Events.IRILLDays2010 page history

CONTENT

CONTACT

UPDATES

MONETIZATION

Web Monetization enabled, thanks for your support!

Web Monetization is not enabled. To learn about it and support this website financially check Monetization.

RIGHTS

  • code

  • content
  • artwork
Recent Changes - Whiteboard / Flashcards

CONTENT

CONTACT

UPDATES

MONETIZATION

Web Monetization enabled, thanks for your support!

Web Monetization is not enabled. To learn about it and support this website financially check Monetization.

RIGHTS

  • code

  • content
  • artwork

IRILL Days 2010 {Events}

IRILLDays2010 04/10/2010 at INRIA

#IRILLDays2010 on Twitter, identi.ca (irill), TwitterStreamGraphs, Collecta, flickr

Attended activities

Day 1

Day 2

Day 1

Michael Tiemann, VP RedHat (9h45-10h30)

Back to the Menu

Venky@RH

  • models of inno
    • Baldwin and Von Hippel
      • transition to collab env.
  • 1/3rd of humanity is connected together
    • Internet is the largest collab env
    • see The World is Flat 3.0
  • compat with indou philsophy of knowledge
    • exist godess of knowledge Sara...
  • theory or practice?
  • policy recommendations

Back to the Menu

Karsten Gerloff, President, FSFE Enlightening! Free Software research in context (10h4511h15-11h1511-??)

  • Karsten Gerloff President of FSFE
  • Karsten on Free Software Breaking chains, building bridges
  • http://twitter.com/kgerloff
  • why do we do free software?
  • software ubiquitous
  • who control all this computer?
  • ref to Lessig on Code is law
  • "Free software need a lobby.", "We wear a suit so that you don't have to."
  • What can researchers do for free software?
    • figures to argue properly
      • number of dev, biz actors, ...
      • to be published under the right license
  • "bad" example of recent CERN move

Georg Greve, Kolab Systems, CEO; FSFE, Board Member: Free Software and the Playing Field (11h1511h??-11h4512h05)

  • http://identi.ca/greve/
  • Georg C. F. Greve Founder and former President of FSFE
  • CEO of Kolab Systems AG
  • still default
  • accounting
    • proprietary license as asset rather than a liability
  • exclusive rights (to exclude others of using)
  • public R&D funding
    • e.g. CERN before
    • rather than focus on competency
  • still misunderstanding of what proprietary software means
    • exclusive ocntrol of proprietor
    • an entire chain
  • mention of the locked-in effect of Switzerland and Microsoft
  • if the model is not understood (proprietary model), how can the opposite model be understood? (FLOSS)
    • redefinition tries
      • fair trade software, bio software, ...
  • quality?
    • in the large project when the wisdom of the crowd applies yes, on the side, maybe not always
  • "Any sufficiently advanced corruption is indistinguishable from incompetence." Georg Greve
    • inspired by Arthur C. Clark "Any sufficiently advanced technology indistinguishable from magic."
  • money is a tool for sharing labor
    • good work costs money
  • differentiate on quality and efficiency, not control (USP)
  • co-existance is not a goal, rather spit-out the "proprietary kool-aid"
    • "technology neutral"
  • several ref to Gartner and their remark on Open Cores (prop. with new packaging)

Back to the Menu

Roberto Di Cosmo, Director, IRILL (11h4512h05-12h45??)

  • importance about pedagogy and education, not just technology
  • entirely new way of developing software
  • foster vendor independence
  • political impact of IT decision and its link with the level of incompetence
  • INRIA valorisation logiciel libre
  • problem of the "white paper" software design which is unrealistic as there already is an ecosystem
  • goal
    • being a catalyzer
    • details of the logo
      • 3 colors
        • developer community
        • industries
        • research and education
    • go further than the publish or perish mindset
    • always updating courses
      • favoritize new material for FOSS Curricula by and for teachers
    • example
      • Mancoosi FP7 project presented right after
    • provide resources
      • classes
      • internships
  • links

