I feel it has been better that the more openly "bad" one of the two has acquired Github, because Google has this "don't be evil" face which is the fake one of its two faces, though many still just fall in for it, forgetting it's just another company with commercial interests.
>It's a pretty stupid motto from a public relations perspective.
Wait a minute. It was a great motto when Microsoft was clearly the enemy. It was a subtle Us versus Oracle, MS...
The motto gave Google massive developer mindshare and support for android and anything Google.
I've got no data to back this up - but I believe Google's key early employees may have joined the company partly because of this motto.
However, with no clear enemy in sight. And with Google's privacy stance, the motto's begun to backfire.
The last sentence from the book Animal farm fits perfectly
>Twelve voices were shouting in anger, and they were all alike. No question, now, what had happened to the faces of the pigs. The creatures outside looked from pig to man, and from man to pig, and from pig to man again; but already it was impossible to say which was which.
Oh, there was a whole round of shitshows with the original typical white male engineer founders, then hypocrisy with the diversity/sjw crowd brought in to save face.
indeed nothing has changed overnight. youre telling me it will stay that way?
the reason 'why' is that GH is no longer independent. there are thousands of comments on all the related posts here on hn for you to figure out why that might be significant..
They also have Visual Studio Profiles/Visual Studio Team Services, which appeared to be them trying to enter the space, but, bar people being forced to have a profile to use VS I doubt many are using the team or project features of that. Then there's one of their dev cash cows: Team Foundation Server, which I assume has been dying a slow death over the past few years due to git dominance in the source control space. I moved my team away from the shit show which is TFS for git (not github) several years back now.
So, I suspect this is shoring up their position, reducing the number of man ours needed in competing, and buying the best product on the market.
I wouldn't be surprised if github got more enterprise project management features, where it's a bit light right now, and perhaps more around management of super large projects. As well as tighter integration with Visual Studio (at the project management level, rather than just the source control level) - to phase out TFS over time.
We left TFS altogether because it was just so terrible, I've lost count of the number of hours lost to that software. We use Gitlab for source control now, Team City for builds, and Target Process for project management.
Perhaps there was some hyperbole in my statement, a more measured statement would be: I suspect that the more modern alternatives to TFS are perhaps starting to create a dent in the TFS cash cow, and M$ are hedging their position. Perhaps with an end goal of some interoperability, or full scale integration of github into TFS over time.
Obviously this is all speculation, only M$ knows why they did it, but it seems like a good bet if they want to be providers of SaaS solutions and one of their existing solutions is under threat.
> Microsoft wanted to be in Github's space but couldn't compete
Did they though? Codeplex predated Github by 2 years and I don't recall there ever being much competition between the two. If anything I think Stackoverflow and Nuget played bigger roles in obviating Codeplex.
Codeplex was launched as a community platform for hosting Windows based opensource projects and libraries. It had forums, bug tracking, and all sorts of niceties but Stackoverflow showed up a couple years after it's launch and became the defacto destination for community support. Around the same time Nuget landed and removed the need for most consumer developers to even visit Codeplex.
GitHub showed up the same time as SO and Nuget and was largely a platform agnostic way to host opensource libraries but wasn't any more appealing than Codeplex to Windows Devs at the time. Since Nuget became the primary means for Windows Devs to consume libraries and Nuget packages could be hosted on any webserver, there was never a driving reason to use Codeplex or Github.
Microsoft launched Visual Studio Online for enterprises three years later without even farting in Codeplex's general direction. A couple years later they added support for Git to Visual Studio and have been moving people to use it for version control over their proprietary solution ever since.
I think Microsoft's purchase of Github wasn't so much to control or insert itself into Open Source, but rather to protect itself from others who might trying to push them out.
Don't you guys think that GitHub was meant to give heads up about acquisition to its subscribers, at least to premium users? Yes this would give away the whole "secret" but if acquirer is some evil corp? Or a foreign entity? I feel like there is an important law missing regarding acquisition of companies who run premium plans.
The only thing that matters here is that they aren't actually owned by MS yet. They have just agreed to be bought soon. So, there's still a substantial time for customers to leave or adapt or whatever before the acquisition is completed.
It's true that it would be worse if one day you suddenly were already a customer of a different company without warning. That's not the case here though.
Side-note: I'm very critical of MS as a acompany and have been critical of GitHub separately, nothing above should be seen as defending them other than on acknowledging the fact that they've announced the acquisition long before it actaully will be completed.
Google buys GitHub. 6 months later it launches a new different code hosting site that is tightly tied with Google services. 3 years later new site and GitHub both shut down.
I think they know better. I mean, how couldn't they? If Microsoft trashed GitHub, that means millions of developers are never going to trust them with anything again. It'd be severe clownery to not see what's clear as day.
That's a good point, but it's one that cuts both ways: the same capacity for change in leadership and therefore behavior that means MS might no longer make the mistakes involved in IE (interested only in market dominance, not in developing/maintaining the potential of the technology) also means that in a decade MS might shift in how they value GitHub.
Out of the list of companies that the article says have expressed an interest in GitHub at one point or another - Microsoft, Google, Amazon, Atlassian and Tencent - I have to say that Microsoft seems the most best fit, especially given their existing work on integrating GitHub with GVFS:
How does Docker even make money? I don't think I've known anyone to pay for it. And with so many people moving to kubernetes for orchestration, I don't think any of their solutions would be useful/profitable.
