Hacker News new | comments | show | ask | jobs | submit login
Why is Atom closed source? (atom.io)
126 points by zeckalpha 970 days ago | hide | past | web | 107 comments | favorite

As far as I'm concerned GitHub is free to handle the licensing of their software however they like, but this opens up an interesting thought experiment:

We now have at least two companies making significant Chromium-based editors, one is Adobe and the other is GitHub.

One of them is developing the core of their editor technology using a fully open-source model (hosted on GitHub, naturally) under a well-understood and widely accepted license, the other isn't (or at least isn't currently planning to).

2 years ago, who among you would have guessed correctly that the one whose editor was open source would be Adobe's?

Adobe's Bracket front-end editor is open-source, yes, but does it have any traction? I have yet to hear any front-end developer friends who even bother to use it in any capacity other than testing it occasionally. I tried it a few times and couldn't justify why I'd want to use it over Sublime Text.

Historically, it appears that closed source front-end editors like Textmate, Sublime Text and others find more success and usage in the market than open-source front-end editors.

"Historically, it appears that closed source front end editors like Textmate, Sublime Text and others find more success and usage in the market than open-source front end editors"

I wouldn't classify Sublime Text as a front-end editor. When I think of front-end editors, I think of Coda, Espresso, and Brackets. Sublime Text is better compared to Notepad++, Notepad 2, Gedit, SciTE, Emacs, and Vim, which are all open source software and extremely popular.

If you're going to market your text editor as hackable and as flexible as Vim and Emacs [1], it's not a stretch for the developers interested in your product to assume it's going to be or should be open source software.

[1] http://blog.atom.io/2014/02/26/introducing-atom.html

"Sublime Text is better compared to Notepad++, Notepad 2, Gedit, SciTE, Emacs, and Vim, which are all open source software and extremely popular."

Fair enough. I use ST2 like Coda/Espresso/Brackets for my own front-end development b/c of the app's agility and abundance of useful plugins for the front-end.

Keeping Atom partially closed source, again pointing to historical marketability, could give it more credence among developers of both ends. There's a reason closed source, proprietary editors are so massively popular vs. open source editors which often can't seem to get meaningful market share (whether Enterprise or Sole Proprietor/Freelancer).

That said, I really wish an open source editor, on the front-end specifically, like Brackets or even Notepad++ would become something I want to use more than ST2, Coda, etc. but there's always something missing.

Atom has potential as a partially open source editor but it remains to be seen. I'll need to be drooling in order to truly make a switch.

> There's a reason closed source, proprietary editors are so massively popular vs. open source editors which often can't seem to get meaningful market share

I'd be very interested to see data that indicates that closed-source text editors have more market share than their open-source counterparts.

I base this on my own personal experience and understanding from a front-end perspective. Maybe back-end devs are different?

I'm quite positive on the other end of development, mainly Microsoft devs, they're almost always using proprietary software like Visual Studio. Sadly, Microsoft dominates the Enterprise and these environments are nearly half the job openings I come across/hear about.

Better than VIM? What do you say that?

You've misunderstood me. I'm saying that Sublime Text should be compared to an editor like Vim, not Brackets. I'm also saying that the folks at GitHub said it was more flexible than Vim.

I think he meant useability (by a newbie).

I have yet to find an editor that makes me more productive than VIM.

While I love Vim and agree that it's probably the most extensible editor available, trying to explain its idiosyncrasies to a new user could take quite a while. Trying to explain why you can't just grab your mouse, highlight a line, and cut it and paste it somewhere else in the document would probably lose a person. "Hey, you're going to want to find the text, go into visual mode, select it, hit :m38 to move it to line 38... "

> Adobe's Bracket front-end editor is open-source, yes, but does it have any traction?

I suspect one of the big detractors from Brackets is that it doesn't [yet] support VIM/other keymaps out of the box (https://trello.com/c/ckAel1Cl/358-118-keyboard-shortcut-prof... ). Additionally it could have set itself apart as a possible thick-client NodeJS debugger, but again not out of the box (http://blog.brackets.io/2013/08/28/theseus-javascript-debugg... ).

