Papers by Philippe Grosjean
arXiv (Cornell University), Oct 13, 2017
Nearly every popular programming language comes with one or more package managers. The software p... more Nearly every popular programming language comes with one or more package managers. The software packages distributed by such package managers form large software ecosystems. These packaging ecosystems contain a large number of package releases that are updated regularly and that have many dependencies to other package releases. While packaging ecosystems are extremely useful for their respective communities of developers, they face challenges related to their scale, complexity, and rate of evolution. Typical problems are backward incompatible package updates, and the risk of (transitively) depending on packages that have become obsolete or inactive. This manuscript uses the libraries.io dataset to carry out a quantitative empirical analysis of the similarities and differences between the evolution of package dependency networks for seven packaging ecosystems of varying sizes and ages: Cargo for Rust, CPAN for Perl, CRAN for R, npm for JavaScript, NuGet for the .NET platform, Packagist for PHP, and RubyGems for Ruby. We propose novel metrics to capture the growth, changeability, resuability and fragility of these dependency networks, and use these metrics to analyse and compare their evolution. We observe that the dependency networks tend to grow over time, both in size and in number of package updates, while a minority of packages are responsible for most of the package updates. The majority of packages depend on other packages, but only a small proportion of packages accounts for most of the reverse dependencies. We observe a high proportion of "fragile" packages due to a high and increasing number of transitive dependencies. These findings are instrumental for assessing the quality of a package dependency network, and improving it through dependency management tools and imposed policies.
Western Indian Ocean Journal of Marine Science, 2022
The transportation of live coral nubbins is a major constraint of the coral trade. Corals must be... more The transportation of live coral nubbins is a major constraint of the coral trade. Corals must be transported within a timeframe of <20 hours using concentrated oxygen to avoid high post-transportation mortality. To understand the effects of transportation water quality on coral nubbin growth and post-transportation mortality, a series of transportation simulations were performed on Seriatopora hystrix. Different water volumes (dry method, 125 ml, 190 ml, and 325 ml) and oxygen concentrations in the gas phase (21 %, 40 %, and 85 %) were tested, and a 24 LED Handy Lamp was added to provide light irradiance of >100 µmol m-2 s-1 in the transportation box. During transportation nubbins extracted calcium carbonate for growth reducing water alkalinity and consequently pH to 6.67. Dissolved oxygen concentration also rapidly decreased to 0.19 mg L-1. The concentration of inorganic nitrogen, particularly ammonium ions, increased after 24 hours and reached very high concentrations after...
Combined Effect of Changing Hydroclimate and Human Activity on Coastal Ecosystem Health (AMOREIII)
AMORE III addresses the dual control of changing human activity and climate on eutrophication in ... more AMORE III addresses the dual control of changing human activity and climate on eutrophication in the Belgian coastal zone and the feedback effect on goods (newly-deployed offshore mussel farming) and services (atmospheric CO2 absorption). AMORE III has among others assessed the geographical spreading and the magnitude of Phaeocystis blooms in the EEZ, the impact of Phaeocystis colony spreading on offshore mussel farming. It combined field observations, laboratory-controlled experiments, and mathematical modelling in an iterative way.
Imaging-PAM -high-resolution imaging fluorometer Ralph et al. (2005) Journal of Phycology To map ... more Imaging-PAM -high-resolution imaging fluorometer Ralph et al. (2005) Journal of Phycology To map the spatial heterogeneity of photosynthesis; Optimized to determine of the effective quantum yield of photosynthetic energy conversion, ΔF/Fm'.
Posidonia oceanica (L.) Delile, the main Mediterranean seagrass species, has high foliar and belo... more Posidonia oceanica (L.) Delile, the main Mediterranean seagrass species, has high foliar and belowground biomass production (Fig. ). Several methods have been used to measure its growth and primary productivity (incubation bells, optodes, biomass and elementary content measurements etc.). A less used method relies on chlorophyll fluorescence measurements with the Pulse Amplitude Modulated fluorometry (PAM) technique. Underwater fluorometers, designed for studying in situ photosynthesis of primary producers, allow performing Rapid Light Curves (RLCs) that provide a reasonable approximation of the photosynthesis-irradiance relationship. The major parameter associated with RLCs is the maximum relative electron transport rate (rETR max ). The electron transport rate is considered as a good proxy of the plant photosynthetic activity. In the framework of the STARECAPMED project, the objective of this work was to assess the seasonal variation of P. oceanica photosynthetic production using the PAM technique. We measured weekly to bimonthly over a one-year period the photosynthetic production of P. oceanica using RLCs generated with a Diving-PAM (Walz Inc., Germany; Figs. 2). The survey was performed at 10 m depth in a pristine meadow (Calvi, France). To obtain reliable and comparable data, the protocol was standardized. Measurements were performed : on the convex middle part of the 3 rd leaf, at zenith, during shiny and calm weather days. Environmental parameters were measured using PAR sensors and lighttemperature probes. RLC parameters were calculated using the Platt, Gallegos and Harrison (1980) or Webb (1974) models (with or without photoinhibition, respectively). Statistical analyses were done in R.
