Debian Med sprint meeting in Southport
After the success of the last Debian Med sprint meeting in Lübeck we were granted the opportunity to organise a similar event in early 2012. This page is here to assist in planning the event and recording outcomes of the meeting.
The main organiser was Tim Booth (tbooth_ceh.ac.uk), the lead NEBC Bio-Linux developer.
Sponsorship
We would like to thank SciEngines for providing the FPGA tutorial on the Friday |
We would like to thank Eucalyptus for sponsoring dinner on Saturday night |
Report on outcomes
The meeting was always intended to follow a loose agenda, taking some initial talks and topics as a starting point for the day then moving to working in small groups. At the end of each day attendees were asked to report their progress to the group. As the meeting organiser I (Tim) have digested and summarised these reports to produce a list of highlights. Also see the original agenda and Andreas' personal highlights below.
Talks and group sessions
The SciEngines tutorial and Friday evening ice-breaking talks went ahead as planned - see the original schedule.
Tim and Soon's talk slides regarding Bio-Linux intro and Galaxy packaging
Laszlo Kajan's talk slides regarding PredictProtein packaging
[please add your slides here]
On Saturday we started with a talk from Andreas on the history and structure of the Debian Med project and Debian in general. This was followed by a Q&A session.
- On Sunday Brian started the day with a presentation on Eucalyptus, their software architecture and company strategy.
Topics and outputs of group working
- Networking, increasing participation
- All attendees agreed this was an excellent environment for networking, asking questions, and discussing projects and ideas
- GPG keys were signed for several new developers - this was co-ordinated by Steffen (Ivo, Sascha S, Martin, others)
- Discussion of systems biology requirements and future planning. Finding sys bio workflows in myExperiment (Steffen, Laszo, Alan, Ivo)
- Created account on Launchpad, joined Deb Med PPA (Sascha S, others)
- Discussion of cloud computing following Eucalyptus talk (Brian, others)
- Software evaluation
- Clinical relevance - evaluating tools for molecular mimicry, emergency annotation of emerging pathogens (Steffen, Matus, ...)
- Taverna exposition and evaluation (Alan, others)
- Looked at Steffens GSOC project (by Dominique Belhachemi) - relates to making EMIs (Eucalyptus images) of Deb Med (Tony)
Looked at turning PredictProtein into a web service. Decided on how to proceed. (Laszlo, Matus)
- Looked at packaging RAILS app and Ruby gems in general - discovered impracticalities. Requires further investigation and will ask on Ruby list. (Stuart)
- Looked at Qiime unit tests - how they should be applied to the existing package (Soon, Tony)
Discussion of getData + BioMaj and possible steps towards compatibility; uses of these tools for Galaxy packaging; publication??; integration with existing packages. Considered the possibility of getting database info from NAR - towards a "semantic" NAR database catalogue. (Steffen, Olivier, Tim, Alan, Laszlo)
- Looked at packaging OMERO for open microscopy. Contains over 800 binary .jar files so this is a major undertaking. Put a message on OMERO community support board asking for a source tarball for each release. (Tony, Andreas, Tim)
- Investigated Maven debian helpers - determined that they need work (Stuart, Tim)
- Looked at issues with packaging API key to Bioportal - decided to query this with site maintainers (Stuart)
- Looked at issues with packaging Seek - maybe a scripted install is appropriate (Stuart)
- New packages, package uploads
- Upload of Phy-Spread package (Andreas)
- Dependency packaging for BEAST - which is a mess (Andreas, Olivier)
- Work on Beast-mcmc but this is a mess (Andreas)
- CHADO (from GMOD) packaging and sponsoring into Debian, including fixes to satisfy Debian policy (Andreas, Olivier)
- PyHMM packaged (Piero, Tim)
- PSCoils packaging and improvement of code to fix Debian compatibility - demonstrating how useful it is to involve upstream authros directly in packaging (Piero)
- prdcchmm packaging - depended on PyHMM (Piero, Andreas)
?TreeVector initial packaging (Brian)
- Initial Rightfield packaging - Java app with awkward dependencies but good progress was made and a working package now exists (Alan, Stuart, Quyen, Tim)
- Packaging of Copasi (Ivo)
- libsbml update in progress (Ivo)
- Initial packaging of Genometools - mainly modifying Makefiles to make things less self contained, build dynamic libs, etc. (Sascha S, Steffen)
- Packaging CSnake which is the build system for Gimias. Almost finished. Patches ready to push to upstream (Martin, Tim)
Started PredictProtein packaging implementation (Laszlo, Matus)
- Ensembl packaging work on libwww-perl (Andreas)
- sofa-framework update (Andreas)
- uploaded phy-spread package (Andreas)
- MMTK packaging work (Olivier)
- CLapack packaging work (Olivier)
- Mentoring, tutorials
- Demonstration of packaging from scratch (PyHMM) on overhead projector (Tim, to Piero + many others)
- Mentoring of Mom (mentoring of the month) student (Andreas)
- Introduction to SVN usage for new contributors (Steffen to Brian)
Demonstration of BioMaj software (Olivier to Laszlo etc.)
