Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merge in https://github.com/datapink/robosat.pink #184

Open
Komzpa opened this issue Sep 20, 2019 · 10 comments
Open

Merge in https://github.com/datapink/robosat.pink #184

Komzpa opened this issue Sep 20, 2019 · 10 comments

Comments

@Komzpa
Copy link

Komzpa commented Sep 20, 2019

Hi,

There's a fork https://github.com/datapink/robosat.pink. How about merging it in and making Olivier, the author, (co)maintainer of the merged thing?

/cc @ocourtin

@ocourtin
Copy link
Contributor

ocourtin commented Sep 20, 2019

Humm, simple solution can only arise since you take all the context dimensions in account...
And so not convinced, about first @Komzpa proposition :)

So about bringing more context informations:

  • First RoboSat release in mid-2018 by Daniel was really impressive,
    and clearly was a game changer in GeoSpatial Computer Vision.
    Kudo to @daniel-j-h !

  • In reaction, my first move, was to became a RoboSat contributor.

  • But as since Daniel left MapBox, my latest PR can’t be applied, and i needed to go on.
    (because DataPink provide commercial support and development on GeoSpatial DeepLearning)

  • So, we began to fork, RoboSat to RoboSat.pink (end of 2018).
    But more than a flavor, it became refactors after refactors, close by the spirit but far away by the code, to initial RoboSat.
    Meaning that there’s no meanigful merging ability anymore. If we would do so, it would be just a pure codebase replace.

  • Few months later (~2019 springtime), Daniel recovers write access on RoboSat, and still maintains it on his spare time.

  • From my (biased) point of view RoboSat.pink clearly becames nowaday a strong enhancement to the initial RoboSat, on all topics: accuracy, speed, standard compliancy, GIS ecosystem friendly, packaging, WebUI, integration tests, modularity, documentation…
    All except one: visibility.

  • I began to realize recently that keeping a so close naming, for two projects as RoboSat and RoboSat.pink, is confusing for people.

  • On a more broder way I’m convinced that all the meaningful Open Source AI4EO solutions (Raster Vision, EOLearn, Solaris, RoboSat, RoboSat.pink, LabelMaker) would better share a common code base, rather than all trying to do more or less the same by their owns.
    And so saving dev time to achieve what could make again a real difference.

  • Historically, mutualisation, happened several time in FOSS4G ecosystem, for instance:

    • ~2006, OpenLayers became the only WebGIS client (since LeafLet years after arise)
    • ~2010, QGIS became the only really option for a decent GIS desktop solution

    And it happened because of:

    • People involved in prior projects were mature enough to look for code source mutualisation
    • Clear and practical governance rules were defined (and followed)

At this point, i see mainly two options:

  1. Renaming RoboSat.pink to something less confusing
  2. Mutualisation with other Open Source AI4EO teams on a common platform

Please make noise if second option is also an option for you...

cc @lewfish, @nrweir, @azupanc, @vincentsarago

@daniel-j-h
Copy link
Collaborator

daniel-j-h commented Sep 20, 2019

Adding from my side here

  • Don't forget @bkowshik who was working with me in private before opening up the original robosat release :)
  • @ocourtin and I are in contact and share ideas and high-level improvements and the projects live a happy life next to each other - no hard feelings ❤️
  • I am no longer with Mapbox and do not have full ownership or admin rights any longer. We thought about just moving robosat to its own project scope and then merging robosat.pink and robosat under this new project. By now the two codebases have diverged quite a bit, though.
  • I work on robosat in my spare time in evenings and on weekends now training on my gpu rig in my living room. I no longer have the capacity to move this project forward fulltime.
  • @ocourtin is way better at making noise than me, for example catch him this weekend at SOTM in Heidelberg (https://2019.stateofthemap.org/sessions/7ZXRXB/) or watch the recorded talk later!
  • I know @batpad and @geohacker are working on unifying some machine learning for geo pipelines with their work at DevSeed

Hope this clarifies the situation in public once and for all :)

@ocourtin
Copy link
Contributor

ocourtin commented Feb 2, 2020

At this point, i see mainly two options:

1. Renaming RoboSat.pink to something less confusing
2. Mutualisation with other Open Source AI4EO teams on a common platform

Please make noise if second option is also an option for you...

cc @lewfish, @nrweir, @azupanc, @vincentsarago

And it's how RoboSat.pink becames Neat-EO.pink
www.neat-eo.pink

@vpicavet
Copy link

Hi, I stumbled upon this issue and wanted to let you know that sadly, Olivier passed away, and therefore neat-eo / robosat.pink would not be developed any longer ( neat-eo source code is now offline ).

@vincentsarago
Copy link

Thanks for letting us now @vpicavet, this is really a sad new I met Oliver last year and was really impressed with the energy he had.

@geohacker
Copy link

Thank you @vpicavet. I had a chance to meet Olivier in Heidelberg and had several great conversations. I'm sad to hear this. His work was really impressive and will be missed.

@daniel-j-h
Copy link
Collaborator

I'm devasted to hear this 😮 Olivier's passion was limitless 🌃

@jqtrde
Copy link
Contributor

jqtrde commented Nov 23, 2020

☹️, sorry to hear this.

@Geoyi
Copy link

Geoyi commented Nov 23, 2020

Thank you for sharing the news, @vpicavet.
So sorry to hear this 💔 and R.I.P Olivier.

@climbthemt
Copy link

RIP Oliver, sad to hear.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

9 participants