Well, they've been losing money for several years now, but they were originally profitable and operated without any VC money for years. If it weren't for the problems that growth has caused them since then I'm not sure they would be looking at acquisition at this point.
sounds like MS is trying to make Git into Clearcase. Typical bunch of FUD stemming from not understanding the tool and how to work with it, and of course the monstrosity of a solution ( "Git at enterprise scale" (https://www.gvfs.io/)) to the rescue:
"Microsoft wanted to move to Git because of Git's features, like its easy branching and its popularity among developers. But the transition faced three problems. Git wasn't designed for such vast numbers of developers—more than 3,000 actively working on the codebase. Also, Git wasn't designed for a codebase that was so large, either in terms of the number of files and version history for each file, or in terms of sheer size, coming in at more than 300GB. When using standard Git, working with the source repository was unacceptably slow. Common operations (such as checking which files have been modified) would take multiple minutes.
The company's solution was to develop Git Virtual File System (GVFS). "
I think it is more that Team Foundation Server is failing to scale and MS needs [their own branded/controlled] "enterprise-scale" VC solution for themselves as
well as for their customers. GitHub with GVFS on top (ie. kind of Sharepoint for VC) fits nicely here.
I think the more likely benefit is that pretty much every engineer knows git, it has become the de-facto standard even if other alternatives like Mercurial are easier to learn and use. In terms of onboarding new engineers, this probably saves them a week where new engineers don't have to ramp up on a new technology. It might make a difference in terms of recruiting new engineers. In terms of understanding their customers, I'd guess that a significant number of customers are using git. It probably also makes it much easier to open-source code.
From what I've heard, TFVC would have scaled just fine, and was similar enough to the internal to Microsoft ancient Perforce fork called Source Depot, but it wouldn't have been particularly agile. TFVC like CSV/SVN has a harder time branching because it is branch-as-path/copy-on-write, for one instance. Scaling a repo that size requires intricate lock systems which TFVC supports, but aren't very developer friendly, etc, as the Windows team was already fighting with Source Depot.
The issue isn't that TFVC wouldn't have scaled, it's that Windows developers wanted git for many of the reasons that most of us use git: cheap branching, distributed/lock-free code development, etc.
GVFS is nothing like Clearcase/Perforce/Source Depot/TFVC. It's still git, it's just offloading some of the .git object database on a machine to servers, just like you might not download every file in your Dropbox/OneDrive account to a local machine.
GVFS doesn't even have to be "centralized". GVFS uses a simple REST API to query object servers and I still don't think anyone has had the need to do it yet, but building a GVFS daemon to pull objects from IPFS (and pin objects to it) is entirely possible.
>TFVC like CSV/SVN has a harder time branching because it is branch-as-path/copy-on-write, for one instance. Scaling a repo that size requires intricate lock systems which TFVC supports
heavy branching and locks is exactly the things which make system less scalable, in all dimensions like devs number, repo size, parallel product versions being developed etc.
>Windows developers wanted git for many of the reasons that most of us use git: cheap branching, distributed/lock-free code development, etc.
exactly my point - they hit scalability limits of TFVC. You just said it using different words. We all using Git because it is [architecturally] the most scalable VC out there (i did development of a major [back then] VC almost 2 decades ago - we were starting to develop bunch of ideas back then which ultimately came out to the wide public in git)
> heavy branching and locks is exactly the things which make system less scalable, in all dimensions like devs number, repo size, parallel product versions being developed etc.
It's not a contradiction. Those things scaled just fine for decades of software development. They aren't "nice", aren't "fun", but they are a scaling solution.
The Windows team used Source Depot ("Perforce") for something like 30+ years, at the scale of the Windows team. There are companies and repositories in the wild in Fortune Xty companies with equally old, equally "scaled" repositories in classic, ugly centralized lock-based repositories.
You and I may not want to work on any of those repositories, but they absolutely scaled in historic fact for a scale that still dwarfs the lifespan of git.
You may confusing workflow flexibility with scalability? Command-and-control workflows are very scalable, but not very flexible. Distributed workflows are very scalable, and very flexible, when built right.
The Windows team wanted a break from classic Waterfall and Command-and-Control workflows. They knew how to scale those and had scaled them successfully for many years. They used the change in source control as an opportunity to change workflows.
TFVC could have done the job if they still wanted that workflow (presumably better than the hodgepodge of very old Perforce code and Perl scripts that was Source Depot, that had "scaled" for decades).
There are still companies and developers happy with that workflow.
The interesting part is that Microsoft wanted "something like git", that enabled those new workflows, but that scaled to what they were already doing in centralized repositories with Windows. Instead of investing in a Distributed-TFS, they decided to just use git, and they put a lot of effort into performance tuning to git which was contributed upstream and which we are all benefiting from now. Yes, they also built a simple optional plugin (GVFS) that is itself open source, and built on simple open REST standards. It's pretty much the exact opposite of building "their own Clearcase" or whatever; they didn't build a D-TFVC for their needs, they contributed directly to the open source one in an open way, exactly as you would want them to.
> Those things scaled just fine for decades of software development.
There is a difference between “were used at scale because there were no known alternatives” and “scaled just fine”, and I think the facts in this case favor the former over the latter.
