# posterior predictive distribution of new observations in Bayesian linear regression

I'm confused on how to sample from the posterior predictive distribution of $$\tilde{y}$$. In class, we've seen that in Bayesian linear regression, the posterior predictive distribution of $$\tilde{y}$$ is a scaled t-distribution. However, when I look on the internet for some examples of code (R2Openbugs, JAGS...) on how to sample from the PDD of $$\tilde{y}$$ they hardly use the scaled t-distribution but they often use the normal distribution:

pred <- alpha + beta * new.x
Y.new ~ dnorm(pred, tau)


What is the correct way to get a sample from the PDD of $$\tilde{y}$$? I can't seem to find any code that shows how to use the scaled t-distribution.