Comment by EnderMB
8 years ago
This was largely my thought behind the move.
Given that GitHub is quite proudly built on Ruby, I can't see them wanting to switch things up from a tech perspective. GitHub is stable, and it's tech stack is capable of staying up despite some major DDoS attacks.
If anything, I think this is an opportunity for Microsoft to introduce themselves to the Ruby and Rails teams, and to finally resolve the issues that stop Windows from being a first-class citizen in the Ruby world. If they can do this through both Windows and the Windows Subsystem for Linux then I think they'll be on to a winner. It's a capture of a much-loved service, and an opportunity to bring a mature set of tools into their domain.
I work for Microsoft, we run systems that are not built on MS technologies. There’s absolutely no push for migration. In my opinion, no-one will pressure GitHub to change their stack, it would be a suicide.
Disclaimer: this is just my personal opinion.
Aside from Wunderlist, who was acquired and running on AWS, and had to switch to Azure and rewrite a bunch of their code to become ToDo.
There is a large difference in complexity here though, and Wunderlist need extra coding for O365 integration so it wasn't just redevelopment purely for a platform shift.
GitHub more complex than todo-list-on-steroids app so a platform change would not make any real sense. MS today may still have some of its old habits but they do seem to have purged a lot of the "not invented here" problem that caused much embarrassment when the first attempts to migrate HotMail over to MS technologies failed. It also has pretty good integration with relevant MS tools (VS & VS.code, etc.).
I expect to see them moving the base infrastructure over to Azure, but non-MS technologies are well enough supported on the platform so that won't require any notable changes to the main codebase of the product itself (though perhaps some rework of the deployment processes to make them more optimal for their new target network?). These days they care a lot more about what runs on Azure than what is written using .Net and even what runs on Windows, and are comfortable releasing their own code using other tech (VS.code being based on Electron being the first example that springs to mind). They'd prefer you used an MS stack from top to bottom of course, but they are more than happy for projects to use other components in/on Azure.
It'll be interesting to see how they would position it alongside TFS, as there is a lot of overlap between the two products. My guess is they'd keep pushing TFS for people who are completely MS shops and GH for people with more varied stacks.
1 reply →
The scale of products you're talking about are vastly different.
Meanwhile, I'd really like for people to stop hating Microsoft just because "Microsoft".
113 replies →
It’s very different when it’s a direct competitor, especially one you’re paying $ to every month. CEOs hate funding their competitors.
I still use Wunderlist and probably will until they forcefully shut it down. Tried the ToDo app, and I don't get why they basically tried to rewrite Wunderlist. The ToDo app was so buggy and crashed all the time not to mention it's lack of features. Hopefully Microsoft learns from this mistake of trying to do a crappy useless rewrite.
I was looking at draft.sh from the Azure open source team and the current (very early) version literally only works on Mac. The only install method listed is Homebrew.
Actually we have release assets for Mac, Windows, and Linux with support for 64-bit and ARM architectures. Someone from the community added Chocolatey support not too long ago. :)
Disclaimer: I am one of the core maintainers of Draft.
4 replies →
Not right away, but eventually MS "old boy network" cancer will metastasize there and push all the good people out, just like it did in every single other acquisition. This problem is not unique to Microsoft: other large companies have it too. But it is more acute, because at MS the old boys network is utterly incompetent, completely entrenched, and the company is 100% top-down. Frontline devs are expected to do as they are told, and are considered to be expendable "resources". In contrast, at SV companies managers are more expendable than good engineers.
So you have inside knowledge of their business purpose in aquiring GitHub? Are you posting this in an official capacity as a representative of the company? Or do you not really know and just elevating your two cents using the name of your employer?
I do have a bit of experience in acquisitions, and revenue is king, for which you have to move fast; changing the stack of an already working system would be suicide. Definitely expect more native GitHub integrations in Azure, and Azure first tools.
I have absolutely no knowledge of the deal.
