Re: Presentation of Debian Med on local Next Generation Sequencing workshop (April)
On Tue, Mar 18, 2014 at 10:12 PM, Yaroslav Halchenko
<debian@onerussian.com> wrote:
> - QA, such as build time testing: cannot be stressed enough on its
> importance for anyone at least whispering about 'reproducibility'.
>
I would definitely prepare good material talking about
'reproducibility'. More below.
>> Any suggestions on something fancy to display would be helpful.
>
> Fancy... hm... I some times like to amuse "reproducibility-eager" folks
> with commands such as
>
> debootstrap --arch=i386 --include=python potato /tmp/potato http://archive.debian.org/debian
>
> which in tandem with schroot could be used to bring Debian release from
> decade(s) back with
>
> novo(potato):~
>>gt; python
> Python 1.5.2 (#0, Dec 27 2000, 13:59:38) [GCC 2.95.2 20000220 (Debian GNU/Linux)] on linux2
> Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam
>
This is interesting. Reproducibility is one reason I believe the
Galaxy Project[1] doesn't favor using a stablished package manager and
are trying to build their own[2][3](or maybe an abstract
meta-package-manager). I wonder if they could integrate something like
this to satisfy their reproducibility requirements.
I lost some of my initial excitement about contributing to Debian Med.
If I cannot use the results of my effort in the system where I
actually do my job, there is not much to gain from my time devoted to
Debian Med. Then there are options like linuxbrew(homebrew port to
linux) with a much lower initial barrier for contributions. Way
simpler to write a formula compared to a DEB.
I believe the Debian way is superior to anything out there. With
higher standards and quality assurances. But I cannot see more people
using it if projects like Galaxy or CloudBioLinux don't fully
embraced. It is quite sad they haven't.
[1]http://galaxyproject.org/
[2]http://dev.list.galaxyproject.org/tool-dependencies-xml-format-td4661410.html
[3]http://dev.list.galaxyproject.org/Better-packaging-for-toolshed-binaries-td4661447.html
[4]http://cloudbiolinux.org/
Best,
Carlos
On Tue, Mar 18, 2014 at 10:12 PM, Yaroslav Halchenko
<debian@onerussian.com> wrote:
> Hi Steffen,
>
> On Tue, 18 Mar 2014, "Steffen Möller" wrote:
>> I was invited to a local (Northern Germany) workshop on next generation
>> sequencing
>> https://sites.google.com/site/nexgenseqmv/home/workshop
>> to give a quick overview on what Debian/Ubuntu/BioLinux can do for them.
>> This is a very friendly environment and besides
>> * explaining how community-run Linux distros work
>> * that Linux not necessarily means desktop,
>> - but may mean virtual
>> - or server and that
>> - all major cloud services feature Debian images
>
> Do not be shy here: by working "in Debian" we are pretty much automagically
> benefiting from the work of more 'cloudy' people cooking up all the appliances
> etc. Thus there is a huge benefit from sharing
> expertise/responsibilities across different teams. That is why work of Debian
> Med is automagically omni-present through-out all possible deployment scenarios
> -- from cell phones (may be not by Debian itself, but by its derivative(s)), to
> servers, clusters, "mainframe" architectures which might not even be generally
> available (s390x, sparc) to regular mortals, and all possible clouds.
>
>> * what Blends are
>> * what packages are available for Next Generation Sequencing
>
>> I am very eager to learn about what the audience expects from our distro(s).
>
> I would also emphasize on unique "features" of Debian such as
>
> - clear open standards (thus packaging is done "the right way" which
> contributes to packages longevity)
>
> - licensing "clearing house": helps with wider adoption and longevity
> of software itself
>
> - QA, such as build time testing: cannot be stressed enough on its
> importance for anyone at least whispering about 'reproducibility'.
>
>> Please kindly mention whatever issue is close to your heart that should
>> be presented. I was very happy about the recent advent of the IGV.
>
>> To have something tangible, I asked Roland from Qlustar.com to
>> demonstrate over the coffee breaks how to set up a distributed work
>> environment with Debian. Tony had found and introduced their
>> technology for the Debian Med Sprint. He kindly agreed, so we will
>> jointly learn about how NGS-savvy wet-lab biologists are approaching
>> us.
>
> heh -- never heard about qlustar... interesting to hear what you learn
> from the coffee break
>
> but also do not forget that Debian itself comes with many open HPC
> related projects such as batch systems etc:
> http://blends.debian.org/science/tasks/distributedcomputing
>
>> Any suggestions on something fancy to display would be helpful.
>
> Fancy... hm... I some times like to amuse "reproducibility-eager" folks
> with commands such as
>
> debootstrap --arch=i386 --include=python potato /tmp/potato http://archive.debian.org/debian
>
> which in tandem with schroot could be used to bring Debian release from
> decade(s) back with
>
> novo(potato):~
> $> python
> Python 1.5.2 (#0, Dec 27 2000, 13:59:38) [GCC 2.95.2 20000220 (Debian GNU/Linux)] on linux2
> Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam
>
>
> if you want to impress them even a bit more -- you could use apt-cacher-ng and
> then debootstrap right there in matter of seconds (depending on your
> harddrive/cpu speed) + pre-cooked schroot and you can really get them back into
> the past ;) JUST MAKE SURE YOU HAVE ENOUGH DISK SPACE (do not ask why it is
> in capitals)
>
>> I would otherwise just dig into the data we are producing locally
>> and show something.
>
> So overall -- press not as much on flashiness (on-a-knee projects might be
> flashier) but rather on long standing standards, quality, pervasiveness, shared
> responsibility, impact.
>
> if really bored, you can even watch my talk a year back, although I think
> I have done much better ;-):
> http://sea.ucar.edu/event/open-not-enough-benefits-debian-integrated-community-driven-computing-platform
>
> My 1c ;)
>
> --
> Yaroslav O. Halchenko, Ph.D.
> http://neuro.debian.net http://www.pymvpa.org http://www.fail2ban.org
> Senior Research Associate, Psychological and Brain Sciences Dept.
> Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
> Phone: +1 (603) 646-9834 Fax: +1 (603) 646-1419
> WWW: http://www.linkedin.com/in/yarik
>
>
> --
> To UNSUBSCRIBE, email to debian-med-REQUEST@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
> Archive: [🔎] 20140319021212.GN28862@onerussian.com">https://lists.debian.org/[🔎] 20140319021212.GN28862@onerussian.com
>
Reply to: