incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matthew Hayes <>
Subject [PROPOSAL] DataFu for Incubation
Date Wed, 18 Dec 2013 22:49:24 GMT
Hi all,

I would like to share our draft ASF incubation proposal for DataFu, a library that makes it
easier to solve data problems in Hadoop and high level languages based on it.

The proposal can be found here:

The source code is available on GitHub:

The text of the proposal is copied below.  Feedback is appreciated!


== Abstract ==

Data``Fu makes it easier to solve data problems using Hadoop and higher level languages based
on it.

== Proposal ==

Data``Fu provides a collection of Hadoop Map``Reduce jobs and functions in higher level languages
based on it to perform data analysis.  It provides functions for common statistics tasks (e.g.
quantiles, sampling), Page``Rank, stream sessionization, and set and bag operations.  Data``Fu
also provides Hadoop jobs for incremental data processing in Map``Reduce.

== Background ==

Data``Fu began two years ago as set of UDFs developed internally at Linked``In, coming from
our desire to solve common problems with reusable components.  Recognizing that the community
could benefit from such a library, we added documentation, an extensive suite of unit tests,
and open sourced the code.  Since then there have been steady contributions to Data``Fu as
we encountered common problems not yet solved by it.  Others outside Linked``In have contributed
as well.  More recently we recognized the challenges with efficient incremental processing
of data in Hadoop and have contributed a set of Hadoop Map``Reduce jobs as a solution.

Data``Fu began as a project at Linked``In, but it has shown itself to be useful to other organizations
and developers as well as they have faced similar problems.  We would like to share Data``Fu
with the ASF and begin developing a community of developers and users within Apache.

== Rationale ==

There is a strong need for well tested libraries that help developers solve common data problems
in Hadoop and higher level languages such as Pig, Hive, Crunch, Scalding, etc.

== Current Status ==

=== Meritocracy ===

Our intent with this incubator proposal is to start building a diverse developer community
around Data``Fu following the Apache meritocracy model.  Since Data``Fu was initially open
sourced in 2011, it has received contributions from both within and outside Linked``In.  We
plan to continue support for new contributors and work with those who contribute significantly
to the project to make them committers.

=== Community ===

Data``Fu has been building a community of developers for two years.  It began with contributors
from Linked``In and has received contributions from developers at Cloudera since very early
on.  It has been included included in Cloudera’s Hadoop Distribution and Apache Bigtop.
 We hope to extend our contributor base significantly and invite all those who are interested
in solving large-scale data processing problems to participate.

=== Core Developers ===

Data``Fu has a strong base of developers at Linked``In.  Matthew Hayes initiated the project
in 2011, and aside from continued contributions to Data``Fu has also contributed the sub-project
Hourglass for incremental Map``Reduce processing.  Separate from Data``Fu he has also open
sourced the White Elephant project.  Sam Shah contributed a significant portion of the original
code and continues to contribute to the project.  William Vaughan has been contributing regularly
to Data``Fu for the past two years.  Evion Kim has been contributing to Data``Fu for the past
year.  Xiangrui Meng recently contributed implementations of scalable sampling algorithms
based on research from a paper he published.  Chris Lloyd has provided some important bug
fixes and unit tests.  Mitul Tiwari has also contributed to Data``Fu.  Mathieu Bastian has
been developing Map``Reduce jobs that we hope to include in Data``Fu.  In addition he also
leads the open source Gephi project.

=== Alignment ===

The ASF is the natural choice to host the Data``Fu project as its goal of encouraging community-driven
open-source projects fits with our vision for Data``Fu.  Additionally, other projects Data``Fu
integrates with, such as Apache Pig and Apache Hadoop, and in the future Apache Hive and Apache
Crunch, are hosted by the ASF and we will benefit and provide benefit by close proximity to

== Known Risks ==

=== Orphaned Products ===

The core developers have been contributing to Data``Fu for the past two years.  There is very
little risk of Data``Fu being abandoned given its widespread use within Linked``In.

=== Inexperience with Open Source ===

Data``Fu was started as an open source project in 2011 and has remained so for two years.
 Matt initiated the project, and additionally is the creator of the open source White Elephant
project.  He has also contributed patches to Apache Pig.  Most recently he has released Hourglass
as a sub-project of Data``Fu.  Sam contributed much of the original code and continues to
contribute to the project.  Will has been contributing to Data``Fu since it was first open
sourced.  Evion has been contributing for the past year.  Mathieu leads the open source Gephi
project.  Jakob has been actively involved with the ASF as a full-time Hadoop committer and
PMC member.

