Homework I

Due Sunday, April 10

Goals of this Assignment:

• develop independent programming skills
• get serious about using a debugging tool
• learn more about the relation between word size and number range, the need to convert back and forth between numeric and character forms of numbers (e.g. 3 vs. '3'), and so on.
• review/acquire basic Linux skills

Problem I:

Write a C or C++ program named matedit, which serves as a matrix editor. With it one can create new matrices and save them to disk, and read in previously-saved ones and edit and save them to disk afterward. Here are the key requirements:

• The command line form is

        % matedit filename nrows ncols
        

  The last two arguments are optional. If they are present, then filename is taken to be the name of a new file to be created, with the specified numbers of rows and columns. If they are absent, filename is taken to be an existing file. The size of a matrix is fixed once it's created; it cannot be expanded or shrunken.

• For simplicity, we'll assume integer entries here. The matrices will be stored on disk as 4-byte signed integers. Note that this means they are NOT stored in ASCII. The number 12, for instance, we be stored on disk as 00000000000000000000000000001100, not as 0000001100000001, 32 magnetic dots on the disk rather than 16. The first two words of the file specify the numbers of rows and columns of the matrix. Storage is in row-major order. (See p.30 of our book if you are not familiar with this notion.) Initial entries are 0s.
• Your program will use the curses (or ncurses) library. There is a tutorial on it cited in our book. Note that in this context, that implies that you cannot use printf() or cout for debugging; you must use a debugging tool to query variables during your debugging sessions.
• Upon startup, the program will display the part of the matrix containing the [0,0] element. The display will consist of a rectangular grid (though no grid lines) of cells of numbers, right-justified, e.g.

            3  12   8
           88  24  -2
         

  I've used 4-space fields here, but you'll need larger ones. Typically there will be a lot of blank space in the display.

• Here are the commands (which, as with vim, emacs etc. do NOT get echoed when typed, and backspacing doesn't work):
  • mc rownum column Enter

    Move Cursor. The action will be to move the cursor to the cell at the specified row and column, at the rightmost digit of the number, e.g. the 2 in the number 12 above.

  • ec somenumbers Enter

    Enter numbers at Cursor. The action will be to enter the specified numbers (could be any number of them, but for convenience let's say there are at most 10), starting at the current cursor position, and in row-major order.

    For instance, consider the matrix displayed above, and suppose the cursor is currently at the 12, and the user types

              ec 5 13 168
              

    Then two things will happen: The cursor will move to the 24, and the matrix display will now look like

              3   5  13
            168  24  -2
           

  • sve

    Save and Exit.

• On the issue of scrolling: Often the matrix will be larger than can be displayed in full on your screen. This is why I used the word "part" in describing the initial display above. The mc and ec commands will move the cursor, and the new position may be "off the map," i.e. in a cell that is not current displayed. In such a situation, your program must update the screen so that the cell where the cursor is located is in fact displayed. Arrange it so that that cell is in the upper-left corner of the screen.
• You are NOT required to put error-checking code in your program. It's a good idea, of course, but you'll already have enough to do without it.

The TA will test your program on several examples of his choosing. But YOU must show the TA how to use the Linux od command to inspect the contents of your files.

Advice:

• Start early! This program seems fairly easy, but it uses a platform that is probably new to you (curses), which always translates to a lot of time spent, and there are a number of subtle problem lurking in there. I don't think even the best programmers among you will find this program easy.
• Make absolutely sure you KNOW the material we've been covering so fair, as it does come into play.

You may be accustomed to having programming assignments that spell things out for you, in the sense of giving you an outline of the program, telling you what modules to break the program into and so on. Most CS professors in our department, and across the nation, do NOT make assignments this way, because you would never develop the skill to "see the big picture" to design programs on your own. So, here and in all my assignments, I simply state what the program is supposed to do, and it is up to you to design your code to make it work as required. Similarly, I do not typically tell you in advance where the pitfalls are; I simply warn that they exist, as I did above in my "lurking" remark.

As stated in our course syllabus, the TA will be happy to give you hints on getting started, etc., but you have to get as far as you can on your own before seeking our help, as that is where you truly learn. Again, this does NOT mean we are discouraging you from seeking help from us; on the contrary, we are happy to help you, but only after you've really thought about it.