Interesting start of a thread, Oliver. I encourage all posters to write a single point per answer to make it easier for others to vote up answers they agree with.
What frustrates you about MATLAB?
453 ビュー (過去 30 日間)
表示 古いコメント
I love MATLAB. It is so quick and easy to write software to do what you want. It has excellent debugging and profiling tools. It is cross platform, making code easy to share (assuming the other people have forked out for the not-so-cheap license). It has interfaces to other software.
However, there are some things about it that irk me. I'd like to hear from other people what things annoy them about MATLAB.
---------------------------------
Because this thread has become so large it is slow to load and navigate. Please post new answers in the second thread.
---------------------------------
45 件のコメント
Steven Lord
2021 年 4 月 2 日
What do you mean by "change the size of a plot window"?
- Change the size of the figure window
- Change the amount of space in the figure window occupied by the axes
- Change the limits of the axes to show a larger part of a plot that's larger than fits in the current view of the axes
- (As a bit of a stretch) Change the width of the line or the size of the markers used to plot the line. [If you hadn't used the word "window" this would be a more plausible interpretation of your request.]
The plotting capabilities of MATLAB do offer a lot of functionality and with that must come some level of complexity. I will admit Handle Graphics does have a bit of a learning curve.
The main way I would recommend interacting with Handle Graphics objects is:
- Find the handle of the object with which you want to interact. This could involve calling plot or other graphics function with an output argument, using the findobj or findall functions, or using the ancestor function.
- Search the documentation for the task you're trying to complete to see if there's a function. Use axis or xlim to change the X axis limits, for instance.
- Use dot notation to change the object's properties if there is no convenience function. f = figure; f.Position = newCoordinates; as an example.
One problem with your suggestion of "clean up this mess of features related to plot" is that the plot function has been around for at least 20 years (probably closer to 30 at this point) and people have written a lot of code that take advantage of those "mess of features". When we introduced the new version of the graphics system in release R2014b we did fix some long-standing bugs, introduced new capabilities, and eliminated or simplified some of the helper functions. Even though we tried to be careful about it and give plenty of warning, some users were still annoyed/upset with us for breaking their code. Doing a wholesale "blow up the system and replace it with something new" would be painful for both MathWorks and for long time users.
回答 (157 件)
Ian
2011 年 3 月 28 日
- GUIDE is a such an underpowered mess. Very poor set of widgets, endless bugs and terrible performance in the Guide editor (try moving sets of elements with the keyboard and weep!). TMW has dragged along hidden support of more UI widgets likes tabs, why can't they just update GUIDE and bring it.them into the modern age? Yair Altman does more in one blog post than TMW does in each year's paired releases to make GUIs better in Matlab!
- UNICODE support -- why is TMW so far behind on this!?
- Graphics quality -- graphics and text should be anti aliased by default and better composited, the pixel vomit is unacceptable on such an expensive product and one where its competitors are so ahead in this regard.
- Poor OS X support. Failure to copy vector figures to clipboard is my biggest gripe.
9 件のコメント
Steven Lord
2022 年 3 月 10 日
According to the release R2022a Release Notes, "MATLAB now uses UTF-8 as its system encoding on Windows®, completing the adoption of Unicode® across all supported platforms."
Jan
2011 年 2 月 16 日
Using a C-compiler not included in the list of known compilers is horrible. "mex -setup" calls an M-file, which calls a PERL script, which creates a DOS batch file, which is interpreted by another PERL script, which calls the compiler through the command line interface.
Strange. A tiny M-file could create the same command line call to the compiler also. But it would be much easier to adjust it to a different compiler.
6 件のコメント
Praveen Kumar
2019 年 6 月 24 日
I had two subsystems(A and B). I want B inports which are conncected to which outports in A. Any ideas.
David Young
2011 年 9 月 28 日
Overriding subsref and subsasgn in a class is spectacularly awkward.
The problem is this. Suppose you override subsref in order to change behaviour of obj(...). Then your subsref is also called for the syntax obj.propname. In order for it to handle this correctly (i.e. behave the same as the built-in subsref) it has to reimplement all the checks for access protections that are normally done by the system. For just how problematic this is, look at Daniel's answer to my question here.
Apart from this major difficulty, it's also inefficient for every subsref to have to switch on the indexing type ('.', '()', or '{}') on every call, and for the system to have to build a struct that includes the indexing type as well as the indexes.
As far as I can see, it would be so much better if the different indexing syntaxes called three different methods.
4 件のコメント
Steven Lord
2021 年 9 月 23 日
As of release R2021b there are three subclasses from which your classes can inherit, one for each type of indexing (parentheses, dot, brace.) Classes can inherit from zero, one, two, or all three of the classes depending on what type(s) of indexing you want to customize. See the Release Notes for more information.
Oliver Woodford
2011 年 2 月 16 日
16 件のコメント
Chris van Halewyn
2020 年 5 月 6 日
Yep. Me too! I've got a Nexys4 board (£200 -not $33,000!) that I got because it is recommended for education/training and got Vivado (free) to program it. Vivado calls out for Matlab HDL tools on installation but they are not available under the home licence which would have been so cool...
Matlab, please sort out home licensing for HDL tools. It's time.
Dan
2017 年 5 月 2 日
Functions and classes defined within a package should not have to import the package or refer to other members of the package with the package name ... the import should be automatic ...
Given the requirement to refer to the local package name within the class or function definitions ...
- It is difficult to rename the package
- It is difficult to place the package within another package
- It clutters up the function and class code with import calls ...
Please fix...
0 件のコメント
Robert Cumming
2011 年 2 月 16 日
plot legends are not always placed in the "best" position - quite often its right slap bang on top of the data...
2 件のコメント
Matt J
2018 年 11 月 13 日
You can, but shouldn't the default legend position be somewhere not covering data?
Csaba
2013 年 12 月 4 日
My current frustration is with the Matlab editor: when the same M-script source is open in both the Matlab editor and another code editor (certain things are more efficient in the other editor), and the Matlab editor detects a change that was made to the file outside, it reloads (which is correct) - but it also forgets about all location information from before the reload, and jumps to the beginning of the file.
I would expect this is just an oversight (how hard would it be to remember the line at which the cursor was before reload - even if the code change that triggered the reload shifted content, most of the time this would be by only a few lines or pages, still much better than starting from the top of the file every_time_ ...).
And if that's not feasible, at least letting the bookmarks survive the reload would be a reasonable work-around, but sadly no, they are erased as part of the reload as well.
Am I the only one having trouble with this?
3 件のコメント
Steven Lord
2022 年 3 月 10 日
The release R2021b Release Notes indicates that MATLAB maintains bookmarks after closing a file. I'm not sure whether or not they persist across MATLAB sessions; I haven't tried.
Oliver Woodford
2011 年 2 月 16 日
2 件のコメント
Wyken Seagrave
2014 年 5 月 14 日
The secret of creating high quality videos on Windows is to use VideoWriter, an object compressed with Motion JPEG, and not movie2avi, which produces low quality video. So it is frustrating that, when you search the Help for "movie", and you find movie2avi, the only reference to VideoWriter is a single word at the bottom of the help page. I think there should be a warning that it has replaced movie2avi!
Patrick
2011 年 10 月 27 日
The figure-export is a pain, if annotation are used they will move about, not all of them and not much and not to any order (at least nothing i could recognise) but sometimes a few mm in the odd direction.
1 件のコメント
K E
2012 年 7 月 25 日
Also, quality of copied figures is poor (blurry lines/text) and requires workarounds like export_fig .
Richard Finley
2012 年 1 月 30 日
I have been a licensed user of MATLAB for the past 20 years. I am also a big Mathematica user and sometimes use Maple and Mathcad and of course I love R and Python. I have enjoyed using MATLAB for my numerical work, but the licensing hassles have driven me to the point that I am considering just dropping it after all these years. I don't mind paying the (substantial) price for my license -- but then just let me do my work -- whether it is on my main desktop or my laptop. I used to just get the update CDs and my license number by email and no hassles. Are you really making more money by stopping piracy with all your licensing hassles? I doubt it. You are driving away your paying customers/user base. I sympathize with this (former) user: http://mybrainextension.blogspot.com/2011/07/aaaarrrrrrgh-matlab.html
Ian
2011 年 3 月 28 日
Lack of OpenCL support: TMW forces us to depend firstly on a toolbox for this functionality, and secondly on the vendor lock-in using CUDA. Mathematica builds this in to its core, not some toolbox, and provides both CUDA and OpenCL support so we aren't forced to a single GPU vendor.
6 件のコメント
Walter Roberson
2017 年 6 月 28 日
I am told that OpenCL is more a framework allowing for the possibility of various actions being possible, but with considerable variability in what the different manufacturers support, and therefore Mathworks would effectively have to do different OpenCL implementations for each manufacturer and to specialize it for different models within the manufacturer.
tom 1line=1bug
2011 年 9 月 12 日
Frustrating factoids after living with Matlab since version 4.x:
1. Ambiguous use of parenthesis ([([( indexing operations and function calling )])]). I wish Matlab's syntactic sugar could differentiate between this two actions, using parenthesis and something else (i.e. like Mathematica's use of parenthesis vs square brackets)
2. Poor quality graphics (already mentioned in detail by other fellows)
3. Excessive "commoditization" or segmentation of functionality that should be part of the core Matlab product (e.g. signal processing, image processing, statistics, and pdes).
