Really hope these talks fall through and nothing comes of it. Too many bad experiences with Microsoft buying something out, mismanaging it, and it goes down the drain. As much as I love GitHub, I won't be using a Microsoft-owned GitHub for any of my company's work.
How many projects are there out there that repositories hosted in GH, run in GH, and have package manifests that directly reference other GH repos. I'm sure there are a lot.
Golang can be the biggest offender in this regard because you import by typing the URL to a github repository... Better hope someone mirrors those projects if they ever go poof. It's not even a generic import to a package[0]. Gotta be fun to have to update all those imports later if a project moves to another host to avoid Microsoft.
Well that's the beauty of git, as long as anyone anywhere has a checkout of the Repository nothing will be lost. If tags are GPG signed you don't even have to put any trust in the surviveing source. That said for mantained code 99.9% of the time you would just use one of the core develops local copies.
In SIT (decentralized collaborative information tracker, https://sit.fyi) there's an issue tracking module and an import tool (https://github.com/sit-fyi/sit-import) that allows to convert all issues and PRs into self-contained issues that can be worked with using SIT issue-tracking,
so at least there's a relatively smooth way out to decentralized issue tracking.
> IMO that's because LinkedIn wasn't a great product to begin with.
Linkedin is way way better than other job sites. For instance, Glassdoor doesn't even support basic keyword search. If you happen to track job ads for C programmers then you'd get all ads which may have a C character somewhere in their text, and forget abou searching for C++ because Glassdoor doesn't recognize the plus character. With linkedin at least basic keyword searches do work.
It's just the stickiest app of all time To me it barely exists. I'm a hirer in the UK.
Before anyone touts how much wider our reach might be for potential hires, can I point out that my little firm has hired people from the other end of Europe directly ("real" friend of friend etc). Oh and just in case you are wondering - we pay and treat everyone equally and equitably.
I'd argue LinkedIn is much worse than it was pre-acquisition, but it was already on that trajectory. The "news feed" and discussions around current events, etc. They've locked down more and more of it without a $30/mo subscription. Privacy controls are a nightmare to decipher.
I think it's significantly worse albeit probably short of totally "screwed up". Just marginally worse in a lot of small ways that add up to a less useful, less efficient experience for me.
just so you know, i know quite a few people at both linkedin and microsoft and for the most part they are still very seperate entities. linkedin is still using google docs and slack for instance. most of the influence microsoft has had on linkedin in smaller isolated integrations. so anything youve seen go downhill in linkedn consumer app is very much due to independent choices/acts by linkedin itself.
As someone with lots of friends at LinkedIn (I used to work there), I can agree somewhat. But I also know that things can be screwed up in non-obvious ways to rank and file employees (Tumblr/Yahoo and Flickr/Yahoo as examples, which I sadly was there for both).
VSTS, Microsoft's hosted TFS solution, gives you unlimited private git repos for up to 5 users, 40 hours worth of hosted builds per month, complete project management, a private markdown based Wiki, and free orchestration of local build and deployment agents.
1. They already have such offering (VS Team Services).
2. Their competitors (GitLab, Bitbucket) already have that offering.
3. I don't think they're interested in GitHub for its revenue, but for its strategic value. GitHub has to protect its revenue model by disallowing free private repos, Microsoft can afford to be more flexible.
4. They've already made similar moves. They acquired Xamarin, which was expensive, and made it free.
It got more bloated... My kids still use an old version to avoid that (they only play with each other so they don't need the latest version that is required to play on most servers online)
Agreed. Buying Nokia of all things stinks of Ballmer. I was listening to an interview where he seemingly took responsibility for the whole stuff up. Not sure how much weight that holds
Buying Nokia wasn’t the issue. Microsoft’s handling of Nokia was the issue. Microsoft is not good at making a physical product and releasing it to the world like Apple and Samsung. The fact many countries had to wait > 12 months to get the latest phone was retarded. Windows phone was awesome, but Microsoft screwed themselves by not doing minimum what it’s competitors were doing.
Huh? Microsoft has several very successful hardware product lines: peripherals (the Microsoft ergo keyboard is the gold standard in many circles), Xbox, the surface line. Maybe they can’t generate Apple hype but they do know how to make consumer hardware.
I don’t think surface qualifies as good consumer hardware. It has been pleagued with driver problems and weird behaviour. But agree on periphericals and Xbox.
The surface line isn’t avaliable world wide on release. Windows phones were released in like 3 markets. Less than surface line. Xbox gets world wide releases.
Phones, people generally want what’s new. A 12-18 month old phone is not new.
I had the top-of-the-line Nokia Lumia 920 (designed just before the acquisition), and the plastic case was too soft and collected dents & scratches like crazy. The matte finish was also too slippery to hold onto (I ended up putting some skateboard tape on it).
OTOH, I found the Windows Phone operating system to be highly responsive and I was able to navigate screens more quickly than the previous iPhone or the current Android I have. The problem was that as third-place in the market, few apps were available for it, and those that were weren't being maintained.
They had a decent Android app, then someone saw that it was too fast/responsive and was using too little memory and took that as a challenge, how can I make skype worst? React Native!
Obviously that didn't pleased people (see rating in Play Store), and they had to provide something for low/mid range phones, and skype lite was born (which is really good one).
The design history of Skype for iPhone was particularly misguided: First it was a native UIKit app, then MS put a lot of effort into making it look and behave exactly like a Windows Phone app [1], down to the unusual gestures. When they gave up on that, the app reverted to a native UIKit design, but just a few updates later we got the current Snapchat clone instead. Needless to say, features regularly got lost during these radical redesigns.
That‘s still very clearly the same Microsoft where a gazillion layers of dumb middle management ruin everything[2].
They undoubtedly did a bad job, but I think that Skype has a lot going against it from external factors anyway.
Everyone I know is using FaceTime or Facebook Messenger. The only time I hear Skype anymore is people using it for podcasts. Obvious anecdata but it seems reasonable that services with better device integration (FaceTime) or social graph access (Facebook) would do better in the long run.
Skype used to be mainstream. It was essentially a cross platform iMessage & FaceTime, almost ahead of its time. It was also efficient in terms of network usage, being peer-to-peer for the most part.
Had it continued to be usable, it would still be mainstream and might have the place of WhatsApp today. Instead, we now have this garbage Snapchat-like UI, a server-based model that doesn't even work (it was supposed to "fix" the P2P model, but for me it's been way worse than P2P which always worked), and a shitty Electron client that melts your battery and looks awful (they killed a gorgeous - especially on Mac - native client to replace it with this shit).
Hangouts is a great example of another horrible UI. The whole tie in to Chrome made it unusable for anyone non technical (where did it go? What did you want me to click? I am in the call, but you're not there... oh you're in another call? How do you hangup? I don't see the window anymore, etc.)
A standalone app would fix this, but there seems to be some strange internal agenda tying it to Chrome... Chromebooks I guess?
I think people forget how bad Skype was in a lot of ways though. The big reason they shifted to the cloud model, for instance, was how unreliable message delivery was on phones with peer-to-peer. I remember when, in order to send images, you and the other person had to be online at the same time.
They made some missteps, sure, but Skype was already a very dated solution.
Haven't really used group chats that much on it, but for call establishment, I stand behind my words when I say it used to be a lot better.
With P2P it would sometimes take a bit longer (10 seconds?) but it would always work. When it wouldn't, it was because of a problem on either one of the endpoints, and switching network connections or rebooting the router fixed it.
With the current model it would just randomly not work (the client also sucks and sometimes hangs in an inconsistent state after the failed call attempts), without anything we can do - it's something with the software and/or servers.
I used to trust Skype calls to actually go through - now even if I see "ringing" I have no faith in whether it's actually ringing on the other side (and if it was, whether the call would actually connect once picked up).
They wanted total control over the data so they converted the p2p supernodes to Microsoft servers... now the government can get what it needs more easily.
It doesn't provide any value to the regular user; most people I know receive lots of emails mostly from body shops but rarely get a job from LinkedIn. The UI is designed to make it hard to configure privacy, and to make you casually spill your contact list. They are always trying to sell premium accounts.
To be fair it should provide some value to a junior dev who doesnt have a Network yet but LinkedIn doesn't even try to hide the fact that you are the product (unlike say fb).
Shameless plug here but I'm also so sick of LinkedIn that I've started building an alternative for developers by myself - https://able.bio/
Still very much alpha but the idea is to have a clean UI and allow developers to write their own insightful articles for the community to read and reward the author with reputation. Like Medium but with more specific language, library, framework topics etc. Then use that to help recruiters who can't code advertise and source candidates for technical interviews. No inbox spamming free for all, they can only contact you when you've applied for a position.
I'd be interested to know what people think of this and if they have any feedback.
So, previously Microsoft used to have a program to integrate new companies, called “venture integrations”. I wonder if LinkedIn is going through VI, or not.
I don’t know, isn’t Microsoft the only software company left? I love Google, but they sell ads, their mission isn’t to make software.
Microsoft is more and more cloud and services, but they still have a developer mentality and they have done some good stuff with open source lately.
If someone is going to buy Github, Microsoft is probably the most friendly to developers. Unless Google or Facebook wants to go full altruist and buy for the credibility.
