46 comments

[ 2.2 ms ] story [ 50.4 ms ] thread
Making their grafana dashboard public is an interesting move.
It is. I love it. Extreme transparency is a norm I hope to see become more and more common in the coming years/decades; extending the open source principles beyond the code.

Gitlab is exemplary there. Mozilla is very good at it as well. Another company I've recently found out about and gained a massive amount of respect for is Buffer. https://buffer.com/transparency

You only have to import a repository once.
In terms of code hosting, you can always have two remotes and push to both. Since this config would only be known on developer machines there's no way to know if GitLab is getting true converts or just a bunch of people kicking the tires while the real work is still happening elsewhere and might never move for real.
My `g pum` alias pushes master to 4 remotes. GNU dislikes GitHub and has 2 gitlab instances plus their primary savannah git hosting. I don't care, GitHub is still the best to work with internally.
Yes, it'd probably be better to look at their Grafana stats on web requests to see if people are staying around.

To me it looks to be up a decent bit, but less than you'd expect from a proper mass exodus event.

What are the reasons for distrusting Microsoft to be good stewards?
(comment deleted)
Some battered spouses never go back no matter how many roses and chocolates they buy.
As far as I can tell, it's because they did some bad things in 1995 and therefore can never be trusted again.
Not just 1995. Windows 10 spyware/crapware is a fundamental violation of the customer's interests in my book.

EDIT: is/was, I've heard they turned some of it down supposedly.

Skype was far more recent than 1995. They have integrated a lot of companies over time and some have done vastly better than others. The risk may be low, but the impact is high enough many are just going to jump asap.
If you lived through that time, you wouldn't be so flippant
Embrace, extend, and extinguish. Microsoft has spent like 4 years building up good will after destroying trust in them for the past 20.

Trust is easily lost and is hard to earn.

GitHub is a company, not a technology. "Embrace, extend, and extinguish" doesn't apply. Unless of course you're worried about them doing that to git itself, not just GitHub.
This is exactly my worry.

1. Embrace Git by buying Github

2. Extend Git, fork git to add Github specific features (Basically integrate https://hub.github.com/ directly into git)

3. Extinguish, by gradually making Github incompatible with the original git. Then stop development on the forked git and push everyone to a proprietary product.

This is complete nonsense. Microsoft has actually _contributed_ to Git. But even outside of that Azure makes them a metric fuck ton of money. They need developer support to continue using Azure which also means Git integration.

They are not going to slowly change git to the point where no one else can use it. Git is opened sourced, by owning GitHub they still can't control Git anyway.

When I left MS about 4 years ago they were working on switching their internal source control over to Git. Not sure on their progress, but that hardly seems like they'd make it incompatible.
Exactly, there are a few MSDN/Channel 9 articles describing the pains of pulling the Windows repo into Git. If anything, Microsoft is going to make Github, and Git, much better. They have the largest code bases in the world, they aren't going to change anything that will disrupt internal processes, but will build tools to help others with massive codebases live better in Github.
I also don't see it - but isn't making Git 'much better' essentially extending it? If they released some kind of tight integration that was actually really useful and only worked well in the github context, don't you think people would use it? I can't imagine what form such a thing would take, but it definitely seems possible.
I get that there are reasons to distrust MS, I'd like to hear the reasons for trusting Microsoft LESS than pre-acq Github.
Do they have a good record with acquisitions? I doubt they'll be evil, but another Skype seems likely, where the product gets worse and weirder for reasons that make no sense.
I agree. Microsoft simply can’t resist pissing on whatever they buy, just to make sure we know it’s theirs. I can see “Microsoft Source Control Pro 3000” coming, in enterprise and smb versions, and with “git” shadowed in their marketing material. Also Skype collaboration and other hallmarks of their heavy handed lazy and ugly integration ideas.
The classic example is Skype but people forget Skype is a really, really old acquisition at this point. Most of the things done to ruin it were done well over 5 years ago.

Let's not forget GitHub already has enterprise products...

Regardless Microsoft has brought in many acquisitions and didn't ruin them. LinkedIn was kept separate and beyond some integration was left untouched. Same with the vast majority of game studio acquisitions. Let's not forget the Xamarin acquisition has been a huge boon to .Net developers as well as open source initiatives.

Microsoft has new leadership. Most of these arguments I'm seeing on HN are all referring to things that happened under much older leadership. This goes to show Microsoft still has some trust building work to do but at the same time they shouldn't be written off either.

Microsoft is to blame for Skype downfall
It might not be about Microsoft specifically.

