[LON-CAPA-users] LON-CAPA changing numbers?

Raymond Batchelor lon-capa-users@mail.lon-capa.org
Thu, 10 Oct 2002 08:39:05 -0700 (PDT)


Hi Gerd,
  Hmmmm... but I still don't understand is how this happens.  I was 
 assuming that he would have been using the sfu server....since his 
username and role are in that domain.  I'm not sure how else the msu 
server gets involved, unless there is a parallel account.  Clearly there 
is some fundamental aspect here that may revolutionize the way I think 
of and use this system... once I understand it.  

- Ray

> 
> Ray,
> 
> No, the randomization is will not look at "parallel" usernames in other
> domains. My initial guess is that the SFU and the MSU servers run
> different versions of Perl, and thus different random number generators -
> we were taken by surprise that the Perl routine was changed right under
> us. The problem has been fixed in the developer release by using our own
> hard-coded random number generator.
> 
> - Gerd.
> 
> Ray Batchelor wrote:
> 
> > I have just noticed something that might be related to this.
> > A user who is presently at msu, but currently has a student role in my
> > course in the domain sfu (boegel1@sfu), appears to be observing
> > different permutations of problems than I observe for him here at sfu.
> >
> > Is this, perhaps, because he has a parallel username in the domain msu
> > which permits him somewhow to access his role in the domain sfu but
> > with a different seed?
> >
> > Very strange.
> >
> > Cheers,
> > Ray
> >
> > On October 8, 2002 08:23 am, Guy Albertelli II wrote:
> > > Hi Mark,
> > >
> > > > I'm trying to figure out the specific problem, as it may have been
> > > > republished at some point, but I'm also looking for the information
> > > > on the random seed creation. This will clue me in as to what has
> > > > changed.
> > >
> > > symb (which means mapfilename + resourcefilename + id number of the
> > >       resource on the map)
> > >
> > > Courseid
> > >
> > > Username
> > >
> > > user's domain name
> > >
> > >
> > > (and as gerd said perl 5.005 versus perl 5.6.1)
> > >
> > > And of course varying the number and order of the random calls can
> > > affect things.
> > >
> > > (I am uncertain about 5.6.0 and the variou 5.005_00? versions
> > > specifically, but I would guess the breakage occurred in the 5.005 ->
> > > 5.6 change)
> > >
> > > > What you're suggesting is that a different machine is giving a
> > > > different random number in certain cases. I remember this being an
> > > > issue with your installation up North of the old capa. Is there any
> > > > chance of this here?
> > >
> > > I have yet to see this if they are running the smae version of Perl.
> > >
> > > > (for example, if I have access servers that are three different RH
> > > > installations, might there be three different random numbers in
> > > > limited situations?
> > >
> > > Unfortunatley yes.
> > >
> > > > I seem to remember seeing a random number check-in - did this fix
> > > > this?)
> > >
> > > Yes.
> >
> > _______________________________________________
> > LON-CAPA-users mailing list
> > LON-CAPA-users@mail.lon-capa.org
> > http://mail.lon-capa.org/mailman/listinfo/lon-capa-users
> 
> _______________________________________________
> LON-CAPA-users mailing list
> LON-CAPA-users@mail.lon-capa.org
> http://mail.lon-capa.org/mailman/listinfo/lon-capa-users
> 


-- 
Raymond J. Batchelor                               batchelo@sfu.ca