So I recently saw someone fork my kernel, add one fairly trivial grid search to it, and post it with a different name but the exact same in-kernel comments and headers without any acknowledgement. Now, my original kernel https://www.kaggle.com/delayedkarma/let-s-add-some-features-for-the-national-holidays() was beholden to another one as well (https://www.kaggle.com/chauhuynh/my-first-kernel-3-699), but I always make sure people know what I'm sourcing my material from.
\n\nWhat does the community think of ethical forking practices? And are there any guidelines to this effect.
", "messages": [ { "id": "436758", "postDate": "12/10/2018 21:58:30", "content": "So I recently saw someone fork my kernel, add one fairly trivial grid search to it, and post it with a different name but the exact same in-kernel comments and headers without any acknowledgement. Now, my original kernel https://www.kaggle.com/delayedkarma/let-s-add-some-features-for-the-national-holidays() was beholden to another one as well (https://www.kaggle.com/chauhuynh/my-first-kernel-3-699), but I always make sure people know what I'm sourcing my material from.
\n\nWhat does the community think of ethical forking practices? And are there any guidelines to this effect.
", "rawMarkdown": "So I recently saw someone fork my kernel, add one fairly trivial grid search to it, and post it with a different name but the exact same in-kernel comments and headers without any acknowledgement. Now, my original kernel https://www.kaggle.com/delayedkarma/let-s-add-some-features-for-the-national-holidays() was beholden to another one as well (https://www.kaggle.com/chauhuynh/my-first-kernel-3-699), but I always make sure people know what I'm sourcing my material from. \n\nWhat does the community think of ethical forking practices? And are there any guidelines to this effect.", "votes": null }, { "id": "436762", "postDate": "12/10/2018 22:39:52", "content": "Just think of Kaggle as a giant, shared pool of open source code. I don't think there are any rules.
", "rawMarkdown": "Just think of Kaggle as a giant, shared pool of open source code. I don't think there are any rules.", "votes": null }, { "id": "436764", "postDate": "12/10/2018 22:40:49", "content": "Not rules really.. rules and open source don't work well together.. more like ethical guidelines...
", "rawMarkdown": "Not rules really.. rules and open source don't work well together.. more like ethical guidelines...", "votes": null }, { "id": "436769", "postDate": "12/10/2018 23:09:40", "content": "Hi Panchajanya,
\n\nYou are welcome to keep your Kernels private if you do not wish for it to be shared or potentially forked. As noted, a public kernel is open-source, and the Kaggle community likes to learn from shared knowledge about a problem statement, whish is where kernel forking may occur.
", "rawMarkdown": "Hi Panchajanya,\n\nYou are welcome to keep your Kernels private if you do not wish for it to be shared or potentially forked. As noted, a public kernel is open-source, and the Kaggle community likes to learn from shared knowledge about a problem statement, whish is where kernel forking may occur.", "votes": null }, { "id": "436772", "postDate": "12/10/2018 23:12:25", "content": "Hey Addison,
\n\nI completely understand that. I just wanted to ask if there are any moral gray areas here. But i suppose once a kernel is indeed public, it's all good.
", "rawMarkdown": "Hey Addison, \n\nI completely understand that. I just wanted to ask if there are any moral gray areas here. But i suppose once a kernel is indeed public, it's all good.", "votes": null }, { "id": "436797", "postDate": "12/11/2018 00:40:17", "content": "I'm thinking maybe add a line at the top to ask people to reference back if they build based on the kernel?\n I think people should at least mention the original post no matter this is 'open source' or not...
", "rawMarkdown": "I'm thinking maybe add a line at the top to ask people to reference back if they build based on the kernel?\n I think people should at least mention the original post no matter this is 'open source' or not...", "votes": null }, { "id": "436799", "postDate": "12/11/2018 00:50:12", "content": "That was basically my point. All papers on arxiv are open source, but that doesn't mean they shouldn't be cited...