The argument is that "scale" is related to quantity in this discussion, not quality. I agree that quality of life is better now with distributed VCS (whether or not a historic alternative existed, which is orthogonal to the discussion at hand), but "scaled just fine" is not a quality judgment on my part in that statement you pull quoted, it is an acknowledgement that quantity was handled appropriately.
>You may confusing workflow flexibility with scalability?
when increasing of scale results in cumbersome jumps through the hoops added to the workflow or the otherwise working workflows become impossible at all - that is a scalability problem. Git's easy branching and multi-repo (local repos) based workflows approach was produced as the solution to the scalability problems of central repository and lock based systems.
Microsoft, love them or hate them, appear to have been turning over a new leaf for the past while. If trends are to be used as indicators, then it seems the trend is that more and more of Microsoft's projects are becoming more and more open, while it feels like Google is becoming more closed.
Yes, Microsoft still has closed source programs (Windows), and yes, Google still has open source programs (Android, Tensorflow), but the general feeling I'm getting from Google is shrinking away from a commitment to open source on practical grounds.
I don't know if I'm ready to say that I have more respect for Microsoft than Google at this point, but I definitely have more respect for Microsoft than I ever have, and that's the direct result of their recent efforts to embrace the developer community at large.
Yeah. That generally how you fool people you are a good person. You do stuff that appears to be good hearted. Companies do that too. Just like politicians.
I can't understand why people today lack the ability to see a scam.
Right -- I don't care what Bill Gates' motives for donating all the money to charity that he donates, because I believe that the money he's donating is doing actual good.
Maybe he feels guilty for some of his actions as Microsoft CEO, maybe he doesn't. Maybe he's cheating on his wife, or doing some other nefarious deed -- but those charities still have that money, and they're still (probably) doing good work with it.
I disagree. We're perfectly capable of judging an act on its own merits.
Orson Scott Card is a good author.
Orson Scott Card is a bigot, and I do not ever wish to give him any money, so I will not buy his books now that I know of his bigotry.
His bigotry doesn't cancel out his talent as an author, and while I might find that I don't enjoy reading his books any more as a result of what I know of him, that doesn't change what my opinion was before I found out about his bigotry.
A good act is a good act, even if it's surrounded by a dozen bad acts.
Companies (and people) are not pure good or pure evil. It's always a mix.
I'm not thrilled with windows 10 ads, but typescript, chakra js engine, dot net core, and others are open source and being developed in the open on github - and that's pretty cool.
In my opinion Microsoft's open source projects are not a scam to trick people into liking them. They're part of a broad realization that projects run like this will see broader adoption and generally will be higher quality.
Companies don't have souls. They're not intrinsically good or bad, but they may act good or bad depending on the leadership team. When the leadership changes the company's behavior changes along with it.
Nadella's microsoft is not the same as Ballmer's and Gates's microsoft. In Gates' day windows licenses brought in the money, and linux and the web were the dominant threats to microsoft's revenue, completely explaining microsoft's behavior around 2000 that earned them their reputation.
Nadella's microsoft however is a cloud services company, with office 365 and azure bringing in an ever larger slice of the pie, and windows reduced to a single digit percentage of the revenue stream. Cloud services are built on open source. It makes complete financial sense for microsoft to be a strong open source backer now.
Good or bad being relative. All companies are bad if you are the thing they consider expendable to make profit. All companies are good if you are a shareholder receiving profits.
Business has parted with ethics a long time ago. Many don't even pay it lip service putting bribes as line items on budgets.
Careful there. There is no consensus that humans do have souls. Personality might have been a better word to pick, since what you are describing better fits the idea of personality, but that is dangerous as well.
But regardless of the specific term we want that separates individuals from companies, there is no denying that companies are more then the sum of the individuals they are made of. Group psyche has been proven to prevail through replacement of every member. Though one should not jump to the opposite conclusion either—that companies’s behavior remains static through a change in leadership.
Let us know when scientists tell us we got a soul :) That will never happen. Nor because it's not true, we do (i think so anyway) but because it's not something you can find by digging around in the body. By definition, the soul is not physical.
> Careful there. There is no consensus that humans do have souls.
"Soul" doesn't have to mean "immortal ghost." Humans have souls tautologically; the word is one way of trying to get at something about our experience with ourselves and others. Personality doesn't quite cover it, psyche might be closer, but soul has the depth.
That said, I agree with the premise that human organizations can have a culture that starts to border if not overlap conceptually with "soul."
> Yeah. That generally how you fool people you are a good person. You do stuff that appears to be good hearted. Companies do that too. Just like politicians.
Companies aren't people though; they're made up of people.
So, where you could make the argument that it's very difficult for people to change, the people who make up a large company change everyday.
There are likely many "good" and "bad" people in a given company for any definition of good and bad. I'm sure the people at any sufficiently large company don't even agree on what is good or bad.
Have you worked inside a company and been any form of leader? There is a culture in all large companies to make things appear a certain way, internally and publicly, while people in charge carry out whatever they are told from their bosses.
Microsoft employees may also believe they are now becoming a nicer company. It would make sense to have that kind of internal communication, and it increases morale and productivity of you believe your work is positive.
>There is a culture in all large companies to make things appear a certain way, internally and publicly, while people in charge carry out whatever they are told from their bosses.
Every large company manages public relations. That does not mean that they manage public relations dishonestly.If a company wants to turn their business model around, even if only out of self-interest, that is likely going to be accompanied by a PR effort.
