The decision has already been reversed Now the code will be moved to slower storage instead of deleted. Details, like how publicly accessible it will be, still unknown.
Everything has a lifetime. https://tildes.net/~talk/109v/thoughts_on_lifetimes_limits_and_tolerances Glad they reversed the decision but it’s a symptom.
Without the numbers, it's a decision that makes sense in a way. It's not sustainable over the long term to host people's data forever, for free unless there's somewhere you're getting that much...
Without the numbers, it's a decision that makes sense in a way. It's not sustainable over the long term to host people's data forever, for free unless there's somewhere you're getting that much value from. For Github, I think they have that - Microsoft gains a lot by having Github continue to be THE place for developers (many whole languages rely on Github now). Gitlab, I don't think, has that kind of critical mass.
But... 1 million? There's no way that's right. That's a laughable amount of money for Gitlab. That's like what 2 or 3 of their engineers cost. Maybe just one if they're very senior. If you really need $1 million so bad just layoff two engineers. Or hell, you don't even have to lay them off, just don't fill in two employees who left as part of the normal employee churn.
Source code is super light, generally speaking - not surprised the savings would only be around that much. To give an order of magnitude comparison, rsync.net quotes $720k/yr for a petabyte of...
It's not sustainable over the long term to host people's data forever, for free unless there's somewhere you're getting that much value from.
Source code is super light, generally speaking - not surprised the savings would only be around that much. To give an order of magnitude comparison, rsync.net quotes $720k/yr for a petabyte of storage.
It'd be pretty unobjectionable if they wanted to impose restrictions on free accounts going forward, but threatening to pull the rug on existing FOSS projects instantly burns a lot of goodwill. Even though they've reversed the decision, a lot of devs, myself included, are going to think twice about using Gitlab in the future: it looks like they've either got the MBAs calling the shots and running rampant, or they're in dire financial straits. Neither bodes well for platform stability.
Especially when the market leader and main competitor has a resiliency policy so extreme that they can use it as a marketing stunt: https://archiveprogram.github.com/arctic-vault/
Look at the fine print though, they didn't back up everything either:
Look at the fine print though, they didn't back up everything either:
any commits between the announcement at GitHub Universe on November 13th and 02/02/2020; every repo with at least 1 star and any commits from the year before the snapshot; and all repos with at least 250 stars.
Totally fair, although I'm sure that would never stand in the way of a good marketing narrative! From a practical standpoint they are also partnered with the Internet Archive and the Software...
Totally fair, although I'm sure that would never stand in the way of a good marketing narrative!
From a practical standpoint they are also partnered with the Internet Archive and the Software Heritage Foundation, both of whom are aiming for full copies of all data as far as I understand it, so I'm still reasonably comfortable in saying there's a genuine difference in ethos between the two companies right now.
I mean, that'd be an easy soul to sell for me, an individual. I get that for even a small sized tech business that a million dries up super fast, however.
for just one million in annual revenue.
I mean, that'd be an easy soul to sell for me, an individual.
I get that for even a small sized tech business that a million dries up super fast, however.
I almost had a heart attack because I misread that Github was doing it. I had some qualms over the years with Github's Microsoft acquisition, so I considered switching to GitLab several times, and...
I almost had a heart attack because I misread that Github was doing it.
I had some qualms over the years with Github's Microsoft acquisition, so I considered switching to GitLab several times, and now I'm glad I never did.
I hope GitLab faces some outcry over this and maybe reconsiders. I have an archivistic philosophy about data (and about everything else), so to me this is a pretty ugly thing to do.
Don't worry, Microsoft will eventually prune as well. It might take longer given the larger pools of backup money from other products, but it'll happen. Until storage becomes free or everyone...
Don't worry, Microsoft will eventually prune as well. It might take longer given the larger pools of backup money from other products, but it'll happen.
Until storage becomes free or everyone self-prunes at the same rate they create, there will always be a need in some fashion.
I personally love Gitlab for providing a great free/cheap self-hosted product in a way Github does not. If they posted which repos they wish to archive it's possible to fully port them onto a self-hosted archival server.
Call me old fashioned but I'll throw support behind damn near anyone offering a viable alternative to a tech giant offering. Re-enforcing the monopolies do not weaken them.