", "rawMarkdown": "That was basically my point. All papers on arxiv are open source, but that doesn't mean they shouldn't be cited...", "votes": null }, { "id": "436803", "postDate": "12/11/2018 00:59:51", "content": "Yes. I saw some refer back and emphasis the original post but some don't. My suggestion is don't let this ruin your day :)
", "rawMarkdown": "Yes. I saw some refer back and emphasis the original post but some don't. My suggestion is don't let this ruin your day :)", "votes": null }, { "id": "436805", "postDate": "12/11/2018 01:05:12", "content": "Haha na.. all good.. it just took me by surprise is all!
", "rawMarkdown": "Haha na.. all good.. it just took me by surprise is all!", "votes": null }, { "id": "436858", "postDate": "12/11/2018 03:11:49", "content": "I agree it is polite to credit the kernels you fork. I rarely upvote kernels that don't credit their sources and I encourage others to do the same. But this is my ethics, not law, and I'm fine with Kaggle being the way it is -- I'm not asking for a rules change, just a norms change.
", "rawMarkdown": "I agree it is polite to credit the kernels you fork. I rarely upvote kernels that don't credit their sources and I encourage others to do the same. But this is my ethics, not law, and I'm fine with Kaggle being the way it is -- I'm not asking for a rules change, just a norms change.", "votes": null }, { "id": "436859", "postDate": "12/11/2018 03:13:42", "content": "Yep.. Rules are bad, ethical norms are usually nice.
", "rawMarkdown": "Yep.. Rules are bad, ethical norms are usually nice.", "votes": null }, { "id": "437159", "postDate": "12/11/2018 13:19:36", "content": "Atleast the person forked your kernel. My kernel got copy & pasted ... what - in my opinion - is not ok.
", "rawMarkdown": "Atleast the person forked your kernel. My kernel got copy & pasted ... what - in my opinion - is not ok.", "votes": null }, { "id": "437284", "postDate": "12/11/2018 17:11:50", "content": "Damn. That is not nice.
", "rawMarkdown": "Damn. That is not nice.", "votes": null }, { "id": "438222", "postDate": "12/13/2018 09:48:46", "content": "Addison,
\n\nI don't think Panchajanya is objecting to forking or sharing code. I think it is more a question of whether etiquette or a code of conduct should be enforced around copying people's work without credit.\nAs you can imagine, it's not a nice feeling to have someone take your work, rebadge it as their own without giving credit to the brain cycles you put into it.\nSure, if they're not happy with the current state, people can make their kernel's private but then the community loses out on a lack of information sharing.\nPlease consider enforcing a code of conduct around this :)
", "rawMarkdown": "Addison,\n\nI don't think Panchajanya is objecting to forking or sharing code. I think it is more a question of whether etiquette or a code of conduct should be enforced around copying people's work without credit.\nAs you can imagine, it's not a nice feeling to have someone take your work, rebadge it as their own without giving credit to the brain cycles you put into it.\nSure, if they're not happy with the current state, people can make their kernel's private but then the community loses out on a lack of information sharing.\nPlease consider enforcing a code of conduct around this :)", "votes": null }, { "id": "438298", "postDate": "12/13/2018 12:46:48", "content": "I think a fork should automatically translate into an upvote, so the original author would receive some credit toward their kernel status. Or alternatively, directly include \"forking\" in kernel expert/master/grand master criteria
", "rawMarkdown": "I think a fork should automatically translate into an upvote, so the original author would receive some credit toward their kernel status. Or alternatively, directly include \"forking\" in kernel expert/master/grand master criteria", "votes": null }, { "id": "438364", "postDate": "12/13/2018 14:37:54", "content": "Yep. A public fork should definitely count as an upvote
", "rawMarkdown": "Yep. A public fork should definitely count as an upvote", "votes": null }, { "id": "448390", "postDate": "01/01/2019 01:31:33", "content": "Great point! I don't think the guidelines for Apache 2.0 are in effect with public Kaggle kernels.
\n\nAs scientists we should take care to cite our sources when using someone elses work. It is a sign of respect.
