The Most Important Social Network: GitHub | Hacker News Reader

The Most Important Social Network: GitHub

7fff.com

160 pointstuke72 comments

To paraphrase Leonard Nimoy from the Simpsons [1]:

The following article is true and by "true" I mean "false". It's all lies but they're entertaining lies and in the end isn't that the real truth? The answer is "no".

We geeks seem to often be susceptible to hype and hyperbole. Someone is really in love with Github and thinks it's the greatest thing ever and it's going to change the world. It's easy to get caught up in your own excitement. I get it. That's fine.

But I have to admit to having some Github fatigue. We've gone through a spate in the last year of "Github is the new resume", "Github will change engineer recruiting" and now "Github is the most important social network ever".

In many cases I don't believe the author is being deliberately "linkbaity" but that's ultimately what it is.

Part of the problem too is that you get a certain about of "bubble thinking" in tech circles. You see this when VCs get excited about Quora thinking it's going to be the Next Big Thing [tm] because "everyone" is using it (meaning "lots of other people in the Valley"). That's what I mean by "bubble".

I play boardgames a lot and it's much like "groupthink" there (an isolated group of players will evolve a play style and view on strategy very different from other such groups).

In all of the above cases the cure is just to get out of the bubble and expose yourself to different influences and views because the end of the road for this kind of thinking is simply stagnation and becoming out of touch.

Github is great. Their engineers are great. Source control is important. Some will be able to use it to demonstrate their work ([2] really resonates with me). All of this is true but let's not go overboard.

[1]: http://www.imdb.com/title/tt0701263/quotes?qt=qt0332688

[2]: http://news.ycombinator.com/item?id=4244420

Misguided title as it may be, article (successfully) really suggests that GitHub is ignored in studies of social networks and group collaboration, and it shouldn't be, given its productive capacity, as opposed to time-wasting alternatives.

Many sociological and even political models could reference it in their development and study.

Love the Nimoy quote. Will be reused.

Related TED talk: Beware online filter bubbles.

http://www.ted.com/talks/eli_pariser_beware_online_filter_bu...

Not disagreeing with you, but here are some ideas on how github could take things to the next level.

0. Get Junio Hamano from Google and several of the top git-core contributors, so they can start releasing open source patches to git that support github's long term roadmap.

1. Come up with efficient custom diff functions for a variety of common file types, not just image diffs but also dbdiffs, mp4diffs, and the like. Github has already done some work in this area, and a little known fact is that git already has some support for custom diffs in .gitconfig:

  https://github.com/blog/817-behold-image-view-modes
  http://man.github.com/git/git-difftool.html
  http://stackoverflow.com/questions/9154110/data-versioning-with-git-custom-diff-for-specific-files

2. Next, integrate git-annex for storing huge files in git into git core, maybe with a configurable option for where to place the files (perhaps on a github-customized API to S3).

  http://git-annex.branchable.com

3. Now, with some thought and work, you could start supporting diff-based workflows in a wide variety of disciplines, including movie, image, and document editing.

4. You might also start expanding the range of github to start including db versioning capabilities. I am convinced that the successor to Postgres will include schema versioning as a first class capability. One way to do this is to allow every table to have a list of historical schemas s_i each associated with the instances v_j that were populated when that schema was in effect. Then all you need are functions that map from s_i to s_{i+1} to migrate historical data forward in time.

Well, that's cool. However, these are just incremental improvements and really don't serve any ideal of github being the most important social network. This would be more for a thread like "What are some bugs we should fix on Github?"

Agreed. Long ago I remember trying to get a non-developer to use cvs, which is much less complicated than git. It's not for everybody.

Which is too bad, I imagine that some version control even in it's simplest form could be good for everybody. At least I know a couple of non-tech types that could use it.

But you're right, it's very complicated.

What's the point of claiming that GitHub is 'The Most Important Social Network'? How can anyone make that claim? I feel silly thinking in the back of my head "But facebook is approaching 1 billion users, how is GitHub even comparable?".

