Apply to be an OSGeo Project #44
Comments
|
Hi, @Robinlovelace ! As I could understand the idea is to submit a description of the project to the OSGeo Incubation Committee to become an OSGeo Community Project. Right? I believe that would be interesting to confirm which packages' authors agree to this proposal. How could I help? |
|
Thanks for getting in touch on this @FelipeSBarros, great you're up for helping out.
Correct. Yes you could certainly help. Next step I think is to draft an email to be sent to OSGeo. If you could put some ideas for that below - what are the key things we should say, that we've had long links with OSGeo packages, GDAL and PROJ in particular but also, dating back to 2000 links with GRASS thanks to @rsbivand. We're an open and supportive community looking to affiliate. Sure there are other things to say. Then someone just needs to send the email and, as I say, happy to do that unless anyone else wants to. |
|
@Robinlovelace, thanks for keeping the ball rolling. I wanted to reply at least to mention I think this project is important. I'd like to help but I'm currently stretched thin so I can't really contribute before the end of the year. Please keep us informed. |
|
Dear @etiennebr and @Robinlovelace : |
|
Great, whoever gets round to drafting a message first - it's on various busy to-do lists, including mine, by the sounds of it! |
|
A heads' up wrt. an ongoing thread on the gdal-dev list. It started with Even Rouault asking which drivers could be dropped from GDAL to ease maintenance, and has widened to cover the struggle independent FOSS developers have to secure income. It strikes me that if we create an OSGeo community linked in addition to R Consortium, we should be able to link for-profit (and other) users of R-spatial packages and whatever mechanism GDAL/PROJ/GEOS devise for soliciting funding. |
|
Excellent idea! |
|
@rsbivand you're probably referring to this thread ? |
|
Great to see this rejuventated. Happy to help. |
|
Yes, and https://lists.osgeo.org/pipermail/gdal-dev/2021-January/053260.html, which started it. |
|
Just reading this in more detail. So your suggestion @rsbivand here
is to apply to be an OSGeo community, as per the opening post? Reading those threads there's no suggestion that OSGeo has greatly helped securing funding. But if there is still 👍 on the idea I can action to take it forward. |
|
I know that OSGeo probably hasn't greatly helped securing funding. The thread touched on difficulties for US corporations to give donations rather than pay invoices. The R Consortium knows how to handle money contributions, better than the R Foundation. It is probably a better channel with knowledge of how to do things. By establishing an OSGeo community we would I think signal that we heard what I think I recall was said by R Consortium, that R-spatial needs a coordination structure, from which flowed the idea of an OSGeo community, a bridge between the two larger fields. Moving on this could also be fed into the GDAL discussion, indicating that we acknowledge the need to try to do something. There are lots of labs, courses, etc., using R-spatial software building on OSGeo core libraries, and ways of letting them help keep things afloat seems worth trying, even if it is mostly promoting the idea of co-responsibility, if that makes sense. |
|
I'm drafting an email here: https://hackmd.io/oiak_4C-SAKIgNneT7w3Pw?both |
|
First draft below (you should be able to edit the doc in the link above): Application to become an OSGeo CommunityDear OSGeo Incubation Committee, We would like to apply, as the 'R-Spatial' community, to become an OSGeo affiliated organisation. We are a diverse group with a shared interest in developing free and open tools for the reproducible analysis of geographic data. R is a popular and rapidly growing language for statistical computing and 'data science'. It is alreay part of OSGeo ecosystem: R ships with the OSGeo Live distribution, integrates with established OSGeo projects such as GRASS, and (now slightly dated) tutorials listed on OSGeo's old website. After a discussion on our GitHub Organisation at github.com/r-spatial, it is clear that closer links could be mutually beneficial. Collaboration is at the heart of open source software and the R community has a long history. The history of R-GRASS bridges, for example, covers more than 10 years and goes in both directions. R interfaces enable OSGeo projects to be accessed, from the command line, to a wide range of people, as illustrated most recently by the qgisprocess package. Perhaps most critically, we as a community rely on the OSGeo projects GDAL and GEOS for data access and geographic operations. We would like to support the ongoing work of these vital components of the wider community that is represented by the OSGeo-affiliated conference series FOSS4G. We also anticipate benefits from being part of the wider OSGeo community and would like to be more active members. Specifically, we would like to initiate the three-step process need to become a full OSGeo project, as outlined on the Incubation Committee web page:
All the best, R-Spatial developers Should I commit this as an .md document somewhere so we can track changes? Happy to send the email but equally happy for someone else to. |
|
@Robinlovelace An .md document for shared editing/suggesting would be helpful. |
|
Thanks for the nudge @neteler. I kept meaning to get back to this and you're input, as someone heavily involved in the successful OSGeo project GRASS, is the ideal motivation for me to get on it. More soon... |
|
Heads up everyone, especially @neteler, @edzer and @rsbivand, here is the letter in editable form: And in an even easier-to-edit form (I tentatively suggest people make edits here as it's quicker): https://hackmd.io/@IGN2QvZIQXKkT2fHo6wX6g/ryoViGWZu How's this as a plan: give people a week or so to make edits and then send this on Weds next week? |
|
I think the draft email is coming together, latest version: https://github.com/r-spatial/discuss/blob/ad039518141d654d2323a513dbc0986fc8c5b695/osgeo-email.md Currently it ends saying "R-Spatial developers:", raising the question, who? That raises the wider question of what is R-Spatial. Important to define it. I can of 3 broad options: r-/rspatial org packages, all spatial packages or something else. Thoughts? Does r-spatial have a formal 'onboarding' process, how do people submit their packages, do we do peer review? In any case I've created a default narrow definition that is the first option above:
Thinking: simplicity is good and it's easier to expand definitions than constrain them, so hoping this is future-proof approach but very happy to hear separate suggestions. And what should we write after 'developers' at the end of the message? One idea on that: authors + contributors to the r-spatial/rpatial packages. |
Another option that I think could be good: anyone who sees this and wants to get involved. Sometimes ad-hoc and works well. |
|
Assuming everyone's happy for me to send this tomorrow - would be great to get explicit 👍 from @neteler, @edzer and @rsbivand - I think it is worth revisiting the definition of 'r-spatial' I see it as r-spatial/rspatial packages at a minimum, something that could be expanded later down the line. I terms of sign-off, I'd suggest just: R-Spatial Developers |
|
Sorry for the late feedback. The email sounds good, but I wonder if we wanted to mention RCosnsortium somehow? After all it sparked this whole discussion. But maybe that was already discussed and I missed it... If so, just ignore me :-) |
Good thinking. Will add and also give a chance for others to add their names/comment - may not be until Friday I send it now as a busy next 2 days and don't want to rush it. |
I don't think that is needed for this application; we can (and should) contact RC again once we are an OSGEO community project. I made some further edits, also pointing to the list of all (?) packages directly linking to OSGEO libraries. The third bullet point (which I edited before from "project" to "community project") is now unclear, and should be removed if what we want is become a community project, rather than a (full) project. That is what we want, right? I also added names, as I think a letter only exists if signed by persons; I now added Robin, Roger, me, Tim, and Robert. Please add your name (here, or in the letter) if you want to be in that list. |
|
The letter is btw in https://github.com/r-spatial/discuss/pull/46/files , the link in @Robinlovelace 's message above is an old snapshot. |
|
Dear all, jumping into this ticket (since i was watching more that one on "working group" where i've provided a comment here #37 (comment) ) As mentioned there i would like to contribute into the inception of such project, but at this stage it is not clear to me what is the exact scope. Do the r-spatial / rspatial intend to federate beyond its current scope of packages they manage? and involve in the loop other developers? Indeed they are many R package initatives dealing with spatial that are not part of the r-spatial/rspatial Github organizations, and that I believe deserve being under the radar of such OSGeo project proposal, at least through the OSGeo project, even if they are not managed in the rspatial/r-spatial communities that are emphasized in the current proposal. Some packages are managed in other Github organizations, sometimes institutional, or individual accounts, for different reasons, sometimes this is done for legacy reasons. Some packages are not even on Github, but elsewhere. Wouldn't be better to set-up a proper community for this OSGeo project proposal, as umbrella to catch R spatial-related packages (wherever their project homepage is)? In case you may want to relate explicitely the OSGeo project proposal to R-Consortium, then