On 17 Sep 2010, at 11:09, Stuart Buchanan wrote: > The project was also a lot larger than it looked at first glance, > particularly the amount of time required to create a good regression suite. >> From memory, I think we spent significantly more time creating the regression > suite than we did actually coding.
Wearing my day-job hat, I can only echo everything Stuart said - unless the old code is very bad, or you already have good regression coverage for it, you will expend considerable effort for little outward (revenue-generating) gain. On the other hard, I have done 're-architectures' where an old design (in C/C++) was totoally replaced with a new one (in pure C++) - that was mostly successful, but took a long time (multiple years), and with understanding that the old design could not possibly deliver on a wide range of areas that would be important for the business. All the comments about regressions still apply - and in the end, some of those areas weren't as business critical as they appeared - but the new code still had to deliver 100% parity with the old. 80:20 rule applies - you can easily get to 80% parity (in features) with the old code - that last 20% .... that'll kill you. There's a reason the banks emulate VMS running COBOL rather than re-writing in C++ :) James ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
