Talk:MATLAB

'Ban in China'?

The paragraph in the section talks about US sanctions. I have not read of any 'ban' here. Besides, the scope of the sanctions is not whole China. According to the paragraph, only a couple of universities have been affected. Thus, I'd call it a misleading section title which needs changing.

193.197.74.254 (talk) 11:53, 8 December 2020 (UTC)

Agree, nor is it sufficiently sourced, nor relevant enough to warrant its own section IMO. This section should be removed. 2001:1C00:1800:F600:FD6C:7CB4:A8B3:24CB (talk) 18:57, 21 February 2021 (UTC)

Examples

My use case is that I do not know MATLAB but want to get an impression as input to the selction of a numerical tool for a project. I like the beginning of the syntax section (which is also an example section) as it is self-contained in explaining the examples. This is done consistently until we get to the magic square M example, where we encounter the line

 (J,I) = meshgrid(1:n)

but the meaning of "meshgrid" is not explained. Thus, I do not really understand that example. Similarly in the 2D sinc example, where we encounter "meshgrid" again, and also other new notation such as ".^" and "eps".

I miss some more advanced examples - like with built-in special functions (if it has some?) how can one handle I/O. Can I interface with external code and examples of how to do that. How about reporting? All in all I miss some more information about the possibilities rather than the limitations. -- Slaunger (talk) 12:34, 4 March 2008 (UTC)

A quick search for matlab meshgrid or matlab eps would find you what you need. Add the info the article if you think like. Dicklyon (talk) 15:41, 4 March 2008 (UTC)

Thank you for responding. I did the same search and found my answers. I do not mind contributing, but I prefer to do it on topics which I am knowledgable about, and secondarily I have reservations at contributing here, as my written English is unworthy en.wikipedia. It was just a suggestion for improvement. -- Slaunger (talk) 20:55, 4 March 2008 (UTC)

C-Mex files

Someone should write something of C-MEX files in MATLAB. It may be helpful to readers.Vice regent 16:47, 15 May 2008 (UTC)

Here is a simple "hello world" example MEX-file:

#include "mex.h"
void mexFunction(int nlhs, mxArray *plhs[], int nrhs, const mxArray *prhs[])
{
    mexPrintf("Hello World!\n");
}

Place it a file example_mex.c, and compile it as: mex example_mex.c (make sure to run mex -setup at least once before). Finally run the new MEX-file in MATLAB as any usual function:

>> example_mex
Hello World

Amro (talk) 20:58, 14 August 2013 (UTC)

History of default data type

Today the default data type in Matlab is the 64-bit double-precision floating point (complex numbers need two times 64 bit). I have been unable to find information about the historical development of the default data type. What was it in the early days of MATLAB? Matlab appeared first on computers on which the standard allocation unit was the byte (8 bit) and I guess 64 bit precision was not possible back then. Does anyone remember or even better know a source for this information? User:Nillerdk (talk) 10:15, 16 November 2008 (UTC)

64-bit "double-precision" floats have been common on computers for a long time, and as far as I can recall, that's what Matlab always had; I started using it around 1990. The programs it was based on were much older, but I think still used the same numeric format, more or less (but before IEEE standard floating point, not all 64-bit floats were alike). Dicklyon (talk) 06:34, 28 November 2008 (UTC)

You can find a bit of history in a blog post by Steve Eddins (a developer at MathWorks): Data Types. Amro (talk) 20:44, 14 August 2013 (UTC)

Programming Language or Scripting Language?

Since matlab code can't be run independent of the MATLAB programming, I am 99% sure that it is a scripting language, and NOT a programming language. Discuss. —Preceding unsigned comment added by 24.10.240.41 (talk) 04:11, 2 January 2009 (UTC)

Well, there is a compiler that allows one to run matlab code outside the matlab environment. -- Jitse Niesen (talk) 15:59, 2 January 2009 (UTC)

If you read the link you provided, you will see that so called "compiled" m-files are completely dependent on the Matlab runtime, which must be either packaged with the executable or downloaded before execution. This equates to packaging the MATLAB engine/interpreter and simply slapping an m-file on the end. Simply put (in non-marketing terms) this is not compiling. This is stripping the engine out of MATLAB and tacking your script onto it. — Preceding unsigned comment added by 108.36.227.199 (talk) 06:33, 26 February 2016 (UTC)

Reading the first sentences of the wikipedia articles on programming languages and scripting languages, I can only conclude that this section was opened with a category error. Scripting languages are a subset of programming languages, so there can't be an either-or. Also, the article on scirpting languages states that scripting languages automate sequences that could be executed one-on-one by a human operator - while it is possible to write such automations in any programming language, MatLab is certainly not confined to that kind of code, although the simplistic source code examples in the article may give another impression.92.218.143.107 (talk) 08:32, 4 November 2020 (UTC)

Since you did not define what scripting languages and programming languages are, in your opinion, your hypothesis is empty. 129.247.247.239 (talk) 15:50, 30 January 2024 (UTC)

Unclear discussion of matrices

This is given as an example in the matrices section but it is unclear to someone without previous knowledge (such as myself) how that command brings that answer.

A(2:4,3:4) ans =

11 8
7 12
14 1

could someone please add an explanation into the article?

THEMlCK (talk) 02:18, 7 April 2009 (UTC)

I added a sentence that should help, right before that example. Dicklyon (talk) 04:09, 7 April 2009 (UTC)

Monkey Analytics

I think this link can be removed. --SiggyF (talk) 09:53, 6 December 2009 (UTC)

NPOV tag

It is pretty clear from the previous discussion dating back for at least 3 years that there will never really be a "resolution" to the NPOV dispute. There are people dug in on both sides. Does that mean that the tag stays on forever? 76.118.23.40 (talk) 18:17, 5 June 2010 (UTC)

I've removed the tag and replaced it with {{criticism-section}}. Dispute tags are not supposed to be badges of shame to be worn indefinitely by articles. This article needs significant work, but I believe it's simply a matter of requiring a competent rewrite to comply with our guidelines on article content and tone rather than something which needs thorough discussion on talk. Chris Cunningham (not at work) - talk 12:47, 7 June 2010 (UTC)

Objects discussion

MATLAB supports classes, however the syntax and calling conventions are significantly different than in other languages, because MATLAB does not have reference data types. For example, a call to a method object.method(); cannot normally alter any variables of object variable. To create an impression that the method alters the state of variable, MATLAB toolboxes use evalin() command, which has its own restrictions.

This isn't strictly true. In the MATLAB object system, there are two types of objects. Objects with value semantics have the property that this example is pointing to in which a LHS argument must be specified to contain an object with modified state after method is invoked.

new_object = object.method(); 

The second type of objects in MATLAB have handle semantics. These objects can have their state modified by the invokation of methods without a LHS argument

object.method();

Extraneous info in introduction

There seems to be some extraneous info in the introduction, regarding the rising importance of MATLAB for continuing education students. This strikes me as possibly worthy of a sub-section (possibly), but not worthy as part of the introduction. No references are given and the concluding sentence, "These stand-alone courses are of special significance in helping older people with backgrounds in engineering, science, and economics remain competitive." seems highly opinionated. The rest of the introduction is fine but I am deleting this portion for now and leaving this comment for open discussion/disagreement. —Preceding unsigned comment added by Sealbock (talk • contribs) 00:09, 4 December 2010 (UTC)

Why there is no toolbox price list ?

Why there is no toolbox price list ? There is not even a single word in the article about how much does it cost. One could think MATLAB is so humanitarian and altruistic.... — Preceding unsigned comment added by 62.206.211.29 (talk) 13:24, 25 August 2011 (UTC)

Ever seen a price list in an Encyclopedia?Billlion (talk) 15:58, 25 August 2011 (UTC)

Comparison of computer algebra systems for example. KudzuVine (talk) 23:40, 28 September 2012 (UTC)

What's secondary programming?