While developers are used to installing features they want, Brackets suffers from the fact that there isn't any built-in package manager and therefore these types of optional features are gated behind a Google search (where most of the top results are questions about support, not the answers).

(Disclaimer, I work for Adobe)

Actually, Brackets has a built in extension manager. You can find a list of extensions here - https://brackets-registry.aboutweb.com/ - and in the product you can go to File->Extension Manager.

All the extensions in the registry can be installed from within Brackets.

Github for Mac, and Windows are not open source. These are tools with tight integrations to Github, and still not open sourced. It's not like Atom is a precedent.

I guess the target market for Atom is those that already use a closed editor like textmate or sublime.

As an emacs user, I would be pretty happy if someone came along, took the super extensibility of emacs (which at first glance atom appears to do) and spun up a sexy, modern, not-lisp-using alternative. The thing is though, my number one priority for a text editor is that it's under a Free Software compatible license. This is simply not a negotiable point for me personally.

I honestly find it pretty amusing that despite the fact technology progresses at such a rapid rate, I'm using a text editor that was first developed a decade before I was even born.

Why not Light Table? It's released under a GPL license [0] and seems to very much be going for "super extensibility". Granted, it does use a lisp, but ClojureScript is pretty nice.

[0]: https://github.com/LightTable/LightTable/blob/master/LICENSE...

I'm an Emacs user and I've evaluated lighttable. I like a lot of what I see, but for now it's lacking some pretty fundamental user interface concepts that are critical in making an editor that's optimized for the long-term user and not just for the newbie.

For example, if I need to move my hand over to the mouse to do some common action, that's not ok. If I even have to move my hand over to the arrow keys, that's seriously suspect.

I would also never use the native OS file open dialogs for the same reason. It's fine to offer them of course. But anybody who's trying to get fast and proficient will rapidly move beyond them. For this, lighttable offers workspaces & the navigation pane, which is still really primitive compared to ido-mode (poor keyboard navigation, no tab completion of partial names, no ability to jump beyond a single, pre-created workspace).

Textmate 2.0 is open sourced (GPL3) https://github.com/textmate/textmate

> I would be pretty happy if someone came along, took the super extensibility of emacs (which at first glance atom appears to do) and spun up a sexy, modern, not-lisp-using alternative.

Like Textadept (http://foicica.com/textadept/) ?

Closed source? Why hack it then? Why waste the effort to learn it or customize it?

It's kind of like the new/gnu reality, where the GPL, designed to protect the authors from commercial exploitation, makes this exact exploitation easier than less restrictive OSS licenses...

Making money off a text editor is going to be hard. Especially when it is meant to be hackable.

For an excellent description of another instance of this sort of issue, see [1] about Canonical and Mir.

Github seems to have things pretty figured out: they have a closed core that they control, with a veneer of hackability, and a lot of hype. You then get people to work for you for free by having them create plugins that, while open source, are effectively controlled by Github, Inc.

The key thing is to make people feel that they're contributing to a community, even though that community doesn't really have ownership of the work. By the time the tension between the community's needs and Github's needs comes out, you've already locked everyone into your ecosystem.

Other notable instances of this phenomenon include the MAGMA computer algebra system, and the creation of SAGE (see [2]).

[1]: http://mjg59.dreamwidth.org/25376.html

[2]: http://sagemath.blogspot.ca/2009/12/mathematical-software-an...

Interesting story of SAGE here. What happened to it? Nowadays it seems that numpy/scipy has more users and community attention.

Sage is going strong. Their new cloud service https://cloud.sagemath.com/ is fantastic (multiplayer ipython omg).

Amusingly, they are also going partially-closed source: http://sagemath.blogspot.com.au/2014/02/what-is-sagemathclou...

SAGE is still going very strong.

numpy/scipy have very many more users, but I don't think that they're really comparable: they're aimed at numerical computation for science, while SAGE aims to be a do-everything tool for mathematics. All of scipy/numpy is actually included in SAGE, and you can interoperate between numpy stuff, and, say, exact rational arithmetic, or group theory, etc.

And you can use R and Octave in a worksheet, too, but my understanding is that the point of SAGE is much closer to a superset of the types of functionalities that you find in SymPy. That is, that it is meant for symbolic manipulation.

I think a lot of people were expecting it to be open source.

I don't have access yet to try it to see how it compares in terms of speed and feel. That said, this information seems to put it more in the category of a reimplementation of Sublime Text with a different tech stack and a much bigger company backing it.

For me it just sounds like a stupid move. I might not have all the details, and I might be missing something. But I would expect some lone programmer or small team to try and commercialize a text editor. In some very rare cases a text editor today might be profitable for a very small team, because of people to afraid of VIM (vimps...).

I would not have expected a bigger and more modern company like GitHub to try and make money off a text editor. It's like building an inferior mouse trap in a world of open source mouse traps. So how this strategy unfolds will be interesting to watch!

I feel like a still-bootstrapped GitHub would have made Atom free and open source and used it as another tool to reach more developers, especially new ones.

But alas, they raised $100M and are probably feeling pressure to find new revenue streams.

The interesting thing is that they would accept pull requests (presumably from the devs they offered that "restricted" access), but might not share the profits from charging for atom.io.

Food for thought.

Disclosure: I'm a founder.

You may be interested in what we're doing over at Assembly: https://assemblymade.com. Open source products that share profits between contributors.

This does look very interesting. Do you have any stats available, or is everything too new?

I'm just wondering if a developer could make a decent monthly living by working 40hrs/week on different projects.

Also, what happens if one of these projects wants to break out and go for VC money? Is each contributor's stake a legally binding "share" of the app?

Lastly, who owns the idea, in a legal sense? Would Assembly receive all revenue and distribute it, in which case it's logical that it would receive all ownership of ideas?

That's really cool and seems like an obvious thing to do. I hope you folks can make it work!

I find that model interesting. It lets me fix any showstopper bugs (possibly with terrible hacks, at first) right now and potentially get it upstreamed later. That is a lot better than any closed source software!

It's much better than closed-source software: they get you to do the work of fixing bugs, but they still keep the money!

Last week, I helped a vendor fix a bug in closed source software after reversing binary diffs. I would have definitely welcomed shared source!

I think either you or I misread that. I assumed "restrictive license" meant anyone could see the source, but the license would prevent you from creating a competing service. Or possible require you to sign over the copyright for any changes.

Something like MIT + Non-commercial would be interesting.

I've actually been looking for a license like that for a while. A combination of the ease of MIT and the spirit of CC-BY-NC, for source code. I don't want to write one myself since I am not a lawyer and would inevitably get it wrong.

Be aware that whenever you use a "non-commercial" license, you might as well keep it under a much more restrictive license, because nobody can use it for any purpose in the open without fear of getting sued.

There is hardly a definition of "non-commercial" uses in any one jurisdiction. There is no way how this plays out in multiple jurisdictions.

CC-NC is a waste of the name Creative Commons. They might as well use "All rights reserved" and then just not go after anyone that copies the material in their basement.

Why not just use the GPL? Only those who are willing to fully open source their product can use your code commercially. I, for one, am quite willing to make that grant to anyone who will go that far.

If you're interested, we have a license generator for source-code over at Binpress that creates licenses based on the conditions you set. Similar use-cases to what GitHub is doing with Atom


It would then be neither Open Source nor Free Software.

Only if you define them to exclude it.

And it just so happens that the Open Source Definition as well as the Free Software Definition specifically do.

Most other commercialized open source products do the same. One could argue that GitHub has done their share for the open source community already, however.

I'm a little confused why Atom isn't free. It certainly couldn't be a major revenue stream for GitHub, which has major investment. I thought this was going to be a project to drive usage, like Chrome for Google, but I guess I was wrong.

Or does not freeing Atom mean that Github is not getting enough revenue from hosting? (Eek, are they dying? Should we move our stuff off to other services? =P)

Git is a distributed VCS though. It wouldn't really matter unless your only copy is on Git and isn't cloned anywhere. (In which case, set up a remote repository on Bitbucket/wherever)

If you're relying on GitHub for Issues/Wiki, then fair enough I suppose. Not as easy to get that data somewhere else.