- Discussion of practical ways to expose software as Taverna service (Alan to Laszlo etc.)
- Demonstration of Java packaging and use of the auto-patch feature of dpkg-source (Brian)
- Sascha S received advice on how to split up a large package (genometools) into six components. Did some modifications to upstream (as the owner) to use standard file locations. Looked into licensing - issues are ongoing but generally looks good
- New prospective packagers ran through on-line tutorial to get initial hands-on experience (Quyen etc.)
- Advocacy, licensing issues
General discussion of DFSG and what it means for companies like SciEngines (Sascha M, Daniel, etc.)
- Review of "nearly free" software (eg. Phylip) leading to discussion on DFSG and legal issues relating specifically to UW and FHCRC software releases - should we have an on-line petition? (all attendees)
- Petition started to reform Phylip license (Andreas)
- Looking at BLIMPS licensing. Soon needs sift which depends on BLIMPS (abandoned). Both have nearly-free license. (Laszlo, Soon)
- CHADO promotion planning now that the package is ready (Steffen, Olivier)
Location/Date/Capacity
The sprint will run from the 27th-29th January 2012 (Friday to Sunday). The main part of event will begin on Saturday morning. Friday is reserved for a tutorial on FPGA computing.
The venue will be the Dukes Folly Hotel Southport, UK. This is an independent hotel that will provide a similar atmosphere to the one we enjoyed in Lübeck-Travemünde, and we will have the place to ourselves for the duration. The hotel has only 17 bedrooms, but to accommodate up to 30 of us working in the daytime will not be a problem. For overspill, other hotels are nearby - see below. For after hours, our hotel has a fully licensed bar and Southport is also well supplied with places to eat and drink. Unfortunately, the famous funfair will not be open at this time of year.
Signing up/Cost
Update: On 14th December 2011 Tim contacted all attendees to confirm dates and room requirements. If you replied to this mail and are showing as "confirmed" in the table below then you have a room reserved in the hotel. If not, you need to contact Tim and say what you need.
Please add your name and other details to the table below. You will be contacted nearer the time to confirm details, so please check you have given a valid e-mail address and arrival/leaving dates. This table will be passed directly to the hotel to confirm reservations. If you have any special requirements you should contact the hotel directly on +44 1704 533355 - they are very friendly!
There is a flat rate per person of £35 per night for bed and breakfast. On the Saturday and Sunday the hotel will provide refreshments plus a full buffet lunch for £10 per person. So if you are staying for three nights the total cost to pay will be £125. For dinner each night and for lunch on the Friday/Monday we will make our own arrangements.
Getting to Southport
The Hotel is not far from the main rail station in Southport - less than 1km. (Click the thumbnail map for a larger version.)
There are three international airports nearby:
Manchester Airport - trains run hourly direct from Manchester Airport Station to Southport, taking around 90 minutes
Liverpool John Lennon Airport - short bus or taxi ride into Liverpool, then trains run half-hourly to Southport
Blackpool Airport - is technically the closest, but there is no direct transport link to Southport, only a bus service via Preston.
For all public transport journeys in the UK, you should check http://www.transportdirect.info.
Overspill accommodation
If we end up filling the hotel, people may want to consider:
Talbot Hotel - Just around the corner, cheap. No on-line booking available.
Rosedale Hotel - Small hotel opposite Talbot. No on-line booking available.
Brittania Prince of Wales Hotel - Large hotel nearer to station. Headline prices are high but promotional offers may reduce the cost to under ~£43 per night. See website for bookings.