If it's a popular expression for some people, it might be worth a separate article and a link. Otherwise, it should be replaced by something clearer, explained, or the section should be removed. Engelec (talk) 17:51, 1 October 2011 (UTC)

No reply six months later, section removed. Engelec (talk) 14:51, 23 April 2012 (UTC)

Section 6 (License)

Section 6 has no citations - is there evidence of this? — Preceding unsigned comment added by 144.212.3.4 (talk) 13:57, 20 January 2012 (UTC)

Colons for submatrices

1. Please describe the colon notation for pulling all elements in a column A(:,ColIndex) or row A(RowIndex,:). Was this an innovation of Matlab (or APL?)?
2. Please describe the government-funded development of MATLAB as an interface to BLAS and LAPack, I believe by Cleve Moler at a State University in New Mexico. The primitive MATLAB used to be freeware in the early 1990s.
3. (Regarding "Easter eggs": In the 1990s, the interpreter responded to "fuck" with "Your place or mine?". Has this been documented?)
4. The article should specify whether MATLAB stores vectors and arrays by row or columsn.

Thanks,  Kiefer.Wolfowitz 17:27, 20 March 2012 (UTC)

Wrt. 2, see [1] where the fortran source code can be found; it can be compiled easily with gfortran (gfortran -o matlab source/*.f). In the readme file, Cleve Moler writes: I distributed a few hundred copies of the source code, usually charging a $100-or-so service charge, and including a letter requesting that it not be redistributed. I never used the term "public domain". But maybe it was made freeware later. Engelec (talk) 15:09, 23 April 2012 (UTC)

Jargon

It is not clear if the sentence "For example, typing in "spy" will get a generate of the spies from spy vs spy." is meant to make grammatical sense or if it is computer jargon. The word "generate" is normally a verbal and not a noun. KudzuVine (talk) 23:37, 28 September 2012 (UTC) Side note: in vsn 2011a, the image is a wolf's head, not a MAD magazine image. — Preceding unsigned comment added by 12.129.98.129 (talk) 12:31, 18 December 2012 (UTC) The grammatical error has been fixed. — Preceding unsigned comment added by 86.143.195.11 (talk) 15:09, 25 January 2013 (UTC)

Both information lacking and superfluous examples

Parts of this article contain information that is i.m.o. not relevant for an encyclopedia. For example,

• A square identity matrix of size n can be generated using the function eye, and matrices of any size with zeros or ones can be generated with the functions zeros and ones, respectively. Does this tell us anything about MATLAB?
• What about 4 examples, even involving arrays in the 'Variables' section? I don't even get to know whether MATLAB is case-sensitive in the entire 'Syntax' section.
• The 'Array' section is way too long. Readers are smart enough to know what happens when the increment operator defaults to 1. The examples hardly clarify anything, and their descriptions are usually saying the same thing multiple times. Logical indexing, however a powerful and characteristic tool in MATLAB, is not included. Neither are Cell Arrays.
• Functions, one of the fundamental structures within MATLAB, do not have an example, and just throw around with vague 'anonymous functions' terminology.
• Do we really need to know how a magic square is created within MATLAB? Sure, let's use a neat little math puzzle as clarification.
• The extremely long examples on how to make a 3d-graph, which is just copy-pasted from the MATLAB documentation. Those plot are not only irrelevant (seriously, wireframe AND surface? Why not go through all 10+ 3d plotting functions in MATLAB?), but also in the wrong section; plots are not GUI; they are output, not interface.

Opinions?92.109.161.68 (talk) 19:00, 13 September 2013 (UTC)

Agree, too many examples. Also, the reception chapter is missing. Shaddim (talk) 20:40, 19 January 2015 (UTC)

Update latest stable release version and date using the Wikipedia latest stable software release template

I'm not sure how to update the Wikipedia article to reflect the latest stable version and date using the Wikipedia latest stable software release template. This Wikipedia article should be linked (http://en.wikipedia.org/wiki/Template:Latest_stable_software_release/MATLAB) so that it works. 146.186.210.67 (talk) 17:49, 6 May 2015 (UTC)

MATLAB Mobile

please add matlab mobile info — Preceding unsigned comment added by Mohammadsdtmnd (talk • contribs) 16:05, 16 November 2015 (UTC)

Article name

Should this article not be moved to Matlab or MatLab as per MOS:TMSTYLE? A.Brudz (talk) 10:14, 8 November 2020 (UTC)

Mathworks consistently refers to this software as MATLAB (I have only ever seen lowercase in function names in official publications). Even if most people use Matlab instead, my reading of MOS:TMSTYLE doesn't lead me to the conclusion that this should be changed. Muxarin (talk) 10:04, 23 September 2021 (UTC)