There is no reason to assume malicious intent behind every PR-campaign because no institution benefits from lying out of principle.
And in the case of Microsoft, this turn-around seems reasonable. Their old, proprietary model isn't making them that much money any more. They're trying to be a developer-friendly platform.
I'm not mistaken on their motives, but nor do I believe that anybody (or any company) is entirely one thing or the other.
On net though, Microsoft appears to be doing more things that I like, but if we're being pragmatic, I should like the things I determine to be 'good' whether or not they are also doing things I consider to be 'bad' -- but on the whole, if they're doing things I like just to make me more likely to contribute to their code, or work for them, or be a customer, then they're still doing more things that I like, scam or not.
>I can't understand why people today lack the ability to see a scam.
While I know that for some people in SV this may come as a shock but not every company is a ploy to use investor's money to pay yourself a fat salary for a year or two before declaring that grubhub for toilet paper is not a sustainable business model, laying off everyone and selling everything you scraped from your users cell phones to whoever will do the most harm with it.
Some companies actually optimize for making a sustainable profit by taking other people's money in exchange for delivering goods and services.
I agree with you except for the section above. I don't know that Microsoft has done anything remarkably "good". Today's Microsoft is centered around their surveillance, keylogging Operating System. It's centered around their deceptive & litigious forced upgrades to Windows 10.
Someone mentioned a piece of software called 'Yammer' (a freemium enterprise social networking service) a few days ago that I'm not familiar with. When I looked into it, I found that Microsoft had acquired Yammer in 2012 for US$1.2 billion.[0]
This is the current result of that acquisition: "The new integrations have implications for the way Microsoft sells Yammer. Microsoft is encouraging customers to migrate to its Office 365 plans. Effectively Microsoft will stop selling Yammer as a standalone service."[1]
I'm sure there are many more modern day examples. I'll have to delve back into Microsoft land to find out (MS refugee for a few years now after jumping ship to the privacy-respecting Linux).
To me, there is little difference between the Microsoft of 20 years ago and the Microsoft today. In many ways, with the addition of user-hostile surveillance and data-hording, Microsoft is far worse than they've ever been.
Agreed, however, we shouldn’t fool ourselves into believing ms would put the wider software community’s needs before their shareholder profits.
The situation with GitHub has radically changed, it is no longer a neutral and independent project which exists solely to provide a service to entire developer community. It is now a tool to leverage more profits for one software company — a company who has proven they care much more about market dominance than the wider tech ecosystem.
I’m not trying to imply some forgone conclusion that GitHub is doomed or that ms will ruin it, but the entire purpose of the project is now completely different than when it was an independent company and we would be wise to keep this in mind.
I particularly think we should be watching and cautious to see what types of changes they try to leverage into git itself.
The economics were aligned with the customers. The whole business model was based on paying for private services using the same great tooling that was available for open source projects. People developed familiarity with open source and naturally clamored for the same tooling within companies. It was a loss leader of sorts.
Now the economic incentives are different. This would be true if Google acquired Github or if Amazon or Facebook or Twitter or any other company acquired it. Since it's not a standalone company, the parent company optimizing for greatest comprehensive profits will make decisions that are not always optimal if GitHub were standalone.
I'm pretty sure that the economics were aligned with getting bought out by someone. Upselling people who used the free product was not making them profitable and they weren't on the path to profitable.
The GitHub story is complex. There was a time years ago when they were profitable: https://signalvnoise.com/posts/2486-bootstrapped-profitable-... They and Sidekiq were the poster children of bootstrapped profitable businesses around open source. Something very clearly changed (most likely raising a boatload from VC firms and the ensuing perils) that flipped the situation.
GitHub was relatively independent in that hosting git repositories was the only product that it sold and it didn't clearly favor any sort of tech stack. Now, MS says that they will not remove any of the existing functionality, but their press release did hint that they would add features to make it easier to launch apps onto MS platforms like Azure.
Sure, but it isn’t profit motive that I think we should be concerned about in this case. Its the entire space in which GitHub exists is now radically different. Until the buyout, they were an independent company who hosted software repos. Now they are owned by a company who has many many many different interests and ms has a tendency to use most of their various software projects to feed into their own proprietary ecosystem.
There are many ways this might be problematic, all of which are obviously just speculation. I’m just trying to say, an enormous part of the software ecosystem is currently in a really precarious spot and we might be wise to consider and work on some alternatives just to give ourselves some wiggle room should ms decide to behave in ways we’ve seen them behave so many times in the past.
Assuming for the sake of argument that they were acting wholly from benevolence anyway, that dynamic would have changed regardless of whether it was Microsoft or Google acquiring them.
The reason I point out the difference in vibe between the two companies is that while sure, there are definitely valid reasons to be skeptical here, I am leaning towards more optimism because it was Microsoft who made the acquisition and not Google.
Yeah I’m not sure any of the big tech empires purchasing GitHub would be necessarily better than the others. It would have been nice to see them bought by someone who we could consider neutral and someone who didn’t have such a long history of using products to leverage market dominance in other areas. But that’s not what happened :/
Im hopeful that ms stays out of it and let’s it remain it’s own entity but practically speaking I just really don’t see that happening.
Fortunately, acquisition of GitHub and the bread and butter income of todays's Microsoft (Office365) do not conflict. In fact, they probably can be brought together, usefully and economically, for Enterprise customers and Microsoft itself.
