my only issue with spending on infrastructure is that it tends to have a high outside lag (time between the policy's enactment and its effect on the economy). plus, the american political system already means that fiscal stimuli have higher inside lags (time taken to enact the policy) than european parliamentary system. I can show you a study/paper that shows that of the infra/education part of the feb.2009 stimulus act, only 11% entered the economy in 2009 and 36% in 2010 - more than 50% will enter the economy in 2011. I do, however, appreciate the broader arguments that infrastructure spending is useful in the future as well and will pay itself off, and it is generally good investment in capital that is needed - however, I wouldn't support it as part of a fiscal stimulus where the primary purpose is 'fine-tuning' the economy.
I'd argue for money in favour of automatic stabilisers - stuff like food stamps/unemployment benefits (during the recession). this money would enter the economy quicker and I think would be better in terms of fine tuning the economy. the same study shows that out of such spending in the feb.2009 act, 32% will enter the economy in 2009 and 41% in 2010.
edit: the paper i mention is http://www.cbo.gov/ftpdocs/102xx/doc10255/06-02-IMF.pdf