difference between repos with and without a ":"

I have only 16 enabled at work, so I am safe :wink:

You are right. Sorry! I am mixing with the other thread I started in another sub-forum, which was more general about the differences between directories with or without “;” in general. I took (a bit randomly) this example here to try to be less abstract.
As for mozilla, I am in fact using mozilla (without “:”) at work and none of the 2 at home.

Anyway, thanks a lot for your replies.

I confess I do not understand why you would need the package manager to tell you when there is a new version out. IMHO its better to just visit the appropriate web site of the scientic package regularly, … most sites typically make an announcement, and one can then check their change history, to see if an update is necessary. How many scientific pages are we talking about?

Still, questioning “the need” of someone else is conterproductive, so instead, let me say I think there are various ways to do this, … a lot depends on the degree of automation you insist on, in your efforts to make it easier.

There was a time when I would use two different software packager managers, but use them differently. During this time, I

  • configured Smart Package manager with many repositories, and Smart would tell me when there were updates available. But then I did not update from smart, instead I took the information on a specific package from Smart and
  • with zypper pre-configured zypper with only OSS, Non-OSS, Update, and Packman, when I saw an update I wanted (as listed as an update in Smart) I would then add the single repos (and only that single repos) into zypper (which is incredibly easy to do with "zypper ar <UR> <arbitrary name of repos>
    ", “*zypper install < application > *” ; “zypper rr <arbitrary name of repos>” and I kept my zypper clean, and no conflicts.
    As an aside, contrary to your experience, I have had to help dozens of users who have messed up their Linux because they had too many respositories, and did not appreciate the associated risks.

But I no longer bother with that dual smart/zpper approach. The reason being after a while, I realized almost all of those minor updates were not important. And the major updates of packages were typically well advertised on various web sites, that I did not need the package manager to tell me of major updates.

It is true, I am not using so many of them, but I do not want to visit their web pages every morning.
However, you make me recall that I think I saw once a firefox extension pledging to signal every change on a list of web pages that you would give it. Maybe that would be a much easier way for me.
However, there is still the issue of packaging for opensuse. These sites will announce their new version but not necessarily packaged for suse, so that the repo saves me the time to find again the options to give at the recompilation or installation.

You are perfectly right, but in my case a minor update can mean a correction of a bug in an algorithm that I use and that would render my results wrong, possibly without my noticing it. That is the problem of big “black box” simulation codes or libraries. However, I must admit that I am probably very paranoiac here since I do not use so much external libraries and most of them are old and tested, but I already spent so much time, several times, compiling and installing the chain gcc/make/gfortran/ATLAS/lapack/BLACS/scalapack/openmpi/ibverbs/eclipse/cdt_plugin/photran_plugin/ptp_plugin, not speaking of sge, fftw, petsc or mumps that I would really appreciate some automatism (Sorry for all these names that you may not know).

But maybe you are right and I am looking in the wrong direction for this automation.