Twitter gets credit for facilitating the arab spring. Facebook has 1000X as many users. Even if we want to be 'work' specific, Yammer is used by over 200k companies for what I assume must be business purposes.

This must be hyperbole. That would make sense.

Guess you should write your own blog post. :-)

"The Most Important Social Network: Real Life"

Good one.

I think his point is that Github is about actual work artifacts, while all other business networks are basically talk about a more or less specific topic - more flexible, but also less focused.

Github is much closer to the topic at hand. There is almost no way around attaching comments to a piece of code.

But Github is about one specific kind of real work. The conversations that happen on Jive and Yammer are also real work, as are the documents made in corporate/public wikis (which are studied extensively) and the work assets made and stored in SharePoint and Google Docs. I don't disagree with the author's premise that GitHub is great and should be studied, but it only covers one small portion of one type of work.

You missed the tiny little word "artifact". The discussions on Yammer and Jive are work related and are part of the process, but are not rarely directly attached to an artifact. You would talk about the artifact, but its not "there". It is produced somewhere else, in the depths of IT. That doesn't mean that Yammer ain't interesting.

SharePoint and Google Docs are not social networks. They are sharing solutions like Dropbox is. Your corporate wiki might qualify, but thats hard to study.

I did not miss the word, just disagreed with his premise that assets are an important distinction.

That being said, wikis are not hard to study, its incredibly common and they are high on the list of the favorite things to discuss among collaboration researchers and have been for years - I've experienced this working with several, but you can also look at research such as http://www.citeulike.org/group/1136/article/114322 to see more.

He uses a metric for importance other than just size.

That said, his points that there were 600000 software developers in the US 10 years ago and more than a million GitHub members now, SO there must be significant developer mindshare seems nonsensical unless maybe if you assume GitHub is completely dominated by Americans, which anecdotally doesn't seem the case to me (I spent a few minutes looking for a breakdown of users by country, but didn't see any).

Also, confirmation bias.

For sure about the confirmation bias, but I would propose that the post is meant to encourage scholars to come up with their own falsifiables and prove it one way or another.

The 600,000 vs 1,000,000 means: A lot of Americans, and the excess are people who are not from the USA or have multiple accounts. But the main point is that in gross numbers many of those 600,000 have accounts.

I think the central point of the story is that github is more like a social network than many of us realize. It's interesting, and could indicate some useful future direction for the service.

if the story were titled "Github - a major social network" or "Github - more of a social network than you may think", would that change your opinion of the story?

It's surprising to me how much of the criticism here on hacker news ignores the central point of the article, and instead focuses in on individual phrases or semantics.

I agree with you that they could have chosen a better title. But I found the article interesting and thought provoking.

I don't think the criticism you're refering to is frequently directly at the articles; They're reactions to link-bait. Link-bait depends on sensationalism, which I think is opposite the goals of HN, so people don't respond to it well. Nevermind that its inherently misleading, which isn't a way to endear readers. It hurts credibility, etc..

Your criticism of the title would be more credible if you also acknowledged the point he was making. As it stands you came across as dismissing the entire article as stupid.

And I wonder if the hyperbole in the title is really link bait, or just the opposite, troll bait. I'm not sure the title helped the article's ranking on HN.

I'd be interested to hear your thoughts on what "troll bait" is in this context. I'm not following.

Because the thing that overthrows Facebook could very well be collaboratively created on GitHub. The conflict between Facebook's business interests and its user's interests is the biggest painpoint and thus most ripe for disruption.

> Note also that Linux powers Facebook, Google and much of the rest of the web, and Linux itself relies on GitHub for collaborative developement.

Now that is not even hyperbole, but just plain incorrect. Linux does have an official git mirror on github. None of the development of Linux is done via GitHub.