Back to the Menu

Mancoosi: managing your GNU/Linux installation, Stefano Zacchiroli (14h00-14h30)

  • mancoosi managing software complexity (FP7 project)
    • "We pursue two main avenues:
      1. Develop mechanisms that provide for rollbacks of failed upgrade attempts, allowing the system administrator to revert the system to the state before the upgrade (work package 2 and work package 3)
      2. Develop better algorithms and tools to plan upgrade paths based on various information sources about software packages and on optimization criteria (workpackages work package 4 and work package 5)."
  • EDOS aimed at improving the stability of a distribution from the point of view of the distribution editor
  • user = sysadmin of a machine
  • killapp = package management om a software distribution
  • http://www.dicosmo.org/space/PlaquetteMancoosi-rotated.pdf
    • competition mode
    • equivalent to routing problem?
  • CUDF Common Upgradeability Description Format
  • DUDF distribution specific equivalent of CUDF
  • MiSC Mancoosi internal Solver Competition
    • CPLEX based (UNSA)
    • importance of the encoding of the problem
  • research trend of "self" optimization through competition in the CS community?
  • Q&A
    • autoconf, gcc toolchain, ...
    • importance in the cloud (with upcoming "eos")
      • for virtual machine to do upgrade plans in a network of interconnected machines
    • cycles of community to model and optimize in 2 sides
      • user -> problem defining
      • dev comm -> to plug solver
    • part of the larger movement especially in CS ?
      • e.g. MLcomp, Kaggle
    • template ? lessons from similar projects ?
      • 1st project by the author and presented at FOSDEM about a year and half ago
        • willing to participate
    • suitable for KM? lookup table?
      • probably not, problems are "too uniq"

Coccinelle at the service of the Linux Kernel, Julia Lawall (14h30-15h00)

  • Coccinelle A Program Matching and Transformation Tool for Systems Code
    • SmPL (Semantic Patch Language)
  • Collateral Evolutions
  • Semantic patches
  • definition of metavariables, subterms, ...
    • research paper in POPL
  • specificities
    • based on regex
    • C, part PHP support (from Facebook)
  • see also A Bug Life (ABL)
  • does this lead to the complexification or simplification of the toolchain?
  • smarter patch series/Quilt/Mercurial Queues by integrating with testing?

Back to the Menu

Ocsigen: rethinking web programming, Vincent Balat (15h00-15h30)

  • Ocsigen proposing clean and safe tools for developing and running Web 2.0 applications.
  • description of the evolution
  • even RoR, GWT and others do not fully anticipated the web as a participatory platform
  • Ocsigen Philosophy Eliom uses Objective Caml
  • checking at compile time, including HTML validity, link, ...
  • live demo (locally) http://128.93.135.222
  • compiler from OCaml bytecode to JavaScript
  • sum-up
    • 1 code client and server
    • high-level expressive primitives
    • static typing
  • Q&A
    • limits of static typing, e.g. Haskell
    • bus actually usable or just at the demo level
    • browser idiosyncrasy
    • OCaml specificities (e.g. mutate ref vs. bus translation to JS)
    • edition=server restart

Back to the Menu

Day 2

Technical session: Emerging projects and technologies (9h30-12h15)

New Trends in Scilab, Claude Gomez (9h30-10h00)

  • http://www.scilab.org
  • Seedea:Seedea/Simulations
  • including Modelica compiler through Scicos Block diagram modeler/simulator
  • preview for Scilab 6.0
    • new kernel for HPC, code generation, debugging
  • applied math algorithms designed for sequential, not parallelism
  • Scilab code currently a property of INRIA who funds it
    • one member of the Scilab consortium
  • License CeCILL
    • CeCILL initiated by CEA, CNRS and INRIA
  • Le Consortium Scilab annonce la création de Scilab Enterprises on Digiteo
    • providing services and support
  • Equalis Your Mathematics Community
  • new Scilab for 2010 = Scilab Consortium + Scilab 6.0 + Scilab Entreprise
    • desiring to be partner with IRILL
  • Q&A
    • lobbying CPGE and Education National, Inspection Generale de Mathematiques (against Mathematica)
    • difference between Mathlab and Scilab
      • "To make money you have to cut the software in pieces." (cf InformationRules)
      • anyway, everything is based on matrices libraries mainly BLAS and LAPACK
    • how did you get successful and gain credibility then trust?
      • using Matlab, estimation of tokens, estimation of alternative
      • being free
    • opinion on NumPy, R, ...
      • working on establishing a link but a different way of using numerical computation
        • Scilab provides an interpreter, is specific to matrices, ...
      • not really aiming at computer scientists