It could work well. We'll see how it goes. I am optimistic.
They make more money and business if they become developer friendly period. If they step back from that they know they will lose out for no true benefit! Microsoft knows this.
> I don't know if I'm ready to say that I have more respect for Microsoft than Google at this point, but I definitely have more respect for Microsoft than I ever have
You expressed it beautifully.
While I'm not a fan yet, I think Microsoft is much more invested in its developer and cloud products than Google. After all, most of Google's revenue still comes from ads.
I've been really confused about the reaction around this. It feels like so many people are just waiting to find out how this is a scam, like this is still the same Microsoft. "Oooh, just you wait, they're tricking you with their PR".
There's just been too many things that I would never think would come to pass to think that opinion is fair, too many different faces. MS seems to have put more effort into going where the developers and customers are instead of trying to wrangle them in.
Google’s main products always were and will be closed source they do have a lot of open source projects but so always has Microsoft especially in the academic scene they just weren’t as prolific and thus impactful.
But essentially the entire Google suite and all their commercial apps and services are close source.
Thank goodness for that. If Google had picked up GitHub I would have had to start a migration strategy since their buyout strategy has included killing off the service in < 2 years.
I find it strange to live in a time when tech people trust MS more than Google, it could be because we think differently about the importance of information tracking, and this is coming from someone that for the past 15 years has used only private/incognito browsers (was a lot of fun back in when Flash was popular and had to manually remove its cookies directory after every session), doesn't have a facebook account, doesn't use social media. But here's something to think about, MS is in a position where they're making a lot of money from closed source software products where a good free open source equivalent would be a direct attack on that revenue stream, which means, there can be many potential conflicts of interest with owning something like GitHub. What do you think would happen in the following scenarios:
- a researcher discovers a zero day security problem in Office/Windows/etc that affects almost all their products in the last 10 years, tells MS about it but doesn't wait for them to patch it for N months or whatever time they'd take so they push a working exploit/proof of concept on GitHub before the bug is patched. Will MS shut down the project/repository and/or sue the people behind it using IP/time information from their own logs?
- MS pushes another Windows 10 update that increases tracking, people write programs to work around/disable that and post it on GitHub. MS is arguing that without tracking they cannot determine who is pirating their Windows so they sue the authors. Can they get access to the IPs/times of when the repository was interacted with without a court order? Can they shut down the project on GitHub before waiting for the lawsuit to reach a conclusion?
- MS develops a new DirectX API that allows for many great things and some very hyped AAA games are already in the pipeline to be released to use it, but the catch is the new API requires Windows 11, a new version that costs money to upgrade to, previous Windows owners cannot use it. A Wine developer implements an emulation of the API on GitHub and releases a DLL build for Windows, suddenly all previous Windows users can play those games. What will MS do about that project?
I feel it's pretty easy to imagine lots of situations where there's a conflict of interest between being a neutral Git repository and being a software company so I can't say I'm very happy with MS purchasing it tho I agree with others that it could have been much worse too :)
The vast majority of projects do not require reliability above what can be simply and cheaply achieved. If/once your project gets to the point where reliability is a concern, that's the exact reason you wouldn't want to rely on a free/cheap 3rd party. The case in point is the one I referred to, GIMP had already seen SourceForge as an issue and had abandoned them over two years before the incident. And lots of other projects have taken on the task of hosting their own. However "gargantuan" the task may be, it's still usually a drop in the bucket compared to the overall project.
140 comments
[ 4.6 ms ] story [ 175 ms ] threadWait a minute. It was a great motto when Microsoft was clearly the enemy. It was a subtle Us versus Oracle, MS...
The motto gave Google massive developer mindshare and support for android and anything Google.
I've got no data to back this up - but I believe Google's key early employees may have joined the company partly because of this motto.
However, with no clear enemy in sight. And with Google's privacy stance, the motto's begun to backfire.
The last sentence from the book Animal farm fits perfectly
>Twelve voices were shouting in anger, and they were all alike. No question, now, what had happened to the faces of the pigs. The creatures outside looked from pig to man, and from man to pig, and from pig to man again; but already it was impossible to say which was which.
> And remember… don’t be evil, and if you see something that you think isn’t right – speak up!
https://abc.xyz/investor/other/google-code-of-conduct.html
the reason 'why' is that GH is no longer independent. there are thousands of comments on all the related posts here on hn for you to figure out why that might be significant..
[1]: https://cacm.acm.org/magazines/2016/7/204032-why-google-stor...
[2]:https://en.wikipedia.org/wiki/CodePlex
1: https://www.visualstudio.com/de/team-services/
i just wanna take that in for a moment. when freaking Microsoft can't compete with you on enterprise software... you've done well.
So, I suspect this is shoring up their position, reducing the number of man ours needed in competing, and buying the best product on the market.
I wouldn't be surprised if github got more enterprise project management features, where it's a bit light right now, and perhaps more around management of super large projects. As well as tighter integration with Visual Studio (at the project management level, rather than just the source control level) - to phase out TFS over time.
TFS also supports git repositories. Do you use TFVC (Team Foundation Version Control) and TFS interchangeably?
Obviously this is all speculation, only M$ knows why they did it, but it seems like a good bet if they want to be providers of SaaS solutions and one of their existing solutions is under threat.
I guess everyone already forgot about code.google.com.