Thanks. I removed the false statement as it was ancillary anyway. That said, in import it was not false. Linux is collaboratively developed using Git, and as Torvalds himself said when he did temporarily move Linux kernel dev onto GitHub, "But hey, the whole point (well, one of the points) of distributed development is that no single place is really any different from any other, so since I did a github account for my divelog thing, why not see how well it holds up to me just putting my whole kernel repo there too?"

I'd not be surprised if he makes a permanent move once GitHub fixes a few things: https://github.com/torvalds/linux/pull/17#issuecomment-56546.... On the other hand, he is a known control freak.

I hate to say it but Facebook is not going to be overthrown, ever. It's planted very firmly. Backed by mega corps and huge international banks.

Facebook is down right scary in the way it has permeated the life of almost every person on the globe. It's monolithic and we have little insight into what they will actually do with all of our data. A little government + facebook cooperation and we'll really be a dystopia.

With that said, I don't use Facebook and hope some more people wake up and smell the coffee too. We need a decentralized network, as that is the only way to avoid corruption.

> I hate to say it but Facebook is not going to be overthrown, ever.

Just rhetorical exaggeration, right? But just in case you're that naive: IBM, Microsoft, Windows, Google. The Roman Empire, the British Empire. Need we go on?

I give Facebook ten years max. It may already be near its peak.

IBM, Microsoft, and Google -- None of them have been overthrown.

The Roman and British empires lasted hundreds of years.

Facebook will be around long after you and I are dead.

It's not another myspace or friendster. It's a firmly ingrained institution that people probably value more than their wallet at any given time.

Overthrown implies a zero sum of land, etc. How relevant is IBM to modern computing? Sure, they're doing some great work in supercomputing, but there's no monopoly they control like they once did. MSFT has a solid monopoly, but their relevance is lessening though still influential in some circles. Google is young/ambitious enough that it has its skunkworks projects - gGlasses is an extension from search, and self-driving cars is from maps/geo data, but it's still not really entirely defensible from their central data monopoly.

"Overthrown" does not mean "disappear", it means "no longer king."

IBM was once the king of computing. It was overthrown by Microsoft. Microsoft was then king, but it is no longer, though those living in the past will point to its continue dominance of the PC, a dominance that is both declining and becoming more irrelevant. Google was crowned the king of the web, with users spending most of their time on google.com, but now they spend far more time on Facebook.

> The Roman and British empires lasted hundreds of years.

We're talking technology, not geopolitical power. The former operates on a vastly accelerated time scale, aka "Internet time". I recommend you read the Innovator's Dilemma, or at least the cliff notes on Wikipedia: http://en.wikipedia.org/wiki/Disruptive_innovation. There's a nice list of things many of which weren't expected to be overthrown so quickly.

I'm wondering if you might actually be a celeb: http://www.youtube.com/watch?v=eywi0h_Y5_U.

You can say that a website that most people on it use for less than five years will outlive us ?

I don't think FB is going out of business anytime soon, but as time passes, I think that it's core value is more like its original vision -- i.e., a social directory. Seems like it'll become a repository of all your social contacts and your virtual passport when signing up for other sites and will cease to be the place where you spend all your time communicating with friends and discovering new things.

Sounds like you're suggesting Facebook may eventually become the next YellowPages. I could see that happening. It's mostly what I use it for anyway—remaining available to people who, since joining Facebook, seem to have forgotten I'm still reachable via phone, email, and IM.

I think that fluid dynamics serves as an appropriate analogy for understanding social networking. Companies like facebook interact with a force that's very similar to fluid flow and equally subject to the conservation laws. As with fluids, it's convenient to view your community as a continuum of users whose desires are well-defined individually, but vary continuously between each other because of socialization.

From this, it logically follows that changing one aspect of the system could cause ``unforeseeable'' consequences in another. I hesitate to say ``unforeseeable,'' because really it's a complete failure by the decision makers to understand their community's intra-member interactions. So, claiming that facebook isn't going anywhere is flawed, because as soon as the users' needs change, altering one service could be like closing a release valve and increasing the temperature of water in a pipe: eventually, there'll be an explosion.