Back to the Menu

Building an R&D ecosystem around GCC: what, why and how?, Albert Cohen (10h-10h30)

  • http://www.ann.jussieu.fr/~cohen/
  • ctuning.org, tools discussed yesterday
  • community > code
  • mechanism thanks to a diagram from MELT project
    • very brief history
  • community specificities
    • based on trust and fair usage through licenses like GPL
    • hybrid of Cathedral the Bazaar
  • example of the price of code merging after long branching with Apple and a lack of understanding of the community structure
    • forked (or... just left)
  • mention of papers from POPL, PLDI, HiEAC, PPoPP, PACT, CAES, CGO, ICS
  • mention of parallelism (OpenMP) but also specialization through GPUs (CUDA) and also Cookbook.Electronics#FPGA
  • explanation of MELT as a way to make "your own pass"
  • MILEPOST GCC
    • gcc.gnu.org/wiki/MilepostGCC
    • gcc-ici.sourceforge.net
  • still important to tune, example of ARM as "the most popular platform today"
  • see also Gregori Fursin from INRIA

Back to the Menu

Component-based programming with Fractal - models, tools and the FraSCAti platform , Philippe Merle and/or Lionel Senturier (10h45-11h25)

Back to the Menu

Dynamic web development, Stéphane Ducasse (11h25-11h55)

  • Seaside.st
    • layered set of abstractions over HTTP and HTML based on Smalltalk
  • objectives
    • remodularisation
    • modular and secure reflective languages
      • Pharo Open Source Smalltalk
  • other frameworks, especially on Javascript, at WebWorkersCampParis
  • examples
  • live demo
  • continuation
    • goto call/answer (procedural return) over HTTP
  • objectif
    • entreprise en Suisse, en Belgique, mais pas au nord de la France
  • Q&A
    • probleme de l'environment non maitrise (reseau lowQ)
    • equivalent Lisp/Guile ?

Back to the Menu

Technical session: Industry and innovation (14h00-16h30)

Hi-Lite: a Verification Toolkit for Unit Test and Unit Proof, Yannick Moy (Adacore) (14h00-14h30)

Missed, extended lunch, noticed they have blackboard in the bathrooms Back to the Menu

Open Source Cloudware initative: towards a free, open cloud, Jean-Pierre Laisné (OW2) (14h30-15h00)

Back to the Menu

Open data and the Data Publica Project, François Bancilhon (15h00-15h30)

Back to the Menu

DORM, the missing tool to persist your deliveries, Pierre Queinnec (Zenika) (16h00-16h30)

Back to the Menu

A short survey of Real Time Linux technologies, Pierre Ficheux (OpenWide) (16h30-17h00)

Back to the Menu

Own objectives

  1. help on that issue
  2. learn this
  3. meet that person

Back to the Menu

Other reviews

  • here

To do

  1. improve Template
    1. take the habit of eliciting the motivation of the organizer
      1. http://www.dicosmo.org
    2. discovered through
      1. Cap Digital
  2. abv
    • CI = Collaborative Innovation
    • inno = innovation
    • env. = environment
  3. recording most likely available under 10 days
  4. pass with : in front!
    • amazingly, everybody did the "correction" mistake!
  5. names of people discussed with...
  6. first conf with that many interpreter/compiler running

Alcatel-Lucent (recall remarks, questions, path to understanding!, ML related project, ...)