I guess free hosting doesn't last forever? This doesn't seem all that bad to me, but backing up old repos to tape (so you can at least download the repo, with a lag) would be a better idea, to...
I guess free hosting doesn't last forever? This doesn't seem all that bad to me, but backing up old repos to tape (so you can at least download the repo, with a lag) would be a better idea, to make sure nothing important gets lost.
We'll see what the real announcement looks like. This leak might be a "trial balloon."
"Of course there are no guarantees it will always be hosted there, but the unwritten rules in open source are that you make the code available and you don’t remove it."
Seems like this "unwritten rules" thing is a way to make up promises that were never actually made.
There’s probably a name for this situation - it’s reminiscent of the tragedy of the commons but not quite that - where the community as a whole is the beneficiary but the burden of payment falls...
There’s probably a name for this situation - it’s reminiscent of the tragedy of the commons but not quite that - where the community as a whole is the beneficiary but the burden of payment falls on some specific unrelated party.
The authors of dormant projects generally don’t benefit from, or even particularly care about, the free hosting. Back when it was active they did, but the fact it hasn’t been touched in a year is a good sign they’ve moved on. The people who care are the people who come to find these long-tail fixes for niche problems years after the fact, and often didn’t even know they would need them until long after that one year mark has passed.
It seems akin to putting an expiry date on Stack Overflow posts because the original asker has long since marked them answered. Sure, the direct participants in the thread might not care, but it’s a huge disservice to the people who would otherwise have made use of that knowledge later.
Even though Steve Jobs was an asshole, he had a great deal of eternal wisdom, like on product vs. marketing people. I think finance people can be lumped in with marketing people because they have...
Even though Steve Jobs was an asshole, he had a great deal of eternal wisdom, like on product vs. marketing people.
I think finance people can be lumped in with marketing people because they have similarly one/low-dimensional thinking about product.
Gitlab really missed for the forest for the trees here.
Think I'll just write one that updates a counter to a hidden file and push that from time to time. I have plenty of dormant projects. Many of then aren't dead, I just haven't had time to work on...
Think I'll just write one that updates a counter to a hidden file and push that from time to time.
I have plenty of dormant projects. Many of then aren't dead, I just haven't had time to work on them all.
While a year does seem like a pretty short amount of time. Im generally glad things are being removed. Although I do think it would be better to define a criteria to archive larger more...
While a year does seem like a pretty short amount of time. Im generally glad things are being removed. Although I do think it would be better to define a criteria to archive larger more substantial projects for longer while deleting one off useless things.
I've come across numerous programs, scripts, and tips from various repos that haven't been touched in 8+ years and still work just fine at their intended purpose. One man's trash is another man's...
I've come across numerous programs, scripts, and tips from various repos that haven't been touched in 8+ years and still work just fine at their intended purpose.
One man's trash is another man's treasure as they say.
I think a good solution would be to generate a list of the to-be-deleted repos and pass them off to the archivist community.
The decision has already been reversed
Now the code will be moved to slower storage instead of deleted. Details, like how publicly accessible it will be, still unknown.
Everything has a lifetime.
https://tildes.net/~talk/109v/thoughts_on_lifetimes_limits_and_tolerances
Glad they reversed the decision but it’s a symptom.
What an unbelievably massive self-own... Imagine trading away your reputation for just one million in annual revenue.
Without the numbers, it's a decision that makes sense in a way. It's not sustainable over the long term to host people's data forever, for free unless there's somewhere you're getting that much value from. For Github, I think they have that - Microsoft gains a lot by having Github continue to be THE place for developers (many whole languages rely on Github now). Gitlab, I don't think, has that kind of critical mass.
But... 1 million? There's no way that's right. That's a laughable amount of money for Gitlab. That's like what 2 or 3 of their engineers cost. Maybe just one if they're very senior. If you really need $1 million so bad just layoff two engineers. Or hell, you don't even have to lay them off, just don't fill in two employees who left as part of the normal employee churn.
Source code is super light, generally speaking - not surprised the savings would only be around that much. To give an order of magnitude comparison, rsync.net quotes $720k/yr for a petabyte of storage.
