Regarding the ‘get it done quickly’ requirement versus the ‘last forever’ requirement, I think there is a need for both. Perhaps another dimension is robustness – i.e. ‘good enough to meet the immediate need’, versus ‘bullet proof because we don’t know how this thing will be hammered in the future and from what angle’. I guess you could also throw in the ‘highly optimised because throughput and scalability is critical’ versus ‘quick enough to deal with the business requirement at hand’.

I guess the point I am trying to make is that there will always be a need for ‘hand crafting’ or ‘precision engineering’, and the trick is to determine when ‘convention’ or ‘configuration’ is appropriate. This is why I am such a big advocate of SOA, because you can more effectively isolate all that hand crafted highly configured detailed stuff, which in my experience is necessary only for a small percentage of the average business system, from the higher level application logic and control, which the majority of developers are (or should be) concerned with.

Back to your central theme, though, the thing I like about conversations with IBM is that it ‘gets’ the bigger picture. To my mind, it is in a fairly unique position to meet the needs of hard-core systems programming requirements (e.g. the guts of front-office systems in the City) and the much broader business application development requirement that depends on developers not being distracted by technical detail that has bugger-all impact on fitness for purpose of the final solution.

There is an interesting contrast here between IBM, coming from the serious ‘configuration’ oriented heritage, and Microsoft, coming from the ‘just do it’ (JDI) heritage, that originated with the push with VB many years ago. I was selling highly structured and high value development tools when VB came onto the scene (circa 1996?), and this totally changed the market (from a vendor perspective), and the nature of the development community (from an end user perspective) – almost democratising software development in a way that we had only seem glimpses of previously.

This is important history that many forget, and the debate around JDI versus ‘serious’ application development was one were having with Steve Mills (head of IBM SW Group, for those who don’t know him) a few weeks ago. The message from the analysts around the table during that meeting was not to dismiss the requirement for JDI, an area in which Microsoft (rightly or wrongly) has traditionally dominated for a long time now (that’s the relevance of the history bit).

One of the things we discussed during this same session was that JDI, agile development, or whatever else you want to call it, has tended to go hand in hand with the ‘good enough’ mentality in the past.

Against that background, I would be interested in whether you see this mindset changing, i.e. whether the majority in development community are becoming more discerning about the quality/nature of the frameworks they work with. It is not so much a question of IBM versus Microsoft, but more one of whether the army of JDI oriented developers out there will start to care about the principles underpinning the stuff they are working with.

I think the stackless stack concept enables business and IT agility; agility is really the key benefit.

Ian