It looks to me like MOS:TMSTYLE is less relevant here than MOS:TMRULES; specifically the bullet point re capitalisation. TomFryers (talk) 22:22, 17 November 2021 (UTC)

Proposed history content

I work for MathWorks, the developers of MATLAB. The current History section is just 1.5 paragraphs long and doesn't include some of the better independent citations available. I'd like to offer an expanded version as a proposed replacement that would grow the history section about 5x, add better citations, and so on. I was hoping an independent editor would be willing to review the draft and approve/implement all or part of it or provide feedback per WP:COI. Lendieterle (talk) 21:22, 14 January 2021 (UTC)

I checked through it and it is well-sourced. I changed the word 'cult-like' to 'strong' as I couldn't access the source for that statement and it was the only portion that seemed to stray from neutral point of view. Otherwise, I think it's great. Brirush (talk) 19:11, 18 February 2021 (UTC)

Alternative Section

See also: Archived previous discussions about "Alternatives"

The current page has an "Alternatives" section that features a long list of competitors. The list is uncited, most of the competitors listed are not really MATLAB's main competitors, and it kind of feels more like a list of plugs/linkbait more than prose-style encyclopedic content. I wanted to ask if such a section was appropriate or could it be removed? I work for MathWorks, so will not edit myself. Pinging @Brirush:, who helped with a prior edit in case they can chime in. BlueAmatero (talk) 20:27, 30 March 2021 (UTC)

After discussion with other math editors, I've removed the unsourced section and replaced it with a sourced paragraph in the appropriate section. Let me know if there are any errors!Brirush (talk) 22:04, 1 April 2021 (UTC)

I removed a growing "alternatives" paragraph. It's just a magnet for people to add their favorite software and history shows it will grow without bound. Comparisons are better served by the pages already linked in the See Also section. Matthew Simoneau (talk) 17:51, 19 January 2023 (UTC)

I think at least GNU Octave should be mentioned somewhere, possibly in the "MATLAB and other languages". This is due to GNU Octave's stated goal of interoperating with MATLAB (incompatibility issues are treated as bugs). It is useful information, even for MathWorks customers, as GNU Octave can be easily included in a CI system, for instance. MATLAB is in turn extensively mentioned on Octave Wikipedia's page (25 matches). Mayeulc (talk) 12:12, 30 May 2023 (UTC)

Release History

Thanks @Brirush:! I just have one more thing I wanted to ask about. I'd like to make a bunch of additions to the "Release history" section as outlined in bold here. I imagine these shouldn't be very controversial edits, even for MathWorks itself to be making, but if you have time to give it a quick lookover and let me know if you have any concerns, your time would be very much appreciated. BlueAmatero (talk) 19:17, 5 April 2021 (UTC)

No, those edits are just fine, I don't see any possible problems with them; they're sourced, neutrally-worded, in line with pre-existing information and fill in missing information. You should be free to make those changes in the main file. Brirush (talk) 22:26, 5 April 2021 (UTC)

Update John N. “Jack” Little’s title

I work for MathWorks and would like to ask an independent editor to update the description text as part of the rollover link under Origins. John N. “Jack” Little’s title was updated from president and co-founder to CEO and cofounder. This update is reflected on MathWorks Founders page[2]. Pinging @Brirush who have shown an interest in the page before. BlueAmatero (talk) 15:08, 20 July 2022 (UTC)

Update infobox to indicate current stable release is R2023b

I work for MathWorks. I request that the infobox be updated to indicate that the current stable release is release R2023b, as listed in the "Release history" section of this page and also in a press release on the MathWorks website (https://www.mathworks.com/company/newsroom/mathworks-introduces-simulink-fault-analyzer-and-polyspace-test-in-matlab-and-simulink-release-2023b.html). Thank you. — Preceding unsigned comment added by 2601:197:1300:4070:3C2C:E46E:DB67:FC6D (talk) 01:40, 15 October 2023 (UTC)

I don't work for MathWorks, but I would also appreciate it if the infoboxes (both of them) displayed the correct latest versions. Unfortunately, I've not been successful. Version information is kept on Wikidata; to add information on a new version, it should be enough to add a new value to the "software version identifier" property (additionally setting the "publication date" and "version type" properties on the new value), and (possibly) purging the cache of the Wikidata item and/or this page. In practice, this is not enough, and the new version never shows up in the infobox.

