I never see an option where I can uncheck the Atomic Identifier once enabled.  The next time I get in a bind I will post the whole work sheet.  But it sure seems related to LHS assignment where a superscript is used.

For sure Maple doesn't like a superscript on the LHS because it gives me an error when I try to do this.  The error goes away as soon as I make it atomic.  But then, as written, I can't go back.

Look at this very simple example:

a^2 is assigned to equal x.  Then I change the a^2 to be atomic.  If one looks in the pull down menu the atomic identifier option is shown as selected.  Fine.  No problem.

But now try and go back to undo this.  Where is the "make this non-atomic" button?????  It seems like I can do this once and then forever the assignment as atomic is permanent.

Thanks,

Scott

Download atomic.mw

All -

I surely appreciate the discussion on this topic.  At the end of the day though, and as a new user of Maple, I am simply trying to identify an efficient workflow that meets my situation requirements.  Since I am new to this, I'm not tied down to an approach that has worked for years back.  But that also makes it a bit painful when I try to walk through what appears to be the ultimate, intended, MapleSoft workflow based entirely upon a GUI based workflow - a flow that is very buggy at this point as claimed by most users I've met on this site.  But I've not given up yet on the GUI flow.  Probably because I'd just go back to using SCILAB/MatLab (and lament the money I threw away on a premature GUI).

One approach I have started to use is to write the code in the .ws format where the code is live and easy to troubleshoot and then cut the procedure, and only the procedure, into a new .ws document and change the format to Maple-1D.  Finally, that 1-D cut-and-paste can be saved as a .mpl file.  Then this can be READ back in.  The .mpl file looks like junk, but it works.  And if I need to edit the procedure, I open the .ws with all the documentation, make the changes to the code and explanations, and then save over the previous .mpl file.

What the above flow does is enable writing a nice documented procedure that is easily legible, equations, formatted text, sectioned, etc., but saved as a Maple-readable command line code set that is almost impossible to read as a human (i.e. no comments, no carriage return, etc.).

Being a GUI engineer to a fault, I'm determined to fight through this unless I find my productivity never improving beyond the CLI approach.

Thanks to all that have commented and guided me.

First, I think you are correct that I simply need garbage collection on a process that loops writing/achiving the same thing to the screen without anything changing for each write.  As far as this being a Java problem or something else, I would be way out of school commenting on this.  My daughter claims Java has an equivalent gc solution.

Secondly - and this is my perspective - the value in Maple is the symbolic capabilities coupled with all of the other features.  A one-stop-shop.  I made this observation before and I believe it was misinterpreted by Acer to mean that I thought that this was all Maple could do.  Clearly that is not true and not my intention.

Rather, there are a multitude of numerical solvers in the market.  Mathematica, Matlab, SCILAB, etc.  But the differentiator is the GUI for Maple.  All of the other products do as good or better job in numerics and CLI programming.  And SCILAB for heaven's sake is FREE and well maintained open source software.  Available on all computing platforms!!!  And in my opinion, just as good as Matlab.  They even have a Simulink equivalent product called XCOS.  Also, just as good.

I think Maple should realize that what sets them apart is solving all of the problems inside one tool.  What i did before with MathCad was to walk through the symbolics and then enter the ASCII equation in SCILAB.  This worked okay, but is obviously cumbersome walking back and forth between symbolics - to numerics - to a final report document with text/pictures.

I was happy that you were able to confirm my observations.  I'm running Parallels 7 on my MacBook Pro and using that as the VM manager.  Windows 7 (even 64-bit mode) is about a factor of 3 slower versus native MAC OS X running the code I listed previously.  Perhaps this is just Windows wrapped under Parallels that is closing the pipe.

Finally, the nature of my job is mostly graphics-entry based computing rather than command line whereas many of my colleagues are stictly command line coders.  I probably wouldn't have some of my complaints/perspectives were I to work primarily in a command line environment.  But then I might not have selected Maple as the tool of choice either.

Cheers...