Attendees
No |
Name |
Organisation (Interests) |
Living in |
Arrival at |
Departure at |
Booked room |
Travelling details |
|
1 |
Steffen Möller |
<moeller> |
U Lübeck (Debian) |
Lübeck (D) |
Thu 26.1.2012 |
Mon 30.1.2012 |
confirmed 14/12 |
|
2 |
Andreas Tille |
<tille> |
Debian |
Wernigerode (D) |
Thu 26.1.2012 |
Mon 30.1.2012 |
confirmed 14/12 |
BE869 Southampton -> Manchester 16:30 |
3 |
Tim Booth |
tbooth_ceh.ac.uk |
NERC (Bio-Linux) |
Wallingford (UK) |
Late Thursday |
Monday |
confirmed |
Car or Rail from Oxford |
4 |
Alan Williams |
alanrw_cs.man.ac.uk |
U Manchester (Taverna) |
Manchester (UK) |
Friday |
Monday |
Confirmed 14/12 |
From Manchester with Stuart+Quyen |
5 |
Tony Travis |
ajt_minke.ukfsn.org |
Unemployed (Bio-Linux) |
Banchory (UK) |
Fri |
Monday |
confirmed 15/12 |
|
6 |
Olivier Sallou |
pre.sur_irisa.fr |
U Rennes 1 |
Rennes (FR) |
Friday |
Monday |
confirmed 18/12 |
|
7 |
Piero Fariselli |
pre.sur_unibo.it |
U Bologna (1D->2.5D) |
Bologna (IT) |
Thur |
Mon |
confirmed 11/1/12 |
|
8 |
Ivo Maintz |
pre.sur_biologie.hu-berlin.de |
HU (Sys Biol) |
Berlin (D) |
Thur |
Sun |
confirmed 07/11 |
|
9 |
Laszlo Kajan |
lsur_rostlab.org |
Rostlab (1D->2.5D) |
Munich (D) |
Thu 26.01.12 |
Mon 30.01.12 |
confirmed 19/12 |
|
10 |
Sascha Steinbiss |
sur_zbh.uni-hamburg.de |
U Hamburg |
Hamburg (D) |
Friday |
Monday |
confirmed 14/12 |
|
11 |
Stuart Owen |
|
U Manchester (Seek) |
Manchester (UK) |
Friday Eve |
Monday |
confirmed 14/12 |
From Manchester with Alan+Quyen |
12 |
Brian Thomason |
pre.sur_eucalyptus.com |
Eucalyptus (cloud) |
Santa Barbara (US) |
Friday |
Monday |
confirmed 14/12 |
|
13 |
H. Soon Gweon |
hyugwe_ceh |
Bio-Linux |
Wallingford (UK) |
Late Thursday |
Late Saturday |
confirmed 20/12 |
Travelling with Tim from Oxford |
14 |
Quyen Nguyen |
pre.sur_h-its.org |
HITS (Seek) |
Heidelberg (D) |
Friday |
Monday |
confirmed 14/12 |
From Manchester with Alan+Stuart |
15 |
Martin Steghöfer |
msteghofer_cistib.upf.edu |
CISTIB/Gimias (workflows) |
Barcelona (E) |
Fri 27.01.12 |
Sun 29.01.12 |
confirmed 18/1/12 |
|
16 |
Matus Kalas |
pre.sur_bccs.uib.no |
U Bergen (BioXSD, EDAM) |
Bergen (N) |
Thu |
Sun morning |
confirmed 11/1/12 |
|
17 |
Sascha Moeller |
presur_sciengines.com |
?SciEngines |
Kiel (D) |
Thu |
Sun |
confirmed 20/1 |
Staying at Carlton House |
18 |
Daniel Siebert |
|
?SciEngines |
Kiel (D) |
Thu |
Sun |
confirmed 20/1 |
Staying at Carlton House |
19 |
William Spooner |
Eagle Genomics |
Babraham (UK) |
Fri |
Mon |
confirmed 9/1/12 |
Staying at Rosedale. Had to cancel at last minute due to deadline |
|
20 |
Jim Procter |
jsur_compbio.dundee.ac.uk |
U Dundee (Jalview,java in debian, osgi) |
Dundee (UK) |
Cancelled |
Cancelled |
Cancelled |
Had to cancel at last minute due to deadline |
Agenda
The meeting addresses the following topics over Saturday and Sunday - please add yours. Work and discussions will happen in parallel in the same room. The schedule is determined on Friday evening when most of us are expected to have arrived on site.