Did they though? Codeplex predated Github by 2 years and I don't recall there ever being much competition between the two. If anything I think Stackoverflow and Nuget played bigger roles in obviating Codeplex.
Codeplex was launched as a community platform for hosting Windows based opensource projects and libraries. It had forums, bug tracking, and all sorts of niceties but Stackoverflow showed up a couple years after it's launch and became the defacto destination for community support. Around the same time Nuget landed and removed the need for most consumer developers to even visit Codeplex.
GitHub showed up the same time as SO and Nuget and was largely a platform agnostic way to host opensource libraries but wasn't any more appealing than Codeplex to Windows Devs at the time. Since Nuget became the primary means for Windows Devs to consume libraries and Nuget packages could be hosted on any webserver, there was never a driving reason to use Codeplex or Github.
Microsoft launched Visual Studio Online for enterprises three years later without even farting in Codeplex's general direction. A couple years later they added support for Git to Visual Studio and have been moving people to use it for version control over their proprietary solution ever since.
I think Microsoft's purchase of Github wasn't so much to control or insert itself into Open Source, but rather to protect itself from others who might trying to push them out.
It's true that it would be worse if one day you suddenly were already a customer of a different company without warning. That's not the case here though.
Side-note: I'm very critical of MS as a acompany and have been critical of GitHub separately, nothing above should be seen as defending them other than on acknowledging the fact that they've announced the acquisition long before it actaully will be completed.
Source: Am LinkedIn Employee. Thats how long it took from public announcement through SEC approval.
"Why do I need a Google+ account to use GitHub?!?"
I look forward to the pull requests from Xbox Live users. Commit message: "You suck, you n00b loser. I f*ed your mom."
https://twitter.com/jntrnr/status/1003869360002220032
Why does such an article not at least mention Google code? The digital memory is short.
https://arstechnica.com/gadgets/2017/11/microsoft-and-github...
If anything, the failing "Docker the company" probably helps sell their competing implementations of the various OCI specs.
sounds like MS is trying to make Git into Clearcase. Typical bunch of FUD stemming from not understanding the tool and how to work with it, and of course the monstrosity of a solution ( "Git at enterprise scale" (https://www.gvfs.io/)) to the rescue:
"Microsoft wanted to move to Git because of Git's features, like its easy branching and its popularity among developers. But the transition faced three problems. Git wasn't designed for such vast numbers of developers—more than 3,000 actively working on the codebase. Also, Git wasn't designed for a codebase that was so large, either in terms of the number of files and version history for each file, or in terms of sheer size, coming in at more than 300GB. When using standard Git, working with the source repository was unacceptably slow. Common operations (such as checking which files have been modified) would take multiple minutes.
The company's solution was to develop Git Virtual File System (GVFS). "
I think it is more that Team Foundation Server is failing to scale and MS needs [their own branded/controlled] "enterprise-scale" VC solution for themselves as well as for their customers. GitHub with GVFS on top (ie. kind of Sharepoint for VC) fits nicely here.
The issue isn't that TFVC wouldn't have scaled, it's that Windows developers wanted git for many of the reasons that most of us use git: cheap branching, distributed/lock-free code development, etc.
GVFS is nothing like Clearcase/Perforce/Source Depot/TFVC. It's still git, it's just offloading some of the .git object database on a machine to servers, just like you might not download every file in your Dropbox/OneDrive account to a local machine.
GVFS doesn't even have to be "centralized". GVFS uses a simple REST API to query object servers and I still don't think anyone has had the need to do it yet, but building a GVFS daemon to pull objects from IPFS (and pin objects to it) is entirely possible.
is pretty contradicting to
>TFVC like CSV/SVN has a harder time branching because it is branch-as-path/copy-on-write, for one instance. Scaling a repo that size requires intricate lock systems which TFVC supports
heavy branching and locks is exactly the things which make system less scalable, in all dimensions like devs number, repo size, parallel product versions being developed etc.
>Windows developers wanted git for many of the reasons that most of us use git: cheap branching, distributed/lock-free code development, etc.
exactly my point - they hit scalability limits of TFVC. You just said it using different words. We all using Git because it is [architecturally] the most scalable VC out there (i did development of a major [back then] VC almost 2 decades ago - we were starting to develop bunch of ideas back then which ultimately came out to the wide public in git)
It's not a contradiction. Those things scaled just fine for decades of software development. They aren't "nice", aren't "fun", but they are a scaling solution.
The Windows team used Source Depot ("Perforce") for something like 30+ years, at the scale of the Windows team. There are companies and repositories in the wild in Fortune Xty companies with equally old, equally "scaled" repositories in classic, ugly centralized lock-based repositories.
You and I may not want to work on any of those repositories, but they absolutely scaled in historic fact for a scale that still dwarfs the lifespan of git.
You may confusing workflow flexibility with scalability? Command-and-control workflows are very scalable, but not very flexible. Distributed workflows are very scalable, and very flexible, when built right.
The Windows team wanted a break from classic Waterfall and Command-and-Control workflows. They knew how to scale those and had scaled them successfully for many years. They used the change in source control as an opportunity to change workflows.
TFVC could have done the job if they still wanted that workflow (presumably better than the hodgepodge of very old Perforce code and Perl scripts that was Source Depot, that had "scaled" for decades).
There are still companies and developers happy with that workflow.