maybe it would be good to mention which spatial projects (although scattered) already received support from R Consortium; Behind the scene, it could be useful as well to inventory projects that are not accepted with the main reason being that RConsortium argued for a more federated approach for spatial R projects. This would be also a good opportunity to exchange further on the different initatives (where sometimes there are clear/evident technical synergies), so the different R spatial sub-communities / clusters can understand each other, and see what is ongoing, what the development perspectives, what are the user communities behind, etc. Last point i'd like to add (Edit): IMHO i'm not sure that R Consortium should be mentioned explicitely for OsGEO. It is surely one known funder of R package initiatives, but they are more, including from national institutions and international organizations that do rely on opensource software and more and more on R software, especially for spatial matters. These organizations do have their own community, and some are contributing (through coding and/or funding), they may be interested as well participating to this project, opening more perspectives for strenghtening the R spatial support. Looking forward to your feedback/thoughs on this, Best regards, |
|
Dear Emmanuel, this could be one of the things we want to clear up, but maybe not before submitting this proposal. I don't have in mind to restrict the scope to these to github orgs, but to use it as a starting point. I think there are two "levels" of R-OSGeo involvement:
They are different because developers of pkgs of the second category typically consult the developers of the package of the first category in case of troubles. Developers of the first category need to deal with the OSGeo developers directly. As you see (here: only addressing GDAL/PROJ/GEOS) a lot of the packages in category 1 are not under the two github orgs mentioned in the letter. There is no need or intention to exclude anyone now, but I think there is a need for some starting point (which software is involved), and to be identifiable (which people). Please feel free to add your name to the list. |
Important point and great to link to this list of packages that interact with OSGeo software directly. There are many more packages in category 2 than category 1 and I think it's good to include both in the 'R-spatial' definition. Looking at the latest version here now https://github.com/Robinlovelace/discuss/blob/patch-1/osgeo-email.md |
Yes I think so @edzer. I've removed that third bullet point now. |
|
In addition to the points made by @Nowosad , this sentence: "As an OSGeo supported community software projects however, we define R-Spatial ..." should have "project" not "projects". OSGeo user name rsbivand if you'd like to add me as involved. |
|
Great to see this is live! Just had a look at https://www.osgeo.org/projects/r-spatial/ In terms of who's involved, please add me and others if possible to the list. I'm robinlovelace on OSGeo also. There is a longer but also incomplete list here: https://github.com/r-spatial/discuss/blob/master/osgeo-email.md that contains the following who signed the doc 2 years ago:
There is a broader question about how we operate as a community but that can wait and for now the main thing I want to say is great work and many thanks everyone involved. |
|
By the way, I indicated involvement by editing my OSGeo profile page, unless @nickbearman got to it first. |
You just motivated me to update my profile, untouched after several years, profile pic added!
Just checked the demo on OSGeo Live and seems someone has already updated it compared with last time I checked https://live.osgeo.org/en/quickstart/R_quickstart.html From my perspective this issue is done so happy for it to be closeed 🎉 although may want to keep it open for a bit for others to comment. |
|
Hi all,
Just catching up on all of this.
@nickbearman great job! Happy to be listed / involved (fcorowe)
… On 29 Aug 2022, at 15:10, Robin Lovelace ***@***.***> wrote:
By the way, I indicated involvement by editing my OSGeo profile page, unless @nickbearman <https://github.com/nickbearman> got to it first.
You just motivated me to update my profile, untouched after several years, profile pic added!
<https://user-images.githubusercontent.com/1825120/187220360-4ac60203-dee3-4552-b5b0-3fa9e22b48bd.png>
Just checked the demo on OSGeo Live and seems someone has already updated it compared with last time I checked https://live.osgeo.org/en/quickstart/R_quickstart.html <https://live.osgeo.org/en/quickstart/R_quickstart.html>
From my perspective this issue is done so happy for it to be closeed 🎉 although may want to keep it open for a bit for others to comment.
—
Reply to this email directly, view it on GitHub <#44 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGZB3IRHF4JIP3V3Y6PD7N3V3TAGJANCNFSM4R4CMHCQ>.