> I hate to say it but Facebook is not going to be overthrown, ever.

For ever is a long time. As eevilspock said, I'll take this as an exaggeration rather than literally. However even then I have to disagree. Virtually any prediction we make (pundits, bloggers, analysts too) certainly will be wrong regarding the tech sector. Disruptive technologies are unknown to us until they are surprised upon the general public. Facebook may stay around for a number of more years, or some new tech will push it into non-existance.

The great thing about guessing about the future is that it is free, and you don't get called on it if you're wrong.

Lastly, the only thing we truly know about the future, is that it will be unlike anything we can imagine today.

MySpace anyone?

100 people creating value >> 1 billion people wasting time.

If you were a designer, you would claim Dribbble instead of Github.

Either way, you would be delusional. Social is just a useful paradigm on the web. More and more products would use it as default. Just like every site now has a search box. Github also has a search box. It is not the most important search engine.

Google has a social network, and it isn't mentioned in the piece. But I'd bet there's little doubt Google is the most important search engine.

You're right about Dribbble, but I'd argue GitHub is for collaboration and Dribbble is for showing off. Certainly people collaborate on Dribbble and people show off on GitHub, but by virtue of how people actually use the platforms I think GitHub is more collaborative and more important (at least in the way the OP posits). Go to http://dribbble.com/, click the first thumbnail, and count the compliments in the comments.

It could be a very important special search engine, like Google Scholar. Sadly, search on Github is also a topic to rant about.

Reading this post I get a nice feeling of joy for belonging to the software community.

Being so entrenched in the software world it becomes easy to gloss over these details. The outside perspective of this article really shines a light on how well the software community collaborates and shares.

I can't think of a single community that even comes close to the level of collaboration we have in software.

Honestly, I had the opposite response. I think the claim that Github is the most important social network is willfully ignorant of the role Facebook/Twitter plays in every industry beyond software development -- or, more accurately, the role Github doesn't play.

I agree. One of my good friends is a realtor. Facebook changed the way she does business. It is the source of 80+ percent of her leads and is, quite frankly, the MOST important social network to her.

Github may be the most important social network for a certain class of open source developer, but the author's attempt to extend its importance to demographics beyond this group makes the author sounds as if they're just trying to stake a claim as an early member of a new "cool" social network.

"Facebook! Twitter! LinkedIn! VK! Renren! These are among the most famous and largest social networking platforms in the world. But are they important? Of course they are. They’ve changed the way humans interact. But let me challenge you and ask: Have they changed the way we work and think? I do think they have, to some extent."

"To some extent" is comically naive. These platforms have played a considerable part in the youth of internet/earth.

Perhaps the author suggests that github is part of the next phase, which is more credible.

Well it all comes down to what metrics you use to decide the importance of social networks, doesn't it.

If you're judging by the ability to interact with your real life contacts online then facebook is probably the best social network(gah!)

If you think that meeting like minded people & having awesome interactions with them is more important, then maybe Google+ is the most important social network.

If you agree with Jane McGonigal that games can help make the world a better place, then WoW is probably the most important social network.

But if you think that collaboration on software is more important(and facebook & WoW are softwares), then GitHub is probably the most important social network. It all depends on what you rate higher.

And putting myself in the shoes of the writer, I think it's the end product that comes out of social networks, is the metric he's pointing to. While other social networks directly affect the lives of more people, much much more than GitHub can ever hope to, the products coming out of GitHub are, or would soon affect more people than any single social network could hope to do.

Someone out there is building the next facebook, the next WoW, the next Linux, or maybe the next Google and it's likely that GitHub will play a part in it.

Keeping that in mind, I'd say that yes, GitHub is maybe the most important social network on the internet.

