[Bio-Linux] Re: Bio-Linux future versions
Dan Swan
dswan at ceh.ac.uk
Thu Feb 19 08:28:53 EST 2004
Hi, I'd just like to offer the following comments.
> 1. It is suggested that the Bio-Linux version can continue on Redhat 9
> or Fedora 1 as in our view it does not make much difference for open
> source community whether Redhat goes commercial and support Fedora
> project. THe Redhat 9 will continue in the name of Fedora 1 (Instead of
> Redhat 10), The linux community is mature and strong enough to support
> its continuity worldwide.
The issue is not so much that of Red Hat's move away from non-enterprise
support, but the fact that the Bio-Linux is having to move away from its
traditional deployment system due to an increase of interest from
external researchers. What we are interested in is the amnout of time
it would take to remaster various distributions. I am the sole
developer on this project, and I am lucky that other members of my group
contribute time to the development of Bio-Linux and support, but if we
are to advance Bio-Linux then time constraints have to be balanced with
making a great product.
> 2. The Bio-Linux Future versions should contain OSCAR, GLOBUS, CONDOR,
> OPEN MOSIX, Sun Grid Engine(All must be included) user should have
> choice to use whatever he needs in whatever contest.
We are committed to Condor, as this is part of the UK Grid Consortium
plan - along with Globus. Whilst it may be possible to include SGE at a
later date, this is not a priority. If you are focused on cluster
development the Bio-Linux team recommend you use BioBew
(http://biobrew.org). BioBrew is cluster focused, the focus for
Bio-Linux is as a scientific workstation - allowing the 2 projects to
comfortably co-exist. Bio-Linux 3.0 comes with Condor preinstalled.
> 3. Bio-Linux should contain almost all Biotechnology, Bioinformatics,
> Medical informatics, Chem-informatics, proteomics, Bio-chemistry,
> chemistry, inlcuding visualization, modelling, graphic multi-media
> utilities and applications.All applications either they are RPM or tar.
> files can be installed in the system instead of converting all
> applications into rpm and making a installable distribution of Bio-Linux
> in CDs. As this will take still some time to take to convert all
> scientific application to RPM format. We are building up Bio-Linux in
> this direction.
Unfortunately this simply isn't possible - we do not have the experience
(for instance) to ascertain what is important in Medical Informatics.
We support genomic and proteomic research teams and Bio-Linux will have
to support these foci as a primary goal. We always accept suggestions
for other packages to be included, but at the very least we try to
provide a broad base of research tools that most bench biologists will
be able to find a use for. Bio-Linux 4.0 should have a properly package
managed set of Bioinformatics software, whether this is deb or rpm. We
will not be including .tgz packages as this is just too hard to version
control. Building the deb/rpm's is going to be a major part of the next
version of Bio-Linux and of course we will make these available to the
wider scientific community as appropriate.
> 4. All databases available opensource should be included in the Local
> server.
Bio-Linux already includes MySQL and PostgreSQL, and will continue to do so.
> 5. With regard to districution of this type of Bio-linux system can be
> made in CDs or hard-discs by cloning the entire system on the hard disc
> or on CDs. Linux has many such utilities. SYstem Imager is very much
> fine when the installation is done at a Local LAN to 100s of nodes to
> build a cluster. But with regard with the internet installation it will
> have still bandwidth problem.
We have to move away from Systemimager for external researchers - the
bandwidth constraints are too much for installs outside the UK.
> 6. We request Dan Swan to send atleast Bio-Linux 3 cloning on CDs by
> using Linux utilities for Backup or system recovery or cloning the
> hard-discs by using free linux softwares.
We have already been through this :) I know that it is technically
feasible to send you a set of bootable CD's with the HD image, but I do
NOT know how this will deal with different hardware. With the
Systemimager installation we at least have a degree of control over the
installation (eg partitioning, SCSI hardware etc). Until Bio-Linux 4.0
comes as a distribution we can only offer installs through the
Systemimager suite. I am aware this is not ideal, but it is the best I
can do for now.
I'd just like to add that it is outside my job description, and even the
remit of our funding to open Bio-Linux up outside the NERC/EG community.
I'm really pushing for this because I believe of the bioinformatics
centred Linux distributions, Bio-Linux is streets ahead of the
competition and I don't want to limit the work we do just to the people
we are contractually obliged support. This is why when Bio-Linux 4.0 is
released, we hope to have a lot more community input - and community
support and the project can take on a life of its own.
All the best,
Dan
--
Dr Dan Swan - Bio-Linux Developer | RHCE
EGTDC, CEH, Mansfield Road, Oxford, OX1 3SR
Tel: 01865 281 658 Fax: 01865 281 696
http://envgen.nox.ac.uk/ | dswan at ceh.ac.uk
More information about the Bio-linux-list
mailing list