4. I wish there was better quality control when rolling out toolboxes. The seemingly lack of coding style/guidelines is particularly annoying. Even toolboxes meant to be used together have inconsistent syntax and seem a bunch of functions "bolted-on" to Matlab, rather than coherent extensions of M language.
5. I hopelessly wish there was first-rate support for functional programming.
6. Very poor support for time series and data structure requiring multiformat date/time indexing.
99nth. Non-specific gripe: Over time Matlab is becoming less and less my primary platform for scientific computing and algorithm development, and more often than not just a target for deployment (meaning re-writing whatever project in the M language) because the customer/client is using it, or because a particular toolbox is a time-saver. I remember the days of yore, when Matlab looked like a hammer and most problems looked like nails. Now I feel that the handle of the hammer is a Baroque wooden carving, richly ornamented, weighting 120 lbs... More surprising is the strange comfort of spending so much more time working with Mathematica's superhuman -indeed extraterrestrial- capabilities for algorithm development and testing, and the "translating" to Matlab upon request. I guess necessity is the mother of strange companions...
David Young
2015 年 5 月 13 日
The inputParser / validateattributes / validatestring collection of function parsing and checking tools.
- You have to write too many lines of code to do even simple things.
- I haven't done any testing, but I wonder about the overheads: all that object, anonymous function and cell array construction and access.
- Despite the complexity and apparent power, there are silly limitations. For example, required arguments must precede optional arguments must precede name-value pair arguments. Given that there are validation functions, there's no need for this.
- There was an absurd change of name of one of the methods of inputParser from addParamValue to addParameter . The documentated functionality appears to be the same, so there was no need for the change, but this is the biggest cause of users of old versions to run into trouble with my code on the FEX.
- The functions don't work nicely together. For example validateattributes always throws an error for invalid inputs, rather than returning a result. That means that you can't combine two calls to it into a single expression for a validation function (e.g. my input could be empty, or a row vector of length 3). I can work round this with a wrapper function that catches the error; but integrated design would have avoided the need.
- Another failure to integrate functionality: validatestring can handle abbreviated inputs nicely, but if you use it as a validation function for inputParser the information as to which possible input whas given is lost, and you have to call it again on the parsed results.
There's more to say, but I hope that makes the point. There's a strong case, I think, for starting again in this area, or failing that for offering a migration route to functions that do work together properly.
5 件のコメント
Andrew Janke
2021 年 1 月 24 日
I don't know about David, but I find your new function argument validation feature to be a big win, and am switching to using it a lot.
My only real complaint there is that the one-arg datetime(x) constructor treats numeric inputs as datevecs instead of datenums. Datenums are more common in all the Matlab code I've seen. So you can't do this and have it work right:
function fcn(dt)
arguments
dt datetime
end
[...]
end
And that there's no single value that you can pass in explicitly to a function to mean "I'm omitting this argument; just use the default", which happens a lot in my codebase because we have functions with like 8 positional parameters and you might want to only specify values for parms 1 and 7.
Knut
2011 年 5 月 5 日
The fact that once you schedule a low-level function, you cannot in any way stop it. Such as conv(zeros(umpteen,1), ones(umpteen,1)). A big red, working "stop" button would have been nice.
I like to run my scripts, insert break-points, then test stuff using command-line on the variables available. In recent versions, I have not been able to define new variables that way easily, having to declare them as global (??)
I would have really liked the core matrix-syntax to be even cleaner (not having scalars being 1x1 2d arrays for instance).
I would have liked some sort of extension of the matrix syntax into tensors.
I really hate the frustration when I go from compact, vectorized, readable MATLAB-code and have to dip my toes into "real" programming languages, where you spend most of your time concentrating on the implementation instead of the functionality :-D
1 件のコメント
Steven Lord
2019 年 8 月 20 日
For the first point, this won't work for all "low-level function"s but the Pause button in the Editor, introduced in release R2016a, may be of use to you as may the Array Size Limit preference introduced in release R2015a.
Yes, I know I'm responding to an eight year old message. I just noticed it and wanted to add the relevant information in case others had the same frustration.
Chad Greene
2015 年 3 月 27 日
I'd like new installations of Matlab to be able to grab preference settings from previous versions installed on my computer. It's minor, but spending 20 minutes sifting through M-Lint settings and fonts and background colors is just enough of a hassle to keep me from upgrading twice a year. I wish upgrading Matlab were as easy as upgrading Firefox.