Wolfram had a Mathematica Player around for a lot longer than Maple. I've been using it for various Mathematica files (.nb) for years since many people would post their Mathematica "notebooks" on-line. I'm not that impressed with cdf since it's pretty tied to Mathematica and seems like just a re-branding of their Mathematica Player.

So, CDF isn't something special, but Maple wasn't first.

I would prefer if Maplesoft would just improve their LaTeX export since we can't make a thesis in Maple document format and journals don't accept them either. Instead, I'm forced to rely on hacks to fix something that's been broken for years.

@FeX141 When I tell simpleLaTeX (using Maple 14) to convert cos(theta_i), I get just \cos(theta_i). I assume that \theta is because of a regex command and I assume that the change in the i to {4 } is because of a regex problem. It's possible that something in the worksheet is affecting the regex, or affecting the expression itself before the simpleLaTeX command.

A latex command on cos(theta_i) gives,

\cos \left( {\it theta\_i} \right)

The basic code for simpleLaTeX just strips out the \left \right commands, also the \it command and converts the \_ to just _ for this code.

@Joe Riel Thanks for the refactored version. I've been going along okay with my original, which I completely admit was a bit of a hack. This newer version is much cleaner. In the new version it seems that user regexes are still applied after the base ones?

I haven't been around MaplePrimes in years so I'm glad to see some work was done on this when I was away.

You're welcome. Maple has no way of knowing that alpha4 is supposed to be \alpha_4 so it can't do the conversion. However you can do it yourself using regular expressions and my export function. Alternatively, I recommend that you use the name alpha_4 which should get converted as you want with my export function.

I'm not sure how difficult it would be, the problem is handling the typesetting of all the different possible expressions. Ideally, it could be handled in the same way that it is typeset for display in the GUI.
Certainly, fixing LaTeX export of worksheets would be very difficult. One possibility is using XSLT on the XML representation.

I've asked Maplesoft on a a few occasions about better LaTeX support and have been told explicitly that they have no plans to improve it. After complaining about the breaking at a fixed length, I was told that some customers actually preferred that. None that I've met.
Maplesoft thinks that improving the worksheets is all that is necessary. However, no journals that I know of accept them, which isn't surprising since they don't have the formatting capabilities of LaTeX.

It seems as if they've done the minimum necessary to claim LaTeX output. I was actually told that people wanted the LaTeX output as it currently is done. One of the answers I got was to increase the line length so it didn't wrap at stupid places, but there is a very small limit on the line length. In the meantime, I have my workaround (posted on my blog) which fixes the line issue (by merging all the lines into one), makes some substitutions for some things and supports regular expressions for customizing other changes.
They've spent all this effort to get the 2D output looking nice so I can't see why they can't transform from those data structures into LaTeX code. It seems like the cleanest solution. I'd also like to see improved LaTeX styles, but I doubt I'll see either one done.

I'm sorry about your back. Rest up and get well.
Implementing a TextMate bundle for Maple is kind of low on my priority list. Maple coding is one of the few things I still use emacs for, but I'm not in a hurry to switch yet. That said, I'll let you know if I implement any new useful features.

One of the things that has bothered me (and others I've talked to) is how Maple displays results. Maple often will display results with leading negative signs, things like
-a + b*x
I'm just making that up, I'm not sure if it actually does it for this example. I'd much rather it be displayed as,
b*x - a
On complex expressions this can be annoying, so I'd really like an option to always try to display a positive term first.
Maple always seems to prefer using cos when doing trig simplification. What if you just want to eliminate the use of tan?

I'm involved with teaching undergrad engineers and mathematics is still a very important part of the curriculum.
With regards to CFD and FEA software, for anything more than simple analyses, understanding the mathematics involved is still important both to set up the problem in the first place and to do post-processing and analysis of the results. Most importantly, without the knowledge of the methods used in the software, the engineer won't be able to critique the result and determine if the result from the software is valid. This is a common problem I've seen with people using the software blindly. The input problem into the program and get a result. However, the wrong input will lead to a wrong result. Without the background how will they know if they've made a mistake?
If you look at Maplesoft's

engineering mathematics survey, hand calculations are actually most common. That said, nothing requires these hand calculations to be simple. I've done hand calculations that are many pages long. I have one derivation that typed up is nearly 100 pages long and most of the work was done by hand since Maple doesn't have the facilities to carry out those calculations. Don't assume that all hand calculations are simple (something that some people at Maplesoft seem to believe according to that press release).

I'm certainly interested in seeing what you come up with when you're done especially since I'm not fond of Maple's LaTeX output.
One of these days, I'll get around to creating a Maple bundle for TextMate (an excellent Mac OS X text editor). I've been going through your emacs code to get ideas.

I regularly use one of the apps that is referenced as an influence,

Quicksilver. It provides a very useful and quick interface to the operating system (launching apps, URLs, appending text, etc.). Certainly something similar could be possible in Maple, the only question is how to set up the interface.
I think that a simple key command can bring up a new window that allows for keyboard navigation of the existing outputs and a right arrow pulls up possible functions to apply to it. This is the most basic form of interaction with Quicksilver.

Thanks, I've been playing with trying to define a diff function, but unfortunately this one doesn't work correctly. This function gives one too many g(a) functions since it applies the T function to the g(a) instead of the h(a). That is, it gives
(T(g(a))*diff(g(a),a)).h(a) + T(g(a)).diff(h(a),a)
when I actually want,
T(h(a))*diff(g(a),a) + T(g(a)).diff(h(a),a)
Somehow I need to tell diff to swap h and g places (and apply/remove T). Basically, I need to control how diff distributes over the dot. I only want it to apply to the second term and I have a rule I can apply to swap the two terms.

Thanks, I've been playing with trying to define a diff function, but unfortunately this one doesn't work correctly. This function gives one too many g(a) functions since it applies the T function to the g(a) instead of the h(a). That is, it gives
(T(g(a))*diff(g(a),a)).h(a) + T(g(a)).diff(h(a),a)
when I actually want,
T(h(a))*diff(g(a),a) + T(g(a)).diff(h(a),a)
Somehow I need to tell diff to swap h and g places (and apply/remove T). Basically, I need to control how diff distributes over the dot. I only want it to apply to the second term and I have a rule I can apply to swap the two terms.

I don't know if a new structure would be necessary, but I've certainly seen that 3d rendering for Maple could be fast.
I think Maplesoft cares quite a bit about backwards compatibility these days, mainly because there is now a large amount of existing code that users have written over the years. Both LinearAlgebra and VectorCalculus define a Vector, but in incompatible ways. It doesn't look like it will be fixed any time soon mainly because of backwards compatibility.