Highlights
关注


投票已结束

投票

It's December 2021. What version of MATLAB are you mainly using?

R2021b (latest official release)
35%
R2020a - R2021a
30%
R2018a - R2019b
17%
R2016a - R2017b
9%
R2015b or earlier
8%
4244 个投票

sasi kala
sasi kala 2021-12-27

I tried with trial version .its awesome and helpful. but trial license over

Hsin-Yu Lee
Hsin-Yu Lee 2021-12-22

I use R2021b to compile programs, but programs fail to open on Win7. It's not recommended to upgrade before MATLAB solve this problem.

David Ackerman
David Ackerman 2021-12-19

Tried to upgrade to R2021b but my late 2013 MacBook pro runs OS 10.14 so I am left with paid-for software that I cannot use. (Can't risk bricking my laptop with later versions of MacOS. I've been warned.)

DGM
DGM 2021-12-16 (编辑时间:2021-12-16)

The only reason I still use R2019b is because I'm active on the forum. If it weren't for that, I'd probably go back to R2015b and save myself a couple gigs of ram and the characteristic lagging, crashing, and broken UI elements.

Rik
Rik 2021-12-16 (编辑时间:2021-12-16)

The local functions in scripts (R2016b) is a real quality of life improvement in my book. The pause during execution (R2016a) is also nice. If you want raw speed you should go back to something like R2007b or even R13 (v6.5). I don't know what caused the sharp jump in speed there, but creation of graphics objects is way faster in R13. Boot times are also excellent. That is my main reason for supporting it in my code, even if it lacks basic features like output suppression ([~,ind]=sort(A) etc).

The next big thing I'm hoping to work with is the functionSignatures.json. Now that the normal (m-file) editor supports them, I would expect more people to create them. I hope Mathworks will soon make something that will allow encoding it a comment block.

DGM
DGM 2021-12-16 (编辑时间:2021-12-16)

Local functions in scripts are very handy, and I was very delighted to discover the generalized array expansion too. For ad-hoc things like what I do for forum answers, both of those are great conveniences ... but I'm not going to publish scripts, and I aim to keep my toolbox compatible back to R2009b, so I still can't escape bsxfun() there either.

Speaking of my dumb toolbox, functionSignatures.json is one of the nightmares on the todo list. Having that info in a comment or something in the file would make things a lot easier to include and maintain (at least with the way I go about things).

Rik
Rik 2021-12-16

Isn't there a function that can run when you install a toolbox? Then you could use that to extract the JSON file from your own comments. I hope this feature will get more attention in the future, as documentation you can access by using a tap on your tab key is perfect. They should also extend it to allow marking char/string inputs as filenames. Native functions have auto-complete for file names, but custom functions don't.

Rik
Rik 2021-12-13

I'm susprised to see the older releases still have substantial active users. Maybe if there were a category below it we could see the effect of the last 32 bit release. (I'm not surprised they have active users, but I don't think I would have predicted 7%)

DGM
DGM 2021-12-16 (编辑时间:2021-12-16)

If I were to base a guess on my recollection of what various users declared in their forum questions, I'd think almost everyone ran R2020a or newer. I imagine the picture painted by the forum observations tends to over-represent users who have fresh student licenses. I also imagine the poll tends to attract people who have a reason to make it known that they don't use the latest release, even if they don't mention why.

J. Ismail
J. Ismail 2021-12-14

Hi Rik, why to be surprised? The Matlab update is simply very expensive for many universities and companies because of the compatibility with other programs like dSPACE and not all of them can afford this price.

标签

尚未输入任何标签。