As I understand it, GitHub make the majority of their money from their on-premises 'Enterprise' product, which is sold with a 'contact us to discuss' price tag. Also they raised a huge sum of money while already being profitable. I wouldn't worry about them vanishing.

Of course they use the term "closed source" instead of "proprietary" because it sounds less scary.

Seems like they want to implement one of those Microsoft "shared source" models, which still count as proprietary.

I expected better from GitHub, disappointed!

GitHub itself isn't open source, nor are its desktop clients, so it is par for the course. On the other hand, if this was from Gitorious...

In an article aptly titled "Open source (almost) everything", GitHub co-founder and current president Tom Preston-Werner writes[0], "Don't open source anything that represents core business value."

He specifically mentions two examples of what not to open source:

Core GitHub Rails app (easier to sell when closed)

The Jobs Sinatra app (specially crafted integration with github.com)

Now, I'd find the argument really stretched thin if you were to tell me that a text editor is your core selling point. Where does a text editor fit into GitHub's overarching mission?

[0] http://tom.preston-werner.com/2011/11/22/open-source-everyth...

GitHub gives away an incredibly powerful service (hosted version control) for free. While hosting it may not be in itself a considerable expense, personnel and R&D is.

They have all the rights to recover their costs in some way. At the end of the day, they are a business.

Making the service free for open source and free software was a strategic business decision. Services like github make money from private repos and a good advertisement for the service is to allow open software to use it for free.

Nobody asked them to spend their money creating an editor. It was their decision. They have been profitable for a while now and deciding to make an editor, charging for it, is not recouping costs for their existing service...

How do you know what their business plan is? They're a big boy company, they can make decisions based off of their own financial analysis. They decided that charging for this editor was more profitable for them as a company. Expecting something to be open source is like expecting McDonald's to give away cheeseburgers.

McDonald's does give away straws, however.

And ketchup.

In Germany they charge for the ketchup.

This is actually standard practice in airport McDonalds in many places of the world.

Why did anyone expect this to be open source? I don't understand this entitled attitude at all.

I think it is because of the expectations set by Tom (mojombo).


He mentioned not to opensource anything that represents core business value. A Text Editor is best set to grow when available free. Examples - VIM, Emacs, LightTable.

That said, there have been hints that it might be paid ("Atom is free till beta") and Github might try to make money through it like Sublime or Textmate.

Though I wouldn't worry much about Opensourcing, I might think twice to pay for a Text Editor - considering that there are really good free alternatives.

While I'm sure you'll be proven correct, I think it's a little premature to add LightTable into that list.

Because "hackable" is right in their headline. And if it isn't really open source, we don't really have the freedom to truly hack on it.

No. This is not the entitled attitude and people definitely do not think that GH absolutely must open source all their products.

The problem begins when you close-source a text editor. For me, this is a deal-breaker. I would not want to invest time and effort in learning a new tool - my primary tool, in fact - that won't stay for more than 5 odd years.

Just take sublime as an example. I've seen so many ST users, right here on HN, drooling over Atom, ready to jump ship. On the other hand, I do not see any Emacs or Vim users saying the same. This is because Emacs and Vim are very powerful editors[1] and there's no point jumping ship to another new and shiny editor that might no longer be there after a few years. Also, I've seen a lot of friction around forums regarding the fact that the author pretty much left ST2 entirely (bug fixes/enhancements) in favour of ST3. Also, the features of a closed source editor might not be well aligned with what the community wants.[2]

[1] I'm not saying that ST isn't powerful - it's okay. But, Vim and Emacs are far more powerful - because you can hack at the source itself and because generally a much larger number of people can contribute to the project (given that not many people are competent enough to actually do that or do not have the time) - [2] and, if you feel the editor/project isn't going in the right direction, you can just create your own fork and if it's good, people will actually migrate to your fork for better usage. And that really is the power of FOSS.