To interact with my real life contacts online, the best social network is email. ;)

Seriously, Github already plays a part in Facebook [0], WoW [1], Linux [2], and Google [3]. No need to wait for "the next". However, its part is quite indirect and it could easily be replaced in all four examples.

The fact that it is based on git makes it inherently replaceable. This is a good thing and probably a big reason for its adoption. If github is down, which networking connections break?

[0] https://github.com/facebook [1] https://github.com/Blizzard/api-wow-docs [2] https://github.com/torvalds/linux [3] https://github.com/google

I think the term "social" is being overused here in the context of a startup, and was probably stuck in the tagline like many buzzwords, with an eye on marketing. The whole internet is social to a point. I think the term collaborative is more apt. The words that tended to jump out to me in the article were collaborative and commentary. He says "GitHub puts the social exchange at the very center", and I disagree. At the center with my view I see a strong tool to manage code that enables great collaboration. When you see a getting started tutorial with Github, the communication aspects are rarely mentioned. In the few projects I have seen getting popular there are usually external forces at work (i.e. Hacker News, Blogs, Large Corporate Support). It may be nice if Github was better at marketing your project. I suspect they may have to be better at it now and the product will undergo many changes in the coming year. That substantial portion of that 100m will likely all go to marketing efforts, not more engineers, which I think is a good thing.

I believe there are opportunities for many Githubs For {insert_industry}

The article focuses on blending communication & code. While it's great to build tools that blend them, we can't pretend that other channels and mechanisms won't be used. Hopefully github will see the importance of doing that rather than forcing everyone to live inside a github world. Back in 2004 I did a class project called Open Sources [1] at MIT that looked at blending public mailing lists with public CVS archives (no github at the time). Years later I see some of these ideas becoming more popular in coding tools -- my take is that visualization of group history and individual contribution will become critical as catalysts of open-based works. These top-down summaries provide a map & narration rather than autistic reverse chronological fine-grain lists, as well as giving a broader overview as to who an individual is and what they have achieved.

[1]: http://smg.media.mit.edu/projects/OpenSources

I would submit that the term "social network" has lost almost all distinctive value.

Github isn't a social network in the way that traditional online social networks have been viewed – it isn't just about "connecting" and communicating the way that friendster, facebook and twitter all are.

Instead, it's an online code repository that has social features – in other words, it's the next step in making our online selves a more effective extension of our off-line selves, doing work, building things, but doing it in the context of a social group – just like we do in the off-line world, more often than not.

This is an example of a niche concept becoming widespread enough that it almost becomes table stakes rather than a notable feature.

That's still pretty cool – but again, I think it means that the term "social network" is losing its distinctiveness.

> That's still pretty cool – but again, I think it means that the term "social network" is losing its distinctiveness.

Good. The term should lose its distinctiveness. Currently, ``social'' closely relates to friendliness or geniality, but its Lockean definition pertains to ``society as a natural condition of human life.'' Although ``societal networking'' may be the more appropriate term in certain contexts (e.g., those outside the scope of friendly or genial conversation), ``social networking'' nevertheless remains correct. The exceptionally broad definition of the word ``social'' effectively portends ambiguity when discussing social networking, so a presumption of distinctiveness isn't warranted.

I don't really agree with the claim, but I'd love to see Github enter other areas of science like Physics, Maths, Chemistry, etc. Not just programming and computer science. That would be great, Github could work as a center of science all around the world.

It could work as a center of collaboration all around the world. Until [Gobbler](http://gobbler.com) came out, I was using Git as a means of version-controlling my music projects (even going so far as to branch and merge the project, my DAW allows me to do this without touching audio files since all changes are saved to a new file). GitHub has shown us that you can create an extremely powerful collaboration tool around Git, all you have to do is design a user interface that's simple, elegant and familiar to those involved. I can't wait to see what other companies do with Git.

Obviously with $100M cash, big things will be coming for GitHub. It is hard to imagine that they haven't thought about other collaborative communities.

Will big things be coming? I hope they use that money to grow on the core things they are good at rather than try to capture markets which don't need capturing.