I'm only a relatively casual user, but doesn't Atlassian mostly buy and then manage existing software products? They didn't develop Trello, HipChat, Bitbucket etc.
Every time I run into an issue with Atlassian software, I find a JIRA ticket from 7 years ago. If they have a capable dev team, it's probably working on projects off my radar?
A lot of people would say Google these days... now that they're starting to move beyond Material Design 1.0 a lot of their products are becoming more visually mature and unique, along with their substantial investment in machine learning to make their software more personalized, etc.
How has Kaggle fared under Google? Just asking. Also, ms has been focusing on open source quite a bit. Largest contributor on Github. Makes sense. It'll also send a message to the tech community. Mostly Google hogs the limelight for open source stuff.
Ignoring Azure, all the major MS products I can think of are at least partially ad supported (even when they release something that doesn’t serve ads, it’s a safe bet that it gathers data for better targeting...)
Hell, Windows is ad-supported now. Ever updated and get your free "gift" of Bubble Witch Saga 2: Electric Boogalo? I would pay double for an enterprise version of Win10 without that crap.
I was at a container orchestration conference the other day, and got the distinct impression that microsoft are not the great satan, and azure is probably a really good choice for cloud service. I can't believe I'm saying that. I'm having dinner with a prolific debian maintainer tonight, maybe he'll beat some sense into me.
All 3 major clouds have consoles that are good in certain things but mostly bad. Azure's console made a big mistake with the vertical blade design, but at least it always loads fast.
The REST APIs and CLI/Powershell interfaces are much, much nicer. The web Portal seems to get revamped every other year and the latest incarnation is not very good.
That would've been my exact reaction while reading this. Good thing I have no large projects on Github.
Well, I guess Gitlab would be happy about the new customers. If they're clever, they'll work on some tools to easen the transition.
From this perspective, Git has made some fundamental things right: Because it only functions as a decentralized system, it strongly reduces lock-in. In the case of Github, you "only" had to port over issues and pull requests. Imagine how great it would be if there was a standard way to save issues in a git repo[1] that was supported by pretty much every provider.
There's no standard, of course, but [shameless plug] I've recently built SIT (https://sit.fyi) -- a tool that allows to collaborate on information in a decentralized, "true serverless" manner. Its first application is (duh!) issue tracking and it has been operational since almost day zero and SIT itself is using it. There's even a GitHub -> SIT import tool (pre-release)
Now, what's also great (did I mention "shameless"?) about the approach used in SIT is that while you can use it with Git (and this is how I've been using it so far), it does not depend on Git's structures but just files -- so it can be easily carried over to whatever might replace Git in a decade.
I've just been thinking: If one manages issues within a separate git repo -- couldn't we also store pull requests this way? I've always been very impressed by how good `git diff` and `git apply` work together, so theoretically, one should just be able to store the result of `git diff master...HEAD` in a new issue. The rest sounds like a tooling problem...
I would understand if they sold, but it would make me sad. GitHub has done a great job. I'd hate to see them have to answer to someone other than themselves.
I agree and I would feel that same way if it were any other large tech company.
So many nice products have passed on in acquisition either because users abandoned them or the acquiring party did. I'm not sure which would happen first here.
If GitHub is financially strapped and looking for a buy out, I hope they're exploring other options for funding.
GitHub clearly needs change of management. I havent seen much inmovation from them compared to GitLab. Microsoft is very strong on innovation since change of management.
GitHub is easy to use, and I’ve found the important features work really well.
I’m all for improvements, but I’m not even sure what I would want. At this point, I’m more worried any attempt at innovation would break the things I like about GitHub now, than I am excited for something new.
I use gitlab for all my private projects and don’t find it to be remarkably better in any way, and the UI always confuses me.
Well, my only request would be that it be more like GitHub.
The only reason for this is familiarity. GitHub is the market leader now and I use it every day at work. It's a lot easier to get used to a new UI if it is similar to the one you are used to. I'm sure if everyone used GitLab at work everyday, it would be fine. The UI doesn't strike me as bad, just very different.
I'm not super familiar with GitLab besides self hosted instances. Even then I prefer the simplicity of gogs. So what do you feel are still be innovations GitLab has but GitHub lacks?
Otoh, GitLanpb seems to have a messy code that is slow and taking much memory by slapping new stuff in frequently. (My impression from reading several comments on HN.)
Our team is actively working on improving the performance of GitLab.com. If you are referring to the merge requests, we're close to converting the merge request view to Vue, this should speed it up considerably.
The list of key MS acquisitions isn't that long and out of several products only Nokia is a definite failure. Skype gotten worse but I wouldn't count it as failure just yet. Remaining big acquisitions (Forethought PowerPoint, Hotmail, Visio, Navision Dynamics, LinkedIn) where rather successful.
Some of us are old enough to remember Microsoft of the 90's and early aughties. "Embrace, Extend, Extinguish" was the name of the game then, many of us haven't forgiven Microsoft for that yet (and nor do we plan to).
Right, I know that, and that's just idiotic behavior. You should concentrate on what is happening today, not what happened decades ago and hold grudges. That attitude leads you into unhappy places.
Yeah, I agree. I would immediately delete my github account if this happened. I don't buy that MS has seen the light and is a different company now. In the last 2 weeks, 3 people have told me their Windows 10 install updated itself and was completely borked. At least one of these people are going out and buying a new Windows machine as a result (thus giving MS a little more revenue). It absolutely disgusts me.
This update has broken computers until the SSDs in question were blacklisted. For people who auto-downloaded the update after that, Windows "only" showed annoying and absolutely unhelpful error messages every single day since the 1803 release.
Many people mention Microsoft's past history as reasons for worrying about this. However, I disagree. I don't think this would be bad because it's Microsoft.
I think it would be bad because GitHub wouldn't be a generally independent entity, with no business interests besides making a product good enough that people will pay for it anymore. Currently, GitHub doesn't really have any incentive to care what tools and platform you are using, or where you deploy your code. To the contrary, they want to appeal to the widest audience possible.
This would all change if owned by Microsoft. GitHubs role would then instead likely instead be to channel users into other Microsoft products and services, especially Azure. Instead of functionality that benefits everyone independent of platform, it will be in their interest to add features that only benefit users of the Microsoft ecosystem. Azure CI, deploy to Azure, .NET integration, LinkedIn integration. I don't think this is the kind of platform most FOSS projects would want to use, certainly not me.
This is what I was thinking also (though you articulated it much better than I would have).
A large company doesn't make an acquisition like this without a business incentive. For them, it's driving people to their products. Maybe making the tooling in github for interacting with microsoft-y products a little bit better. Maybe adding whatever licenses microsoft likes to the front of the default licenses you can pick from when you create a new repo. Things generated by their editors going into your .gitignore by default.
I don't see any reason they'd acquire github and then entirely leave it alone.
Microsoft has embraced Github with it's recent .Net core development, they are one of the largest contributors with some incredible cross platform dev tooling. C# is a pleasure and visual studio code is an excellent, extensible coding tool.
Go forth Microsoft, the standard bashing in these communities is unjust.
I lived through the nineties as well, that doesn't mean I think everything that was amazing back then still is amazing today nor that everything that sucked back then still sucks today...
I think the Microsoft of today is very different from the Microsoft of the nineties.
Personally I'm not so much worried about Microsoft going back to Microsoft'95 but more about the fact that they, like most (all?) other big tech companies, collect more and more data and integrate more and more tightly into our lives. Still I feel that currently Facebook, Google and Amazon are all scarier in that respect.
> I think the Microsoft of today is very different from the Microsoft of the nineties.
If they've managed to change that much, then nothing prevents them to make the opposite turn and return to the monopolistic Microsoft as soon as a new CEO steps in.
That's why I prefer Microsoft to stay away from the products and projects everybody loves. Furthermore, I don't think Microsoft has anything to contribute to GitHub.
> I don't think Microsoft has anything to contribute to GitHub.
Microsoft has experience handling online code repositories and team services. Azure integration has the potential to improve deployment pipelines for many users. Microsoft has lots of experience doing things at scale, and can potentially find cost reduction maneuvers for Github. Visual Studio integration can bring more users to Github. Cooperate users are more likely to upgrade their existing agreement with Microsoft than they are to enter a finical relationship with a new company. Microsoft has just as much or more to offer Github as any other top 5 tech company.
Azure, VS... Yes, I know Microsoft has tons of things to offer to their customers. But not everybody in Github is into Microsoft tech. In fact, most are not. So that's what worries me: trying to tie Github to their products and strategies.
It might happen, but I suspect they will continue to play nice with other tech stacks. Their recent strategy is to provide good tech that can integrate with other tech, to compete on superior products instead of locking in users. They haven't recently extended or extinguished, just embraced, and embracing has done great things for their stock price.
Yes they want... yet it has been Windows only for a year and a half, so far. That's the problem with Microsoft buying big things like Github. It turns people not using their flagship OS into second class citizens.
> then nothing prevents them to make the opposite turn and return to the monopolistic Microsoft
Well, the world has changed and there is no longer any opportunity for a 90s Microsoft. So I'm pretty sure that unless the world changes back to the 90s, Microsoft wouldn't survive such a turn.
