lcolladotor.github.io - this post

How to ask for help for Bioconductor packages | L. Collado-Torres

Lcolladotor.github.io Spined HTML

How to ask for help for Bioconductor packages | L. Collado-Torres Toggle navigation L. Collado-Torres Home Publications Talks Blog CV Students Teaching Contact Search How to ask for help for Bioconductor packages Mon, Mar 6, 2017 4 min read rstats tl;dr Please post your question at the Bioconductor support website https://support.bioconductor.org/ and trammels the posting guide http://www.bioconductor.org/help/support/posting-guide/. It's important that you provide reproducible lawmaking and information well-nigh your R session. Recently I have been getting increasingly questions well-nigh several packages I maintain. It's unconfined to see increasingly interest from users, but at the same time most questions lack the information I need to help the users. I have moreover gotten most of the questions via email, which is why I am writing this post. As of today, I will no longer wordplay questions related to my Bioconductor packages via personal emails. This might sound harsh, but hopefully the rest of this post will convince you that it's the weightier thing to do. You might moreover be interested in the nuts of using derfinder, regionReport or recount, among others. The Bioconductor project is a polity project and it benefits from users interacting in public venues. When a user asks a question at the Bioconductor support website, they are providing information that future users might be interested in. That is, the user (U1) is contributing information to the overall documentation virtually the Bioconductor package they are asking a question about. Ideally, a new user (U2) can then read through the question U1 wrote, trammels the solution, and move on. This is one of the main reasons why we (developers) want questions to be well documented. There are a couple of quick things that U1 can trammels that will make their question much increasingly useful to the community. Session information One of the strengths of Bioconductor is that all the packages have vignettes and lots of documentation. The packages are moreover checked regularly and must pass some tests. That moreover ways that packages can transpiration frequently, at least increasingly commonly than CRAN packages. There's moreover the widow complexity that at any given point in time there is a release workshop and a minutiae branch. This ways that there are many variables and saying that you are using the "latest version" doesn't midpoint much to the developer. All of this information and increasingly is part of the R session information. That is why I and others request users to post their session information. It's very easy to get, simply run the pursuit code: ## Install devtools if needed # install.packages('devtools') ## Reproducibility info library('devtools') options(width = 120) session_info() The output might be too long to post in the Bioconductor support website. The easy solution is to save the information you want displayed in a gist. Then simply add the gist link in your question. Note that you need to have the link under "text" formatting and not "code".Lawmakingto reproduce the error If U1 includes the session information, their question will be pretty good, but not platonic yet. Many of the questions I've been asked do not include lawmaking for me to icon out the word-for-word steps of what they were doing. A lot of times I can infer pieces of what they were doing from their unravelment of the problem. But doing so takes quite a bit of my time and effort, and is still not perfect. Now imagine that U2 is reading through the question: they would probably get lost! There is a wide range of things that U1 could have done. To help the developer, the weightier thing is for the user to include the lawmaking that lead to the error. The lawmaking should include how the data was loaded, so that the developer can run it themselves and trammels in increasingly detail what went wrong. This ways providing a small subset of the data or using some publicly misogynist data. I realize that writing lawmaking that reproduces the error is not easy. But it helps a lot for learning increasingly well-nigh R and Bioconductor. I can tell you that I went through the same process, and in my wits you can find out what you are doing wrong by writing the reproducible code. Extra Here are some other tips that are useful. If you run traceback() immediately without getting the error and include the output in your question, that would be great. It makes it easier to trammels at what point the lawmaking failed and produced the error. Recently when I ask questions myself, I include the "non-evaluated code" (clean lawmaking in your script) and "evaluated code" (think of the R console: a mix of lawmaking and output). The non-evaluated lawmaking makes it easier for others to copy-paste the lawmaking into their R session without having to deal with any formatting issues (example). If you encounter a new error, post a new question instead of "replying" to the first one. Introduce yourself. Be polite. By now you should be ready to post some unconfined questions! Thanks for contributing to the Bioconductor community. Want more?Trammelsother @jhubiostat student and alumni blogs at Bmore Biostats as well as topics on #rstats. Help Related Use subconscious wide arguments for user-friendly functions Trying to reduce the memory overhead when using mclapply Please enable JavaScript to view the comments powered by Disqus. comments powered by Disqus © 2011-2018 Leonardo Collado Torres under (CC) BY-NC-SA 4.0. All thoughts and opinions here are my own. · Powered by the Academic theme for Hugo. × Cite Copy Download