• Use vi or jot to edit the .bashrc file in your home directory. Add the following two lines

  # list directories in columns alias ls 'ls -C'

Save and quit. The next time you log in the alias will take effect. If you want it in effect during your current session, type the command:

source .bashrc

in a the shell window (don't forget the leading period!). This will cause the shell to reread its configuration file.

A path is the sequence of directories through which one must move to reach a particular file. Your home directory is reached by traversing the path: /home/your loginname.

• If you want to move or copy a file to a directory other than the one that is currently your active directory, you must provide the mv or cp commands with the path along which to operate.

• For example, suppose you have in your home directory the file 'somedata' and you wish to copy it to MY home directory. You must supply the path to my home directory as part of the command:

  cp somedata /home/rfortjr

  If you have permission to write to my directory, the command will be executed.

• If you leave off the path, and issue:

  cp somedata rfortjr

  you will find a new file, called rfortjr, in your home directory, containing the same things as the file somedata.

• For directories that are subdirectories of your home directory, it is not necessary to give the full path; just the path from your $home on down is needed.

  To copy file somedata to directory DATA, which is a subdirectory of directory PCMODEL, which is a subdirectory of your home directory:

  cp somedata PCMODEL/DATA

  will work just fine. Using /home/yourname/PCMODEL/DATA doesn't hurt; it just wastes your time.

Yup, lots of ways. Here's one. Do an ls command in the directory where the files are, but redirect the output from ls to a file. Thus

ls > files.dat

Instead of the file listing appearing on the screen, you will create the file files.dat, containing the file info. Now edit it with vi or jot to suit yourself, paste the whole file into your message, and send it off.

Essentially, the answer is that I haven't bothered to tell them. Linux workstations talk to each other about the time (without human intervention), and quite literally argue about which one of them is in charge of keeping the time for all of them. Log files contain statements like, "Katahdin is candidate time master." and "Magic is slave to OldBlue". The result is that if the master machine drifts off time, all of the others follow. So far, I haven't found any way to stop this except to shut down the time daemons. I'm working on it.

Until we can afford a printer for the Modeling Center, here are the options.

• For a data file, such as pcmod.out, ftp to your account from a PC. Get the file. Load it into Programmer's File Editor or NotePad, and click on print.

• For a picture file, you also can ftp the file over to a PC. There, fire up Paint or other graphics program, read in the structure, and print.

A little discussion of file formats is in order. All of the data, input and output files, mail files, and so on that we work with under UNIX are plain text files - ASCII files, in computer acronymese.

• The only formatting information they contain is an endofline/linefeed character.

• These are the only kind of file that more, vi, jot, and our other file handling tools know about.

• A document word processed on a PC is full of control characters, embedded characters that tell your display and your printer what should be italicized, what type face to use, where the tab stops are, and on and on.

• Word processors no longer display these characters (25 years ago, they did), but when you look into the file with 'more', 'more' tries to interpret everything it sees as an ASCII character. The result is gobbledegook.

• Even plain text files look different on UNIX machines than on PCs, because PCs use carriage returns (down one line, back to left margin) at the end of each line, whereas UNIX uses line feeds (down one line).

  That's why we have PFE (Programmer's File Editor) for your PCs: it can translate and save text files in either PC or UNIX format.

Because the axial conformer is in a minimum on the potential energy surface - it's just not the deepest minimum.

• In fact, there are many more minima as well; for example, two twist boat conformers, one with a pseudoequatorial methyl and one with a pseudo axial methyl.

• A simple minimization, recall, searches conformation space while watching the first and second derivatives of the energy with respect to the Cartesians.

  As long as the change corresponds to downhill motion, the search continues. When an inflection point is found, searching ceases.

• Think of the potential energy surface as a planet of rugged hills and valleys. When we create a structure by drawing it, and then start a minimization, we are essentially dropping the molecule from an airplane high above this terrain.

  There is no guarantee that it will land in the deepest valley; much more likely is a landing in the valley most directly under the point where we drop it. We created an axial structure, so the valley most directly under it is the minimum energy axial conformer possible.

• This is why we must be careful to use many different starting points if we are trying to find the "global minimum", the deepest valley on the entire molecular planet.

The various computational models within SPARTAN all offer the choice of conducting a Single Point calculation or a Geometry Optimization.

• The former means simply calculating the energy of the molecule as it was input, without altering the geometry at all. This feature allows us to evaluate structures that may not be minima, or to generate a set of molecular orbitals for a structure minimized elsewhere.

• Choosing Optimization iniitiates a search for the minimum nearest the point of entry, just as with PCModel. In fact, PCModel offers a stationary point calculation; it just doesn't call it that. Choose MMX-E from the Minimization screen.

So far, with either PCModel or SPARTAN, we have been examining molecules in the gas phase (isolated, no solvent) and nearly vibrationally unexcited (that is, essentially at 0 Kelvin).

• When we do a calculation of the vibrational spectrum of a molecule, using SPARTAN, we will also obtain its thermodynamic properties, including the so-called zero-point energy, which we can use to correct the calculated energy for residual vibrational energy at 0 K.

• SPARTAN also offers the opportunity, as we shall see, to compute solvation energies, and explore at least in a general way, the influence of solvent on structure. Recall that PCModel allows us to set a dielectric constant other than that of a vacuum, but this option does not reliably model conformational changes with solvent.

• You might wish to explore this latter point with PCModel.
  • Try computing the gauche-anti- energy difference for, say, 1,2-dibromoethane with D set to its default value of 1.5. Then recalculate with D = 80 (water). The energy difference should change significantly.

  • Now look at the two structures calculated for each value of D. How much are they different?

In the course of an optimization, SPARTAN needs to refer many times to the same set of integrals. When 'direct' is checked, these integrals are recalculated each time they are needed.

If it is not checked, the integrals are written to a scratch file on disk the first time they are calculated, and read into memory whenever needed. Because disk accesses are very slow compared to memory accesses, 'direct' is much faster, except for very small molecules.

You can get a list of SPARTAN's 'built-in' basis sets, by selecting 'Ab initio' from the HELP menu.

• To use one that is not on the default setup list, select 'Custom' for your basis set on the Ab Initio Setup dialog box. An additional small window will appear.

• Type in the descriptor for the basis set you wish. For example, to use 3-21G* with diffuse functions, type in 3-21+G*.