\n\nI also think it is decent to follow the golden rule:
\n\n\"One should treat others as one would like others to treat oneself.\"
", "rawMarkdown": "Great point! I don't think the guidelines for [Apache 2.0][1] are in effect with public Kaggle kernels.\n\nAs scientists we should take care to cite our sources when using someone elses work. It is a sign of respect.\n\nI also think it is decent to follow the golden rule: \n\n\"One should treat others as one would like others to treat oneself.\"\n\n\n [1]: https://www.apache.org/licenses/LICENSE-2.0", "votes": null }, { "id": "448584", "postDate": "01/01/2019 15:20:36", "content": "I disagree, I explained why in a similar discussion in another competition. I may fork a kernel, try it, and decide it is not worth it. Why should I upvote it then?
\n\nThis said, I totally agree with the post: sources should be acknowledged, and added value should be clearly identified.
\n\nedited: removed the offending example as the fork author acknowledged the source.
", "rawMarkdown": "I disagree, I explained why in a similar discussion in another competition. I may fork a kernel, try it, and decide it is not worth it. Why should I upvote it then?\n\nThis said, I totally agree with the post: sources should be acknowledged, and added value should be clearly identified. \n\nedited: removed the offending example as the fork author acknowledged the source.", "votes": null }, { "id": "448604", "postDate": "01/01/2019 16:22:54", "content": "I understand the situation you describe, but I would argue by and large people fork codes that are either well written or at least written with some merit. then again, I think a better alternative would be linking number of fork to kernel status (expert/master/grand master etc) - this way you can make a separation to between \"forking\" and \"upvote\"
", "rawMarkdown": "I understand the situation you describe, but I would argue by and large people fork codes that are either well written or at least written with some merit. then again, I think a better alternative would be linking number of fork to kernel status (expert/master/grand master etc) - this way you can make a separation to between \"forking\" and \"upvote\"", "votes": null }, { "id": "448797", "postDate": "01/02/2019 05:24:25", "content": "StackOverflow recently nagged me \"You haven't voted on questions in a while; questions need votes too!\" when I upvoted an answer to a question. Maybe Kaggle could have a similar nag whenever a user forks a kernel to remind them to consider upvoting.
", "rawMarkdown": "StackOverflow recently nagged me \"You haven't voted on questions in a while; questions need votes too!\" when I upvoted an answer to a question. Maybe Kaggle could have a similar nag whenever a user forks a kernel to remind them to consider upvoting.", "votes": null }, { "id": "449130", "postDate": "01/02/2019 17:40:41", "content": "Unfortunately Kaggle is a microcosm of the world at large.
\n\nSo people will plagiarise, take and not upvote, fork and not upvote - yes people who do this aren't very nice. As you cannot control selfish people there isn't any point stressing over it. I use kernels to force me to clean up my code and make it reusable where necessary so that I can use it in future competitions - that is how I win as I am not bothered about medals until you can swap them for snacks! ;)
", "rawMarkdown": "Unfortunately Kaggle is a microcosm of the world at large.\n\nSo people will plagiarise, take and not upvote, fork and not upvote - yes people who do this aren't very nice. As you cannot control selfish people there isn't any point stressing over it. I use kernels to force me to clean up my code and make it reusable where necessary so that I can use it in future competitions - that is how I win as I am not bothered about medals until you can swap them for snacks! ;)", "votes": null }, { "id": "449140", "postDate": "01/02/2019 17:56:12", "content": "You may want to support this change request: https://www.kaggle.com/product-feedback/76394
", "rawMarkdown": "You may want to support this change request: https://www.kaggle.com/product-feedback/76394", "votes": null }, { "id": "454736", "postDate": "01/12/2019 05:40:04", "content": "For all it is worth, I think about forking as bookmarking, I fork a kernel to study it closer when I have time. It does not mean that I would like what I see there when I do look into it. Hence my forking is not an upvote. Having said that, I usually upvote those kernels that I found useful and well thought through.
", "rawMarkdown": "For all it is worth, I think about forking as bookmarking, I fork a kernel to study it closer when I have time. It does not mean that I would like what I see there when I do look into it. Hence my forking is not an upvote. Having said that, I usually upvote those kernels that I found useful and well thought through.", "votes": null } ], "comments": [ { "id": 436762, "author_name": "mconway", "author_url": "", "post_date": "12/10/2018 22:39:52", "content": "Just think of Kaggle as a giant, shared pool of open source code. I don't think there are any rules.
