Why ASCEND?

A personal statement

When you talk to students and faculty about computational tools, programming and so on - sooner or later they all gravitate (or seem to) towards programs that are mostly GUI's - with the hard work being done underneath by routines written in C (or even Fortran, yea, Fortran still lives for numerical computations).
I do not have a fundamental objection to GUI's - though I believe that students nowaways are clueless about the command line and are almost paralyzed when they have to deal with anything without a GUI. It is almost as if they want What you see is What you get for everything.
My transformation to a distaste for GUI for GUI's sake began when I had to typeset equations - the software I had access to were Wordperfect for DOS and soon after Microsoft Word for Windows. Wordperfect for DOS was actually not terrible and almost tolerable, Word for Windows (do not remember what version it was ... it was many years ago) was terrible. Wordperfect was still ruling the world then. Add to this the fact that a few times, Microsoft Word destroyed documents that I had spent hours working on (and yea, saved them from time to time, so I thought) and I asked myself - There must be a better way to write equations than what I had.
I then discovered TEXand LATEX- and wonder of wonders, it was freely available - entire books for beginners, all sorts of help on the internet (comp.text.tex) and so on. I started using TEXand LATEXand have not looked back since. In fact, I truly am at sea when I have to write equations using a GUI - seems convoluted to me, I know, I am an oddball. When I tell people I write using an ascii editor (yes, even vi) and then compile the document into a dvi or a pdf file and then look at it, they think I am strange and like pain. Fact is - it is logical, to me and I am comfortable - add to the fact that the documents look beautifully typeset - equations look like they are supposed to look.
One year, I started teaching chemical engineering design. The word simulator started appearing - Aspentech, Design II, ChemCAD and so on. Programs that had a GUI wrapped around routines that did all of the calculations with access to as much thermodynamic data that was out there. Students were mesmerized by these programs that seemingly could do any chemical engineering calculation by pulling a few icons from here, there and connecting them and ask the program to RUN. Except, that for the most part, students stopped thinking and accepted what these programs gave as output without questioning. It is not the fault of these programs ofcourse - that's what the teacher is there for. Yet, I wondered.
None of the commercial simulators we could afford did any dynamic simulation (they were all steady state). They were black boxes for the most part and the student interacted with icons and lines and arrows and such.
I discovered ASCEND. I downloaded it - started it - windows started opening up - I started poking around - looking at examples (they were all in ascii!!! - so yea, I could see what the heck was inside the models!) ... Believe me, I am not an expert whatsoever in ASCEND - but was certainly hooked after a few days.
I figured that the only way I will learn enough about this program was if I started writing my own example problems - and initially I took problems for which there were solutions available (so I could check my answers!). I made several attempts at teaching ASCEND in my undergraduate classes (Mass Transfer, Thermodynamics ...) - sadly, the program did not make a huge impact on many (it did convince John White in 1998 sufficiently, that he wrote the models, scripts for a rather complicated problem from Felder/Rousseau.
I have used ASCEND on and off for many years now. I still struggle with many aspects of it's powerful features - for example, using the built in thermodynamic data through the models that come with the distribution. Oftentimes I get lost as one model calls another which calls another and with the different variables and flags and so on. For a while, I kept struggling to understand the structure and while I have a better idea, I wish I knew it as well as I think I need to, to write better, compact models.
But - I said - hey, why not use it in the way that I can understand? I started writing models that were a collection of equations and methods - all together. I did not bother with lower limits, nominal values or anything like that - surprisingly, I ran into very few problems - the program seemed to merrily go along and give me the right answer!
I then started using units properly - instead of calculating by assuming that T is a factor, I started declaring T is a temperature, P is a pressure and so on. I even defined new variables following the way they are defined in atoms.a4l and so on.
To quote a part of the ASCEND HOWTO manual - once you have defined the equations, the methods - you can solve the problem inside out, right side up - or whatever. So, so true. Try writing and solving cubic equations of state for pressure, volume or whatever using MathCAD or Maple (sure, they work) and look at how easy it is with ASCEND - there is nothing that comes close to the ease with which you can solve sets of equations (linear, nonlinear) as easily as with ASCEND. PERIOD.
Now, if I could only figure out how to add differential equations to the mix and integrate them, I'd be in heaven. And I understand, it is coming. I cannot wait! Krishnan K. Chittur

File translated from TEX by TTH, version 3.70.
On 17 Jan 2006, 17:02.