Hi,

> I just learned about Docker and would like to share this info with you:
> https://www.docker.io/
> 
> The idea is that you have a very basic ubuntu installation just like
> in a virtual machine, but it interfaces to the pre-installed system
> so that image files become very small - that's what I
> understood. This could be a very nice container for
> reproducibility. What do you think?

I think it is nice because it is not redundant, but this implies that
the host system has to remain stable if we want to always benefit from
the small size of filesysten of this virtual machine. So for me this
defeats goals of long-term conservation. It seems more adapted to
people who want to have and use many lightweight vm environments,
with different settings.

Moreover, Docker.io or not, I think the vm strategy only solves the
problem of system dependencies at a given instant. I still see no
argument or study ensuring that a virtual machine created now will be
usable 5 years later. And I know some virtual machine images and
configurations from 5 years ago that can not be used today without a
manual reconfiguration.

-- 
Nicolas LIMARE - CMLA - ENS Cachan        http://limare.perso.math.cnrs.fr/
IPOL journal                                            http://www.ipol.im/
-> image processing, reproducible research, open science

Attachment: signature.asc
Description: Digital signature

--
IPOL - Image Processing On Line   - http://ipol.im/

contact     [email protected]          - http://www.ipol.im/meta/contact/
news+feeds  twitter @IPOL_journal - http://www.ipol.im/meta/feeds/
announces   [email protected] - http://tools.ipol.im/mm/announce/
discussions [email protected]  - http://tools.ipol.im/mm/discuss/

Reply via email to