", "votes": null, "replies": [ { "id": 436764, "author_name": "delayedkarma", "author_url": "", "post_date": "12/10/2018 22:40:49", "content": "Not rules really.. rules and open source don't work well together.. more like ethical guidelines...
", "votes": null, "replies": [] } ] }, { "id": 436769, "author_name": "addisonhoward", "author_url": "", "post_date": "12/10/2018 23:09:40", "content": "Hi Panchajanya,
\n\nYou are welcome to keep your Kernels private if you do not wish for it to be shared or potentially forked. As noted, a public kernel is open-source, and the Kaggle community likes to learn from shared knowledge about a problem statement, whish is where kernel forking may occur.
", "votes": null, "replies": [ { "id": 436772, "author_name": "delayedkarma", "author_url": "", "post_date": "12/10/2018 23:12:25", "content": "Hey Addison,
\n\nI completely understand that. I just wanted to ask if there are any moral gray areas here. But i suppose once a kernel is indeed public, it's all good.
", "votes": null, "replies": [] }, { "id": 438222, "author_name": "joeycorea", "author_url": "", "post_date": "12/13/2018 09:48:46", "content": "Addison,
\n\nI don't think Panchajanya is objecting to forking or sharing code. I think it is more a question of whether etiquette or a code of conduct should be enforced around copying people's work without credit.\nAs you can imagine, it's not a nice feeling to have someone take your work, rebadge it as their own without giving credit to the brain cycles you put into it.\nSure, if they're not happy with the current state, people can make their kernel's private but then the community loses out on a lack of information sharing.\nPlease consider enforcing a code of conduct around this :)
", "votes": null, "replies": [] } ] }, { "id": 436797, "author_name": "reputation18", "author_url": "", "post_date": "12/11/2018 00:40:17", "content": "I'm thinking maybe add a line at the top to ask people to reference back if they build based on the kernel?\n I think people should at least mention the original post no matter this is 'open source' or not...
", "votes": null, "replies": [ { "id": 436799, "author_name": "delayedkarma", "author_url": "", "post_date": "12/11/2018 00:50:12", "content": "That was basically my point. All papers on arxiv are open source, but that doesn't mean they shouldn't be cited...
", "votes": null, "replies": [] }, { "id": 436803, "author_name": "reputation18", "author_url": "", "post_date": "12/11/2018 00:59:51", "content": "Yes. I saw some refer back and emphasis the original post but some don't. My suggestion is don't let this ruin your day :)
", "votes": null, "replies": [] }, { "id": 436805, "author_name": "delayedkarma", "author_url": "", "post_date": "12/11/2018 01:05:12", "content": "Haha na.. all good.. it just took me by surprise is all!
", "votes": null, "replies": [] } ] }, { "id": 436858, "author_name": "peterhurford", "author_url": "", "post_date": "12/11/2018 03:11:49", "content": "I agree it is polite to credit the kernels you fork. I rarely upvote kernels that don't credit their sources and I encourage others to do the same. But this is my ethics, not law, and I'm fine with Kaggle being the way it is -- I'm not asking for a rules change, just a norms change.
", "votes": null, "replies": [ { "id": 436859, "author_name": "delayedkarma", "author_url": "", "post_date": "12/11/2018 03:13:42", "content": "Yep.. Rules are bad, ethical norms are usually nice.
", "votes": null, "replies": [] } ] }, { "id": 437159, "author_name": "springmanndaniel", "author_url": "", "post_date": "12/11/2018 13:19:36", "content": "Atleast the person forked your kernel. My kernel got copy & pasted ... what - in my opinion - is not ok.
", "votes": null, "replies": [ { "id": 437284, "author_name": "delayedkarma", "author_url": "", "post_date": "12/11/2018 17:11:50", "content": "Damn. That is not nice.
", "votes": null, "replies": [] } ] }, { "id": 438298, "author_name": "yifanxie", "author_url": "", "post_date": "12/13/2018 12:46:48", "content": "I think a fork should automatically translate into an upvote, so the original author would receive some credit toward their kernel status. Or alternatively, directly include \"forking\" in kernel expert/master/grand master criteria
", "votes": null, "replies": [ { "id": 438364, "author_name": "delayedkarma", "author_url": "", "post_date": "12/13/2018 14:37:54", "content": "Yep. A public fork should definitely count as an upvote
", "votes": null, "replies": [] }, { "id": 448584, "author_name": "cpmpml", "author_url": "", "post_date": "01/01/2019 15:20:36", "content": "I disagree, I explained why in a similar discussion in another competition. I may fork a kernel, try it, and decide it is not worth it. Why should I upvote it then?
\n\nThis said, I totally agree with the post: sources should be acknowledged, and added value should be clearly identified.
\n\nedited: removed the offending example as the fork author acknowledged the source.
", "votes": null, "replies": [] }, { "id": 448604, "author_name": "yifanxie", "author_url": "", "post_date": "01/01/2019 16:22:54", "content": "I understand the situation you describe, but I would argue by and large people fork codes that are either well written or at least written with some merit. then again, I think a better alternative would be linking number of fork to kernel status (expert/master/grand master etc) - this way you can make a separation to between \"forking\" and \"upvote\"
", "votes": null, "replies": [] }, { "id": 454736, "author_name": "calabist", "author_url": "", "post_date": "01/12/2019 05:40:04", "content": "For all it is worth, I think about forking as bookmarking, I fork a kernel to study it closer when I have time. It does not mean that I would like what I see there when I do look into it. Hence my forking is not an upvote. Having said that, I usually upvote those kernels that I found useful and well thought through.
", "votes": null, "replies": [] } ] }, { "id": 448390, "author_name": "carlolepelaars", "author_url": "", "post_date": "01/01/2019 01:31:33", "content": "Great point! I don't think the guidelines for Apache 2.0 are in effect with public Kaggle kernels.
\n\nAs scientists we should take care to cite our sources when using someone elses work. It is a sign of respect.
\n\nI also think it is decent to follow the golden rule:
\n\n\"One should treat others as one would like others to treat oneself.\"
", "votes": null, "replies": [] }, { "id": 448797, "author_name": "peterhurford", "author_url": "", "post_date": "01/02/2019 05:24:25", "content": "StackOverflow recently nagged me \"You haven't voted on questions in a while; questions need votes too!\" when I upvoted an answer to a question. Maybe Kaggle could have a similar nag whenever a user forks a kernel to remind them to consider upvoting.
", "votes": null, "replies": [] }, { "id": 449130, "author_name": "scirpus", "author_url": "", "post_date": "01/02/2019 17:40:41", "content": "Unfortunately Kaggle is a microcosm of the world at large.
\n\nSo people will plagiarise, take and not upvote, fork and not upvote - yes people who do this aren't very nice. As you cannot control selfish people there isn't any point stressing over it. I use kernels to force me to clean up my code and make it reusable where necessary so that I can use it in future competitions - that is how I win as I am not bothered about medals until you can swap them for snacks! ;)
", "votes": null, "replies": [] }, { "id": 449140, "author_name": "cpmpml", "author_url": "", "post_date": "01/02/2019 17:56:12", "content": "You may want to support this change request: https://www.kaggle.com/product-feedback/76394
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "436758": "So I recently saw someone fork my kernel, add one fairly trivial grid search to it, and post it with a different name but the exact same in-kernel comments and headers without any acknowledgement. Now, my original kernel https://www.kaggle.com/delayedkarma/let-s-add-some-features-for-the-national-holidays() was beholden to another one as well (https://www.kaggle.com/chauhuynh/my-first-kernel-3-699), but I always make sure people know what I'm sourcing my material from. \n\nWhat does the community think of ethical forking practices? And are there any guidelines to this effect.", "436762": "Just think of Kaggle as a giant, shared pool of open source code. I don't think there are any rules.", "436764": "Not rules really.. rules and open source don't work well together.. more like ethical guidelines...", "436769": "Hi Panchajanya,\n\nYou are welcome to keep your Kernels private if you do not wish for it to be shared or potentially forked. As noted, a public kernel is open-source, and the Kaggle community likes to learn from shared knowledge about a problem statement, whish is where kernel forking may occur.", "436772": "Hey Addison, \n\nI completely understand that. I just wanted to ask if there are any moral gray areas here. But i suppose once a kernel is indeed public, it's all good.", "436797": "I'm thinking maybe add a line at the top to ask people to reference back if they build based on the kernel?