L'objectif de cette etude est de determiner comment il est possible de coupler des donnees is... more L'objectif de cette etude est de determiner comment il est possible de coupler des donnees issues du FlowCAM analysees a l'aide de Zoo/PhytoImage avec des mesures realisees en cytometrie de flux (a enregistrement des profils optiques - Scanning Flow Cytometry), sur les memes echantillons des systemes d’observation et surveillance en manche et sud Mer du Nord. Dans cette approche preliminaire, nous avons collecte une serie d'echantillons au large des stations du SRN-REPHY d’IFREMER (Dunkerque, Boulogne sur Mer et Baie de Somme) ainsi que sur la radiale de la Baie St Jean (Wimereux-Slack) du LOG. Ces echantillons ont ete numerises avec un FlowCAM et un cytometre de flux de type CytoSense. Les premiers resultats de cette comparaison sont detailles dans le livrable. L'analyse complete grâce au FlowCAM et a Zoo/PhytoImage est detaillee dans le livrable. Afin d'avoir un meilleur point de comparaison entre l'analyse d'image et les profils en cytometrie de flux, ...
Ce livrable detaille les travaux realises pour l'evolution du logiciel Zoo/PhytoImage afin d&... more Ce livrable detaille les travaux realises pour l'evolution du logiciel Zoo/PhytoImage afin d'en optimiser l'usage dans le cadre de l'etude d'echantillons de phytoplancton de maniere generale, et dans le cadre de son utilisation operationnelle pour le monitoring des eaux cotieres tel que realise par le REPHY a l'Ifremer en particulier. Zoo/Phytoimage permet d'analyser des echantillons de plancton fixes numeriquement, c'est‐a‐dire, sur base d'images obtenues a l'aide d'un appareil specialise comme le FlowCAM ou le FastCAM (voir livrable n°3). La classification supervisee (machine learning) permet de classer de maniere automatique les particules imagees dans les differents groupes taxonomiques, et d'en deriver ensuite des statistiques sur l'echantillon tout entier : denombrement, biomasse et spectre de tailles par groupe taxonomique. Deux changements majeurs ont ete introduits en 2015 dans les calculs realises par Zoo/PhytoImage : - l...
The various models for assessment of fisheries dynamics and evaluation of management strategies a... more The various models for assessment of fisheries dynamics and evaluation of management strategies are currently implemented in separate software programs and their respective input and output formats are often incompatible although many are performing similar tasks. Most of these packages provide basic analysis tools (model estimation, graphing, result reporting) that are already available in various software platforms. Comparing the results of such models is difficult and requires exporting them to an environment that has more efficient analytical tools. Moreover, integration of such different models into a single simulation environment that allows evaluation of the whole fishery system has been impossible.
Le systeme couple FlowCAM/ZooPhytoImage est devenu un outil veritablement operationnel en 2014. C... more Le systeme couple FlowCAM/ZooPhytoImage est devenu un outil veritablement operationnel en 2014. Cependant, pour qu'il soit totalement adapte aux observations du phytoplancton realisees dans le cadre du reseau d’observation REPHY, et afin de mieux repondre aux sollicitations presentes et futures concernant l'evaluation de la qualite des eaux littorales et marines dans le cadre des exigences europeennes, telles que la DCE et la DCSMM, des nouvelles fonctionnalites doivent etre integrees aux outils existants. C'est pourquoi, differents axes d'evolution ont ete proposes par l'UMONS et Ifremer pour adapter, a la fois l'appareil de numerisation et le logiciel ZooPhytoImage aux contraintes definies par le REPHY. Premierement, la version 5 de Zoo/PhytoImage contient de recentes nouveautes telles que le developpement de routines pour importer et analyser les donnees automatiquement, et la refonte de l'interface graphique pour une meilleure ergonomie et une simplic...