If it is a deal breaker for you go away from the deal. That simple. And if you do not want to invest a couple of hours learning a new tool I doubt you will invest dozens of hours fixing it anyway.

This "it is not open source, I cannot hack it" sounds a lot like "I cannot change a battery on my phone any more" whining: many complain about it, few ever had a need or done that even if it were possible.

> If it is a deal breaker for you go away from the deal.

If it wasn't already abundantly clear from my previous comment, that is exactly what I'm doing.

> And if you do not want to invest a couple of hours learning a new tool I doubt you will invest dozens of hours fixing it anyway.

First, my comment was in response to the person claiming that "those not using Atom because it's not FOSS" are having an attitude that they're entitled to the editor being open-source. I was merely trying to point out the reasons why some of us expected a text editor (not any other software but specifically a text editor) to be open-source.

Second. Since you've entirely glossed over what I was trying to say, I'll explain it again (I should use easier English, I suppose).

It is evident to me that you are absolutely oblivious to the philosophy that drives FOSS. For a software to be open-source isn't important for me because I can make changes to it - it's because anyone can make changes to it. Let me put it in easier terms: If there are a large number of people using a software, and, that software happens to be open-source, it will generally lead to the betterment of the software as time goes on. Why? Well, it's because it is not at all imperative for me to change the code - in fact I know I'll pretty much never need to patch, say, Emacs - for the software to become better because I know that there are other people who are doing so - thereby leading to a better piece of code. Examples of this kind of behaviour is spread all over the place. The Linux kernel is a pretty good example. So is git. Now, because an editor is the primary tool of a programmer, you can be sure that a very large number of people are going to use it. If even a fraction of those people contribute to the editor's source, you can be pretty certain that the editor is evolving according to the needs of the community (and not according to the organisation owning the software) since that fraction of the people contributing code will represent pretty evenly the general demands of the community. So, the editor will become better over time, evolve in a way that is representative of the demands of the community and won't die out.

Now, I'm not saying that Atom isn't a good product. Indeed, from the looks of it, it appears to be a very well rounded text editor. But again, it's not open-source and hence, the benefits that I talked of in the last paragraph won't apply to Atom. This means that the project might die out in 4-5 years. Then, there's the fact that Emacs can do much much more than Atom right now and I expect Emacs will be around 20 years from now [1]. These facts lead me to the obvious conclusion that there is no incentive for me to switch to Atom. Had Atom been open-source, then, it being a pretty great editor already, I know that it would have been much more likely that Atom would still be around 10 years from now (Why? See the last paragraph!). Then, I might have made Atom a second (actually third, second is Vim) editor that I'm proficient in. But alas! That is not to be.

Anyway, I truly hope that you understand why it is that I (and many others like me) want the text editor to be open-sourced. It's not because I will make it better - it's because I know that there are others, far more competent than me, who will, and, do so according to the wishes of the community using the software.

[1] Extrapolating from the fact that it has been around for 20 years already.

> If it is a deal breaker for you go away from the deal. That simple.

OK if he talks about why, since this is a discussion site?

For one thing because Git is open source, and who would expect a profitable open source company to attempt to commercialize an html5 text editor?

GitHub itself is proprietary. Don't confuse the DVCS with the software hosting service that made it popular.

GitHub is based around open source software. Ruby on Rails, Git, various other tools in deployments. Working inside open source browsers mostly. Breathing the open source community on its site.

For me, GitHub is an open source company even if they don't open source all of their code...

I don't equate git with github, I just don't like bloating my posts with too much nitpicking.

A ton of companies are based around open source software. That's kind of the beauty of it. It has enabled the proliferation of businesses and has sped up innovation.

But by your logic, the NSA must be an open source government agency. After all, the Snowden leaks revealed that their technology (including XKeyscore) is built on open source software.

"Working inside open source browsers mostly."

Every website and web application is open source now? If you can find a way to arbitrarily lift server-side code from any web application, you'd be hailed as the next technological messiah.

No, GitHub using open source software does not make their platform open source. They do give back to an extent, sure. But that's hardly anything to celebrate by itself.