You are receiving this because you were mentioned.
|
|
Thanks @fcorowe unfortunately you have to add yourself, I can't do this for you. You can do this by editing your OSGeo profile page - shout if you need help. |
|
Cool - done!
… On 30 Aug 2022, at 09:43, Nick Bearman ***@***.***> wrote:
Thanks @fcorowe <https://github.com/fcorowe> unfortunately you have to add yourself, I can't do this for you. You can do this by editing your OSGeo profile page - shout if you need help.
—
Reply to this email directly, view it on GitHub <#44 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGZB3ITJFSXSDB6Y77JMW4LV3XCTNANCNFSM4R4CMHCQ>.
You are receiving this because you were mentioned.
|
|
Thanks for the comments. I have tweeted at https://twitter.com/NickBearmanUK/status/1564537622302756864 |
|
I think it might be worth adding a README page to the GitHub organisation to give us a bit of a landing page (see #55). This is partly related to an email Jody sent to me and circulated to the incubation list:
The implication being we should have a code of conduct. On the readme.md file we could include the text from the OSGeo Project page: Clearly I want to avoid getting too bureaucratic but I think we do need to tick a couple of boxes for OSGeo. |
|
Maybe a final sentence something like: "We also understand R-Spatial as extending to responsibility for supporting and guiding the authors and maintainers of packages depending on these core infrastructure packages for their functionality." Improvements welcome! |
|
Sorry for the late reply. You can also add me as a contributor. Do I need a osgeo account for that? |
No problem, yes you do need an OSGeo account to be added as a contributor. Any problems setting one up, let me know. |
|
Another email from Jody - I think we might have jumped the gun slightly, but it's not a major problem. This is a bit of a long list, apologies. I have replied, and will add in more details. Some questions for us:
Yes - we can add this Does anyone have a good example? I can supply one if not
I will check and confirm I want to say 'yes' to this, but need confirmation from someone who is more familiar with the project
Any more links to add? These are all by project, so I think what you have is fine. I will check.
I will check about this. We might rearrange to a more logical layout (OSGeo R-Spatial Project libraries, and links to other libraries). How does this sound? Then we can say 'All OSGeo R-Spatial libraries are open source, check each library for license details' or something similar.
We have a range of individuals - can we add these similar to the 'Who's involved' list? Or do they have to be added to OSGeo as an explicit Service Provider? I will also be adding myself as a service provider. See here for details: https://www.osgeo.org/community/getting-started-osgeo/add-service-provider/ There's also some SF things to think about:
We can work on this @edzer I know you had a fairly long discussion about this, but I can't remember what the outcome was. What are you views on this? Can we just add the license text into the relevant file? Or have I missed something key?
I'm not sure what he means by this. Do you know?
Yes, this is something we can add |
|
Any comments on the above? I know there is a lot there, sorry! |
Licenses are clearly stated, and linked to the original texts, on the CRAN landing pages. The R community doesn't try to copy around standardised license texts. As an example: https://cran.r-project.org/web/packages/sf/index.html states GPL v-2 or MIT (pick your favorite); the LICENSE file mentioned there just states name & year, which are compulsory parts of the MIT license. Can we mention this somewhere? I don't see any point in trying to copy parts of the license information. |
|
Jody simply does not understand that we are not a single piece of software, but a cluster of > 1000 software packages written in R, meeting CRAN standards, with a subset listed in the Spatial, SpatioTemporal and Tracking CRAN task views. Licence conditions are set and policed by CRAN. The "headers" are the top lines in files expressing copyright, CRAN assumes that the copyright is that of authors and contributors in DESCRIPTION unless otherwise stated. |
|
I have not forgotten this, but it is towards the bottom of my list. Just as a note, we had a similar issue with license for a JOSS article: JOSS have x license requirements and CRAN have y license requirements! See here for discussion: jlacko/RCzechia#55 And here for the overall review: openjournals/joss-reviews#5082 |
Would it be helpful to create a r-spatial community at Zenodo in order to group and visualize packages which their authors wish to flag as 'r-spatial'? Zenodo serves to preserve package releases (their source code) for the 'very' long term and provides DOIs. BTW such a community cannot serve as a package installation repo, it's just for collecting and preserving. This can be automated in GitHub, i.e. triggering a Zenodo deposit (next version of same record) when creating a new release. Such a Zenodo community – essentially a curated collection of Zenodo records – could group packages from the two GitHub organizations and perhaps other packages not residing in github.com/r-?spatial (maintained under an author's profile, in GitLab, ...).