✔ ✔ Free (open source) software written in R and Java specialized to classify zoo-and phytoplankt... more ✔ ✔ Free (open source) software written in R and Java specialized to classify zoo-and phytoplankton digital images ✔ Machine learning (supervised) classification ✔ Adaptable to analyze any plankton image, e.g., images from the FlowCAM...
The total alkalinity variations, due to the important calcification together with the closed syst... more The total alkalinity variations, due to the important calcification together with the closed system design, are now discussed in the text at p13:l-390-396. The paired design allows, at least, for parallel and identical changes in both acidified and control conditions, which is a good point. We claim that such changes occur in many other closed systems with lots of calcifiers... but they are simply not enough documented in many cases. Yet, it is a weak point where much effort is still required to obtain more stabilization of this parameter over time.
Marine Environmental Research, 2020
Dynamics of the subsurface (2-3 m) mesozooplankton (i.e., > 200 μm) in the Bay of Calvi (Corsica,... more Dynamics of the subsurface (2-3 m) mesozooplankton (i.e., > 200 μm) in the Bay of Calvi (Corsica, France) were explored, combining time series (2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) of 14 zooplankton groups, wind gusts, water temperature, nitrate and chlorophyll-a. Zooplankton data was obtained through image analysis. While contrasted group-specific seasonal patterns were observed, the most productive zooplankton annual event occurred in April (spring peak), concentrating on average 25% of the total annual abundance. A "typical" year was defined based on the annual succession of different community states, highlighting particular years (2007, 2015 and 2012), mainly characterized by weak spring peak. Environmental influences on the interannual variability of zooplankton were explored and while relationship between chlorophyll-a and zooplankton abundance was unclear, the availability of nutrients (December-March), potentially mediated via the wind regime (October-January) seemed to be essential to the occurrence of the spring peak. Additionally, we observed an influence of temperature, with winter thermal thresholds (between 12.1 °C and 13.4 °C) conditioning the spring peak. Also, the occurrence of lower annual abundances after 2010 was synchronous with the sharp increase of seawater warming trend, especially regarding winter temperature (0.30 °C.year -1 ). Finally, winter North Atlantic Oscillation (NAO) was found to be correlated to both winter water temperature and spring peak abundance, which suggests large-scale processes to impact regional zooplankton community.
Empirical Software Engineering, 2018
Nearly every popular programming language comes with one or more package managers. The software p... more Nearly every popular programming language comes with one or more package managers. The software packages distributed by such package managers form large software ecosystems. These packaging ecosystems contain a large number of package releases that are updated regularly and that have many dependencies to other package releases. While packaging ecosystems are extremely useful for their respective communities of developers, they face challenges related to their scale, complexity, and rate of evolution. Typical problems are backward incompatible package updates, and the risk of (transitively) depending on packages that have become obsolete or inactive. This manuscript uses the libraries.io dataset to carry out a quantitative empirical analysis of the similarities and differences between the evolution of package dependency networks for seven packaging ecosystems of varying sizes and ages: Cargo for Rust, CPAN for Perl, CRAN for R, npm for JavaScript, NuGet for the .NET platform, Packagist for PHP, and RubyGems for Ruby. We propose novel metrics to capture the growth, changeability, resuability and fragility of these dependency networks, and use these metrics to analyse and compare their evolution. We observe that the dependency networks tend to grow over time, both in size and in number of package updates, while a minority of packages are responsible for most of the package updates. The majority of packages depend on other packages, but only a small proportion of packages accounts for most of the reverse dependencies. We observe a high proportion of "fragile" packages due to a high and increasing number of transitive dependencies. These findings are instrumental for assessing the quality of a package dependency network, and improving it through dependency management tools and imposed policies.