I did not say their platform is open source. I did say instead that the character of their main business puts them firmly in the open source space. Not even counting their own github account with open sourced software they created...

And Ruby on Rails was born inside the closed source project which remains closed source.

Git is not GitHub.

For a lot of Git and GitHub users, that distinction is of minimal importance.

How is expecting something to be open source an entitlement? I guess in your world view, it's an entitlement to be free and alive.

It's an entitlement because developing something like this requires an absurd amount of time and money. Why should they work for free. Do you work for free? Would you consider your boss entitled if he expected you to work for free?

What's my view of OSS? I'm thankful for it. Yes, it's free work - they're working (although doing it for pleasure) and giving it away as a gift. I guess that's the difference, I view it as a gift, others seem to think it's immoral to work for pay.

Edit: Please enlighten me, what exactly is immoral about charging for software that takes thousands of man hours to create?

Did they have to build git? Did they pay for it? What about ruby? No? Rails?

I don't get what point you're trying to make. Is it wrong to sell software that has some piece of open source in it? It's not like they cobbled a bunch of open source projects and placed a license check on it. Do you realize that virtually every application uses some open source library?

Funny, Rails was born as side product of closed source software.

So what's your view on Open Source/FOSS projects where people contribute and write code and share it for free without asking for retribution ? Do you consider that "free" work ? Don't you benefit from that anyway ?

Open source isn't about working for free. It's about pooling resources in the community to provide a baseline from which competition can begin. Open source improves the efficiency of the overall system, allowing innovators to focus on innovation, not basic functionality (like operating systems, text editors, web servers, web frameworks, programming languages, virtual machines, etc.)

Google sells advertisements, so they make a free browser that makes the web fast and secure. They improve search to make the web more useful for everyone. They help get third-world countries online. Why? Because when the internet grows, Google grows. They're not working on search "for free".

GitHub might have taken a similar view with Atom: by creating a freely-available, polished, advanced, and easy-to-use programming environment, they might help foster the growth of a whole new generation of hackers that would use GitHub to easily store and share their code, right there, baked into the editor. Universities might pick it up, and it might be featured in coding academy curricula. Student projects might involve writing plugins for it, extending it, and making it their own: often the first task of an artisan is to make and customize his or her tools of the trade. And GitHub would be there, at the center of it all. Rather that some advanced hacker tool shrouded in mystery, GitHub would be built-in to every new coder's first experiences with programming: an iconic brand synonymous with that eye-opening rush of seeing your code work for the first time.

That doesn't appear to be their plan, though.

Open source isn't about entitlement. It's about making core building blocks of software available to all because everyone shares an interest in those core pieces, and it's cheaper/faster/better to work together on shared code that doesn't contribute to market differentiation than mucking around trying to build an maintain a bunch of secret, closed-source code that everyone is going to need anyway. Open source frees up resources to solve more interesting problems.

Editing text isn't a particularly interesting problem: the best models for it were developed decades ago, and the main innovations since were fonts and color themes. Text editors are a prime candidate for open sourcing, and most good ones are, including huge IDEs like Eclipse, Netbeans and IntelliJ. To say nothing of Emacs, Vi(m), nano, mg, TextMate, LightTable, Brackets, Notepad++, Kate, GEdit and dozens of other sophisticated, open-source text editors.

Expecting GitHub (of all companies) to err on the side of closing up the source for a text editor (of all products) seems strange.

Don't they know that if you want to open source something but don't want any contributors you're supposed to use GPL v3 with a CLA?

As I recall, Github was making money from day one, and has been profitable for nearly as long. I'm sure that fact has contributed significantly to the quality of their products thus far.

The business model around a desktop application is bound to be different from the web services, and probably isn't even finalized yet. I'm just looking forward to trying it out to see if I like the editor or not. That bar is high enough, given that I'm pretty comfortable with my workflow as it is. To expect that something this new will be polished enough to make me alter my day-to-day process, but do so without any viable business model, seems rather unlikely.

A free/open version would be great, but I'll feel better about adopting it if the project is clearly profitable.