The interesting part is that Microsoft wanted "something like git", that enabled those new workflows, but that scaled to what they were already doing in centralized repositories with Windows. Instead of investing in a Distributed-TFS, they decided to just use git, and they put a lot of effort into performance tuning to git which was contributed upstream and which we are all benefiting from now. Yes, they also built a simple optional plugin (GVFS) that is itself open source, and built on simple open REST standards. It's pretty much the exact opposite of building "their own Clearcase" or whatever; they didn't build a D-TFVC for their needs, they contributed directly to the open source one in an open way, exactly as you would want them to.
There is a difference between “were used at scale because there were no known alternatives” and “scaled just fine”, and I think the facts in this case favor the former over the latter.
https://blogs.msdn.microsoft.com/devops/2017/05/30/optimizin...
when increasing of scale results in cumbersome jumps through the hoops added to the workflow or the otherwise working workflows become impossible at all - that is a scalability problem. Git's easy branching and multi-repo (local repos) based workflows approach was produced as the solution to the scalability problems of central repository and lock based systems.
Yes, Microsoft still has closed source programs (Windows), and yes, Google still has open source programs (Android, Tensorflow), but the general feeling I'm getting from Google is shrinking away from a commitment to open source on practical grounds.
I don't know if I'm ready to say that I have more respect for Microsoft than Google at this point, but I definitely have more respect for Microsoft than I ever have, and that's the direct result of their recent efforts to embrace the developer community at large.
I just hope Microsoft doesn't inflict the moral equivalent of XSLT, XPath, and XML-RPC proxies on us.
I can't understand why people today lack the ability to see a scam.
Doing good things is not a scam.
Maybe he feels guilty for some of his actions as Microsoft CEO, maybe he doesn't. Maybe he's cheating on his wife, or doing some other nefarious deed -- but those charities still have that money, and they're still (probably) doing good work with it.
it’s necessary for the good to precede the bad in a way that demonstrates the good was to prevent the detection or objection to the subsequent bad.
using good deeds to atone for past transgressions is different in a way that is subtle.
Orson Scott Card is a good author.
Orson Scott Card is a bigot, and I do not ever wish to give him any money, so I will not buy his books now that I know of his bigotry.
His bigotry doesn't cancel out his talent as an author, and while I might find that I don't enjoy reading his books any more as a result of what I know of him, that doesn't change what my opinion was before I found out about his bigotry.
A good act is a good act, even if it's surrounded by a dozen bad acts.
so those good deeds aren’t poisoned by the bigotry.
if you found out that he set out to make stellar art for the explicit purpose of softening people to his bigotry, then it would be a smoke screen.
I'm not thrilled with windows 10 ads, but typescript, chakra js engine, dot net core, and others are open source and being developed in the open on github - and that's pretty cool.
In my opinion Microsoft's open source projects are not a scam to trick people into liking them. They're part of a broad realization that projects run like this will see broader adoption and generally will be higher quality.
Nadella's microsoft is not the same as Ballmer's and Gates's microsoft. In Gates' day windows licenses brought in the money, and linux and the web were the dominant threats to microsoft's revenue, completely explaining microsoft's behavior around 2000 that earned them their reputation.
Nadella's microsoft however is a cloud services company, with office 365 and azure bringing in an ever larger slice of the pie, and windows reduced to a single digit percentage of the revenue stream. Cloud services are built on open source. It makes complete financial sense for microsoft to be a strong open source backer now.
Business has parted with ethics a long time ago. Many don't even pay it lip service putting bribes as line items on budgets.
Careful there. There is no consensus that humans do have souls. Personality might have been a better word to pick, since what you are describing better fits the idea of personality, but that is dangerous as well.
But regardless of the specific term we want that separates individuals from companies, there is no denying that companies are more then the sum of the individuals they are made of. Group psyche has been proven to prevail through replacement of every member. Though one should not jump to the opposite conclusion either—that companies’s behavior remains static through a change in leadership.
"Soul" doesn't have to mean "immortal ghost." Humans have souls tautologically; the word is one way of trying to get at something about our experience with ourselves and others. Personality doesn't quite cover it, psyche might be closer, but soul has the depth.
That said, I agree with the premise that human organizations can have a culture that starts to border if not overlap conceptually with "soul."
Companies aren't people though; they're made up of people.
So, where you could make the argument that it's very difficult for people to change, the people who make up a large company change everyday.
There are likely many "good" and "bad" people in a given company for any definition of good and bad. I'm sure the people at any sufficiently large company don't even agree on what is good or bad.
Microsoft employees may also believe they are now becoming a nicer company. It would make sense to have that kind of internal communication, and it increases morale and productivity of you believe your work is positive.
Every large company manages public relations. That does not mean that they manage public relations dishonestly.If a company wants to turn their business model around, even if only out of self-interest, that is likely going to be accompanied by a PR effort.
There is no reason to assume malicious intent behind every PR-campaign because no institution benefits from lying out of principle.
And in the case of Microsoft, this turn-around seems reasonable. Their old, proprietary model isn't making them that much money any more. They're trying to be a developer-friendly platform.
https://www.engadget.com/2017/04/05/microsoft-explains-windo...
On net though, Microsoft appears to be doing more things that I like, but if we're being pragmatic, I should like the things I determine to be 'good' whether or not they are also doing things I consider to be 'bad' -- but on the whole, if they're doing things I like just to make me more likely to contribute to their code, or work for them, or be a customer, then they're still doing more things that I like, scam or not.