They have a unique position in that almost 100% of its users can only say good things about it (bar maybe Torvalds).

I want small things from Github, glorious small things.

With this level of funding, they can try to achieve both small and big objectives simultaneously.

I would argue that 10 million+ people have gotten jobs through LinkedIn [1]. And getting a job equates to at least $40,000 a year in income, and a direct improvement to people's lives. The existence of LinkedIn has been a part of ~$400 billion a year to the world's economy. I think Github is a tiny tiny tiny fraction of that in terms of its real-world effect on peoples lives.

[1] LinkedIn has 161 million user accounts, and assuming 5% of them have gotten jobs through the services which could be low. Around 4 billion candidate searches a month are done there. http://press.linkedin.com/about

That assumes LinkedIn helps companies/workers fill positions/get jobs 1 year sooner than they otherwise would have (if the pay stays the same).

While I'm sure Andreessen Horowitz feels some tech-world pride in this investment, they are still a venture capital firm. This article is a bit too sensational, self-loving to really portray the importance of the investment.

I think TechCrunch nailed it in their article today, where they wrote, "Think of it as a filing system for every draft of a document." Github right now is limited to code-sharing. But it's potential is so incredibly huge. Like the article stated, imagine applying it to PSDs, Word Docs, Excel sheets, any document imaginable. IMO that would be THE killer enterprise prodcut.

We didn't write it.

It's not really limited to code sharing - any version of a text file can be versioned in the same way (e.g. the raw LaTeX code of revisions of a book or other document).

Granted, this is still largely of interest to STEM geeks, so extending its capabilities to binary format versioning might lead in very interesting directions, as you point out.

If Github can figure out how to serve customers who want to do version control on cat pictures, then I think the link title would have some merit.

The most important social network: GitHub ... (as long as you're a developer)

Clearly my mom has no reason to know github even exists.

>Clearly my mom has no reason to know github even exists.

Your mom probably also has no reason to know that the Haber—Bosch Process exists, but it's still one of the most important scientific developments of last few centuries, even to her.

You're comparing github to the Haber-Bosch process??!

While I think that is ridiculous, your claim appears to imply that even though my mom is not aware of github, its existance is as important that a major part of our civilization's sustainability relies on it?

No, thats just plain stupid.

Anyway - my original comment was that github is only valuable as "the most important social network" to a very narrow portion of the populous.

The funny part about this is that most software that the generally public deals with regularly is not on Github anyway and even the ones that are existed just fine for many years before Github so it's not as if there is some contingenct on Github or something.

>You're comparing github to the Haber-Bosch process??!

No, don't be dense. The point of the analogy is that something's importance to a person is independent of that person's knowledge of that thing.

Oh that's just ridiculous. I mean there's a point of "jumping the shark" here even though I hate that phrase. I mean come on, it's not that great. It's arguable that this whole attitude about Github is massively pretentious and overblown. I don't think we can compare Github to the development of agriculture, pasteurization, classical mechanics, etc.

I didn't compare github to any of those things. I simply pointed out that "X doesn't even know about Y" is not equivalent to "Y isn't important to X".

I think your choice of wording was confusing some of us.

It's frustrating that one cannot make an analogy these days on the internet without some people assuming that you're making a comparison of the analogues. If a comparison seems unreasonable, it seems to me like the obvious thing to do is reread it and see if it makes sense as an analogy instead.

The problem is that an analogy's illustrative power is increased by using an extreme analogue, so if you interpret a good analogy as a comparison, it will always look like hyperbole. On the other hand, intentionally misconstruing analogies as comparisons is a great dark-side rhetorical technique, if you're into that sort of thing.

I think this issue is compounded by the "skim this and spend .02 seconds on critically think about it before moving on to the next data-byte online" syndrome.

I believe the thrust here is that Source Code Management has given us a template for Source Content Management.

Reply on news.ycombinator.com