I first thought that this was due to problems with the references on Wikidata, perhaps because only referenced property values would be considered, but it isn't so. I am stumped and would, like my anonymous friend above, kindly ask that a more experienced editor ensure that the correct current release (R2024a) shows up in the infoboxes, and explain the process here on this talk page as well. Thank you. Mjaðveig (talk) 16:40, 28 March 2024 (UTC)

Example code in the functions section

The box of example code in the functions sections is not relevant. It looks like it's inherited from a section about the colon operator in array indexing, illustrated with an image manipulation application. It could usefully be replaced by an example of a function definition if it isn't just removed completely. 46.208.239.4 (talk) 20:06, 18 March 2024 (UTC)

Version history and WP:NOTCHANGELOG

There is public value in the release history of long-running and popular software like MATLAB or Firefox.

I consider A list of every version/beta/patch is inappropriate. Consider a summary of development instead to be as fulfilled in an exemplary manner since the numerous prereleases and later updates which would multiply the length are not covered. MATLAB release notes are usually several pages long; thus the Wiki table is only listing a few highlights. (1)

Is Use reliable third-party (not self-published or official) sources in articles dealing with software updates to describe the versions listed or discussed in the article. an exclusion criteria rendering a table illegal for Wikipedia or is this to be understood as an advice? Firefox early version history for example nearly exclusively uses official docs from Mozilla. (2) --Bernice Heiderman (talk) 22:56, 5 July 2024 (UTC)

'Public value' is not the standard we use on Wikipedia. There is also public value in listing the phone numbers of service providers, but our article on Plumbing does not contain a list of phone numbers for local plumbers. The community has determined that some things are simply off topic here. Release histories (and phone listings) are such things. Wikipedia is a large site and there is a lot of maintenance that needs to be done - if you have found other places on the encyclopedia that contain information that shouldn't be here, that is not a reason to make the problems worse by adding such on this article as well. (WP:OTHERSTUFF) MrOllie (talk) 23:20, 5 July 2024 (UTC)

@MrOllie: Please reply to (1) and (2). I've used public value in the intro not trying to making a point with it.--Bernice Heiderman (talk) 23:33, 5 July 2024 (UTC)

My reply stands. MrOllie (talk) 23:35, 5 July 2024 (UTC)

@MrOllie: I'm a non-native speaker of English. Shall I publicly dissect your answer to find any statement about (1) or (2)? You are just referring to 'public value', talking about phone listings which neither I nor the policies ever mentioned and mention the maintenance burden,

but the maintenance burden at other places of Wikipedia is simply off topic for this article. A version history is by its nature stable information.

The community involved in this article has generally been in favor of a balanced version table (see above Talk:MATLAB#Proposed_history_content Talk:MATLAB#Release_History). @Brirush: judged that proposal as sourced, neutrally-worded [and] in line with pre-existing information.--Bernice Heiderman (talk) 00:05, 6 July 2024 (UTC)

I have answered your post, and I decline to repeat myself. Thanks! The other discussion topics you have just pointed to are completely irrelevant. MrOllie (talk) 00:07, 6 July 2024 (UTC)

@MrOllie: Therefore The community has determined that some things are simply off topic here is your only assertion? That's both vague and uses very simple language.

In WP:NOTCHANGELOG after the advice NOT to produce Exhaustive logs of software updates [a] the concretization A list of every version/beta/patch is inappropriate [b] follows. If [a] would mean a general ban of version histories no concretization [b] would be necessary. That's exactly the approach: Listing only the main biannual releases, but not beta/patch versions (YES to Release Notes, NO to Changelogs).

