We set out on the project intending to create something 'generic' that could be used in a variety of contexts. However we've uncovered a number of questions, some of which are technical, relating to the construction of the original code and architecture.
In discussion at the end of last year, our developer Mike suggested the system could be made ‘generic’ to a lesser or greater extent in one of two ways:
(1) ‘Shallow’ – well coded, open source codebase designed according to best practice (get rid of all hard-coded stuff, elimination of duplicate code, object orientated design, n-tier architecture, stored procedures instead of embedded SQL etc)
The risk-benefit equation is different for each option: it’s something that we need to discuss and come to a collective decision on. In essence, (1) has less risk but (2) offers much wider benefits and it would certainly be much easier to hit the target of making the system “into something that can offer a 90% best fit for generic placement activity” .
Given the time and budget constraints of the project, in the end we opted for (1) while recognising the desirability of (2).
Posted by CIePD Team on 05 February 2013, 1:35 PM
| Comments (6)