Yishay Mor, >Throughout my career as a software engineer, and as an educator of such, I was >brought up on the postulate that modularity is the holy grail of design. Software >should be built in encapsulated, reusable, independent bits. > >My recent experience in using software construction as a means of exploring >mathematical ideas has led me to rethink this position. Not that I would like to >denounce modularity, but I think I need to redefine (for myself) its boundaries and >limitations.
I agree. The advantages of modularity in hardware is very well argued in "Design Rules" by Baldwin & Clark. However, one of the main benefits they found for hardware (replaceability of components) might not be applicable to software (which tends to stay put once written). The "Goldilocks Conjecture" argues that there is an optimal module size that minimizes faults in the code. However, the data is scanty and the correlations poor. I know developers who used to split code up into separate files to minimise recompilation time during development, or because they had exceeded the capacity of the editor being used. I guess with todays computers these issues are nolonger driving forces in the structuring of software. derek -- Derek M Jones tel: +44 (0) 1252 520 667 Knowledge Software Ltd mailto:[EMAIL PROTECTED] Applications Standards Conformance Testing http://www.knosof.co.uk ---------------------------------------------------------------------- PPIG Discuss List ([EMAIL PROTECTED]) Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss Announce admin: http://limitlessmail.net/mailman/listinfo/announce PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/
