As I understand it, this project tries lots of different approaches to the protein structure problem, and the complicated WU names would describe that to anyone who knew a lot more about the subject than I do.

But then I wonder, how is it htat the same app does all these different algorithms? Is one of the input files a program in a form of pseudocode? Is it all done with that horrendously long command line? (*)

How big would a change need to be in your approach that it would need a new app to implement it?

Not that it makes any practical difference whether I know or not, just curious

River~~

(*) linux users will see what I mean if they start a shell (terminal window) and type

ps ax|grep rosetta

---

As I understand it, this project tries lots of different approaches to the protein structure problem, and the complicated WU names would describe that to anyone who knew a lot more about the subject than I do.

But then I wonder, how is it htat the same app does all these different algorithms? Is one of the input files a program in a form of pseudocode? Is it all done with that horrendously long command line? (*)

How big would a change need to be in your approach that it would need a new app to implement it?

Not that it makes any practical difference whether I know or not, just curious

River~~

(*) linux users will see what I mean if they start a shell (terminal window) and type

ps ax|grep rosetta

yes--it is all done with that horrendously long command line! our philosophy has been to develop one common code base and add modules for different types of problems. for example, for a given input starting backbone structure, the -design command line finds the lowest energy sequence for the structure, while the -relax command line keeps the sequence the same but varies the backbone conformation to find the lowest energy nearby structure. the advantage of having one common code base is that all of the problems we are approaching have common features that the core algorithms in rosetta have been optimized to deal with. also, because the core algorithms are shared, an advance in one area, say structure prediction, can immediately carry over to improvements other areas such as protein design.
when we test new approaches, we implement them within the rosetta code base; when we update the app it is because we have added many new methods we want to test. I'd guess we won't be sending out a brand new app until (if/when) there is an interactive version to test.

---