The integrations in this expression have until recently been the source

[EngLearner]

Source: "Markov Chain Monte Carlo in Practice" by W.R. Gilks, S. Richardson and D.J. Spiegelhalter.

They give an integral in their book (which they denote E), and then this text follows:

The integrations in this expression have until recently been the source of most of the practical difficulties in Bayesian inference, especially in high dimensions. In most applications, analytic evaluation of E is impossible. Alternative approaches include numerical evaluation, which is difficult and inaccurate in greater than about 20 dimensions; analytic approximation such as the Laplace approximation (Kass et al., 1988), which is sometimes appropriate; and Monte Carlo integration, including MCMC.

MCMC is what they recommend for evaluating E, which means to me that until recently it was difficult or impossible to evaluate E, but now that MCMC is available, the integral E can be relatively easily evaluated. If that is the case, then why is the present perfect used and not the simple past? I would've written:

The integrations in this expression were until recently the source of most of the practical difficulties in Bayesian inference, especially in high dimensions...

Is the simple past "were" correct at all in this case? If so, does it mean something different from what the present perfect "have been" does?

 

[jutfrank]

Is the simple past "were" correct at all in this case?

Yes. It's not 'incorrect'.

If so, does it mean something different from what the present perfect "have been" does?

Yes. For me, if you use a different aspect, you get a different meaning.

The present tense here is used because the speaker is thinking of the recent past as the present moment. That is to say that the concept of 'now' encompasses the recent past.

 

[EngLearner]

The integrations in this expression have until recently been the source of most of the practical difficulties in Bayesian inference, especially in high dimensions.

The integrations in this expression were until recently the source of most of the practical difficulties in Bayesian inference, especially in high dimensions.

Do both versions imply that now the integrations in this expression are no longer the source of most of the practical difficulties in Bayesian inference?

 

[jutfrank]

Yes. In fact, the implication is very heavy due to the phrase 'until recently'.