Aquaculture, 2017
Since a couple of decades, coral aquaculture has been developed in many countries to face an incr... more Since a couple of decades, coral aquaculture has been developed in many countries to face an increasing live coral market and to support conservation of endangered natural coral reefs. To evaluate the potentiality of communitybased coral aquaculture in Madagascar, we experienced suitable farming techniques using the species Acropora nasuta and Seriatopora caliendrum. Survival and growth rate of the nubbins were monitored during wet, warm and dry, cold seasons. To determine economical feasibility, the coral market was investigated and the yields were also calculated using the technical, biological and social parameters of the production. Coral nubbins were reared in situ at appropriate conditions. Coral nubbins reared during the wet, warm season showed a final survival rate of 67 ± 6% and 57 ± 4% respectively for A. nasuta and S. caliendrum, while in the dry, cold season, the survival rates were of 85 ± 7% and 69 ± 1% respectively. A. nasuta had a significantly higher survival rate than S. caliendrum during both seasons. During the wet, warm season, growth rates were 0.46 ± 0.16% d -1 and 0.54 ± 0.16% d -1 respectively for A. nasuta and S. caliendrum. In the dry, cold season, A. nasuta had 0.63 ± 0.18% d -1 of growth rate, while S. caliendrum grew 0.65 ± 0.15% d -1 . Significant difference was observed between both species during the wet, warm season, but not during the dry, cold season. Furthermore, both species grew faster during the dry, cold season. These results are in the range of reference values for corals. The activity can be profitable from 25 coral nubbins sold per month. Profit can already be perceived from the second year and a total of more than EUR 27,000 earned after 5 years of developing project, for an initial investment of EUR 1978. Marine animals wholesale companies and biodiversity conservation NGOs seem to be the appropriate clients for this form of aquaculture on Madagascar. Statement of relevance: This paper provides new form of coral aquaculture: the community-based coral aquaculture. The socially and environmentally responsible production of coral is among the benefits of this new economically viable form of aquaculture.
2016 IEEE 23rd International Conference on Software Analysis, Evolution, and Reengineering (SANER), 2016
When developing software packages in a software ecosystem, an important and well-known challenge ... more When developing software packages in a software ecosystem, an important and well-known challenge is how to deal with dependencies to other packages. In presence of multiple package repositories, dependency management tends to become even more problematic. For the R ecosystem of statistical computing, dependency management is currently insufficient to deal with multiple package versions and inter-repository package dependencies. We explore how the use of GitHub influences the R ecosystem, both for the distribution of R packages and for interrepository package dependency management. We also discuss how these problems could be addressed.
Biogeosciences Discussions, 2014
The design and evaluation of replicated artificial mesocosms are presented in the context of a th... more The design and evaluation of replicated artificial mesocosms are presented in the context of a thirteen month experiment on the effects of ocean acidification on tropical coral reefs. They are defined here as (semi)-closed (i.e. with or without water change from the reef) mesocosms in the laboratory with a more realistic physico-chemical environment than microcosms. Important physico-chemical parameters (i.e. pH, pO 2 , pCO 2 , total alkalinity, temperature, salinity, total alkaline earth metals and nutrients availability) were successfully monitored and controlled. Daily variations of irradiance and pH were applied to approach field conditions. Results highlighted that it was possible to maintain realistic physico-chemical parameters, including daily changes, into artificial mesocosms. On the other hand, the two identical artificial mesocosms evolved differently in terms of global community oxygen budgets although the initial biological communities and physico-chemical parameters were comparable. Artificial reef mesocosms seem to leave enough degrees of freedom to the enclosed community of living organisms to organize and change along possibly diverging pathways.
Journal of Statistical Software, 2013
The concept of structured, interwoven code and documentation has existed for many years, but exis... more The concept of structured, interwoven code and documentation has existed for many years, but existing systems that implement this for the R programming language do not tightly integrate with R code, leading to several drawbacks. This article attempts to address these issues and presents 2 contributions for documentation generation for the R community. First, we propose a new syntax for inline documentation of R code within comments adjacent to the relevant code, which allows for highly readable and maintainable code and documentation. Second, we propose an extensible system for parsing these comments, which allows the syntax to be easily augmented.
Proceedings of the 2015 European Conference on Software Architecture Workshops, 2015
This paper explores the ecosystem of software packages for R, one of the most popular environment... more This paper explores the ecosystem of software packages for R, one of the most popular environments for statistical computing today. We empirically study how R packages are developed and distributed on different repositories: CRAN , BioConductor, R-Forge and GitHub. We also explore the role and size of each repository, the inter-repository dependencies, and how these repositories grow over time. With this analysis, we provide a deeper insight into the extent and the evolution of the R package ecosystem.
2014 IEEE International Conference on Software Maintenance and Evolution, 2014
The R development community maintains thousands of packages through its CRAN archive network. The... more The R development community maintains thousands of packages through its CRAN archive network. The growth and evolution of this archive makes it more and more difficult to maintain packages and their interdependencies, and the existing tools that aim to help developers in this process no longer suffice. We propose maintaineR, a web-based dashboard that allows CRAN package developers to understand and deal with the implications and problems raised by package updates. The dashboard complements existing tools, such as the R CMD check tool, by providing additional support such as the visualisation of package dependencies and reverse dependencies, package conflicts, cross-package function clones, and so on.
Uploads
Papers by Philippe Grosjean