I invite you to reach out to more parties interested in this article (like @Lendieterle and BlueAmatero:) and find a broad consensus here. In the meantime I will restore to the prior state.--Bernice Heiderman (talk) 01:06, 6 July 2024 (UTC)

This is exactly the kind of situation that WP:NOTCHANGELOG is supposed to prevent, a long table of irrelevant version history. You are close reading and trying to find loopholes, but that really is not going to work here. (WP:LAWYER). The WP:ONUS is on you to get consensus to keep content in the article. You should not restore it. MrOllie (talk) 01:25, 6 July 2024 (UTC)

@MrOllie: I have neither added the words Exhaustive (1) nor every (2) in WP:NOTCHANGELOG. „Irrelevant version history“ is a personal judgement, proving that you are interpreting the policy. Regarding consensus I cross-reference to Talk:MATLAB#Release_History again.--Bernice Heiderman (talk) 02:20, 6 July 2024 (UTC)

A years-old question from a Mathworks employee about updating the article is completely irrelevant to this discussion, today. What is relevant is that we have a policy that specifically calls for the removal of version histories and change logs. Cherry picking individual words of the policy will not somehow cause it to mean the opposite of its plain meaning. MrOllie (talk) 02:30, 6 July 2024 (UTC)

@MrOllie: That's what I don't understand. It's not we (..we have a policy..) – you are enforcing the removal of each and every version history despite WP:NOTCHANGELOG refers to ([Common sense must be applied regarding ..]) the level of detail to include: There is no level of detail to include after a complete removal.--Bernice Heiderman (talk) 02:43, 6 July 2024 (UTC)

The appropriate 'level of detail' is the prose history section already present in the article. That is the 'summary of development' suggested by the policy. MrOllie (talk) 02:44, 6 July 2024 (UTC)

@MrOllie: WP:NOT has not decided whether version histories shall use prose or tables. But I'm certain there have been discussions about that – may you please point me to them?--Bernice Heiderman (talk) 02:50, 6 July 2024 (UTC)

If you have general questions about Wikipedia policies (or just in general), WP:TEAHOUSE is a good place to ask them. MrOllie (talk) 02:54, 6 July 2024 (UTC)

@MrOllie: Thanks, but I'm not interested in general chatter, but a specific question regarding this article.--Bernice Heiderman (talk) 13:17, 6 July 2024 (UTC)

Reading WP:NOTCHANGELOG, I see that the emphasis is on avoiding exhaustive, long lists - not on eliminating the list altogether. I would find it helpful if a short list of when particularly useful features were added - possibly as a list, possibly as a table, possibly as bullet points, possibly as prose - whatever. The product is about 45 years old and has definitely changed over its lifetime. We do not want to give the impression that MatLab in 1990 was the same as MatLab in 2024.

For me, the addition of SimuLink was massive. Then the push for entire life cycle auditing of requirements, code and unit tests circa 2010 - something aimed at commercial users that no other competitor offered then.

Also the method of compiling to an executable for either Windows or an embedded platform has dramatically changed multiple times over the years. In the 1990s it translated m-files to C, and compiled that with a large library into an EXE (I used Turbo Debugger on it). In the 2000s they just put all the M-files into a self extracting ZIP file that also included a runtime interpreter (no editor or command line) and just ran it (Turbo Debugger became useless, need to debug). In the 2010 they went back to translation to C (yay! debuggers useful again). But this might be of interest only to embedded developers like me.  Stepho  talk  01:27, 7 July 2024 (UTC)

Major developments are already covered in the history section, as I mentioned above. The first compiler for Matlab that you bring up is already mentioned, for example. I believe some version of SimuLink (albeit under a different name) has been available since the 1.0 version. MrOllie (talk) 01:40, 7 July 2024 (UTC)

I came here originally expecting to defend the old list, but after reading the Wikipedia policies it does seem it was out of place.

However, I agree with you that more history would be beneficial. In fact, the current 'Recent History' section is missing citations and is underdeveloped. I you, @Stepho-wrs, wanted to add more about it, I think it would be very useful to describe in prose some of the more influential changes. I think the standard should be the second- or third-party sources you can find; if it was big enough for people to talk about it in reliable sources, it should go in. Brirush (talk) 20:49, 8 July 2024 (UTC)

As a long-time Matlab user, the version history list was the main reason that (repeatedly over the years) I visited the page. It was very, very useful. I understand the table may not be desirable in the main Matlab page, but could not the version history, which has a lot of merit in its content and gives a feel for the evolution of Matlab, be put into a separate page and linked to from the main page?

Tuppence worth. 2user2 (talk) 16:59, 26 July 2024 (UTC)

I'm afraid that this simply isn't what we do here - the community has determined that such tables are off-topic for the encyclopedia. That would also be the case if it was placed onto another page. MrOllie (talk) 17:22, 26 July 2024 (UTC)