[Rprotobuf-commits] r559 - papers/rjournal

[noreply at r-forge.r-project.org]

Author: murray
Date: 2013-12-17 03:12:56 +0100 (Tue, 17 Dec 2013)
New Revision: 559

Modified:
   papers/rjournal/eddelbuettel-francois-stokely.bib
   papers/rjournal/eddelbuettel-francois-stokely.tex
Log:
Also will want to cite Hadley's Split-Apply-Combine JSS paper as that
pattern works much better with RProtoBuf and you remove the
restriction that all phases must be done in R.



Modified: papers/rjournal/eddelbuettel-francois-stokely.bib
===================================================================
--- papers/rjournal/eddelbuettel-francois-stokely.bib	2013-12-17 02:09:42 UTC (rev 558)
+++ papers/rjournal/eddelbuettel-francois-stokely.bib	2013-12-17 02:12:56 UTC (rev 559)
@@ -72,6 +72,16 @@
  address = {New York, NY, USA},
  keywords = {collection, communication, cross-language, garbage, managed, memory, model, object, rpc, runtimes, shared, synchronization, transparent, type-safe},
 }
+ at article{wickham2011split,
+  title={The split-apply-combine strategy for data analysis},
+  author={Wickham, Hadley},
+  journal={Journal of Statistical Software},
+  volume={40},
+  number={1},
+  pages={1--29},
+  year={2011},
+  publisher={Citeseer}
+}
 @inproceedings{Sumaray:2012:CDS:2184751.2184810,
  author = {Sumaray, Audie and Makki, S. Kami},
  title = {A Comparison of Data Serialization Formats for Optimal Efficiency on a Mobile Platform},

Modified: papers/rjournal/eddelbuettel-francois-stokely.tex
===================================================================
--- papers/rjournal/eddelbuettel-francois-stokely.tex	2013-12-17 02:09:42 UTC (rev 558)
+++ papers/rjournal/eddelbuettel-francois-stokely.tex	2013-12-17 02:12:56 UTC (rev 559)
@@ -42,6 +42,11 @@
 between integers and floating point.  Large numbers of JSON messages
 would also be required to duplicate the field names with each message.
 
+TODO(ms): Also work in reference to Split-Apply-Combine pattern for
+data analysis \citep{wickham2011split}, since that is a great pattern
+but it seems overly optimistic to expect all of those phases to always
+be done in the same language.
+
 This article describes the basics of Google's Protocol Buffers through
 an easy to use R package, \CRANpkg{RProtoBuf}.  After describing the
 basics of protocol buffers and \CRANpkg{RProtoBuf}, we illustrate
@@ -125,7 +130,7 @@
 \section{Dynamic use: Protocol Buffers and R}
 
 TODO(ms): random citations to work in:
-
+q
 We make use of Object Tables \citep{RObjectTables} for lookup.
 Many sources compare data serialization formats and show protocol
 buffers very favorably to the alternatives, such