I'd say that Microsoft is worse than most other companies because it caters to governments, not consumers. Skype and MSN mysteriously worked in China at a time when no other service did. What's their incentive to push back against LE backdoors when they're not an edgy consumer brand anyway?
Corporations only exist as legal constructs. I'm not trying to be pedantic. I'm just saying that it is always going to be the case that a business, any business, will work within the law and with lawful government because they ultimately depend on contract and property law.
All companies have to comply with the law, but Microsoft sells lots of software to governments. They aren't going to put up a fight over end-user privacy if they make more money from gov't contracts than from privacy-conscious consumers. The financial incentives are completely different than for e.g. Apple, and I don't see how that is ever going to change.
All companies, including Apple sell lots of products to the government because, next to Walmart, its the largest employer in the country. It has a $4T budget.
Apple doesn't give a s* about your privacy. They claim to give a s* about your privacy. Those are different things.
If you want this to change you have to get involved with legislation to change the rules/law. That's the only option. None of them "care" about you or really anything, because they aren't human. Companies respond to changes in the law which bound the approved set of behaviors. That's it.
And the more Apple doubles down on protecting its users, the bigger this financial incentive becomes. Of course that doesn't stop me from supporting better legislation (GDPR was a great start).
I reviewed the data that Microsoft has on me and there is nothing in there that makes me feel uncomfortable. Facebook on the other hand.. not only is the data wrong but if it gets public I can see it damaging me
Microsoft has been changing a lot. Embracing linux in their cloud offering, offering unix as a subsystem on Windows, open sourcing their work. Anyone who likens the current MSFT to anything resembling what it was in the 90s hasn't been paying attention to the current leadership.
Oh Hey! We see you're browsing AWS issues, GitHub recommends using Azure (GCP?) for the best developer experience, click [here] to begin your free trial!
Automatic updates are great, I use it on most of my dev servers
I could even see that defaulting auto update to "on" is a good move too. If you know what you're doing you can turn them off.
I haven't adminned a windows machine since '98 -- can you really. Or turn off automatic windows updates?
As for any telemetry stuff, that should be opt in, and have a target server to send to, which you can point to your own box if you want to.
Microsoft today is different to Microsoft in the 90s though. In the 90s it was a software company that used embrace extend extinguish to maintain and extend its monopolies, but it wrote software and the software was the product.
Now it seems it uses the "user is the product" ala Google, Facebook etc. I'm not aware of apple doing that. Still I guess inevitable as companies penny pinch more and more.
Microsoft is a big company, that's a different team with a different set of leadership. The guy in charge of Windows stepped away. Windows 10 is the last version of Windows basically. Not sure if that means someday they may release a distro (crazy I know) but it's interesting nonetheless.
I know about that one, but I mean one that'd be consumer driven in some way. I know the ship has sailed but I'd love to see Microsoft create an Android / iOS competitor, and maybe something open source would give them an edge in that regard. It would be best if it were somehow privacy / security focused. Microsoft has the resources to make amazing products the issue is and I've mentioned it before, the company is ever so big, with ever SO MANY PRODUCTS and projects. We know they've got quite a bit going on on the user facing front, but I can't imagine how many internal tools they've got that they rely on as well.
Forced updates can be bad, for example I narrowly missed having a bricked workstation when the April update was released and it borked machines using Intel 600p SSDs. I hadn't used that machine for a few weeks and it had been powered off; by the time I finally started using it again Microsoft had blocked the update for machines with that hardware.
Anyway, I can manage the limited telemetry via hosts and I can defer updates (which I will do from now on), but my biggest gripe is forced installation/reinstallation of previously removed apps and games. I run Windows 10 Pro and I end up with a few variations of Candy Crush games, Minecraft, Autodesk drawing apps, and others after initial OS installation and after every major update so far. This is while using an offline account, and even if I used my Microsoft account on this workstation I've never purchased/downloaded any of those apps in the past.
I get that "Pro" at this point means home power user, and companies should be using Enterprise, but Microsoft shouldn't be pushing any software back onto the machine once the user has uninstalled it. That's user-hostile and takes control out of the hands of the one person who should be in complete control of the system.
Once full, non-buggy support for Ryzen APUs arrives in Linux, I'm switching Windows 10 to a VM and going back to Linux on this machine as the daily driver OS.
2. Windows will restart the computer to update, potentially interrupting processes.
3. Users should have freedom to use software in the way that they choose, and not have actions forced on them by the software.
I understand that many windows users are far from power users and the forced updates provide security updates. My point of view is anyone willing to give up liberty for security will receive neither.
3. I agree. I don't give a shit about what users do to their machine, but when they launch a DDos to my servers, I blame them for not updateing their OS. In Windows XP history, there were many viruses/trojanes/whatever that were only possible because people didn't update.
I wouldn't mind forced updates if they were only used for critical security updates. They are not; Microsoft abuses the privilege, and therefore may not be trusted to have it at all.
> I understand that many windows users are far from power users and the forced updates provide security updates
I think this largely was only true of the period between the internet becoming a thing and the smartphone becoming a thing. Nowadays there's phones, tablets, and chromebooks that serve the needs of those people better and in a more convenient form factor. Desktops are for people who do work, create stuff, or are hobbiests.
I think a lot of the reasons computers suck today is because of this idea that they need to cater to some strawman drooling moron of a user.
Those devices only take you so far. They are consuming devices only, similar to a television. People still have and use normal PCs and will do so in the future. (Notebooks included of course)
That's kinda my point though. The strawman user these kinds of arguments prop up as the reason everything needs to suck is pretty much 100% consumer. There might be a few things, like typing documents and printing that aren't entirely covered by smartphones, tablets, and chromebooks, but that's a rapidly shrinking set of things.
Doubt that. "Desktop" is defined by mouse, keyboard and display. A chromebook satisfies this. I have yet to meet a user who uses a smartphone instead of a PC to do stuff that are notably easier to do on a PC. (while having both)
I would also bet that every user would chose a keyboard over a touch interface to type text.
Maybe you have a point, but I don't feel like connecting a bluetooth keyboard and mouse to a smartphone makes it a desktop. Regardless, I don't think those OSs are good desktop OSs either. I still maintain that this is a poor excuse for having shitty tools.
No, forced updates are also bad. I own my computer, I built it, and I paid for Windows. Microsoft shouldn't get to decide how I run it (I can't turn off Windows Defender permanently) or when it updates (you can only delay them so long before it forces you to update). If their updates were stable it would maybe help the argument, but they routinely break my system such that I lose time that was intended to be spent on playing a game. They also conveniently reset some of my settings which always tends to favor the way Microsoft would like for me to use my computer. Or they reinstall fucking Candy Crush.
For god's sake, even Apple, the prime example of knows-better-than-you lets you never update, if that's what you really want.
I saw one happening on a PC in a conference room two weeks ago at the very moment a talk should begin. Nobody could do anything with the pc until the update was over. Great user experience.
I recently had to do the forced April Windows 10 update, and not only did it take 10 hours, the UX was horrific... e.g. Random CMD prompt windows popping up etc.
If that is allowed to pass QA in Microsoft, then it kind of scares me what can happen if they were to acquire GitHub.
I'm not enthused about the idea of Github being bought. But on what grounds do you suspect that the Microsoft of the 90s is just lurking, like a conscious spirit, to reveal itself after a decade?
CEO Nadella started working at Microsoft in 1992, but the executives who coined and lived the EEE phrase are all out. Is Nadella a sleeper agent? If not, what entity at Microsoft will subvert his leadership?
> what entity at Microsoft will subvert his leadership?
I agree, it seems that both Gates and Ballmer have found new callings in life, which has left Nadella with pretty much control of the company. I remember almost feeling sorry for the guy when he took over, what freedom could he possibly have between those two? A lot, it seems in retrospect.
I've mentioned this a number of times, some of Microsoft's open source tooling is under the very permissive MIT license. In the case of GitHub it just means competitors will rise up and replace GitHub. Remember when SourceForge was the GitHub of the day, and then it became bad, and finally it's acceptable with the new owner. People will outright ditch GitHub if evil is truly done to it, but Microsoft doesn't want that, they'll do everything they can to not screw up GitHub.
Whoa, hang on a second. SourceForge? Acceptable with new owners? Have you tried to download anything from the site recently? They may not be bundling crapware with downloads anymore but it's still an ad-laden slow-loading pile of trash unusable website. (I don't think I'm being hyperbolic there, to be honest!)
I lived through the 90s too. It's been 20+ years since the peak of the Microsoft that gave them the reputation you're reflecting.
If "I lived through the 90s" was how I measured technology and companies, I'd sure as hell not be using Apple products - their technology was pretty terrible then, and the company was incoherent. Would any rational person say that the terrible tech and incoherence then translates to how one would accept what they produce in 2018? Some companies change, some don't (see: oracle). It seems a healthier approach is to measure the behavior and state of a company in the present.