This is a little off topic, but I _really_ like the discuss.atom.io forum platform. Snappy, intuitive, clean. There are a few visual subtleties that would make reading large blocks of text difficult, but seem well suited to the short lengths of most comments (text is #333333, lines are spaced extra far apart).

Looks like Jeff Atwood is still creating great products. I even felt a nostalgic whiff of his Microsoft infused sense of style (the <code> font-family is "Consolas, Menlo, Monaco, "Lucida Console", "Liberation Mono", "DejaVu Sans Mono", "Bitstream Vera Sans Mono", "Courier New", monospace, serif", in that order)

So it can never be forked?

The one GitHub project that can't say "fork me on GitHub".

As someone mentioned on the atom forum, all public github projects need to be "forkable" as per GitHub's ToS.

Expected that they need to make money with it. But then, how is it different from cloud9 ? (http://c9.io) I thought they would try to have it open source (as it's non-core to their business) and use it as a trojan horse to sell more private repos.

But they could do that with free to use I guess. And it makes sense, then no one is allowed to removed the github-as-a-backend part, and people have to get the paid plans to develop closed-source software.

It's nothing like c9.io. Atom is a desktop app and relies on Node's FS modules...

It is. But my guess is that it will work on both desktop and web. We'll see.

I am not sure about why so many people are surprised. Although GitHub promotes open source, it's not open source itself. And it's fine. At the end of the day, GitHub is a company. They need money to do the great things it does.

While Open Source is great and it's contribution to the world is amazing, we shouldn't forget that it's a small niche thing, if we are talking about % of programmers contributing.

> I am not sure about why so many people are surprised.

Because the whole premise of this editor is that it's "hackable". It isn't truly hackable if it isn't truly open source.

I interpreted "hackable" in this context to mean "personally customizable". Even if it's closed-source, that doesn't preclude an extension/library ecosystem from forming around the core Atom app, does it?

Except we would all be stupid to invest a lot of effort into said ecosystem when Github can pull the rug out from under it at any time, and Github will be the only one monetizing it.

Those of us who've been around longer have seen this play out many times. Free software isn't just dogma, it's very practical self-defense.

"They need money." Who doesn't? But it does not immediately follow that a software has to be restrictively licensed. It has to be part of an overall strategy. Most Open Source contributions actually come from companies pursuing a money making strategy. In the case of GitHub very successful even!

Equating open source to "not for profit" is just a misunderstanding of both business and open source. One of the main reasons for businesses contributing open source is a solution for the problem of the necessity of in-house development and the un-necessity of an unprofitable/risky in-house software marketing effort with all the overhead that entails.

It sounds like they might be planning to charge for the software, include (limited?) access to the source with purchase, and put a no-derivative-works-or-redistribution clause in the license.

Actual FOSS would be nice, but non-free software + source still beats regular closed-source proprietary software.

Interesting to hear. Not sure how this half-open/half-closed source will work. Something like Red Hat? Then what about a CentOS equivalent for Atom?

This is a lot more like Android, where the code itself is open but the core services which make it useful are hosted and not open.

Nah, Android is the other way around. The core Android platform (the AOSP) is free software, but the Google services are proprietary.

This looks like they're going to allow people to view the source code, but not actually use it. My impression is that it would be similar to Assembla's not-really-open-source license [1]

[1] https://www.assembla.com/wiki/show/breakoutdocs/Singleton_De...

Sounds like providing source code under a proprietary license. That's not Red Hat's model.

No matter what they say, no matter what you say, and no matter what reason itself says, it is eminently comedic that GitHub!!! would do a project that isn't plain open source.

As to why Atom, I'd bet this was something one or two senior guys just had an itch about, and they sold the project.

Really? The main source of revenue (GitHub.com) isn't open-source either, companies are free to promote open-source whilst being entirely closed and proprietary, GitHub releases almost everything freely, and encourages the same – there's no hypocrisy that I can see.

There are some reasons that not everything is open-source here: http://tom.preston-werner.com/2011/11/22/open-source-everyth...

Making it free and open source, with great GitHub support as well, would also be a good strategy :)

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact