[gmx-users] Pull Code

[Justin MacCallum]

Hi Michel,

the pull code isn't so bad as to be unusable. There are a number of
groups, including ours, that are using it quite extensively. All I meant
is that if you are using one of the features that we don't use, it is
possible that it could be buggy as it hasn't been tested. You are
welcome to submit bugs and suggestions to me and I will do my best to
fix them as soon as I can.

>   Thanks for your answer, although it is a bit discouraging. I think 
> that code that is "in a bit of disarray" or even "likely horribly 
> broken" as you qualify it, should not be part of an official release, 
> for the mental sanity of your users. The pull code (as described in the 
> manual) was the main reason to choose GROMACS for my project...
> 
>   I have however a few additional questions :
> 
> 
> 1) In case I manage to use 'starting structure' anyway: What kind of 
> constraint is applied on the pulled group? Is it determined by the AFM 
> 'forceconstant' parameter?

The starting structure generation option applies a constant velocity
specified by translation_rate.
> 
> 2) Concerning the question (3) in my previous email quoted below: Yes, 
> the AFM pulling should be at a constant rate - the one specified by 
> 'pullrate'. But when there is a CMM removal, the position of the spring 
> is apparently not properly updated, and the pulling rate resulting in 
> the trajectory is different from 'pullrate' - and maybe not even constant.

Again, I'm not sure exactly what you're trying to do. The spring is
moved at a constant rate in absolute coordinates. Do you want it to be
moved at a constant rate relative to the reference group?

> 3) You mention that "GROMACS has been undergoing a lot of development 
> recently which makes it harder to work on a feature like the pull code". 
> Has version 3.1.4 already undergone those changes? Do they make it 
> impossible to repair the pull code?

No, they haven't, but once the changes are made it would likely require
modification of the pull code anyway. If you want, I can send you a
version of 3.1.4 that we have been using. However, we have only been
using the AFM and umbrella sampling options, and we don't have the
periodicity problems you talk about for AFM - so no guarantees.

> 4) The big question for me now is: is it worth that I try to debug or 
> re-implement this myself... Do I have a chance? I need your opinion 
> David... I have some experience in programming but I'm by far not an 
> expert, especially with big packages.

That's up to you I guess. As I said I will do my best to fix any bugs or
add any features you need, but I can't guarantee what time frame they
will be done in.

Cheers,
Justin