Re: macs. Yes. MacOS releases up to 1999 lacked features common in other systems, like full preemptive multitasking and memory protection. Windows NT had that in the mid 90s, and various Unix flavors had it for decades. I liked the UI on the macs vs windows95/98/NT or Linux/Solaris, but the stability of macs was a joke. Everyone was pretty happy when they cleaned up NeXTSTEP and turned it into the basis for OSX - that was a very smart move.
> Microsoft has embraced Github with it's recent .Net core development, they are one of the largest contributors with some incredible cross platform dev tooling.
Not going to deny this, but that's really the reason I'm not comfortable with the idea of a MS owned Github. Microsoft doing open source development is great. Microsoft trying to influence open source and FOSS development as a whole, not so much. Them aquiring a central hub of such development efforts doesn't signal good things to come: The users of Github have nothing to gain from this, whereas MS has, and there's a non-zero chance of them ruining the thing for everyone in the process.
Microsoft has been relying on Github to share all of its UWP samples for the last few years. This year, they added all their developer docs to Github so they can be directly addressed and updated by the community. They've done a good job of responding to PRs and issues. I write Hololens software and I could not have achieved such rapid development without their examples and developer outreach.
However, Github is one of my favorite services, professionally and for hobbies, and I really don't want significant changes. I also use VSTS (Microsoft's code+PM online service) and it is a beast. They don't need to be heavily intermingled.
They are however embracing open source in a manner that feels very sincere to me.
For the record, the so-called "spyware" is telemetry which exists only in the SDK for .NET Core and not the actual runtime that you would ship to the user with the software. It is anonymous data, completely open source so it really is only what they say it is, and the first time you use it it tells you how to disable it.
I wonder what Linus Tovalds is thinking right now. If I remember correctly, he wrote Git because Perforce was forcing him to use Microsoft software licences.
Git was developed as an OSS clone of BitKeeper (which was originally provided freely to Linux kernel developers under the agreement that they not develop a clone).
Git was not developed as a clone of BitKeeper, it has a very different design (snapshot-based content-addressed filesystem, versus the SCCS "weave"). If anything, git was a (much faster) clone of Monotone, minus the heavy cryptography focus which made Monotone so slow.
UWP is native, you can code rich GUI in C++ without any GC involved.
Currently it only works for MS platforms (Windows incl. ARM, xbox, embedded) but they might make it cross platform in some future. Technically the API is not that complex, i.e. unlike WPF it’s not that hard to port from the current Direct3D to Vulkan or GLES or Metal.
Probably news to the Win32 devs of the world that their technology is "not the same as real native like C++." What distinction do you mean to make here?
Yeah, let's evaluate corporations for decisions made decades ago by people no longer at the company. At least spend 15 minutes reading about the progress MS has made in recent years before throwing around shade.
Well, you know, this wasn't just an "insane" guy at the top who now left the company and everything's cool. Companies have culture, and while we can agree today Microsoft's not the same Microsoft from 10 years ago, we shouldn't forget their history so soon.
> this wasn't just an "insane" guy at the top who now left the company and everything's cool
Satya Nadella has completely re-organized MS inside-and-out since taking the helm in 2014. It is a VERY different company than it was even 5 years ago. In March [1], Microsoft even went so far as to officially disband the Windows department and merge it into something called "Experiences & Devices".
It's different, but not very different. They still have a lot of lock-in like Outlook (ActiveSync), DirectX, exFAT and so on and so forth. They show no sign of dropping lock-in, except in markets where they are way weaker than competition.
How is DirectX a lock-in? Vulcan/OpenGL are available. Using Outlook/ActiveSync is no more a lock-in than depending on GSuite (or ANY business productivity suite), and I fail to remotely even understand how exFAT is a problem. Weak arguments.
It's MS only. It's more obvious in cases like Xbox, where MS don't support anything except DirectX. I.e. developers are forced to use it if they want to release for Xbox.
exFAT is an issue since it's a patented filesystem that MS pushed as a standard for SD cards. You can't freely use it (unless you just ignore those patents on your own risk).
ActiveSync doesn't have commonly available FOSS clients, so you can't use it on Linux easily. It's highly irritating (since it's widely entrenched in corporate environments). And etc. and etc.
The progress they've made? Baking in more spyware into Windows and other products than ever before? Taking more control away from computer users than ever before?
You can proliferate the doublespeak if you want but I'll call it what it is.
Like rebooting computers whenever they feel like it. Reenabling previously disabled options etc. Your head is in the sand but you don't have to live this way.
The one thing that I like about Github is that everything is public by default and you have to pay to keep your code private. I feel like a positive consequence of that business choice was the opening up of way more software. Gitlab, while being very awesome doesn't have that same rule and neither does Bitbucket. It's my assumption that this is the biggest reason there's so much open source code flowing right now.
It makes me wonder whether I'll be able to find as many random interesting projects just by searching for things anymore.
Though there is indeed much more software open source lately, I feel that is mostly a consequence of the free and easy availability of open source hosting.
I still remember the old days where most opensource projects were doing self-hosting or using sourceforge. Self-hosting is something that most developers that just want to build something nice won't consider. And using sourceforge back then was also atrocious (they didn't exactly have an easy to use UI, if I remember correctly).
This makes me think that what contributed the most to there being so much open source, is the easyness of creating your own repository on GitHub. With just a few clicks you have one.
Seems like the view here is pretty negative. I feel like most people here haven't worked with Microsoft dev products recently. They're getting better. I guess the crowd here isn't typically in the .NET world.
The fact it's getting better is obviously a positive thing, but will it stay that way in the future? no one knows.
I like the fact Github is independent of all the big players. I really, really, really hope it stays a neutral ground.
Also, if acquired by MS, I'm not that afraid for github itself. However, what is at risk in my opinion is all the integrations around github. There is a rich and striving ecosystem surrounding github. Tools like appveyor, travis-ci, coverall, readthedocs, codacy are really useful, ease a lot development and are infinitely easier to setup/maintain than a Jenkins box plus a few slaves. What would happen to this ecosystem remains to be seen.
That's the point. No one acquires GitHub so no one can eventually corrupt it with partisan "my stuff is better than your stuff" idiocy.
GitHub works because it effectively belongs to "no one" as far as the big players are concerned. Sure they might themselves go bad one day but there is no point in hasting this for some two cents Azure/AWS/GCP mode.
If the acquisition goes through, VS Code is probably going to kill Atom, which is my go-to editor for many tasks. VS Code may be better overall, but it's not for me, and I'll have to change my workflow.
Particularly with how they interact on GitHub now, it's quite impressive. I've reported issues for docs.microsoft.com, participated in discussions on Visual Basic development, etc. and while there's still definitely that corporate veil to some degree, they're extremely responsive, they're very active in responding to requests and discussing fixes.
I generally now hope that whatever Microsoft thing I'm dealing with somewhere ends up linking to GitHub, because everyone on the side of the company that's working with GitHub is pretty on point.
Which is what gives me some hope that they could manage this acquisition right: They're already doing GitHub the way GitHub should be done.
Even when it leads from GH to internal email chains, they're really good about keeping those on the GH issue in the loop... often from only parallel issues.
Issue filed in node sql driver... azure tech notices the issue, raises to sql team... sql team dev starts replying and working to resolve... recommends patch to upstream project in github... issue resolved.
Oracle would NEVER do something like that. It was actually very impressive, and that's just my one anecdote, and I've seen many others.
Well it's not like the reason for being negative is unfounded. MS has a long history of running services into the ground, and their recent 'we love open source' charade brings back strong memories of E3 for many folks.
I have never seen any good .NET code in the wild. A typical story IMO is to come into a workplace where they are trying to dump legacy .NET systems as soon as possible because they are incredibly brittle. .NET seems to be one of the biggest software scams of all time. Great vendor lock-in though. Visual Studio, MS-SQL, LINQ, Entity framework etc... the only solution is to rewrite everything.
That's very subjective. C# has tons of useless keywords that you just know some arrogant dev is going to use because they can, which then no junior can work with. It also has a worse ecosystem for writing server software (while obviously a better ecosystem for writing Windows applications) and is less suitable for cross-platform work.
It is subjective, but things like genetics and reflection I think are better in C#. I am not sure which keywords you think are "useless." Either way though, they're quite similar. There is no innate quality about C# where the code has to be terrible and the only hope is a total rewrite in something else.
I have hands-on experience with Unity3D and .NET Core 2. Both are cross-platform and worked relatively well. I’ve built MacOS, iOS, Android software with Unity, and Linux software with .NET Core.
I’ve heard Xamarin is also good for mobiles but haven’t personally used it.
Can you elaborate on this "keyword" statement? I find the statement very confusing especially when comparing it to C, modern C++, modern Java or any functional "magic" languages.
The second part is true. .NET does not share the same wide ecosystem as Java.
I've done legacy work in both and I'd never choose Ruby over C#. It might be a little faster to write the code the first time around, but maintaining a big legacy project stinks.
Also, none of Atwood's reasons is "I don't like working in C#." And several of the concerns in this post are obsolete.
Well who cares about that, really. I'm not talking about how pretty the code is; I'm talking about the ease of maintaining it. Ruby code frequently requires digging through the method and methods it calls to figure out the expected input.
