
"When today's purists reflect on the ones that remain, like those in Drive safe, Go slow, She sure fooled me, He spelled my name wrong, and The Moon is shining bright, they may hallucinate a grammatical error and promulgate prissy alternatives" :D
I've been meaning to read "Sense of Style" but I was afraid it would be a giant rant against excessive prescriptivism. Most of the interesting things I saw were about exposition and classical style.
@Gabriel Does anyone here know other CEOs who write and care about documentation this much? Hell, I know too many developers that think writing documentation is beneath them.
@Pickett I'll keep that in mind if I run into the librarian again. Good other examples he mentions. @Searke I follow him on Twitter, so I remember hearing about it, but I haven't checked it out. I read Blank Slate and Stuff of Thought over some college summers and then Better Angels after, but not his newest.
@Silvia Looks fine to me. I never noticed the "excluding uppercase letters" for "@" in the StringMatchQ docs before. That doesn't affect your example, but I just always assumed it would match "A" too.
If they remove it, I'll probably end up using their ~~___ syntax even if it is more verbose.
But I actually don't know if I've ever used it for something that's not covered by StringStartsQ and StringContainsQ now. Except for all files like Silvia mentioned.
You can still use Regexs. You just have to use RegularExpression. That shouldn't be too verbose. What's really important is that regexs and normal strings are separate. We aren't bash or perl.
@J.M. Probably depends on how many times it has to be interpreted. Doing StringCases across a list probably only once, but Select with a StringMatchQ inside of the filter function is probably re-interpreted every time.
Anyway most of the time I use StringExpression, more intuitive grammar and more expressive, and possibly faster.
I shall go to sleep now. Nice talking to you guys! See you!
Sorry my memory fault. It's not anything about compiling regex, but a translator from StringExpression to regex: Semantic ` PLIDump ` stringExpressionToRegex
@Searke Unfortunately, he is a CEO, not a developer, not QA, not DR. Let me put it this way, no one cares whether I can cook high-quality steaks, as long as I don't do my job.
@Searke Cleve Moler of Mathworks seems pretty awesome. His blog doesn't need to always be about himself and he has made 2 great online books. I guess I am alone on this, but I just can't take S.Wolframs complete lack of discussing Mathematica as a team effort, or that many, many other systems before and since have similar technology. Maybe it is just grating to my ear.
Hypothesis: for some reason now the kernel makes a copy of packedArray before passing it to the LibraryLink function.
Is this correct? If yes, why would it do this?
I wish Todd Gayley were around ... don't see him post much anymore.
Also @xslittlegrass, please read above. I know you dug deep in LibraryLink.
OK, some other things. @Searke I'm really happy about `DistanceMatrix`: it performs very well. It's the kind of "convenience function" I like to see added. I'm hoping by the time the [[Experimental]] tag is gone it's going to perform even better on large packed arrays.
@Gabriel I was actually thinking of him when I was typing.
@Szabolcs I'll look at it. But I don't know much about packed arrays and all the random things that unpack them.
@Gabriel It's advertising. Most advertising is awful.
@Szabolcs I'm glad to hear it performs well. I didn't have anything to do with it, but from what I know there was a lot of conversation about how to make it perform well
@Searke One nice thing is that it seems to be optimized for several distance measures. But for Euclidean distances I can still get a several times speedup with LibraryLink + OpenMP, that's why I said I hope it will be sped up even more.
Regardless, it's likely the fastest way to do this computation in pure Mma now, which is exactly how it should be for a builtin like this.
Oh well. I'll forward it to the developer then as a suggestion myself
Tech support has to go through a lot of email from many people - some not reasonable.
2
So they (and by they.. I mean myself included when I did this) can have trouble responding to a ticket like this if it isn't clear what is being requested.
Tech support at wolfram research is the only job where you can get yelled at for not remembering how to do residue calculus over the phone and have to walk someone through debugging some unicode related nightmare over a chat window at the same time.
I take that back. The worst thing I can remember was a researcher at a pretty prominent institution who having already used the variables "t", "tau", the gothic mini t, the irish script hand for t, and all the other t looking alphabetic characters that exist
proceeded to being using a character from ancient aramaic that looked nearly exactly like tau or something
There is a famous anecdote about Barry Mazur coming up with the worst notation possible at a seminar talk in order to annoy Serge Lang. Mazur defined $\Xi$ to be a complex number and considered the quotient of the conjugate of $\Xi$ and $\Xi$: $$\frac{\overline{\Xi}}{\Xi}.$$ This looks even bette...
@Gabriel He's a fun guy. It's also nice that not a few important methods of numerical linear algebra were his stuff.
@Szabolcs Yeah, it is nice. Hopefully it will eventually support all the metrics Mathematica currently has, someday.
@Searke As in, the sort of stuff we'd be closing with swift justice were it asked over here? :)
@Searke I think it's one of those cases where "too little knowledge is dangerous"; they don't realize that "not engineered for specs" applies to numerical methods as much as it does for materials.
@Searke :) yeah I need to stop reading the marketing stuff ... I am like a flame to the fire. Instead I just need to remember all the fricken cool and smart people I have dealt with in the this community and from Wolfram the company :)
@J.M. yeah I just started reading his blog seriously. I is great. I have learned so much.
Here, I will give a Global B-spline Curve Fitting according to the least-squares method.
LEAST SQUARES CURVE APPROXIMATION
Assume that $p\geq 1, n \geq p$, and $Q_0 , ... , Q_m (m > n)$ are given.
We seek a $p$-th degree non-rational curve
$$C(u)=\sum_{i=0}^{n}N_{i,p}(u)P_i \quad u \in[0,1]
$$...
Title says it all so let me show it on Git example.
Let's say you have your git executable on PATH:
KillProcess @ process;
process = StartProcess[$SystemShell];
WriteLine[
process,
"cd path\\to\\repository\\directory"];
WriteLine[
process,
"git push origin master"];
Pause[.1];...
