Open Source Growing At an Exponential Rate

sipmeister writes "Two computer scientists who work for enterprise software giant SAP have shown that open source is growing at an exponential rate. Not only is the code base growing exponentially, but also the number of viable projects. Researchers Amit Deshpande and Dirk Riehle analyzed the database of open source startup ohloh.net and looked at the last 16 years of growth in open source. They consistently got the best fit for the data using an exponential model. Relating this to open source market revenue, Desphande and Riehle conclude that open source is eating into closed source at a non-trivial pace."

Comments

Re:Viral License?

[Anonymous Coward]

The GPL concerns free software, no open source software. This study has no bearing on how 'viral' the GPL is.

The code base is growing

[Soleen]

This is sad that code base of Open Source projects is growing exponentially. Projects become fat ugly and unmanageable. It is also getting harder to debug, port, and even use such programs. http://suckless.org/ has several programs that do their job every well and yet very managable. For example window manager: dwm less than 2K lines of code, is the most feature complete WM I've seen. I've been using it as my main window manager for over year, and was very happy with it. There are few good CLI applications availble that hold approach of been efficient and useful and almost no GUI applications.

Re:Extra Extra

[argent]

Yah, I'd like to see them do the same curve fit for commercial software. If it's not also exponential I'll eat your hat.

(not mine, it's icky)

What is growing?

[Anonymous Brave Guy]

“Measuring programming progress by lines of code is like measuring aircraft building progress by weight.” — Bill Gates

The rest of us got over this particular naive metric years ago. The fact that lines of OSS code produced are growing exponentially doesn't tell us anything useful about how much useful stuff can now be done with OSS.

Moreover, the rate of growth now is not the interesting thing. The total volume of serious OSS is still relatively small, and so is its growth in absolute terms. The future potential is far more interesting to explore.

For example, if (as TFA tells us) packaged OSS generated revenues of $1.8B in 2006 and this was around 0.7% of total revenue generated from all packaged software sales, then I disagree with the article's claim that the OSS revenue was not trivial compared to the market as a whole. In business terms, 0.7% market share is nothing. On the other hand, if you also say that the OSS revenue is doubling every year while the total remains roughly constant, and you have evidence that this will continue giving exponential growth, then your data suggests that in a few years the OSS revenue very much will be significant.

However, I'm struggling to find data to support those claims on a first quick look at TFA. The pretty pictures just show that the volume of code is going up, which doesn't tell us anything about the value (economic or practical) of what's being written, nor what the future trends for that value are likely to be.

Re:What is growing?

[epine]

The pretty pictures just show that the volume of code is going up, which doesn't tell us anything about the value (economic or practical) of what's being written, nor what the future trends for that value are likely to be. Neither is the GDP a particularly good measure of economic progress, since the figure is quite happy to add a mess to the cost of cleaning up the mess and then tell you that you are quite wealthy.

LOC has the same problem: it will add lines of code creating a bug to lines of code working around the bug.

The purchase of an SUV adds to the GDP more than a less expensive vehicle. The SUV adds yet more to the GDP when it burns more gas to travel the same distance. If that SUV rolls over on the highway two years after purchase and causes one of the occupants to collect $1m in heath benefits through insurance, the GDP rockets upward yet again. GDP has an extremely dim relationship to the *value* of the activity it measures. What you can infer is that the society is wealthy enough that people (some people) actually *have* million dollar health insurance packages, and there is a medical establishment capable of delivering that service.

But the same is true of LOC: you can successfully infer from a project having 1m LOC that the project probably has more than a single core contributor.

In fact, prior to the sub-prime collapse, American economic health metrics were an orgy of double counting. Five to ten years from now the press will be writing stories about how the market has *returned* to the level of 2006, having conveniently forgotten that the numbers from 2006 were fictitious to the point of fraudulence.

If you are taking one figure more seriously than the other, just because one is denominated in dollars (and hence more "real"), you aren't thinking clearly.

Goolsbee remarks late 2006:

The one true dark spot on the US picture is our totally unsustainable fiscal position.

First, you should disregard the official numbers because of the accounting. Some time ago the government got tired of people seeing how much they were actually promising to spend so they switched to cash accounting. Nothing counts as a cost until it is actually spent. So the social security system is bringing in tons of money which it was supposed to use for your retirement. But they don't have to actually start spending money on your retirement for a few years sooooo, they can count the cash coming in as revenue and not count the what they will owe as expenditure. They are almost literally charging money to a credit card and calling it income.

Now if you take the total value of what we are on the hook to pay and the amount that we will raise in taxes, do I really need to tell you that they don't add up? Boy do they not add up. The latest numbers indicate that the net present value over the next 75 years is almost $70 trillion. According to some budget experts, by the standard of a business, the nation is bankrupt. But hey, it's only money.

Re:What is growing?