I think that's the point. I'm personally worried Microsoft may have a very narrow view of what GitHub is. A lot of tech hosted there doesn't even run on Windows, or isn't even related to software. (legal, political, docs, guides, etc.)
what makes you think that? they killed off their own code hosting site in codeplex and moved everything to github. they are one of the largest if not the largest open source contributors in the world. and .net core is setting itself up to be a major player in cross-platform application development. it has a chance, if microsoft brings over uwp, wpf, and windows forms to .net core, to take over cross-platform gui development. with the windows subsystem for linux, windows 10 can run most linux-based software. they've developed the hottest ide at the moment on top of github developed technology.
You mention a lot of stuff that's in the same '.NET world' skellera mentioned, but the currently dominant GUI platform is the web, and neither is going to do take over. There's also still embedded software, hardware description, databases, emulators/simulators, competing virtual machines and GUI frameworks, a whole bunch of non-tech stuff, etc.
They do get a lot of this stuff, but I feel like whatever they want to do with Github is more geared towards average Joe .NET developer. Possibly making the platform less attractive for everything else through neglect.
The new Microsoft has made it clear that they want to support as many areas as they can. Azure does not have any lock-in of a specific platform (other than Azure itself in some ways) and Microsoft is constantly trying to get their tools to work in Linux and Mac. While they obviously would prefer you to be on .Net, they realize that lock-in ultimately hurts them.
Just give it a little more time. They are turning around years of company culture and they're doing a great job at it. If VS Code is a sign of future products, I'm excited to see what they come out with down the road.
not really. half the stuff i mentioned is all about the web. that’s the core market for .net core and visual studio code. visual studio code is certainly not .net centric at all. the in the box language support primarily targets web languages.
and i don’t get why you are mentioning those other things. what do they have to do with anything regarding microsoft understanding github? microsoft is no doubt and expert in all of those fields anyway, at least internally.
and what cross-platform gui framework would compete with a .net core cross-platform implementation of uwp, wpf, and windows forms? nothing, in my opinion. people would flood the .net ecosystem to get away from things like c++ and qt and the horror of writing script-based GUIs in python.
Since the Core era, not all .NET applications are built for Windows either.
Considering you bring up the second part, it makes me wonder if maybe outside of the obvious (owning GitHub), perhaps they're after the extra talent to bolster their SharePoint product.
Azure team gives a shit about "not invented here". They are exclusively interested in your cpu cycles and bytes in memory. Who organizes the cycles and bytes, they do not care.
Surely, there are preferred stacks (e.g. in the high level PaaS areas), but this includes node, python and .NET.
PS: one save bet: Java will never be on their support list :). But only for legal reasons :)
Depends on who you talk to... I'm pretty sure it's MS developers that drove the integration of git repositories into TFS (many were already using git locally for local branches and commits with TFS repos). The move to GitHub for all MS OSS was probably developer driven too.
The Office team is probably more than happy to not be hamstrung to windows only. Particularly Office 365. Azure has been using a lot of non-ms/windows tech from the start.
I'm sure there's some in the recently re-org that used to be Windows still have issues, but the old guard is largely gone now.
I constantly remind myself that HN does not actually represent what most customers want and/or need in the real world.
Most routine jobs/contracts at the coalface are still Java/.NET (possibly with a bit of Kotlin thrown in) - I even see companies looking for PHP developers on a weekly basis.
They are definitely alive and kicking, though if you judged based on this website alone, they all ceased to exist 10 years, replaced by Golang, Elixir and Clojure.
Mild exaggeration (and it's not a criticism of the people here), but you get my drift. People here like to talk about new/shiny/cutting-edge - and that's fine, but it's also not necessarily what customers/clients want or need.
C#/.NET in particular, though not perfect, is one of the most pleasurable languages/frameworks to work in.
Many visitors to HN are concerned with selling their skills to various customers, whereas others are concerned with the best technologies for building a successful startup. If you're selling your skills to a legacy institution and want to switch to selling them to a startup, looking at tech discussed on HN could be useful.
The fact that the Microsoft of the past 2 years is "getting better" at something they've tried to do for many decades is a great reason to be worried about them taking ownership of a product they are unfamiliar with.
e.g. "GitHub is really starting to improve again" - Devs in 2050
Why do you think "they are unfamiliar with it"? On the contrary, Microsoft has a lot of stuff ongoing in GitHub. Seems many critical people in this thread don't have a clue.
It's actually really interesting. Seems like internally, MSFT has kind of been let loose and no longer has to adhere to the more strict style Ballmer was a fan of. We've seen this with them hiring OSS talent (Miguel de Icaza, who was a pretty big advocate of trying to make MSFT less proprietary and locked down).
Externally though, the image looks completely different to those who don't really care for what's happening at MSFT - they just seem like an outdated Apple. That's slowly starting to change, I think the Verge recently put out a piece claiming they've switched positions. However, a lot of major consumer problems (Windows Update, Skype) have plagued the company tremendously. No amount of loving Typescript or VSC is going to change that.
1. LinkedIn's dark UI patterns are as bad as they've ever been. Bringing the same to GitHub or otherwise LinkedIn-ifying it seems like a very real risk.
2. Ramping up telemetry. Key loggers in Windows 10, opt-out telemetry in .NET Core, seems impossible to get VSCode to stop phoning home.
That one was really sad. Sunrise was an excellent product. I'm actually tempted to recreate it myself. I'm sure it would be possible to commercialise something like that.
They made a lot of money, but were not successful. Business software companies become irrelevant decades before their revenues go away (eg IBM). Microsoft lost its future relevance under Ballmer.
Ballmer may have hurt microsoft, but they are still alive and well. The truth of it is that for a lot of things they do, there just isn't a competitor, nor will one arise soon. No one is replacing excel. As for windows -- there are a couple of niches where mac or linux could step in, but for most of them, that just isn't an option. Azure, typescript, vscode, ms office, office 365 are massively successful.
It's a big company and hard to blame an individual for mistakes. But they totally missed out on being the mobile OS of the future under him. And they ended up getting a late, slow start to cloud computing (minus SAAS).
I used to like Skype, but they've made a bunch of terrible choices.
Their integration of Microsoft Messenger and Microsoft IDs into Skype has meant that I now have two or three contacts for some people, and I have to switch which contact I send messages to depending on what device they are currently using.
They've also broken other basic functionality like sending messages - even with a solid internet connection on both sides, it will often refuse to send messages leaving them in a 'Sending' state.
Other times i'll be in the middle of a conversation and I won't get a reply to certain messages that appear to have sent correctly. Hours or days later I'll get a confused "What?" message from that person, to which I've got no idea what they're talking about, and then they send me a screenshot showing a half dozen messages suddenly showing up from previous conversations.
Then there's situations where I'll get messages on one device, but not on another (eg phone). So I've got someone pissed at me because I'm not responding to them, when the notifications are all piling up on my phone or another laptop, or whatever.
All in all, Skype has become so unreliable that I will only use it to make calls, because at least that way I can be sure the messages are getting through.
Really? It's become nearly unusable for me. I used to have it open all the time and talk to lots of people. Now it feels closer to an infection on my computer that I only pull up when I have to talk to someone on it and then promptly struggle to logout and make sure it's truly closed.
I haven't bashed Microsoft for a long time, and I won't any time soon. However, this development has me worried.
There's a big difference between engaging the developer community with great products and purchasing important services that the developer community relies on.
Microsoft has succeeded many times with products and services they've built themselves, and with products they've purchased, but how often have they succeeded with services they've purchased?
Most of those are products. Of the services you mentioned, I would only call Hotmail a success in the sense that there wasn’t a net detriment to users because of the purchase.
Maybe the founders are fed up with running it. I can understand you want out, and it's very hard to find people to replace you, and much easier to sell it to a big entity that is already used to do something like that.
Maybe the founder are fed up with running it. I can understand your want out, and it's very hard to find people to replace you, and much easier to sell it to a big entity that is already used to do something like that.
I can't help but be insanely sceptical. Microsoft has a massive legacy of Embrace, Extend, Extinguish and I don't want the GitHub ecosystem to suffer from the same reoccurring pattern. Call me jaded by Microsoft's past behaviour, but I just cannot forget the numerous times Microsoft has absolutely ruined thriving ecosystems by using the same tricks.
Heck Microsoft killed their own competitor TFS and pushed people to migrate to Git, and uses GitHub themselves for hosting several products. Microsoft also has full Git support in Visual Studio Online (their enterprise competitor to GitHub), Visual Studio, and VS Code.
Microsoft would likely use GitHub to advertise their development products, tools, and services like Azure. I doubt they'd want to "extinguish" GitHub because it just means less eyes on their adverts.
They extinguished Skype. They will manage to do the same with Github eventually. They are the biggest user already and would therefore be forced to cater to their own needs first over those of the rest if the community. This would drive the service squarely into the ground by doing everything right.
This is negative value. The only reason the big companies host their source code on GitHub is because it's independent. They will stop doing that tomorrow if GitHub is acquired and it will lose all its value.
As someone who has always despised git for its complexity and being a movement more than a new technology I have been waiting years for this. Let git fracture and die. Maybe then we can get tools that work for us instead of the other way around.
