303 assignment 2A

CSE 303, Autumn 2009
Homework 2B: anagram++

Due Tuesday, October 20, 2009, 11:30 PM
50 points total

For Part 2B, you have a choice of what to do. Also, since I am announcing this two days late, it will now be due on Tuesday, October 20, 2009 @ 11:30PM.

Choice I: for those of you who (probably) were writing scripts for the first time and feel that you "got it done" but don't feel that it's clean enough, good enough, fast enough, etc.
Choice II: for those of you who don't feel that there is much improvement left in your 2A scripts.

Both choices have the same deadline, earn the same points, etc.

Choice I:

The objective of this choice is for you to do what we as programmers should do more often -- rewrite our programs, because the process of writing them the first time teaches us better ways to write them again later on.

Rewrite -- I recommend from scratch -- your solution to 2A. Added requirements:

It must deal consistently with non-alphabetic characters under the following rules

If allgrams.sh receives anything other than precisely one parameter, where that parameter contains only alphabetic characters, it should return "Error: ill-formed input"
Well-formed parameters match precisely those entries in the dictionary that have the same alphabetic characters (in any order) ignoring case (and contain only alphabetic characters -- that is, no dictionary entry with any non-alphabetic character can ever be returned as an anagram). Although the anagram match itself is case-insensitive, the original dictionary entry must be output with its case unchanged. That means that if there are two entries in the dictionary that are identical except for case, both would be returned as anagrams of any matching parameter.

You must revise test-allgrams.sh to demonstrate each of the varying cases: normal usage, ill-formed input, etc. Your test cases must show that your program respects, precisely, the rules above.
You must include, as a text file named differences.txt, a description of the key changes you made from your original submission for 2A to your new solution for 2B. At the highest level, I want you to explicitly convey what you learned in rewriting the scripts. You should include any changes in your basic approach (algorithm), why you made different decisions, and what you think are the biggest improvements in the new version.

Note: if you had a strong solution to 2A, patching it to deal with the (relatively) minor added requirements will not likely give you enough to explain in this differences file -- so you should consider doing Choice II, in those cases.

Excessively slow performance -- roughly, over a minute to process a normal input properly -- should be eliminated. If you cannot achieve a minute, then you must explain (in differences.txt) why you think the script is slow.

Grading will be based on a combination of

the correctness of your scripts both for normal and ill-formed input
on the clarity of your script
on the efficiency of your script

credit will be deducted for programs slower than described above, although credit will be added in those cases for good descriptions of why performance is poor, and
no added credit will be given for programs faster than described above

and on the quality (not quantity) of your differences.txt description.

Choice II:

The objective of this choice is for you to push your shell script programming skills farther. Look at the site http://wordsmith.org/anagram/ -- which creates anagrams of multiple words from multiple words. For example, if you enter "computer science" you get a (large) number of anagrams including the following (for which I've reordered a few words for fun):

Comic Cup Entrees
Necrotic Scum Pee
Teen Cop Rice Scum

Your assignment is to provide, in batch form, the same basic service as this anagram server. That is, you must accept multiple words as input and find anagrams comprising multiple words in the dictionary. I have not done this, so it might be harder than I think (and I don't think it's trivial). You should use the same dictionary as in 2A (I couldn't find quickly what dictionary this anagram server uses), and you should follow the non-alphabetic character rules listed in Choice I above. (You will, of course, allow multiple parameters as input to your script, to enter multiple words.) Call your script multigrams, and provide a test-multigrams script for testing.

Grading will be based on a combination of

the correctness of your scripts for normal input (don't worry about handling ill-formed input)
on the clarity and documentation of your script
and to a much lesser degree, on the efficiency of your script -- the server itself is slow in many situations.

You are allow to add extra bells and whistles, such as those allowed in the advanced query engine on the server (http://wordsmith.org/anagram/advanced.html). However, bells and whistles that are included at the cost of basic features will hurt your grade, not help it.

Turn-in information

Both choices will be turned in via Catalyst (for which I will try to get the link right this time). Follow the same directions as in 2A for creating a hw2b.tar.gz file to submit.