[EWIPlayer]

Has anyone thought to ask what closed source is doing? Is it growing exponentially as well? Is it growing faster? Without something to compare it to, saying that OSS is growing exponentially is about as significant as saying it's growing linearly and all the best programmers have long hair.

so is my bank account

[nguy]

My bank account is also growing exponentially, at 1% interest. That doesn't make me rich any time soon.

Exppnential growth is a meaningless property since many things grow exponentially, many of them quite slowly. What matters is the growth rate and any upper limits to growth.

Re:Competition

[AKAImBatman]

Sorry, I must have missed that memo. How many major name web services actually make money today?

A lot more than you think, apparently. My last two employers have provided services over the web in the Financial and Health Care industries. They're both rather well-off from that business alone.

A more visible example would be news and blog sites. Quite a few of them make a killing off of advertisements. Their profit models are more difficult to maintain than direct service costs, I'll grant you, but many do well for themselves in spite of the challenges facing them.

On another note, I did just occur to me that I may have caused some confusion by using the term "web services". A lot of people think "SOAP" when they hear that term. While I do know a company or two who charges for access to their SOAP interface (basically, a really fancy remote database interface), I was referring primarily to the delivery of business services over the web. My apologies for any confusion. :-)

Re:I for one

[francium de neobie]

Why waste time welcoming us when you can contribute and become an overlord yourself? ;)

Re:Signal to Noise Ratio

[pembo13]

Please explain how this is not true for non open-source software. I don't see failed projects/programs being at all unique to open-source. At most, it is just more visible.

Re:Viral License?

[howlingmadhowie]

i find complaints about the GPL being viral somewhat amusing, seeing as it is invariably closed-source software which is viral and forces everybody else to buy it if they want to interact with it. the GPL however produces free software which everybody can interact with as they wish.

Re:Viral License?

[IBBoard]

Only if your virus is of the "make you freer, healthier and happier" type.

Re:Viral License?

[jrumney]

You can make up hypothetical situations as much as you like, the fact is that if a library is useful, and there has been a GPLed library available for years, then someone somewhere will be selling a commercial library that does the same thing, which you can use in your proprietary project. Even if that were not true, there is no sense in crying about the fact that you can't profit from other people's software without giving something back.

Re:Viral License?

[Bert64]

If this code you want to reuse is GPL, then the author clearly didn't want you packaging up his code into a closed source game and selling it...
You still have the choice of releasing it as GPL and still selling it, most games players won't go to the trouble of downloading and compiling the source themselves.

And how is this worse than proprietary software? I doubt any closed source vendor would allow you to package up their code as part of your product either...

Re:Viral License?

[aproposofwhat]

Poor little student - what better way to learn than to reimplement the algorithms?

If you want to make money, put the work in!

What this study highlights to me is that despite the protestations of the patent / copyright lobbies, free software promotes innovation, rather than the profit motive.

As I say - if you want to charge, then put the work in - it's not as though it's hard.

Re:I for one

[hey!]

And poetically, "Operation Overlord" was the code name for the allied invasion of Normandy, which signaled the beginning of the end of the European war.

Of course, we are much more at what Churchill would have termed the "end of the beginning" stage when it comes to free software, and in that spirit I offer a Churchill quotation that is rather apt:

This is no war of chieftains or of princes, of dynasties or national ambition; it is a war of peoples and of causes. There are vast numbers, not only in this island but in every land, who will render faithful service in this war but whose names will never be known, whose deeds will never be recorded. This is a war of the Unknown Warriors; but let all strive without failing in faith or in duty, and the dark curse of Hitler will be lifted from our age."

Of course, it's not precisely true that "their deeds will never be recorded", at least if they are using source control as they should.

Re:Is it really a good thing?

[RAMMS EIN]

Open source is growing. Twenty years ago, virtually nobody had heard about it. Ten years ago, all the cool kids were talking about it, but few people were using it, and companies usually used it in secret. Now, open source is virtually everywhere. All the major players in the computer world are using it and advertising that they are. It's in many home routers. Most organizations I have been to in the past years have at least a Linux box somewhere. The one company I've visited that didn't was a Microsoft shop that developed using the latest Microsoft tools and a bunch of open source libraries. Few people know what open source is, but more and more people have interacted with it in some way.

Linux, Apache, Firefox? The number of people using those is enormous. Perl, PHP, and MySQL are huge, too. And now Java is going open source, which means that a huge part of commercial software development will be done using open source (to the extent that this wasn't true already; think JBoss, Ant, et al.)

Last, but not least, open source is on the desktop. And I don't just mean the odd geek who runs Linux on his desktop. I've already mentioned Firefox, but let's not forget that everybody who uses a Mac uses open source.

Really, open source is all around us.

Subtle critical flaw in your logic there ...

[Zero__Kelvin]

"You still have the choice of releasing it as GPL and still selling it, most games players won't go to the trouble of downloading and compiling the source themselves."

Actually, only one person needs to download and compile the source (for each platform.) He can then post the resultant binary for others to download and run.