@Kitonum 

thanks. of course you're both correct!

...long day.

ok it looks like

simplify(cc,symbolic);

seems to work with a few simple examples i tried. I was going to delte this question, but i guess i will leave it in case it's useful for someone else.

@vv  @Rouben Rostamian

While we're at it... exporting plots via cmaple interface is also does not work. (cmple interface can be very useful if one, say, wants to run code in batch mode, produce plots and save them - for example on a cluster, remote headless machine, etc).

Regarding exporting to pdfs and bounding boxes. Some time ago, I had some luck manually editing the pdf post export to fix some issues with the bounding boxes. This can also be done programatically via a script, etc. But surely it's a last resort "solution"...

I agree, it would be nice to get export working reliably, and maybe more sophisticated plotting abilities (eg: color bar for contour-filled, density plots, etc)

Hi @Stephen Forrest 

my take on this would be that by default, a call to latex() should:

1) always use \frac{a}{b} only

2) add no extra (latex) white space characters.

I agree with @Leo Brewin  a/b can be useful in inline equations, however, this if often done for shorter expressions and those are easy to type anyway. The real power of a usable latex() command would be for the long expressions/matrices/etc that are really painful to type out.

Regarding (2), I do sometimes use white space characters, but this is often done when defining commands or special characters, and very rarely to "manually typeset" equations.

Having some of these things user-settable via an argument passed to latex could also be an option, as you suggest.

@acer 

the docs for "interface(prettyprint)" claim:

"....Note: Subexpression labeling is on, by default, for prettyprint=2, and off, by default, for prettyprint=3."

but when I set interface(prettyprint=2), I do not see the subexpression labels, only with prettyprint=1.

Am I missing something?... or is this outdated docs?

syntax highlighting in 1d input (i.e the red input text).

... the holy grail for me would be more configurable interface and much more sophisticated editing/navigating capabilities: something like emacs or vim keybindings that would not requre one to reach for the mouse or for the arrow keys constantly when editing a maple worksheet.  (something like jupyter notebooks for example, where a substantial enough subset of vim key bindings can be used).

.. but i realize this is unlikely to happen. Perhaps actually interfacing the maple kernel (even at cmaple level) with jupyter notebooks would be a reasonable first step here. Something i might play with if i ever get the time.

@acer 

nice touch on that last one in terms of Ks... not obvious at a quick glance form the one before it would have such an elegant form.

@Preben Alsholm 

thanks, but that's the point i was trying to make; this function seems out of date and not documented well (i.e. these days a "vector" has a meaning of a "Vector" type, and not "array", if someting returns a Matrix, then LinearAlgebra functions shoudl be able to operate on it, also different returned objects depending on 'method', at least to me, is not obvious,  etc)

it was a dummy example i showed.. in my case, A(t) is large, time dependent, and my DEs are heterogeneous, so doing a MatrixExponential as you've showed is not enough.

ok i guess to get back my initial conditions i need to do:

evalf(subs(t=0, M. subs(t=0, LinearAlgebra:-MatrixInverse(M)).X0));

this wasn't obvious to me from the docs - but maybe this is standard enough that it should be clear?

@Paul 

thanks for looking into this.

Once you figure out what the problem is, please post here if there are any workarounds (especially if the fix somehow doesn't make it into 2018.2).

@Leo Brewin  

great - thanks for posting that. I will play with it.

@Leo Brewin  

nice!...

i could look through your code, but easer to ask: did you implment a parser of maple expressions (i.e. resutls) into latex? or are you using maple's "latex()" command.

If it's the former, and if it spits out something that's more akin to what people write manually, it would be *exremely* useful to wrap your code into something easily accessible from maple so that one can get a reasonable latex version of maple's expressions.

would save many people a lot of manual typing!.. right now (IHMO anyway) maple's 'latex()' command is unusable.

@Daniel Skoog 

yup, I'm in the worskeet mode.

looking forward to 2018.2 then.

thanks

@Daniel Skoog 

thanks for your answer.

>> Since you can create a new "sub-section" by simply inserting a section inside of a section, we felt that it was redundant functionality and elected to remove it in order to simplify the interface.

But if my cursor is inside a section and I go to "insert section", I do not get a subsection where my cursour is, but instead my cursor is moved to passed this current section and a new section is created. Is that a bug then?

Currently, I found the only way to create a subsection is to hightlight some code/text (while cursor within a section) then click on a ">" - like button in the toolbar. 

At this stage I would honestly recommend to just export your data and use another tool for plotting. For anything but most basic plots, plotting/exporting is *extremely* limited in maple. For a comparison take a look what other packages can do. For example, matplolib.
With a couple of lines it should be straightforward to export your data and then plot in (say) python. The flexibility that comes with that (i.e. multi-plot figures, latex, arbitrary layouts, export formats, consistency, etc) is unparalleled.

(I don't mean it to sound very negative here - I like maple and use it all the time; just that when it comes to exporting/making "production quality", or any non-trivial for that matter, plots, it's still in the bronze age)