I enjoy Mercurial as much as the next person. But git is an engineering feat simply because it works well enough for so many use cases. Also, if git is the worst part of your tooling, that's some great tooling and a good problem to have.
It's just software. My experience with git was about a month long phase where I devoted a few hours a week to learning it. Now I can use it but I still have to google commands for anything other than the basics. That's the problem with git. Git without google would be unusable. That's why I consider it to be extraordinarily poorly designed. You've seen git commands, I'm not going to go look up some esoteric command just to prove a point to you. The point is a good program would not force you to google everything, it would just work.
I consider git to be more of an advertising feat. It fascinates me how many people drank the kool-aid.
For the comparisons of this acquisition to Skype, yes Microsoft bungled the Skype acquisition.
But they have been killing it with Typescript and Visual Studio. I feel like that at such a large company, the Typescript recent history matters more than the Skype history. Two very different areas with different people in charge, etc.
As much as I love VSCode, you live in a buble. A communication app affects way more people than a programming language or a text editor. Even my mother uses the former, while only a fraction of all the programmers I met in my life use the laters.
And MS did messed up on other major products or companies
- Rare video games. They did freaking golden eyes and conkers bad day and they manage to kill it.
- they messed up Wn10, which is still not as good as Win7 after all those years but spies more on you and forces update at the worse moment on you
- the latest outlook is barely usable. Sometime mails can't be sent. Sometime it freezes or crashes. It's very slow, the search sucks and I never find the buttons I'm looking for.
- they killed Nokia and the windows phone despite all the investments the community did in it. And it was considered a good product.
- they are milking minecraft like disney is milking SW, so that all the creative community has moved to private servers.
I mean, we are talking about a company that had a monopoly on the media player and the web browser on the most used OS in the world. The player and browser, the most used software ever by the non tech saavy users.
And to this day, none of their versions is remotely as good as the competition.
Many of your criticisms may be justified, but this one reads a bit silly to me:
> monopoly on the media player and the web browser on the most used OS in the world
I know it's a trope to say that IE was a monopoly, but the longest view of these stats I can easily find comes from W3Counter[1] and at what appears to be its market peak in 2007, IE6 and IE7 had 50% and 17% of the market.
OK, so the combined share was downtrending since the actual peak in 2004, according to ADTECH data from Europe[2]...
Huh. There are stats that go way further back, to before IE was dominant, and Wikipedia says it actually peaked in 2002, with 96% of the market.[3] I stand corrected?
I started my reply to make a point, and upon doing the basic research found that it appears I don't know what I'm talking about. These stats are really interesting, though!
FWIW, I have not been primarily a Windows user since I was about 14 years old. That was in 1999, when Netscape was still a player in the market, and I began my switch to Linux. I guess I forgot (or blacked out, it was traumatic) the fact that there was a point in time in the Browser Wars when IE was effectively declared the winner. It's been a steady downhill trend since then!
Aaand this is the weakness of centralizing everything in Github. Inevitable that it gets acquired by something with other motives (or is forced to exploit its central position in some way because business). The fact it may be Microsoft that seals the deal is just a cherry on top.
We need proper decentralization.
There could still be aggregators for browsing projects and bookmarking but the structural weakness of keeping all the eggs in one basket will never go away. No matter how much wishful thinking you throw at it.
> There could still be aggregators for browsing projects and bookmarking but the structural weakness of keeping all the eggs in one basket will never go away. No matter how much wishful thinking you throw at it.
The Web is decentralized, and so everybody’s using Google to discover it. The problem with decentralization is you still need global entry points, i.e. some sort of centralization.
As a long time GitHub user I really do hope these talks do not result in an acquisition. I would really like GitHub to be independent from OS vendors, or vendors of platforms, ecosystems, compilers, IDEs etc.
I want GitHub's goal to be a good GitHub. Not a good part of the Microsoft ecosystem. Those may initially be the same, but it takes very little for them to not be the same.
At first they will just inject random commits that don't change any files but inform you about the latest price offerings on One Drive and Azure hosting. Later on if you're still on the "free" plan some of the commits will modify your hosted software to show their ads at runtime. Have fun!
> will do away with GitHub's current, open nature.
Apparently the Gnome folks concluded that GitHub was already too closed to even consider them in the competition for the question "Where will the Gnome project live in the future?"
GitLab won that competition, btw, and icrosoft wasn't even part of that calculus.
With current trends in Microsoft's Developer Division, it's possible that GitHub could end up way more open if Microsoft acquired them. GitHub's core business is GitHub and they need that to stay closed to stay competitive. GitHub under Microsoft could have less financial pressure to keep closed.
Is it likely? I don't know. It is an interesting possibility though, that would certainly change the game.
There is precedent. Prior to Microsoft acquiring them, Xamarin was closed-source and licensing was $1000 per platform per year. At the Build conference in March 2016 Microsoft announced (1) that they were acquiring Xamarin, (2) that pricing would be dropping to zero, and (3) the Xamarin platform code would be open sourced as soon as possible after the acquisition.
This seems, to me, to illustrate perfectly how Microsoft still doesn't understand open source as an idea nor as a culture. The whole movement exists as a reaction against central, corporate control.
Large companies think that, by acquiring important hosted open source infrastructure, they'll somehow acquire the users or act as gatekeepers. But open source tooling is so decentralized these days that is not easy to be a captive user. The cost of switching is so low -- plenty of other comparable services supply collaborative utilities for git -- that, for most projects, a weekend spent altering URLs and configs should suffice.
If such a deal goes through, Microsoft will spend millions on the acquisition and, simultaneously, alternative providers who appear more independent (Gitlab, BitBucket, etc) will see a surge in new accounts.
And I, for one, think it'd be a good thing for open source. Every time an OSS sugar daddy turns on its users -- SourceForge, BitKeeper, Oracle, SCO, etc -- the community has grown more robust, more decentralized, and more resilient.
Does anyone have insight into what they're been working on recently at Github? I feel like the user experience hasn't changed/improved in the last couple of years at all. It feels to me like there is a lot of low hanging fruit to make a ton of money. For one they could integrate or replicate the functionality of Bountysource and let people donate to users. There is also tons of room for improvement in licensing like making copywriter-assignment frictionless and letting people dual license their software (so I could at a click of a button buy a proprietary license for someone's GPL'd library)
People are being overly pessimistic. This would be a potentially great deal. GitHub is very successful at what they do and I don't think MS would change that. More likely they would use GH as lead-in to selling cloud services the same way Google uses GitLab.
Some of my pessimism is not just with Microsoft, itself, even though I do have that. I think an argument can be made that we should not want all the decent software companies to be bought up by the same big five giant companies, no matter who it is.
If you've watched the last couple Build Conferences (Microsoft's annual developer conference) they frequently demonstrate their CI/CD pipeline based on Github checkins, not VSTS checkins (just takes a webhook invocation and all the source-hosting sites support sending those out on check-in or merge).
What does GitHub's choice of cloud have to do with Microsoft's aptitude to take over an existing, huge, critical piece of the open source world? From a business perspective, or an operational one?
I have the opposite opinion. I think every few months I see an interesting feature being added that are slowly making github a better platform. Just for the last few months we've seen a lot of improvements for Enterprise, improvement to gitHub pages, saved replies, team discussions, security alerts. I love the pace they have to release meaningful improvements.
lol, it can be slow on larger projects but its innovative and complete. i really like the packages system/community. always open to suggestions, whats better?
AFAIK the timeline of profitability is somewhat like this:
- really profitable after introducing private repos
- burning quite some money for growth after they took VC money
- really profitable again after they changed their pricing to per-seat pricing
Most companies leave startup status long before they are profitable. It has much more to do with the stability and maturity of the business model and whether it can be scaled.
I don’t know why you’d willingly be against competition though. GitLab makes GitHub better, but GitHub makes GitLab better. Why isn’t that a good thing?
It's getting depressing to see that mid tier companies are all now reliably getting swallowed up by larger ones. It is starting to feel like the golden age of technology - where brand new small companies emerge and become giant ones - is waning. Were Google, Amazon and Facebook actually the last of the tech giants to emerge and from now on, it's just a procession of watching every other company merge into the existing 5 or 6 big companies? I guess it is inevitable but it still makes me sad if we are transitioning to a more mature, but fundamentally more boring phase.
I really hope this doesn't happen. Microsoft design and services are all over the board, not in a good way. I'm sure they would add more features, or make it more enterprise-y, which could have benefits, but I would definitely move to a self hosted setup if this went through. It's nice to have diversity of brand/company personalities in the tech ecosystem.
I run Windows. I have respect for Microsoft. A lot of people do not, however, and the sense that Microsoft is ritually unclean will drive many of them away from github. It's the best thing that's happened to BitBucket and Altassian ever. If I have a choice between MS tools and Altassian tools I will usually take MS. (Visual Source Safe really did suck, but Visual Studio is not that bad...)
Also, Microsoft acquisitions do not seem to go well. Look at Skype, LinkedIn, etc.
I'll be really honest. This would lead me to start looking at alternatives for all of my projects very quickly.