I’m commenting on the tech stack discussion, given my experience in the company and if GitHub joins, I’m 99.99% sure they will stay on the tech stack that they have today. Look at LinkedIn, it is still running on Scala.
3 replies →
Except minecraft...
Microsoft bought one of my favorite pieces of cloud software from a few years ago Deis[1]. With that they also got Helm[2] with the purchase. They are doing GREAT with Helm and are going in a different direction that looks super cool as they mothball Deis called Draft[3]. They are moving away from the OS company they used to be and betting heavily on cloud technologies and I think this Github purchase makes sense. Github has been stagnate for years. MS is embracing open source in a way they haven't before, and I think they are doing so in a way that is going to surprise people.
In NO way am I a Microsoft fan boy. I've been windows free going on a decade. I run Linux Mint and OSX as my primary desktop environments. Apple is burning me hard, the way the computing world is going to change in the next couple years, cluster technology is going to be at it's core and we are going to see some very different things grow out of it. I'm as shocked as anyone to see MS play nice with linux and especially contribute how they have to Kubernetes; which I think is the largest open source project in the world right now?
What if MS dumped resources into world class CI tools to go with Github? What if they made a Github open source module and would let you federate your content? I could see this being a really interesting thing. They could also screw us all, but under their current management I think they are getting ready to be competitive in an emergent environment that can't exist without open source.
[1] https://deis.com [2] https://docs.helm.sh/using_helm/ [3] https://github.com/Azure/draft
> What if they made a Github open source module and would let you federate your content?
Wow, that's optimistic! I'd be happy if they just keep it neutral.
Fact is all the big vendors publish and collaborate on github, this purchase threatens that ecosystem. And we are more likely to see a message of private code hosting sites, than a solid federation system for source control.
I don't disagree with you on the threat to our ecosystem, but I think Github has been treading water for a while. I'm really excited to see Gitlab step up, especially with the traction they are getting from the diaspora, but I don't know that they will. There are a ton of reasons people won't take Gitlab seriously as a replacement right now. I hope they find a way to take their tools and this new attention and produce what we all really want.
I think at this exact moment, there is a really interesting space that can be filled by a FOSS Github alternative. I think a new player might be better equipped to offer it. Maybe someone can build something off of Keybase's[1] git services.
I am sad to see an independent voice go, but I don't think Github has been able to stay competitive and IMO this could be a good thing. That remains to be seen, but regardless of Gitlab's ability or what MS might do, I think there is a vacuum left that I'm hoping we can fill as a community and I'll put my effort and dollar behind whatever shows up to do it.
Someone should ask Linus what he thinks about git federation. Maybe he can save us from ourselves again.
[1] https://keybase.io/blog/encrypted-git-for-everyone
Helm was contributed to CNCF last week: https://landscape.cncf.io/cncf=hosted,graduated,incubating,s...
Draft (which started with Deis and is continuing with Azure) continues to get active development: https://landscape.cncf.io/grouping=landscape&landscape=appli...
I don't know how I haven't seen CNCF before, but can you give me (and anyone else who might not know) a brief summary of what they do?
Their page makes it clear that I should know, but doesn't give me an obvious place to click to get a clue.
EDIT:
Found some markety things here.
https://www.cncf.io/
------
What is CNCF?
CNCF is an open source software foundation dedicated to making cloud native computing universal and sustainable. Cloud native computing uses an open source software stack to deploy applications as microservices, packaging each part into its own container, and dynamically orchestrating those containers to optimize resource utilization. Cloud native technologies enable software developers to build great products faster.
5 replies →
Reminder: That didn't stop them from converting HoTMail over to NT/Exchange.
How many times did they try and publicly fail first? Two or three? Excellent validation for choosing FreeBSD and Apache for your 1999 startup. At least two announcements that it was "complete" turned out to be lies.
Years later the back end was still on FreeBSD and Solaris, but the front end was on Win 2000 using Windows Services for UNIX.
Twenty years ago!
It took them about 20 years to do so also.
> Given that GitHub is quite proudly built on Ruby
And for that matter, GitLab as well :-)