That RoR isn't typical for government is much of the point. They've been spending 20x too much for one-tenth of what they should have: moving to traditional web tech is a pretty brilliant step forward.
As regards scaling, it's a valid concern but one I suspect they'll have considered. A good number of brilliant people work there (I've visited and know the crowd).
Is it not the intention to replace directgov, rather than integrate with it?
Yes. I am considering the integrations which are going to have to be moved from direct.gov.uk to their new Rails platform.
Re:The UK government's Github
by justincormack on Thursday August 30, 2012 @12:37PM
(#4455436)
direct.gov has been a monumental failure, which is why it is being canned (at vast cost due to contractual terms, I believe). If the "traditional integration paths" are broken they will need fixing too.
It doesn't really match the traditional integration paths that they're going to have to deal with across direct.gov.uk and I'm genuinely not sure it'll scale up that high if they are pushing people down that route.
We all know what happens once a year when we have to do our tax returns and that's all on a huge Java EE cluster apparently.