Timeline for Announcing a change to the data-dump process
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 17 at 23:40 | comment | added | Journeyman Geek | In 'this' context though the redistribution is implicit - all these cases are about redistribution in different forms. I recently came across a defunct FOSS LLM dataset that contained info from SE dumps. That would be an interesting case but I am waiting on the next set of clarifications before I ask yet another question(tm) | |
| Jul 17 at 16:51 | comment | added | Abdul Aziz Barkat | @Michaelcomelately your second statement is correct, the license doesn't require redistribution, but if one does distribute they need to use the same license. | |
| Jul 17 at 16:07 | comment | added | Michael come lately | "are they welcome to continue to distribute their transformations?" If it's truly CC-BY-SA, isn't that compulsory? "ShareAlike — If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original." Or is it saying that iff distributed, distributions must use the same license? | |
| Jul 14 at 3:28 | comment | added | Journeyman Geek | @ZachLipton In these cases - someone can request the dumps - as per meta.stackexchange.com/questions/401324/… . As long as one or more people are fine with requesting a full dump (and we're yet to see the process, and how all of this does/doesn't scale) and is willing to share a download somehow, there's no need. | |
| Jul 14 at 3:24 | comment | added | Zach Lipton | Also it appears that preparing something like SEqlite would require clicking through 364 separate download forms (assuming "additional checks for bots" manages to successfully prevent automation), which poses a rather large practical impediment. Saying that someone is "welcome" to do something if only they click through hundreds of forms instead of the single download available now is...not particularly welcoming. | |
| Jul 13 at 21:42 | comment | added | Zach Lipton | @Philippe I'd appreciate if you're able to clarify: are they welcome to get the data dumps, or are they welcome to continue to distribute their transformations? One of the conditions shown in the screenshot is "I will not transfer it to others without permission from Stack Overflow." You also state that if someone doesn't follow that condition, the company has the option to decline to provide them with future dumps. Projects like SEqlite amount to transferring a format-shifted version of the data dumps to the public. How much transformation is required to make transferring the dumps ok? | |
| Jul 13 at 12:33 | comment | added | Philippe StaffMod | @JourneymanGeek - All three of those are more than welcome to continue to get the data dumps. We certainly can't and wouldn't consider declining to provide them to someone based on something that a downstream user MIGHT do. | |
| Jul 13 at 11:16 | comment | added | Journeyman Geek | ... I didn'r recognise you in the comments, but yeah, I think there will be community backups :D | |
| Jul 13 at 10:08 | comment | added | Journeyman Geek | It wasn't meant to be exhaustive, but rather illustrative :D. Aka, there's probably a ton more | |
| Jul 13 at 9:59 | comment | added | Restore The Data Dumps Again | Add seqlite.puny.engineering to the list | |
| Jul 12 at 23:04 | history | answered | Journeyman Geek | CC BY-SA 4.0 |