---
title: Eboard 31  Project introduction
number: 31
section: eboards
held: 2017-11-10
---
CSC 151.03, Class 31:  Project introduction
===========================================

_Overview_

* Preliminaries
    * Notes and news
    * Upcoming work
    * Extra credit
    * Quiz 10
    * Questions
* Quiz
* About the project
* Group composition and construction

### News / Etc.

* Please try to keep the room straight.  
* Quiz 9 returned.

### Upcoming Work

* Reading for Monday: [Association lists](../readings/association-lists)
    * Yes, it's ready!
* No writeup for class 30.
* No writeup for class 31.
* [Exam 3](../exams/exam03) 
    * Prologue due TONIGHT
    * Exam due Tuesday the 14th.
    * Cover pages due Wednesday the 15th.
    * Epilogues due Wednesday the 15th.

### Extra credit (Academic/Artistic)

### Extra credit (Peer)

* Eat at ISO food bazaar.
* Pioneer weekend walkins at 6pm today.

### Extra credit (Misc)

### Other good things

* Showvember tonight
* Drag

### Quiz 10

Notes

* Did not go well.
* So we'll go over it.
* I'm now dropping the *two* lowest quiz grades in calculating your final
  grade.

Problem 1

```
(cons
  (cons 5 10)
  (list 15
        20
        (cons 25 null)))
```

See whiteboard for drawing.

* Note: Drawing from inside out is generally a good strategy.

Result in Scheme

```
'((5 . 10) 15 20 (25))
```

Find the smallest value in a vector of real numbers.

We need to have a helper that keeps track of `pos` and `smallest-so-far`.
We are going to write a kernel.

* Named let
* Write a separate helper procedure
* `letrec`

```
(define smallest-in-vector
  (lambda (vec)
    (let kernel ([pos 1]
                 [smallest-so-far (vector-ref vec 0)])
      (if (>= pos (vector-length vec))
          smallest-so-far
          (kernel (increment pos) 
                  (min smallest-so-far
                       (vector-ref vec pos)))))))
```

How do we fix this so that it doesn't make the smallest inexact
if it's not already inexact?
* Keep track of the position rather than the value.
   But Sam is too lazy to implement that.
* Write better preconditions.
* Write tests for exact and inexact
  Sam really is lazy.
* Write inexact->exact in the base case.  No!
* Rewrite min

### Questions exam

Where should I start on problem 3?
  : 1. Make a list of all the things you need for it to be a BST.
    Two options: Empty tree.  Nonempty tree.
    i. Is it a node?  ii. Does it contain a string?  iii. Is the left
    tree a BST? (Trust the magic recursion fairy.) iv. Is the right
    tree a BST? (TTMRF); v. c; vi. d.
  : Use `and` to group things together for the nonempty tree case.

### Question in prep for next quiz

Quiz
----

Ooh!  A tree problem.  That will be helpful when I solve problem 3 on the
exam.

About the project
-----------------

* Do something interesting of your choice within constraints
    * New and different data set
    * Requires some processing/munging/whatever you want to call
    * With a novel algorithm that does something to help you understand 
      the data
* Examples
    * Literature classifier
    * Text generation
    * Something similar to cancer cell identification.
    * A set of tools for dealing with a data set. (E.g., ...)
* Goals
    * Time-boxed: 8-10 hours per person over two+ weeks 
    * Ideal group size: 3 (or 2 or 4 or 1)
* For a week from Monday: Find the data set and Write a proposal
    * Describe set
    * Describe your intended goals with the data set
        * What you think you can do
        * What would satisfice
        * Reach goal
    * Describe the algorithms you will be generating

Group composition and construction
----------------------------------