Thursday, August 28, 2008

Technical Investment

A recent post on Technical debt on InfoQ got me thinking about technical investment.
In the age of YAGNI, ominous PM's and tight budgets, often time is not spent on investing technically in the team or the business. We build stuff as we need it and if we can reuse something we try to, but often end up hacking it to fit our new needs, if the shoe does not quite fit.
At my current place of work we are "post live" and in a situation where investing technically makes great business sense. Our project is some what of a flag ship project as the company is
A) New to .Net, coming from a Delphi background
B) New to NTier distributed development, coming from a rich client with db connection
C) Looking to move to an SOA architecture across the company to more cleanly delineate the business logic get some reuse out of the business processes.

Because of these factors there have been mistakes that have been made. The application we have built works but it is high in technical debt. It is just holding itself together, push it just a little and it will come crumbling down.
A lot of this is to do with hurried framework design and best practices ignorance. We have acknowledged our mistakes, but that is not enough. Over the next few months this application stack is to be used as the basis for other NTier development in the new and growing .Net department, we need to make sure the other teams don't copy our mistakes. We need to make it easy for them to succeed!

IMO we need to start with re factoring our core libraries, specifically Data Access, Logging and Service Boundaries. These are areas that will be needed for all of the applications in the future so why not get this stuff right.
Are we using WCF as intended?
Are we getting the most out of NHibernate?
Can we reduce the noise from our non business aspects that have leaked in to our code?

Is this Gold plating? Well, yes i guess it is and i would not recommend it normally. However given the circumstance i think the cost of not getting this rights is too great.

So how would I do it if i was in the same position tomorrow as the guys were at the start of this project? Maybe that's the next post...

No comments: