Re: [Rd] optional package dependency (enhances)

From: Allen S. Rout <asr_at_ufl.edu>
Date: Wed, 20 Jan 2010 15:07:28 -0500

tl;dr: Repackaging in native formats dramatically improves user access to packages.

>> On Wed, 20 Jan 2010 20:23:51 +0100, Uwe Ligges <ligges_at_statistik.tu-dortmund.de> said:

> 1. You probably mean "Depends" rather than "Requires".
> 2. You forgot "LinkingTo"

Heh, thanks! I'll add them.

> All binary packages on CRAN are checked, we do not need a signature.

Since you don't generate e.g. RPMs, every packaging system is forced to re-do this work. I was talking about that signature, the one on the RPM.

> Recursive reverse dependencies are also checked after updates. Hence
> you do not need to worry at all. Be sure we did most of the work
> already.

Some of us care a lot about automatically understanding the installed packages on a system. This lets us do things like cleanly and consistently produce a new one "like that". This goal is very important to many sysadmin types.

It is intractible to ask, e.g. R, Perl, Python, Ruby, Mozilla, etc... essentially every application that came up with a plugin or module system, "Hey, what do -you- think is installed". It is also extra overhead to expect each system administrator to go learn CPAN, jpp, ppi, gems, CRAN, etc... it adds myriad extra steps to the analysis and configuration of machines.

On the other hand, many R developers appear to live in a world of machines which are unique flowers, and they expect that other machines with which they interact will be nurtured in the same individual way. This is not "bad", it is just "different"; but the architectural decisions which are completely reasonable in a world of individual gardens place restrictions on those who would like to plant 100 or 1000 machines, all identical.

If we on the borders connecting R to [distribution in question] can easily repackage R software, then the utility of R to the world will increase. I think this would be a good thing.

I understand that it's not a big priority for that fraction of R-devel who lives amongst their gardens, and I'm not offhand expecting some mass conversion here. :) But I wanted to make the case.


R-devel_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel Received on Wed 20 Jan 2010 - 20:11:28 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Thu 21 Jan 2010 - 04:50:17 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-devel. Please read the posting guide before posting to the list.

list of date sections of archive