|
|
As a side note, most rspatial/r-spatial packages repos are already fully archived by the Software Heritage archive. mapview: But, AFAICT, Software Heritage offers no specific "community" feature. |
|
Thanks for drawing attention to Software Heritage @rCarto. It seems that the Software Heritage archive harvests everything from GitHub and other platforms and in some way serves as a mirror to it (providing all commits). Indeed, no community or release-DOI feature. It reminds of rdrr.io, e.g. https://rdrr.io/github/r-spatial/mapview/ (though this is latest-version-only I think), and adds the long-term preservation. |
I never do GitHub releases, I find it a hassle and not worth the effort. What is wrong with releasing packages on CRAN? Why does every release need a DOI, and what is the value of the DOI if it doesn't point back to the CRAN release? |
|
BTW: It might be worth applying |
|
A Zenodo deposit is certainly no substitute for a CRAN release (Zenodo has no native integration with R), it is complementary at most, but its metadata can point to the CRAN release and are version-specific. Version-specific DOIs are relevant for stable referencing since each version at Zenodo has its own page, hence URL and DOI. Compared to CRAN, Zenodo merely adds long-term preservation and referencing. With relation to collections of spatial packages, a Zenodo community may have the risk to further complicate the landscape of already existing overviews, such as the Spatial CRAN Task View, although archival is not the focus of the latter and a Zenodo community would at most be a subset of it. Also, implementing Zenodo does take some extra (one-time) effort to setup for a package maintainer. On the positive side, I think it could be a way to visualize packages as belonging to OSGeo R-Spatial, which was the primary reason to post the idea here. |
|
@florisvdh are you volunteering? |
|
@edzer I think it first has to be sorted out whether the OSGeo R-Spatial community members are interested to make use of a Zenodo community. But if so, then yes I'd be happy to assist. For example by making such a community, and providing some guidance to package maintainers, such as a template An example of an R package Zenodo record is https://doi.org/10.5281/zenodo.7181250. Another thing to add is that Zenodo communities can up to now not be curated by multiple people. That is to change, probably this year (zenodo/zenodo#810). |
|
I just noticed that https://www.osgeo.org/projects/r-spatial/ is up. So we are on board. Does anyone know how to edit the SEO tags? Where might the sources to the website live? It isn't https://github.com/OSGeo/osgeo. Reason for asking: the header includes SEO tags to |
I think this issue is well and truly fixed. Great work everyone involved! I'm not sure how to edit tags Roger, any ideas @nickbearman? |
I have access to the page so can make any updates. Tags now removed - let me know if there are any other changes. Happy to call this issue 'closed'. However, it is worth noting we are not officially a 'OSGeo Community Project'. We are in a no man's land of 'Project Type = None' (as listed in the WordPress backend of the site). I think we never managed to square the circle of OSGeo being setup for an individual projects, and r-spatial being a group of projects (see Jody's comments pasted in above). Happy to hear others views, but I suggest we stick with things as they are for the moment. One way forward would be for individual libraries to register with OSGeo as Community Projects - and I'm potentially happy to help with this. |
Building on previous discussion this is a specific suggestion to follow the guidance here to become an OSGeo Community Project: https://wiki.osgeo.org/wiki/OSGeo_Community_Projects
To become a project I think we need to
Looking at the archives it seems no request has been made for an R-Spatial Community project: https://lists.osgeo.org/pipermail/incubator/
Please comment on this thread to avoid going further off topic in #37. I'm up for giving it a 1st bash unless anyone else involved in the community would like to. I think as a starting point 'the repos in the r-spatial github org that are on CRAN' is fine and we can cross the bridge of how to add other packages such as raster/terra when we get to it and ask it as an open question (I'm in favor of including those assuming package authors agree).
The text was updated successfully, but these errors were encountered: