State Level Convention for Academicians on Free Software in Research and Teaching - K. Gopinath IISc

Page created by Everett Stevenson
 
CONTINUE READING
State Level Convention for
Academicians on Free Software in
    Research and Teaching

           K. Gopinath
               IISc
Rationale of state convention on
        computing freedom in education
              and research in S&T
●   Science a great success story for last ~500 yrs
    ●   Technology also!
●   Crucial "genes": observation and its close cousin
    experimentation, attempted refutation of
    hypotheses advanced through careful
    experiments, and generalization.
●   For this to succeed, free exchange of ideas and
    data important.
How does this happen?
●   face2face meetings
●   presenting papers in confs & research societies
●   publication of papers in journals and conferences
●   publication of data in monographs
As volume of scientific work becomes large
●   Need indexing, reviewing and archiving
●   Need specialized services eg. Science Citation Index
    ●   If you want to find out most recent work in cloning of
        sheep, first find a “keystone” paper in cloning
        –   eg. the paper describing Dolly, the first cloned sheep
    ●   Locate keystone paper in SCI: it will list all articles till
        today (recent!) that have referred to the keystone paper
Free Software Closely Follows Science!
●   Exchange of ideas as in science
●   Free software adds a new and critical dimension:
    that of sharing code.
●   Code can incorporate executable models of some
    part of reality.
    ●   Eg: in predicting weather, ocean codes developed. Such
        codes incorporate an understanding of the various
        component interactions including some very specific
        local ones.
●   Sharing code => sharing models of reality!
●   Across the globe betw researchers who never met!
Why is sharing important?
●   For complex problems facing us today (such as
    food, health, pollution, global warming), need
    for transparent and effective flow of information
    ●   amongst scientists themselves, and
    ●   between scientists and society
    so that effective action can be taken.
●   Free software is a necessary lubricant in this
    process as software embodies the executable
    version of some these societal models.
Sharing
●   Free software can be used as is (such as the
    statistical package R) or used as a way of
    communication between researchers and teachers
    on current research topics.
●   A researcher comes up with a prototype sw and is
    later modified and augmented by many others.
    ●   eg, for alignment of genetic codes, a free sw called
        Muscle is used while HMMER is used for studying
        genetic sequences.
    ●   HMMER current version is the result of many
        discussions and collaborations between the original
        developer and many researchers around the globe. It
        would not have happened without freely distributable sw.
Freedom and Free SW
●   A great need for freedom in science and
    technology both
    ●   from an education perspective +
    ●   from a research perspective
●   Free software nowadays playing a critical role.

Hope this conference furthers this important goal!
Free Software and its Manifestations

               K. Gopinath

        Indian Institute of Science
A Related Idea
●   Knowledge has to be free
●   Knowledge does not decrease when shared!
        subhashita:
    ●   ``Cannot be snatched away by a thief,
        cannot be snatched away by a king,
    ●    cannot be divided among brothers,
        not heavy either.
    ●    If spent daily, it always keeps growing.
    ●    The wealth of knowledge is the most precious of all.''
●
    Paper and Printing made copies easier post 15th century
●   Knowledge in mathematics, natural sciences free for all!
    ●   India's past contributions in mathematics, grammar, ...
What is Free Software?
●   It is an idea of how deliverability of software
    should be framed so that complete control rests
    in the hands of the end user
●   So the source of the software is needed!
    ●   Good for academics!
●   It can be shared also with others!
●   Richard Stallman's “The four freedoms”
    ●   Run sw, study/change sw, distr sw
    ●   Distr changed versions
●   Last 15 years or so: free software good for
    industry too!
Operating Systems Success Story!
●   GNU/Linux
●   BSD family
●   OpenSolaris
●   “The adoption of free software to resolve incompatibility
    between the economic need for provider diversity and the
    engineering need to avoid product diversity is fairly unique
    across all industry. It is not a coincidence that the most
    important and successful free software is the system software
    such as the operating system kernel and middleware such as
    low-level software libraries, many system tools, compiler tool
    chain, X server, windowing systems, GUI toolkits, web, email
    and other services, databases, and even a dot-net workalike''
    Slightly rephrased from Andrew Morton (a key developer for
    Linux kernel), OLS 2004
The Power of Openness!
●   Paul McKenney and Dipankar Sharma (IBM) '07
    ●   Closely involved in the development of the RCUpdate
        feature in the Linux kernel
●   Their understanding of this feature and its
    interactions with the rest of the kernel deepened
    more in the few years (5) it was in the Linux kernel
    than the 15+ years it existed in a commercial
    operating system where it originated first.
●   This happened due to the diversity of uses the basic
    mechanism was subjected to and tested against by
    completely unknown developers across the globe.
●   Similar to the scientific system of enquiry (attempt
    “falsification” of theories)! A success story of the last
    few centuries.
Other examples
●   Java language and libraries
●   Google's Android mobile: the full stack
●   All new research ideas typically start out as free
    software nowadays
    ●   Yahoo!'s Hadoop, Zoo, ...
    ●   Biological genetic sequence modelling: Muscle (for
        alignment of sequences), HMMER (for studying genetic
        sequences), ...
    ●   Climate change studies: eg: GISS 3-D general
        circulation models (GCMs) and coupled atmosphere-
        ocean models for simulating Earth's climate system
●   Almost all socially committed programs
    ●   Medical records: OpenMRS; disaster mgmt: Sahana
Technical Ideas Also Need to be
                 Become Free!
●   Patent System developed so that an invention
    becomes accessible to society
    ●   Inventor has right to exploit it exclusively for a fixed period
    ●   But inventor has to disclose in sufficient detail so that a
        competent person can duplicate it after exclusive period
    ●   Works well for most technical areas that deal with material
        goods rather than abstract ideas
●   But not for software!
    ●   If non-obvious, software often based on (deep)
        mathematics that is free as per the patent system
    ●   Also an engineering product that often encodes “simple”
        or obvious processes in technical or business processes
Patent System
●   Breaks down for software
●   Often has given patents for obvious ideas
●   “Software Patents” use a strategem of
    embodying abstract ideas in physical devices
    for patentability
●   Indian patent system luckily still does not allow
    such strategems but many attempts to make it
    close to the US system that allows software
    patents
●   Data=code in CS => software patents tread on
    copyright system, more confusion!
Hardware, Protocols also Want to be Free!
●   LinuxBIOS, OpenFirmware, OpenSPARC
●   Internet protocols open for all from the start!
    ●   Most successful set of protocols across the globe
    ●   Most Internet software free!
●   When protocols or interfaces not open (proprietary)
    => failure typically!
    ●   eg. many proprietary GIS systems but “islands” typically
    ●   eg. Indic computing/e-gov: too many non-open solutions
    ●   Only market dominance prevents failure but a
        competitive marketplace does not allow this for too long
        a period!
●   Cloud Computing Needs to Become Free!
    ●   Too little control (sw/data) in the hands of user
Other related ideas
●   Open Access Publishing
    ●   eg. PLoS Biology, BioMed Central
    ●   10-15% of the 20–25,000 peer-reviewed journals
●   Open Publishing (Wikipedia, Wikis)
●   Creative Commons
    ●   content creators take control of how they choose to
        share with others
●   Right to Information
    ●   Governmental processes or documents should be
        freer!
    ●   In India from 2005, US '67
Major Indian Issues
●   Free software in Indian educational system
●   Free software for Indic computing
●   Free software in E-Governance
●   Free software for Geospatial Analysis
●   Free Software in Accessibility Computing
●   Free software in Mobile devices
●   Free Software Business Models
●   Open Standards
●   Free software in UID Project?
    ●   Proprietary technologies???
A Small Example of Free Software
               Synergy
●   IIT Madras developed Acharya system for
    Indian Languages
    ●   A Qt based application
    ●   Better than UNICODE encoding
        –   Current snippets or texts on Internet often wrong!
    ●   Makes sorting, search and other operations simpler
●   IISc here has taken this software and made it
    run on S60-based mobiles (Nokia, some Sony
    Ericsson). Demos by Anand the developer
●   Plans to use this Indic software in OpenMRS,
    etc for storing health records on mobiles
Free Software Needed Now More
               Than Ever?
●
    Major Environmental issues of the 21st century
    such as pollution, climate change!
●   Complex systems have many interdependent parts
    with no one (individual or a company) having a
    complete picture
    ●   Any software handles only a small part
●   If proprietary, only the two owners can put two diff
    programs to work together or standards needed
    ●   Economic incentive needed (+ time, staff)
●   If free sw, source avlbl. May be possible to put
    them together by anyone! Scalable model!
Global Warming
●   Due to Solar flares?
●   Due to CO2?
    ●   Industry, transportation, respiration/forest cover,
        microorganisms in oceans and on land, ...
●   Due to Methane?
    ●   Cows, agriculture, permafrost, ...
●   Affected by and Effects
    ●   Atmosphere, clouds/rainfall, ocean currents,
        geophysical processes, glaciers, coastal cities
    ●   Biodiversity: Corals, “good” and “bad“ microorganisms,
        productivity of agriculture
Climate Change Software
Need (for example)
●   numerical methods, compilers, data ontologies,
    visualization software, embedded systems software,
    as well as
●   coordination across Internet using wikis, Wikipedia,
    social media, reputation systems, etc
Need transparency and effective flow of
information between scientists, and between
scientists and society so that effective action can
be taken.
Free software is a necessary lubricant in this
process as software embodies the executable
version of the models.
India Energy Situation
●   GRIM!
●   Only coal avlbl in plenty
    ●   Nuclear power will take time
    ●   Oil supply: no reliable longterm partners
    ●   Heavy competition from China that has succeeded
        in cornering deals and winning them against India
●   Clean coal technologies needed by India
    ●   SOFTWARE NEEDED!
        –   Advanced algorithms, new techniques
    ●   Can this software be developed as research
        projects by students and faculty?
    ●   Can this be done as free software so that easier to
        work together?
Another Major Future Issue
●   Mobiles more affordable than PCs
●   Free software based mobile systems!
●   Mobiles for distributed coordination
    ●   eg. study groundwater depletion, pollution,
        epidemiology, traffic congestion, health monitoring, etc
    ●   Reduce energy usage, pollution
●   Need incentives for truthful pooling of data
    collected by cheap sensors (“mobiles”!)
    ●   Lots of brand new software has to be written!
●   Start a movement on Open Hardware for mobiles
    and Free Software on such hardware?
Conclusion
●   Free Software has significant momentum
●   Has the potential to further open up society
    ●   But a proper incentive system needed for deeper
        gains and to solve continental scale problems
    ●   Just like the “copyleft” mechanism in GPL to avoid
        freeriders!
●   Students, Teachers, Activists, Reporters,
    Developers, Industry support, Free software
    businesses and models all needed!
You can also read