After years of watching Microsoft gobble up and assimilate companies, I'd rather jump ship before they start the extinguish portion of the cycle. Mojang is starting to leave the extend part with updates that are pushed out early for the "bedrock" version of minecraft.
I'd be interested to learn why there's the perspective that the LinkedIn acquisition hasn't gone well. I feel like most of the stories I've read about it have been pretty positive.
I am sure that some journalists interview PR people from Microsoft and some self-promoters who promote the idea of promoting themselves on LinkedIn and say it is great.
I think they've gone down the road of harvesting.
I don't know if anybody is left at LinkedIn rather than life coaches, scam artists, offshore development companies, self-promoters, "business development" people, salesmen who can't sell (maybe that is what "business development" means?), recruiters that can't recruit (ex. they have been contacting you for years because you once said something about Cold Fusion and they badly need somebody to move to Nome, Alaska but they never hire remotes...), people who seem to write 20 articles about data science a day that verge on word salad (you'd think if they knew that much they could write a better neural network to generate the text.) And people who are gullible enough to think that there is some value in sending somebody an InMail or using all of those search features they sell so you can do more refined searches of life coaches, self promoters, ...
Numerous design problems with how the forums work mean that good discussions get drowned out. The same fraudulent messages get spammed that were being spammed six years ago and they won't do, don't do, or can't do anything about it
LinkedIn has closed off most access to the API and you cannot get anything interesting out of it... except all of the email addresses of your contacts so if you do that "Lion" thing you will probably need to change your email address at some point. (I think many people don't know they get so much spam because they have a LinkedIn account.)
I am on the threshold of nuking my Facebook account now that I've disconnected all of the sites that I need to log in that use Facebook. I still haven't decided if I nuke my LinkedIn account or just delete 95% of my contacts.
TFVC is being outphased. Microsoft doesn't want to openly say it but everything in MS' docs is pushing Git. When you have an on-prem TFS installation, the default version control for new projects is Git. Visual Studio Team Services doesn't allow to import anything into TFVC, only into Git. Visual Studio lists Git first, when you want to add version control. Visual Studio Code makes working with Git a lot easier than working with TFVC.
TFS as the development process management solution will live on as an on-prem service, but TFVC is unofficially feature-complete and in the maintenance phase.
Oh yes... i especially love it when my co-worker locks 10 000 files and goes on 2 week vacation... can't merge for days
I also love it how it detects changes on unchanged files ( they look at last modified timestamp) and you can't undo the unchanged unless you install TFPT and google for a bit for the command.
I also love it how when doing merges sometimes you have 10 000 conflicts even tho the files are unchanged :)
TFVC is bad and it should die... I'm glad they took up on git... now all i have to do is train my team on how to use git :)
Intel has followed Apple into the rabbit hole of style over substance. I know so many people who want to do neural networks and they want to buy a mac and sorry you can't do that because Apple picked the wrong horse in the GPU race and it might be a few more years before they refresh their models.
Back in 1984 Apple released the first microcomputer that came without a software development kit. Since then it has all been about the computer using you than the other way around, a proud tradition of bay area companies such as Google, Facebook, Palantir, etc.
An acquisition would mean the inevitable split of support between the MS toolchain and the OS toolchain, which would eventually be left to die. Today's MS is no different than yesterday's MS.
I'm not sure that's true, but it is true that MS would have motivation to leverage Github to promote other (licensed) MS software.
At a conference years ago, a Microsoft rep told me over beers that almost everything his division did had the ultimate objective of selling more SQL Server licenses. That's why SQL Server is a required component of almost all MS enterprise software.
So you might for instance see a SQL Server backend option, and then "Github Enterprise" sold with that configuration.
MS just want you using Azure and don't care whose toolchain you use. If they bought GitHub it would be entirely so they could get people _not_ using other MS products onto Azure. So there is no way they kill off the OS toolchain and turn it in to another VSTS.
Give it time. MS does what it has always done to competitors (even to MS internal products) that it acquires. Takes awhile sometimes and you younger kid will get to live through what others already have.
It's history, not condescension. Weird to take it so personally. If you are older than me, you already know the pattern, meaning you're being disingenuous (or deliberately ignorant) to dismiss it.
I'm happy that Microsoft has gotten so on board with git recently and made contributions to the project, but this kind of consolidation still makes me nervous, for multiple reasons.
Unrelated to that incident, but I feel their web UI to be extremely slow lately. Working with PRs (or "merge requests" as GitLab calls them) and CI pipeline is a pain, you're looking at 1 second response times.
To me it doesn't look like it's a front-end issue. I am mainly complaining about the HTTP response time. It seems extremely slow on both pipelines and merge requests, and Vue ain't gonna fix that I'm afraid.
I don't know if it's an optimisation issue with the software itself, or if it's just GitLab.com being overloaded..
I can't understand why Github hasn't built out a clear path from repository hosting, to continuous integration, ending in low-friction renting of cloud instances.
I don't want Google or Amazon to buy them, and that almost leaves Azure.
It seems like something they could have done themselves a few years ago, though. Puzzling.
Considering that software powers most of the modern economy, it's pretty sad how small the market is for quality software development tools. Programmers get paid high six figures, but balk at paying for quality debugging tools. Visual Studio, despite being a strong debugger and exceptional piece of technology in its own right, is not a real revenue driver for MS, it's just an exceptionally sophisticated way to nudge programmers to Microsoft's libraries and platforms.
Github has generally been platform agnostic (except for git itself), and it would sadden me to see Github start to compromise those values and start favoring one platform or another. (One could easily see Github / Visual Studio plugins, enabled by default, with built-in subtle nudges towards the MS Cloud platform).
Possibly Amazon. They tend to leave acquired companies alone to manage how they prefer, apart from Bezos micromanaging how to cut costs and restructure logistics, which is often a good thing. Plus Amazon (like Microsoft or Google) also has economy of scale opportunities for hosting GitHub on AWS.
I do believe Amazon would interfere less with what users like about GitHub than Microsoft would.
Apple. I know it's not Apple's style but imagine if Apple started expanding outside of it's bubble like Microsoft has. If they really went all in it would be amazing.
I use mostly Apple products, for 25 years, but I would not want Apple to own GitHub.
Microsoft has always been a developer-tools company. They are many other things besides, and I have had some nasty experiences with their sales reps (who would say anything to close a deal), but I have always felt that at its core, Microsoft is a developer-driven organization.
Apple likes developers too. They taste like chicken. And a cloud-based developer institution would let Apple destroy two populations - cloud operations and developers - with one asteroid.
Yea but I don't think it's continued focus that will help here. Apple has continued it's modus operandi and quality has decreased like you mentioned. Granted, acquiring GitHub is most likely not the right fix, but I think something about Apple's particular way of operating is holding it back.
Not too keen on GitHub being acquired. If Microsoft wants to do something good here, they should give GitHub enough money to form a Foundation similar to Mozilla. That would be the best way GitHub could pursue its mission.
This is our chance to gain a GitHub+LinkedIn integration. Just imagine... your LinkedIn profile and endorsements will auto-populate based on commit & PR activity.
Don't really agree with most comments trashing this hypothetical deal, bringing up the Skype precedent.
Its been what, 2 years since the Linked-in and Xamarin acquisitions? Nothing seems to have changed. Microsoft just seems to know that the desktop is no longer the cash cow that it used to be, office still is but may not last long, so its a good idea to diversify its portfolio.
I foresee a great future... for Gogs [1], Gitea [2], Gitlab [3] and other Github-alternatives. I was a late and somewhat reluctant adopter of Github as I don't like the idea of centralising so much of the free software infrastructure and putting it in the hands of one company, even a supposedly 'friendly' company like Github. I mirror everything I do on Github to a local Gitea-instance as a backup and exit strategy for when Github becomes a liability. If these talks go through and Microsoft does take over the latter will become true and I'll remove my accounts there.
I wonder how projects like Electron are figuring into the interest that Microsoft has in GitHub. How would Microsoft move forward with such a project if they bought GitHub? Would its continue to be based on Chromium? Would future versions of Office follow VSCode and be build as Electron apps?
I was in DevDiv at the time of the launch of VS Code (but on a different team): The pressure was on to ship a cross-platform editor ASAP to support the new cross-platform .NET infrastructure. This was very much an "IBM PC"-project with the Directors tasking a small team to get something done and to disregard the NIH culture. Atom was chosen as the platform because it was already close to what they needed with the right license. Heck, the rush to ship was so intense that they screwed up the name (I forget the story exactly, but "VS Code" was not the name the MSFT naming department came up with, but the name they initially chose turned out to be used elsewhere and they didn't have the time to come up with something trademarkable so they just sighed and shipped as "Visual Studio Code").
Anyway - I think the VS Code experience helped with Microsoft's institutional "fear" of pushing ahead with a JavaScript ecosystem elsewhere - I think the knock-on effects contributed towards the Office org's recent announcement to add JS as a first-class macro language in Excel, I just wish it was TypeScript instead.
This has me wondering, what Microsoft wants GitHub for? It's not like they can't use it without owning it, and I doubt they want to improve it as it is. Probably they want control before integrating it with their products.
