[Rd] NOTE: unstated dependencies in examples

Uwe Ligges ligges at statistik.tu-dortmund.de
Fri Oct 14 21:30:34 CEST 2011

On 14.10.2011 20:53, Nick Sabbe wrote:
>> So other people from the administration tell you which software to use
>> for teaching? And what happens if they tell you only Excel is
>> available?
>> ... weird ...
> On the contrary, I should say: very common. We have hardly any control over
> what is installed on the application servers at our university (let alone
> how well it is installed).
>> You can still use the old version there, if the new version depends on
>> R
>>   >= 2.14.0. The old version will stay in the binary repositories for
>> old
>> versions of R and in the package archives of the source repository as
>> well. I think Duncan explained that already.
> Then what about new features that don't depend on parallel?

What I do not understand (both in the reasoning of this thread and in 
practise): How do you install the package with its new features if you 
cannot install a new version of R with its new features? And why is it 
so important to get the new features of the package but not to get a new 
version of R with new features and many bugfixes?

If it is possible for you to install a new version of that package in 
your labs, you are also able to install a new version of R, right?

> Maintenance hell
> is born, just because you effectively force somebody to fork (True, the R
> builders are not the ones forcing people to stay on some "old" version, yet
> this scenario is not that strange that it should be ditched without regard).
>>> A second point is that the package would not *depend* or anything on
>> R>=
>>> 2.14.0.
>> But it depends on it: it won't pass the checks for R<  2.14.0.
> Somewhat silly if it only depends on it because of the check, no?

Yes, somewhat, but actually I still do not see the reason why the new 
package should be able to work under R < 2.14.0 (see above).

Also note that the package would be accepted on CRAN as is, if you 
declared "parallel" as a Suggests, as far as I understand Jari. At least 
binaries for Windows for old R versions will be built, since I am 
checking with
on Windows.  Therefore, I believe (I haven't seen the package) this 
discussion is meaningless anyway.

Uwe Ligges

> Nick Sabbe
> --
> ping: nick.sabbe at ugent.be
> link: http://biomath.ugent.be
> wink: A1.056, Coupure Links 653, 9000 Gent
> ring: 09/264.59.36
> -- Do Not Disapprove

More information about the R-devel mailing list