My matrix package actually only provides semi-persistent (a technical term in the data structures world, I guess) matrices. A fully functional one would probably just use a double layered hasheq.
Basically, you can always read any older version of the matrix, but you only modify the latest modification. Also, reads to an old version are O(n) where n is the distance (in writes) to the latest but reads to the current version are O(1) Jay On Wed, Oct 31, 2012 at 6:38 PM, Danny Yoo <[email protected]> wrote: > On Wed, Oct 31, 2012 at 5:12 PM, Gregory Woodhouse > <[email protected]> wrote: > > I've been thinking about whether how to implement matrices in Racket. > One obvious option is just to use a vector > > > Oh! That's related to something I was coding up called "multidim" to > do multi-dimensional matrices. Here is the source: > > https://github.com/dyoo/multidim > > I was going to ask the list for a code review to figure out what a > nice API would be. No documentation yet, but does have test cases. > > > > > ;;a matrix with data stored in the vector data > > (struct matrix (rows cols [data #:mutable])) > > > If we're going to use a single mutable vector for data, then we don't > need to mark the field itself as #:mutable. We'd need it if we were > to use set-matrix-data!, but I don't think you're doing that, right? > > > > > Is there a preferred way of doing this sort of thing? > > I'm not sure. I think it really depends on the application. If we > need persistent structures, then probably something like Jay's matrix > package might be helpful: > > > http://planet.racket-lang.org/display.ss?package=matrix.plt&owner=jaymccarthy > ____________________ > Racket Users list: > http://lists.racket-lang.org/users > -- Jay McCarthy <[email protected]> Assistant Professor / Brigham Young University http://faculty.cs.byu.edu/~jay "The glory of God is Intelligence" - D&C 93
____________________ Racket Users list: http://lists.racket-lang.org/users
