This is easily doable with a teachpack: 1. how to create a teachpack http://docs.racket-lang.org/htdp/index.html?q=teachpack
2. the key is to document the (a) structure type definition and (b) the data definition you export. Example: #lang racket (provide ;; A DR is (make-dating-record String String Number Boolean String) ;; interpretation: (make-dating-record f l dob b c) denotes a person with ;; first name f, last name l, date of birth dob, gender b (true for female, false for male) ;; and a comment string c ; DR -> String ; extract first name from dr dating-record-first-name ; DR -> String ; extract last name from dr dating-record-last-name ; DR -> Number ; extract dob from dr dating-record-dob ; DR -> Boolean ; extract sex from dr dating-record-sex ; DR -> String ; extract comment from dr dating-record-preference ;; [List-of DR] the-dating-db ) ;; -------------------------------------------------------------------- ;; IMPLEMENTATION ... On Nov 17, 2013, at 2:24 PM, john_burnette wrote: > I hope this in the spirit of the question. My students are working with > structures, a list of several hundred randomly created members of a dating > database. They are happily working to test whether members of the database > not only share common interests, but are also satisfying what the member is > looking for in a dating match. > > It's a nice review exercise because they're doing a lot of list processing, > automatically generating letters to members telling them the good news that > they have matches, well lots of fun stuff like that. > > Currently I have them just cut and paste in the database file I give them, > but I do thing it would be more realistic if they accessed it with a > "require". It's working ok without doing this, but I'm thinking of taking > that next step - offloading not just the database but the structure > definitions as well. > > So having said that I guess I'm obligated to give a reason why. I think doing > so really drives home the point about data driven design. Far too many of my > students have taken it upon themselves to rewrite the structures and of > course as long as they were generating their own examples of "clients" > everything worked fine. When they are given the larger database to work with > (using the data structures they should have been using) this fall apart. > > Now, yes, I know, if they were carefully following the design recipe is a non > issue, and perhaps this serves as a good lesson. Still, putting them in a > place which forces them to use the data structures as written would probably > be a good thing. > > > > Sent from my Verizon Wireless 4G LTE Smartphone > > > > -------- Original message -------- > From: Norman Ramsey <[email protected]> > Date: 11/17/2013 1:27 PM (GMT-05:00) > To: [email protected] > Cc: Racket mailing list <[email protected]> > Subject: [plt-edu-talk] Using separate files as modules with HtDP > > > [Shifting to plt-edu-talk] > > > > At Sat, 16 Nov 2013 12:44:03 -0500, Norman Ramsey wrote: > > > > >> On their next assignment, my students will reuse code they have built > > >> for binary search trees. I would prefer that they place the old code in > > >> a different source file than the new code. I tried doing this using > > >> "require", but I cannot figure out how to get "require" to load a file > > >> that is written in Intermediate Student Language. Is there another > > >> mechanism I should try? Is the thing I want to do even possible? > > Matthias Felleisen replied: > > > Part of the HtDP philosophy is that I want students to ask for > > concepts. For example, once you have written the same list template > > over and over again, you want them to say "isn't there a way to > > avoid writing this same schema repeatedly? and you're ready with "I > > am glad you asked, here are loops defined in ISL and they are what > > you want." In the same spirit, I want them to ask for modules and > > components. > > As teachers, how do we put students in situations where they are > stimulated to ask for modules and components? Your list example is > apropos, and I have certainly written this data definition many times > this semester. But my students have not once written 'require', nor > have they ever seen *me* write 'require'. Over the course of the > term, there few functions that they see two or more times (search in a > binary tree might be one), and no functions that they see defined over > and over to the point of tedium. > > To put some numbers on it, students might get 40 hours of lecture in > which to become bored and irritated by seeing the same data definition > over and over. But they might get only 11 programming assignments, > and if repetition is the technique we use to get them to ask for > things, that repetition is going to crowd out something else? > > So again, what situations do you put students in that motivate them to > ask for components and modules? > > > > My personal preference is to (1) not have code from one assignment > > be critical for the next one ... [this] is important for weaker > > students. > > I see merits both ways, and I tend to do some of each. In real life, > code from last week can be critical for next week. And I want to > militate against an experience that is too common among university > students: code is built, used, then thrown away and never examined > again. > > But by having code from one assignment be critical to the next, I now > place myself in a difficult position: > > - I want them to reuse their earlier code. > > - I have told them that programming by "clone and modify" is not > acceptable. And I have asked N times in lab, "what part of the > design recipe says to copy the solution to a previous problem and > start editing it?" > > - I have given them no mechanism by which they can reuse a previous > solution. > > For this term, I think I can't work my way out of this one: I can use > Matthew's workaround to get `provide`, but I have no infrastructure > that would enable them to submit a solution spread across multiple > source files... > > I'd like to hear other teachers' thoughts on using modules in the > first course. It is considered an important part of our current first > course, so if I want to get my colleagues to use the HtDP > infrastructure, I will have to have a better story about it. > > Are you using modules? How do you motivate them? If you allow your > students to spread a program across multiple modules, what > infrastructure do you use to manage submissions? > > > Norman > ________________________ > PLT Educators talk list: > http://lists.racket-lang.org/plt-edu-talk > ________________________ > PLT Educators talk list: > http://lists.racket-lang.org/plt-edu-talk
____________________ Racket Users list: http://lists.racket-lang.org/users