It might just be about github no longer being independent.

Or it might more to do with the $7.5 billion sale price, and a wish not to join in a game of "you have to be on github because everyone will look for you on github because everyone else is on github".

(So much the same reason as some people leave Facebook: a wish not to be someone else's product.)

(comment deleted)
Skype and arguably Nokia are two I can think of from the top of my head.
You haven't read the Halloween document yet?
Looks like in the low 10s of thousands of projects over a period of around 7-8 days. Would be interesting to see how many GitHub stars that equates to, or if you assigned a metric to a GitHub project by counting number of unique contributors, and then aggregated that across the Gitlab import data.

This might mostly reflect the values of projects or single maintainers who place a super high priority on Microsoft-avoidance. Many more people might migrate though after waiting a while to see if Microsoft begins trying to push any agendas through GitHub.

I for one am super happy to stay with GitHub as long as no part of it starts to feel like there is any type of Azure agenda, VS agenda, etc. Keep my tooling, CI, deployment, etc., choices separate from the git platform, and I'm fine with it being a Microsoft subsidiary. The minute that changes, I'll move all my code.

It takes time to organise changes for a team/project, surely that would suggest most early transfers will be individuals repos. Also people are likely to shift smaller projects first to be sure there are no complications.
I have never seen one of these Silicon Valley “boycotts” have any effect whatsoever.
I'm pretty sure if Microsoft still had 95% PC OS market share they would be a much bigger (crappier) company today than they are.

If a group of people get annoyed with a product they will flock to the next available one and drive that business to become competition. So while you don't get that instant gratification of being pissed and getting results you do force companies to make changes.

People suffered through IE for years, and when Chrome came around it practically took IE's market share overnight because people were so willing to blindly abandon IE.

But that’s one product beating another. It happens all the time. I’m talking about a departure from a product for ethical / philosophical reasons.
The relevant data is how many new projects (open source) will open on github in the future. I’ve delted my account and no longer collaborate on project who are left on github.

What’s important is open source.

Microsoft is an acceptable option to many still, no matter how bad it proves to be, over and again.

Its easy to forget that Microsoft hasn't actually taken over control of GitHub yet.

The big announcement was that they have entered into an agreement to acquire GitHub, after slogging through (perhaps) months of legal, accounting, and regulatory paperwork.

There very well may be more bumps in that graph, as the news sinks in, and when the deal actually closes.

I'd be curious to look at the list ported over and see if anything moved really mattered.

The whole boycott was nonsense anyway. Microsoft has only agreed to buy GitHub; they can't actually buy and own it for about 6 months. So even if you had a legit fear of Microsoft, why would you even bother yet? Anything of importance on GitHub is going to take a while to move over anyway which is why I suspect the majority of these "moves" were people moving over repos that no one cares about.

Microsoft's cloud business is up and coming and is already worth many billions of dollars. They are not going to jeopardize user trust, which could harm Azure, by messing with your little GitHub project.

There may be some people migrating from GitHub to boycott Microsoft, but most of the old guard free software folks weren't on GitHub to begin with.

I think the biggest concern is just a lack of trust that Microsoft will fuck things up out of incompetence rather than malice.

Another rationale is that, even if you reason it through and conclude that Microsoft isn't going to screw it up at all, the mere risk of GitHub being a SPOF is enough to support some alternatives. That might be a risk that existed pre-acquisition, but it's definitely come to attention.

Also, GitLab has been getting better and better for awhile and capitalized on this really well from a marketing standpoint; let's give them some credit instead of assuming it's purely a negative reaction to GitHub and Microsoft.

> but most of the old guard free software folks weren't on GitHub to begin with

This is important to remember. Some people were also openly hesitant to GitHub, but used it anyway due to its network effect. This seems like a perfect moment to reevaluate.

> Anything of importance on GitHub is going to take a while to move over anyway

Actually, not really. The importer is very good, issues and PR/MRs get moved over. The only adjustments have to be to CI if you used Travis.

Original title:

The "mass exodus" from GitHub to GitLab: 10 days later

I was already started do moving to my own (a-la VPS on digitalocean) this just added some extra velocity to it.

its not that i dont like github, i do. But keeping internet decentralized is important (imho).

Also given that I live in Turkey, government can block all github just because someone shares bad stuff about current leaders in gist. (ps. they did block google drive because someone posted some photographs/documents)

Many people here using somehow their own vps, just adding push-hook to sync with other github/gitlab is not a trivial thing.

tl;dr: push to your repo, sync with cloud providers' one.