While I know that for some people in SV this may come as a shock but not every company is a ploy to use investor's money to pay yourself a fat salary for a year or two before declaring that grubhub for toilet paper is not a sustainable business model, laying off everyone and selling everything you scraped from your users cell phones to whoever will do the most harm with it.
Some companies actually optimize for making a sustainable profit by taking other people's money in exchange for delivering goods and services.
I agree with you except for the section above. I don't know that Microsoft has done anything remarkably "good". Today's Microsoft is centered around their surveillance, keylogging Operating System. It's centered around their deceptive & litigious forced upgrades to Windows 10.
Someone mentioned a piece of software called 'Yammer' (a freemium enterprise social networking service) a few days ago that I'm not familiar with. When I looked into it, I found that Microsoft had acquired Yammer in 2012 for US$1.2 billion.[0]
This is the current result of that acquisition: "The new integrations have implications for the way Microsoft sells Yammer. Microsoft is encouraging customers to migrate to its Office 365 plans. Effectively Microsoft will stop selling Yammer as a standalone service."[1]
I'm sure there are many more modern day examples. I'll have to delve back into Microsoft land to find out (MS refugee for a few years now after jumping ship to the privacy-respecting Linux).
To me, there is little difference between the Microsoft of 20 years ago and the Microsoft today. In many ways, with the addition of user-hostile surveillance and data-hording, Microsoft is far worse than they've ever been.
[0] https://en.wikipedia.org/wiki/Yammer
[1] https://venturebeat.com/2016/09/26/microsoft-is-killing-yamm...
The situation with GitHub has radically changed, it is no longer a neutral and independent project which exists solely to provide a service to entire developer community. It is now a tool to leverage more profits for one software company — a company who has proven they care much more about market dominance than the wider tech ecosystem.
I’m not trying to imply some forgone conclusion that GitHub is doomed or that ms will ruin it, but the entire purpose of the project is now completely different than when it was an independent company and we would be wise to keep this in mind.
I particularly think we should be watching and cautious to see what types of changes they try to leverage into git itself.
Now the economic incentives are different. This would be true if Google acquired Github or if Amazon or Facebook or Twitter or any other company acquired it. Since it's not a standalone company, the parent company optimizing for greatest comprehensive profits will make decisions that are not always optimal if GitHub were standalone.
https://www.crunchbase.com/organization/github
Sure, but it isn’t profit motive that I think we should be concerned about in this case. Its the entire space in which GitHub exists is now radically different. Until the buyout, they were an independent company who hosted software repos. Now they are owned by a company who has many many many different interests and ms has a tendency to use most of their various software projects to feed into their own proprietary ecosystem.
There are many ways this might be problematic, all of which are obviously just speculation. I’m just trying to say, an enormous part of the software ecosystem is currently in a really precarious spot and we might be wise to consider and work on some alternatives just to give ourselves some wiggle room should ms decide to behave in ways we’ve seen them behave so many times in the past.
The reason I point out the difference in vibe between the two companies is that while sure, there are definitely valid reasons to be skeptical here, I am leaning towards more optimism because it was Microsoft who made the acquisition and not Google.
Im hopeful that ms stays out of it and let’s it remain it’s own entity but practically speaking I just really don’t see that happening.
It could work well. We'll see how it goes. I am optimistic.
You expressed it beautifully.
While I'm not a fan yet, I think Microsoft is much more invested in its developer and cloud products than Google. After all, most of Google's revenue still comes from ads.
There's just been too many things that I would never think would come to pass to think that opinion is fair, too many different faces. MS seems to have put more effort into going where the developers and customers are instead of trying to wrangle them in.
But essentially the entire Google suite and all their commercial apps and services are close source.
it makes zero sense for MS to look at private repos because the moment they are discovered everyone would move off the system.
they also run outlook.com and office online. if they got caught spying they'd lose all corporate customers
- a researcher discovers a zero day security problem in Office/Windows/etc that affects almost all their products in the last 10 years, tells MS about it but doesn't wait for them to patch it for N months or whatever time they'd take so they push a working exploit/proof of concept on GitHub before the bug is patched. Will MS shut down the project/repository and/or sue the people behind it using IP/time information from their own logs?
- MS pushes another Windows 10 update that increases tracking, people write programs to work around/disable that and post it on GitHub. MS is arguing that without tracking they cannot determine who is pirating their Windows so they sue the authors. Can they get access to the IPs/times of when the repository was interacted with without a court order? Can they shut down the project on GitHub before waiting for the lawsuit to reach a conclusion?
- MS develops a new DirectX API that allows for many great things and some very hyped AAA games are already in the pipeline to be released to use it, but the catch is the new API requires Windows 11, a new version that costs money to upgrade to, previous Windows owners cannot use it. A Wine developer implements an emulation of the API on GitHub and releases a DLL build for Windows, suddenly all previous Windows users can play those games. What will MS do about that project?
I feel it's pretty easy to imagine lots of situations where there's a conflict of interest between being a neutral Git repository and being a software company so I can't say I'm very happy with MS purchasing it tho I agree with others that it could have been much worse too :)
[1] https://github.com/W4RH4WK/Debloat-Windows-10 [2] https://github.com/10se1ucgo/DisableWinTracking
IMHO the cost of self hosting your projects has been too low for many years now to give control of your web presence to a third party.