Re: Function size

On May 23, 2012, at 3:49 PM, Ashley Sheridan wrote: > > I'm of the same mind. Generally I'll split a function if I'm reusing more than a > couple of lines of code. I only split a "large" function if it's actually doing > several things, if it happens to need 200 lines to perform one 'step' then I'll leave > it as is. While I do prefer my functions to fit into a single 'screen', it rarely happens > quite like that, because I move from screen to screen with different resolutions, so there's no > constant limit for me. > > As a rough example, on a random selection of 27 functions taken from a controller on a site I > worked on I get these general statistics: > > Functions: 27 > Mean lines: 22.5 > Mode lines: 3 > Max lines: 218 > > The function with 218 lines is a large switch, and it doesn't make sense to do it any > other way, because it would actually end up less readable. I see you and I are like minds in many ways. I had one large switch block that had 255 different cases. Oddly enough I was parsing a "Tiger" data file (USGS survey data) that contained 255 different record types. Each record type required a different function to parse the data and render it's portion of the overall map. That lead to me create a linked-list that held the memory addresses of both data and function. That way simply accessing the linked list coupled data to function and drew the map. It was neat. I find it also neat, while I'm not an expert on the subject, eliminating the need for 'switch' and 'if' statements via extending classes in OO. Shiplu provided a link, which I found interesting: http://www.youtube.com/watch?v=4F72VULWFvc It showed how one can eliminate such conditionals, but at the same time it massively increased the code to preform 1 + 2 * 3. :-) And to others, I don't need comment on how I missed the point -- I didn't. Cheers, tedd _____________________ tedd@sperling.com http://sperling.com