\n I think people should at least mention the original post no matter this is 'open source' or not...", "436799": "That was basically my point. All papers on arxiv are open source, but that doesn't mean they shouldn't be cited...", "436803": "Yes. I saw some refer back and emphasis the original post but some don't. My suggestion is don't let this ruin your day :)", "436805": "Haha na.. all good.. it just took me by surprise is all!", "436858": "I agree it is polite to credit the kernels you fork. I rarely upvote kernels that don't credit their sources and I encourage others to do the same. But this is my ethics, not law, and I'm fine with Kaggle being the way it is -- I'm not asking for a rules change, just a norms change.", "436859": "Yep.. Rules are bad, ethical norms are usually nice.", "437159": "Atleast the person forked your kernel. My kernel got copy & pasted ... what - in my opinion - is not ok.", "437284": "Damn. That is not nice.", "438222": "Addison,\n\nI don't think Panchajanya is objecting to forking or sharing code. I think it is more a question of whether etiquette or a code of conduct should be enforced around copying people's work without credit.\nAs you can imagine, it's not a nice feeling to have someone take your work, rebadge it as their own without giving credit to the brain cycles you put into it.\nSure, if they're not happy with the current state, people can make their kernel's private but then the community loses out on a lack of information sharing.\nPlease consider enforcing a code of conduct around this :)", "438298": "I think a fork should automatically translate into an upvote, so the original author would receive some credit toward their kernel status. Or alternatively, directly include \"forking\" in kernel expert/master/grand master criteria", "438364": "Yep. A public fork should definitely count as an upvote", "448390": "Great point! I don't think the guidelines for [Apache 2.0][1] are in effect with public Kaggle kernels.\n\nAs scientists we should take care to cite our sources when using someone elses work. It is a sign of respect.\n\nI also think it is decent to follow the golden rule: \n\n\"One should treat others as one would like others to treat oneself.\"\n\n\n [1]: https://www.apache.org/licenses/LICENSE-2.0", "448584": "I disagree, I explained why in a similar discussion in another competition. I may fork a kernel, try it, and decide it is not worth it. Why should I upvote it then?\n\nThis said, I totally agree with the post: sources should be acknowledged, and added value should be clearly identified. \n\nedited: removed the offending example as the fork author acknowledged the source.", "448604": "I understand the situation you describe, but I would argue by and large people fork codes that are either well written or at least written with some merit. then again, I think a better alternative would be linking number of fork to kernel status (expert/master/grand master etc) - this way you can make a separation to between \"forking\" and \"upvote\"", "448797": "StackOverflow recently nagged me \"You haven't voted on questions in a while; questions need votes too!\" when I upvoted an answer to a question. Maybe Kaggle could have a similar nag whenever a user forks a kernel to remind them to consider upvoting.", "449130": "Unfortunately Kaggle is a microcosm of the world at large.\n\nSo people will plagiarise, take and not upvote, fork and not upvote - yes people who do this aren't very nice. As you cannot control selfish people there isn't any point stressing over it. I use kernels to force me to clean up my code and make it reusable where necessary so that I can use it in future competitions - that is how I win as I am not bothered about medals until you can swap them for snacks! ;)", "449140": "You may want to support this change request: https://www.kaggle.com/product-feedback/76394", "454736": "For all it is worth, I think about forking as bookmarking, I fork a kernel to study it closer when I have time. It does not mean that I would like what I see there when I do look into it. Hence my forking is not an upvote. Having said that, I usually upvote those kernels that I found useful and well thought through." }, "source": "meta" }