I have not been blogging much from here have I, it is not that the sessions are not that interesting, but no single item has been giving me an huge urge to write.
As I said in my last post I think this is a conference of best practice ideas and as such you tend to pick up a useful nugget here and there which you store away for future use. This is particularity relevant as at present I am reviewing our engineering process to improve our software development life cycle.
Like many companies we use a variety of tools beyond Visual Studio such as CruiseControl, nUnit and our own home grown work tracking system. I have to consider when it is advantageous to swap these for the new features in Visual Studio Team System 2008. Being pragmatic this is always going to be a slow migration, these is little point investing time in moving an old project that barely still under maintenance to a new system. In fact I have chosen to only move our active projects from our old SourceSafe based system to TFS at a major release point, e.g. V1 to V2, snap-shoting it at this point and not bothering to bring over all the change history.
The tools round the edge is another question. If you, as we do, have an investment in nUnit and CruiseControl for projects is there any good case to rework everything to MSTEST and TFS Build? In the long term I think the answer is yes, to get a unified end to end solution, but it is hard to justify the time to do an ‘instant’ swap over, so again it will be slow move. Especially when you can use a combined system e.g. have old testing nUnit and new ones in MSTEST pulling it all together with CruiseControl which can happily access the TFS SCC, build using MSBUILD and run all the testing frameworks.
Anyway Roy Osherove is tuning up his guitar for a session on testing to time to go….