Microsoft has been investing in developer services for a number of years - they acquired hockeyapp (mac/ios crash analytics), and have a whole bunch of other tools under the azure brand https://azure.microsoft.com/en-us/services/
It's almost worth re-asking the question as...what does Azure want Github for, to try and strip away thinking about MS's windows/office products. There could be some crossover between github's Atom efforts and VS Code, and MS already worked on projects like Git virtual filesystem (GVFS) for their own needs ( https://github.com/Microsoft/gvfs )
GitHub do for developers what Microsoft want to do for the rest of business through SharePoint.
Maybe they could add Office 365 editing to GitHub and pitch it as the successor to SharePoint. Or try to integrate it with their other products in some kind of Frankenstein mess.
They seem to be doing much better with these things lately though. MS Teams is incredibly good. VSTS is pretty nice as well. I'm frankly unsure why they'd take on Github, other than better integration points into Azure services, CI/CD and other tooling for commercial/enterprise users.
They want to join the social networking data with linkedin and office 365, yammer, enterprise Skype, etc, etc.
They have a clear long-term strategy of owning professional social networks, both within and across companies.
That way they can sell you access to your professional network, and your employer access to info you don’t even realize exists about you, and also sell it insights into its own org structure.
They already have a bunch of tooling to make it easy for users on VSTS to ship their software to Azure, but VSTS users are probably 100% already in the Microsoft ecosystem.
Buying GitHub and then adding tooling to make shipping from GitHub to Azure easy (and the obvious choice) would help them push Azure more to companies not currently using other Microsoft products.
Because much of the free software world uses it as a coordination place - it's a weak, undefended target. If they were to buy it and quickly shut it down, the free software world would be in a tailspin for quite awhile. Microsoft could use the delay to push for legislation across the world to force developers to have a Microsoft(Linkedin/Ripple enabled) license attached to their (Microsoft) global ID. Within a year or two it could be illegal, worldwide to possess non-Microsoft licensed crypto software, then eventually all software. Especially look for software relating to the boot process/BIOS related stuff - that's another weak point right now, but there are probably others.
If SendGrid and Twilio could IPO, then surely GitHub can. Keep incentives aligned and do that instead, please. If Microsoft acquired GH it would a non-neutral platform, and I would be off to Gitlab immediately.
.NET is great and the openness is great, I don't really see why so many people are up in arms about this. Microsoft is a large user of github. Maybe this will push you all to use gitlab ;)
I don't understand why all the bashing, there is no better time to cash in. Gitlab in few years time will make Github worth less.
Gitlab.com and self host GitLab offer the best of both worlds. Gitlab right now still isn't as good as Github, but it is good enough for many things. And we see improvement coming in every few months. Not to mention Microsoft will likely move the whole Github infrastructure to Azure.
HOT TAKE: The things that drove Skype to change from the Golden Age Of Skype were largely market-driven things like the rise of the smartphone that killed pretty much every other popular communications platform at the time (AIM is dead now, too) and there is no evidence that an independent Skype would have weathered the market shift any better than a Microsoft-owned Skype did.
Microsoft has been the largest contributor to Git for a while. Getting to a point where all of Windows is building out of Git repos has caused them to solve problems that Git has never faced previously.
For all the silly unicorn IPOs of companies which don’t necessarily have a clear product or revenue, GH seems like the type of company that could successfully go public and actually succeed. It would fend off potential suitors for a while too.
I don't think anyone really understands what Github is. It's more like a spirit then a thing. I bet Microsoft is prepared to pay a lot for the goodwill, they don't care about the product/tech.
GitHub being bought buy Microsoft is like the opposite of what it represents. If Microsoft wants to pay for goodwill they should drop a big chunk of cash on forming a GitHub Foundation similar to Mozilla.
If this goes through, it's going to impact users of Atom and VS Code - one of the two has no reason to survive. I wouldn't be surprised if the former is discontinued then.
That would be kind of awesome, but I don't know how the github culture would meld with a big corporation. It would be an interesting case study for other similar start ups.
Didn't github already deliberately kill their own culture a few years back with policy changes and a layer of middle management?
edit: This isn't snark. Didn't they legit cut a bunch of perks, refocus everyone's priorities, and hire some people to start cracking whips? Or was I reading the public accounts of that situation wrong? Got bad scuttlebutt?
I hadn't heard that. That stands to reason. Business people actually detest that "flat organization" model. They really get off on hierarchy and power.
I think you are saying that they shifted from the full holocracy ambition to one with a more traditional management element because it wasn't working as well as they wanted.
Aren't dev tools are going cloud. MS might have just bought a bunch of servers with nasty dependencies on Jenkins in the hopes it results in more deployments to Azure.
E3 -- Embrace, Extend, Extinguish. Stems from at least the 1990's. The MS behavior as opposed to the term can be argued to pre-date that decade.
I'll just remind people who may think I'm being old, crotchety, and living in the past, that this has been a repeated pattern of theirs that has resurfaced so often and reliably as to become considered by many to be a core component of MS culture. Repeated including in the face of DoJ prosecution and monitoring.
Almost nobody who even worked at Microsoft in 1996 still actually works at Microsoft. You are bringing up something said in one meeting from when floppy disks were still pretty important to the computing world.
I can't fathom a single statement from 1996 about a computer company that is still accurate today, and EEE falls into that category as well.
But with the latest news that MS is looking at buying GitHub demonstrates nothing has changed at MS.
Instead of helping and pushing the industry forward like Google does instead we have them hurting the industry. They buy GitHub and we will have a mess as FB, Google and the others that contribute to the greater good of the industry will have to go else where and we have a fragmentation mess. Google gave us map/reduce and K8s and so many incredible papers and a big one is VP8 and VP9 stopping the mpegla from trying to extort ridiculous license fees.
Or finding all the big security vulnerabilities like Shellshock, meltdown, Cloudbleed, Heartbleed Spectre and so many more and sharing and helping the entire industry.
I swear MS just does not want us to have nice things.
If Microsoft buys github I will be pulling all my repositories, deleting my account, and finding any software that I use on github and quickly cloning it and importing it to some competing server(Probably gitlab)...and I would encourage everyone else to do the same.
edit In the meanwhile I would recommend filing an issue on every single project as you would normally file issues on, asking what their migration plan in the event of a Microsoft takeover is, before, not after this happens.
Leave it alone! Github pricing is perfect for all incomes. They'll slap on a few features, create a tiered payment system and charge $20 for the first tier.
100% will take my code somewhere else if this acquisition goes through. Go acquire Atlassian if you want to "diversify" your portfolio. Go stick it up your Azure MS.
Is there a truly decentralized peer to peer alternative to github or gitlab? And please no lectures about how git is inherently decentralized. Obviously there is something missing that causes us to rely on github etc.
- Github user account migration/integration with outlook.com
- A much-internally-championed Github/Office/LinkedIn integration effort created by some sharp but inexperieced program manager fresh just out of Washington State, that happens to require Github project descriptions to be authored in .docx format, launched as an improvement from that primitive format made by a certain mac fanboy.
- An effort to migrate the Github infrastructure to Azure cloud on Win 10 that ties up 85% of all development resources for the next five years and ultimately fails.
And that’s from when they weren’t allowed to use BSD or Linux internally and Windows Server was rudimentary. Microsoft would have no problem migrating Github to Azure quickly.
Has Google ever killed an acquisition or product of Github's scale? Most products that come to mind were niche things like iGoogle, or things that got deprecated/integrated into other products, e.g. Buzz, Wave, Picasa
Google Reader is the product that seems to leave the most bitterness, but that too seems to have been relatively niche.
Google killed Google Code in 2016 and moved much of their code base to Github. Maybe they'll rebrand Github back to "Code", but outright killing it seems unlikely.
Github is a niche product. Average people do not use it at all, its utility value is entirely niche. The average person will never know what Github is or what it does. It's useful to less than 2% of the US population, and less than 0.5% of the global population.
They've killed / destroyed / butchered large acquisitions before, yes.
Metaweb / Freebase cost them a reported $200-$300 million.
Feedburner cost them $100 million.
Slide.com cost them $228 million.
Motorola was a $12 billion disaster. They might as well have killed it.
DailyDeal cost them $114 million.
Wildfire Interactive cost them $450 million.
There are a couple dozen more in the $30m to $100m range.
Please no. I spent months to find strategies to limit my google product usage between mails, search, phone, calendar, contacts, maps, streaming, dns, adsense, cdn and analytics. It's a struggle. They are everywhere. I really hope they don't eat github too.
I mean the other option is to just not sell and if money is a problem adjust pricing. I find their plans to be incredibly cheap for what they offer so they definitely could have room to grow in that regard.
If you want to get Google search results and not give them any personal info or create a search bubble, try startpage.com or ixquick.com (same company). They have built a privacy-focused search engine that rides on Google's results. I'm not affiliated, I just enjoy using it.
If you don't trust Startpage you can try DuckDuckGo with a !g modifier, it will search Google (mostly) anonymously for you.
As far as I understand, the !g modifier doesn't search anonymously for you, it just redirects you to the google results page. It's useful when you know that ddg won't give you a result as good as google.
You can also use the !sp flag, that redirects you to the startpage result page, without passing through google at any time, so that's what you should use instead of !g.