As a special feature, just like with bigger conferences, only better, this year brings us a 'hands-on' tutorial on FPGA computing for application acceleration. It is planned to take up all of Friday and this way only enriches and not disturbs the remainder of the meeting. Respective attendees are expected to arrive already on Thursday.
- How do we see Debian Med changed / do we actively change Debian Med with what we learn from
- Interaction of with Ubuntu (MOTU) maintainers
- Interaction with bioinformatics research groups
- having their work packaged by us
- packaging it themselves
- packaging their third party tools
- Interaction with service groups in academia and industry
- Review of latest developments
Cloud computing progress - priority tools for Cloud (eg. Cloudman)
- BOINC
- GPU / FPGA application acceleration
- Workflows
- Infrastructure with Taverna and Galaxy
- Can we find ways to present such more visibly?
- What packages are missing in Debian to render new kinds of workflows a reality with Debian?
getData / BioMaj
- first packages using it
- standardising interaction between databases and packages/Debian policy - (consider extensions to APT? - Tim)
- better integration with above workflows
- effectiveness in Cloud/cluster environments
- comparison with the data management configuration in Galaxy
- other further developments?!?
- Review of what came up on the mailing list
- Renaming of user-executable files to fully comply with Debian Policy for scientific packages (like omitting .pl suffix)
- License checking session: Lets do a collaborative try to write upstream of non-free programs to consider relicensing
Personal Agendas
Andreas Tille
See also Blog posting
- Ensembl packaging specifically working on installation in unstable without any conflicts (=trying to solve libwww issue)
- problematic package is actually liblwp-parallel-perl (need more input from previous maintainer)
- Beast-mcmc finishing
- last problematic JAR mtj.jar is a mess of other included JARs based on f2j blas/arbapck JARs with no visible source
Continuing with current MoM; presenting MoM to the audience to gather more potential students
- Kept on mentoring the student
- Finishing sofa-framework which needs to be updatet to new version
- Bug squashing of Debian Med team maintained packages
- Solved #657827 and #648167
- If time permits work on some Beast related phylogeny packages
- Uploaded phy-spread package
- Checking status of Debian Med tasks
- tasks tools are broken because of technical changes in Debian Infrastructure, needs sorting out
- Spontaneous mentoring of people
- Introductional talk about Debian and Blends structure
- Mentoring Quyen, Laszlo and others about gpg keys
- Working on Blends sentinel bugs overview
- Terribly blocked by non-functional fetching translations from ddtp.debian.net which is down. Caring for alternatives to at least get tasks pages updated again but no success so far.
- Helping others packaging and sponsoring
- Sponsored chado (gmod suite) prepared by Olivier Sallou
- Helped Charles getting snappy-java using Debian packaged library (see below external wishlist)
- Worked with Ivo on copasi
- Helped Martin with Python packaging
Had a look with Tony Travis into OpenMicroscopy noticing that it is a beast containing > 800 binary JAR files inside the source and will take a talented and very patient Java expert to cope with all this stuff
- Looked into packages prepared by Piero
- License checking
Wrote ePetition to free Phylip
External wishlist
Schedule
- Thursday: Arrival of participants / trainers for FPGA tutorial
- Friday:
- 9-19: FPGA tutorial, breaks for coffee and lunch as it fits
- 19-21: dinner
- 21-22: icebreaking 7-15min talk series: how Debian helps today and after this sprint
- Laszlo and Piero: structure prediction
- Stuart and Ivo: system biology
- Tim and Soon: Bringing the Galaxy server to Debian
- 23-24: discussions and planning of synchronisation points on Sat/Sun
- Saturday/Sunday: roundtable work, planning, discussion and writing
- unplanned discussions over breakfast
- teams start whenever breakfast ends
- joint lunch
- coffee+tea+biscuits?
- joint dinner
- more teamwork
- Monday: joint breakfast and group travel to airports
Related meetings
We hope that after two successful meetings this event will continue in subsequent years.
It's worth noting for those not in the know that there is some symbiosis between this sprint event and the annual BOSC Codefest which is held before the ISMB conference. Though the meetings are quite different in focus, they have recently engaged many of the same people and projects.
See also
Other upcoming Debian sprints
Debian Med developer's portal