It'd be pretty unobjectionable if they wanted to impose restrictions on free accounts going forward, but threatening to pull the rug on existing FOSS projects instantly burns a lot of goodwill. Even though they've reversed the decision, a lot of devs, myself included, are going to think twice about using Gitlab in the future: it looks like they've either got the MBAs calling the shots and running rampant, or they're in dire financial straits. Neither bodes well for platform stability.
This instantly makes me never want to use them again.
Especially when the market leader and main competitor has a resiliency policy so extreme that they can use it as a marketing stunt: https://archiveprogram.github.com/arctic-vault/
Look at the fine print though, they didn't back up everything either:
Totally fair, although I'm sure that would never stand in the way of a good marketing narrative!
From a practical standpoint they are also partnered with the Internet Archive and the Software Heritage Foundation, both of whom are aiming for full copies of all data as far as I understand it, so I'm still reasonably comfortable in saying there's a genuine difference in ethos between the two companies right now.
I mean, that'd be an easy soul to sell for me, an individual.
I get that for even a small sized tech business that a million dries up super fast, however.
Might change your calculus if you were already pulling in $200M/yr. 😉
I almost had a heart attack because I misread that Github was doing it.
I had some qualms over the years with Github's Microsoft acquisition, so I considered switching to GitLab several times, and now I'm glad I never did.
I hope GitLab faces some outcry over this and maybe reconsiders. I have an archivistic philosophy about data (and about everything else), so to me this is a pretty ugly thing to do.
Don't worry, Microsoft will eventually prune as well. It might take longer given the larger pools of backup money from other products, but it'll happen.
Until storage becomes free or everyone self-prunes at the same rate they create, there will always be a need in some fashion.
I personally love Gitlab for providing a great free/cheap self-hosted product in a way Github does not. If they posted which repos they wish to archive it's possible to fully port them onto a self-hosted archival server.
Call me old fashioned but I'll throw support behind damn near anyone offering a viable alternative to a tech giant offering. Re-enforcing the monopolies do not weaken them.
I guess free hosting doesn't last forever? This doesn't seem all that bad to me, but backing up old repos to tape (so you can at least download the repo, with a lag) would be a better idea, to make sure nothing important gets lost.
We'll see what the real announcement looks like. This leak might be a "trial balloon."
Seems like this "unwritten rules" thing is a way to make up promises that were never actually made.
There’s probably a name for this situation - it’s reminiscent of the tragedy of the commons but not quite that - where the community as a whole is the beneficiary but the burden of payment falls on some specific unrelated party.
The authors of dormant projects generally don’t benefit from, or even particularly care about, the free hosting. Back when it was active they did, but the fact it hasn’t been touched in a year is a good sign they’ve moved on. The people who care are the people who come to find these long-tail fixes for niche problems years after the fact, and often didn’t even know they would need them until long after that one year mark has passed.
It seems akin to putting an expiry date on Stack Overflow posts because the original asker has long since marked them answered. Sure, the direct participants in the thread might not care, but it’s a huge disservice to the people who would otherwise have made use of that knowledge later.
Yes, fair enough. To some extent they could figure this out by looking at downloads, but you never know when someone will find a project useful.
Why do I get the feeling that the finance and marketing heads have taken over? Every new release is just an absolute step back.
Even though Steve Jobs was an asshole, he had a great deal of eternal wisdom, like on product vs. marketing people.
I think finance people can be lumped in with marketing people because they have similarly one/low-dimensional thinking about product.
Gitlab really missed for the forest for the trees here.
Sounds like someone needs to write a bot that submits an issue on each of the dormant repos.
Think I'll just write one that updates a counter to a hidden file and push that from time to time.
I have plenty of dormant projects. Many of then aren't dead, I just haven't had time to work on them all.
git commit --allow-empty
Of course git will let you commit an identical tree!
Thanks for the suggestion which otherwise I would never have thought to even look for.
I was thinking it could raise issues on ALL of the dormant projects, not just ones that I own.
While a year does seem like a pretty short amount of time. Im generally glad things are being removed. Although I do think it would be better to define a criteria to archive larger more substantial projects for longer while deleting one off useless things.
I've come across numerous programs, scripts, and tips from various repos that haven't been touched in 8+ years and still work just fine at their intended purpose.
One man's trash is another man's treasure as they say.
I think a good solution would be to generate a list of the to-be-deleted repos and pass them off to the archivist community.