Re: Nested R-expressions in r..eal
"Pukite, Paul R (US SSA)" <[email protected]>
| Newsgroups | gmane.comp.ai.prolog.swi |
|---|---|
| Message-ID | <[email protected]> |
Wouter said: " Btw, for those who have used the older R interface<http://www.swi-prolog.org/pldoc/package/R.html>in the past, I have found r..eal to be a significant improvement over that one (link to paper on r..eal<http://bioinformatics.nki.nl/~nicos/pbs/padl2013-real.pdf " I agree that it is much better. However, I still use the older version because it is thread-safe. If you use SWIPL in a web-served configuration, one needs to be careful in which server thread that R is called in. With the new version, R is a dynamic library and so is not re-entrant safe against being called by different threads. The old version used spawned R instances so although slower and more resource hungry this was guaranteed thread-safe as each thread had its own instance of R to communicate with. I have a simple example of using Nicos's original R..eal interface in a web server configuration here: http://entroplet.com/context_demos/navigate Click on the "Use R .." link This is slower than having R as a dynamic link ready for service requests. This doesn't address the question of nested calls but we do support the idea of using R in Prolog apps, as R is the right tool for certain numerical and graphing applications.