=== Homogeneous Developers ===

The current core developers are all from Linked``In.  Data``Fu has also received contributions
from other corporations such as Cloudera.  Two of these developers are among the Initial Committers
listed below.  We hope to establish a developer community that includes contributors from
several other corporations and we are actively encouraging new contributors via presentations
and blog posts.

=== Reliance on Salaried Developers ===

The current core developers are salaried employees of Linked``In, however they are not paid
specifically to work on Data``Fu.  Contributions to Data``Fu arise from the developers solving
problems they encounter in their various projects.  The purpose of Data``Fu is to share these
solutions so that others may benefit and build a community of developers striving to solve
common problems together.  Furthermore, once the project has a community built around it,
we expect to get committers, developers and contributions from outside the current core developers.

=== Relationships with Other Apache Products ===

Data``Fu is deeply integrated with Apache products.  It began as a library of user-defined
functions for Apache Pig.  It has grown to also include Hadoop jobs for incremental data processing
and in the future will include code for other higher level languages built on top of Apache

=== An Excessive Obsession with the Apache Brand ===

While we respect the reputation of the Apache brand and have no doubts that it will attract
contributors and users, our interest is primarily to give Data``Fu a solid home as an open
source project following an established development model.

== Documentation ==

Information on Data``Fu can be found at:

== Initial Source ==

The initial source is available at:

== Source and Intellectual Property Submission Plan ==

 * The Data``Fu library source code, available on Git``Hub.

== External Dependencies ==

The initial source has the following external dependencies that are either included in the
final Data``Fu library or required in order to use it:

 * fastutil (Apache 2.0)
 * joda-time (Apache 2.0)
 * commons-math (Apache 2.0)
 * guava (Apache 2.0)
 * stream (Apache 2.0)
 * jsr-305 (BSD)
 * log4j (Apache 2.0)
 * json (The JSON License)
 * avro (Apache 2.0)

In addition, the following external libraries are used either in building, developing, or
testing the project:

 * pig (Apache 2.0)
 * hadoop (Apache 2.0)
 * jline (BSD)
 * antlr (BSD)
 * commons-io (Apache 2.0)
 * testng (Apache 2.0)
 * maven (Apache 2.0)
 * jsr-311 (CDDL-1.0)
 * slf4j (MIT)
 * eclipse (Eclipse Public License 1.0)
 * autojar (GPLv2)
 * jarjar (Apache 2.0)

== Cryptography ==

Data``Fu has user-defined functions that use MD5 and SHA provided by Java’s``Digest.

== Required Resources ==

=== Mailing Lists ===

Data``Fu-private for private PMC discussions (with moderated subscriptions) Data``Fu-dev Data``Fu-commits

=== Subversion Directory ===

Git is the preferred source control system: git://

=== Issue Tracking ===

JIRA Data``Fu (Data``Fu)

=== Other Resources ===

The existing code already has unit tests, so we would like a Hudson instance to run them whenever
a new patch is submitted. This can be added after project creation.

== Initial Committers ==

 * Matthew Hayes
 * William Vaughan
 * Evion Kim
 * Sam Shah
 * Xiangrui Meng
 * Christopher Lloyd
 * Mathieu Bastian
 * Mitul Tiwari
 * Josh Wills
 * Jarek Jarcec Cecho

== Affiliations ==

 * Matthew Hayes (Linked``In)
 * William Vaughan (Linked``In)
 * Evion Kim (Linked``In)
 * Sam Shah (Linked``In)
 * Xiangrui Meng (Linked``In)
 * Christopher Lloyd (Linked``In)
 * Mathieu Bastian (Linked``In)
 * Mitul Tiwari (Linked``In)
 * Josh Wills (Cloudera)
 * Jarek Jarcec Cecho (Cloudera)

== Sponsors ==

=== Champion ===

Jakob Homan (Apache Member)

=== Nominated Mentors ===

 * Ashutosh Chauhan <hashutosh at apache dot org>
 * Roman Shaposhnik <rvs at apache dot org>
 * Ted Dunning <tdunning at apache dot org>

=== Sponsoring Entity ===

We are requesting the Incubator to sponsor this project.

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message