search

ropensci-archive / dataone-restful

:no_entry: ARCHIVED :no_entry:
0 stars 0 forks source link

Project status / Tasks #4

Open emhart opened 10 years ago

emhart commented 10 years ago

At the DUG, and there's been a few mentions about tools for D1 in R so a few questions about the R Package came to mind. I went poking around the repos, and I'm curious about a couple of things that I'm hoping @cboettig or @mbjones might be able to weigh in on.

mbjones commented 10 years ago

@emhart emhart The official package is the DataONEorg/rdataone package. It is the one from which the 1.0 version was released to CRAN, and is the version on which I have been (slowly) making changes to move to supporting a native client using httr. The actual httr branch was experimental and was merged into master, and so is there now only for historical reference (I dislike the practice of removing branches after merge).

The 1.0 version is working fine, so in that sense no tasks remain for it to be useful. That said, the issue tracker lists tasks remaining for the 2.0 release, which would include a fully native rewrite. The main tasks there are encapsulated in DataONEorg/rdataone#6, DataONEorg/rdataone#2, and DataONEorg/rdataone#3.

@cboettig helped stimulate this work, which is why he created the rOpenSci version of the repo. I've incorporated a similar approach in the MNode and CNode classes in the DataONEorg/rdataone repo. Maybe @cboettig can comment on where he wants to go with this rOpenSci version of the repo. I'd hope we could maintain a single main repo.

I'm hoping we can work on this at the Open Science Codefest that I mentioned to you. I'm planning that for September and hope to have an announcement out this week.

karthik commented 10 years ago

At some point we'll only have one repo, not two. So we should discuss removing the rOpenSci one at some point if it's outdated.

cboettig commented 10 years ago

yeah, we should probably delete github.com/ropensci/rdataone when Matt's package is released, if not before. (I only hesitate right now since our current package is mildly useful as a java-free alternative until a working version of the other package is released. I noticed that an r dataone package interface is one of the things listed on the NCEAS hack-a-thon website ideas)

On Mon, Jul 7, 2014 at 9:40 AM, Karthik Ram notifications@github.com wrote:

At some point we'll only have one repo, not two. So we should discuss removing the rOpenSci one at some point if it's outdated.

— Reply to this email directly or view it on GitHub https://github.com/ropensci/rdataone/issues/4#issuecomment-48203733.

Carl Boettiger UC Santa Cruz http://carlboettiger.info/

karthik commented 10 years ago

@cboettig: for the time being we can make this repo private in case anyone finds it confusing to see two repos (although so far no one seems to have noticed or complained). That way everyone that needs access to it can have it and still be able to push changes etc.

@mbjones: Would you be opposed to moving your rdataone repo over to our organizational account? We request this from all rOpenSci package maintainers since it allows us to curate everything in one place rather than have packages scattered all over GitHub. It also allows for greater discoverability since people come to one package then browse around our org.

mbjones commented 10 years ago

@karthik I guess we could consider moving the package. But it seems like you'll run into this issue more as you grow. DataONE, like rOpenSci, is trying to promote the idea of our network, and so its helpful to maintain some degree of consistency in our presentation. All of our other repositories are hosted in DataONE github and svn repositories. We're planning on migrating our many other svn repos to the DataONEorg github account. Once we do, people will see that there are other DataONE clients available for python, java, the commandline, and other environments. So its nice if those are grouped together. Are you willing to help contribute to the package in its current location, or is the rOpenSci branding a show stopper? We can certainly add rOpenSci to the README and other places where we provide attribution. If it is a problem, then I'd have to talk over that move with the DataONE team.

cboettig commented 10 years ago

I don't see any reason for Matt's repo to move at this time. If the deprecated ropensci version is causing confusion, certainly making it private or removing it makes sense.

Matt, having all the DataONE (and/or NCEAS?) repos on the DataONE GitHub account would be great; as I mentioned earlier; sounds like you've figured out a strategy to have the svn ones still on GitHub as well? I certainly think that helps for discover-ability / lowering the barriers of contribution. Even though I started on svn I'm not sure I'm up for merging large code with it ...

Carl Boettiger http://carlboettiger.info

sent from mobile device; my apologies for any terseness or typos On Jul 15, 2014 7:07 PM, "Matt Jones" notifications@github.com wrote:

@karthik https://github.com/karthik I guess we could consider moving the package. But it seems like you'll run into this issue more as you grow. DataONE, like rOpenSci, is trying to promote the idea of our network, and so its helpful to maintain some degree of consistency in our presentation. All of our other repositories are hosted in DataONE github and svn repositories. We're planning on migrating our many other svn repos to the DataONEorg github account. Once we do, people will see that there are other DataONE clients available for python, java, the commandline, and other environments. So its nice if those are grouped together. Are you willing to help contribute to the package in its current location, or is the rOpenSci branding a show stopper? We can certainly add rOpenSci to the README and other places where we provide attribution. If it is a problem, then I'd have to talk over that move with the DataONE team.

— Reply to this email directly or view it on GitHub https://github.com/ropensci/rdataone/issues/4#issuecomment-49115824.

karthik commented 10 years ago

@mbjones It's definitely not a deal breaker. Your reason for building a NCEAS/DataONE network on GitHub sounds perfectly reasonable to me. We would be happy to continue contributing to the package it its current location.

mbjones commented 10 years ago

OK, great. Then let's leave it, but I will add ROpenSci as a co-maintainer, and make sure your names are listed in the authors list.