The Javascript Trap
gnu.org
gnu.org
And as for the commercial viability of open source software, consider 37signals. Their entire tool set (Rails) is open source, and they make millions. Keeping Rails closed source was not necessary for their success. When you pay for a 37signals app you pay for their reputation and quality. If someone else made the same application a year earlier, it wouldn't sell nearly as well. Saying that you need to keep your software closed source is overvaluing the software. That's only a small piece of what makes a successful product. Hell, 37signals probably could open source Basecamp today and they would still make millions.
Why? RMS's movement would benefit most from convincing us, as developers, to contribute to FOSS. I pray that being a foolish-looking extremist isn't the best way to do that.
Based on the growing size and productivity of the FOSS community, I really don't think RMS needs to change what he is doing. Programmers have already produced billions of dollars of free software and will continue to do so. It's an amazing display of our (the hacker) culture.
Where this becomes useless is in his demonizing proprietary software, because even though proprietary software takes many of your rights away, it has some great benefits. Since it doesn't have to be free as in beer, a person can quit their day job and focus solely on writing that program, leading to better designed and better implemented software. It also creates a stable brand that people can evaluate, and clear responsibility for issues. Without a means of monitizing software, we probably wouldn't have the fit and polish systems that we do.
I love free software, but there are two RMSisms that bother me: 1) The GPL is actually less "free" than a BSD style license because of his campaign to promote free software. 2) His view is entirely supply side (except for the roundabout means of creating personal wealth by forcing others to use the GPL on their software) to the point where it seems to me like free software has to fall out of trees or large corporations (which I would rather due without).
RMS: I love the idea, but I want to see more!
Neither does FOSS software, you realize. Let's say Google charged money for Google Earth. Now, if they open-sourced the client, they would still make money from it, because they'd still be the only one with the license to the map tiles themselves. There are many situations like this, where the software is actually worthless without some resource (perhaps just your brand) that you provide alongside it. You can make plenty of money selling water (a FOSS product)--but remember, it's the bottles that people are buying; the water is obviously worthless when you just pour all over them (or release the source to them.)
> The GPL is actually less "free" than a BSD style license because of his campaign to promote free software.
Here's a thought-experiment that I think shows you RMS's world-view. Pretend, for a second, that AI existed. The GPL would give it rights (it would be "free"), while the BSD license would allow it to be sold into slavery ("non-free.")
Conclusion: The BSD license gives the freedom to to be sold. (Really-free).
The GPL guarantees that I can obtain the sources of your modifications, making it more free for me than the BSD license.
So rare that this debate actually gets to the real point, which is that the BSD license and the GPL optimized for two different "freedoms". Ne'er the twain shall meet.
Which is larger, a company of 100 selling $500 million a year of goods, or a company of 1000 selling $50 million a year of goods? Well, it depends on your metric. Two perfectly valid metrics produce two different results.
You don't know the guy, you don't have any thing on him. He doesn't force anybody to use his licenses. He is just warning users of a proprietary software trap, that's all. If you happen to serve this kind of website it's stupid to just get into the name calling and instead perhaps state clearly your JavaScript isn't free software. Add a license, you need it anyway to cover yourself.
What average user of a webapp even understands what a software license is? Why would they care either way if they can see comments in the javascript or not??
Obviously you need to make clear if your javascript is open source or not, and both should be respected decisions.
I think the majority just care if something solves their problem, rather than if its open source or not. Linux doesn't beat windows because it's open source, it beats it because it works far better.
One of the reasons it works better just happens to be because it's open source, which enables far more people to contribute than otherwise etc etc.
Are people seriously going to stop using webapps that don't provide their source code unobfuscated? Of course not.
RMS isn't being extremist in this case at all. Your only two options are to fully enable or disable JavaScript. That's not much of a choice because if you enable JavaScript you let in some company's proprietary code to be run on your machine. Imagine you're coding a Free Software Javascript library and the proprietary competitor's product is run on your machine. They can prove that you've run their code, and maybe you looked at it and now you're setup for copyright infringement.
If you fully disable Javascript with NoScript, then you won't be able to run "safe" Javascript.
Two choices that are both extreme. The real extremism is in the web browser that doesn't allow you to create a blacklist/whitelist or to use replacement JavaScript.
Sure, most people using GMail or Google docs are programming open source javascript libraries on the side..... Very common I'm sure :/
If you're programming an open source js library and you're worried about a webapp suing you, don't use it?
His overriding concern as an ongoing basis is that people should have the freedom to tinker with software - to be able to read, improve, share the programs we are running.
He's not arguing that you're falling into a personal, small scale 'trap' like you will get your finger stuck or something, but that whole groups of people who care may not have noticed this set of events sneaking past them.
Or that this is yet another area where we should care, but don't, and that should change so that if you do care, you can do something about it.
No end users care.
That maybe true, but it shouldn't be true.
I think the majority just care if something solves their problem, rather than if its open source or not.
Exactly the problem - this week I saw an end user called us up asking us to help get some information out of a program they run. Luckily, it runs on a popular database backend. Unluckily, it's proprietary closed source software and there is no documentation of what the database tables and fields mean or how they are used.
They bought it because it "solved their problem". Now that lead to other problems. Problems which are both completely predictable (information stuck in artificially limited proprietary system) and avoidable (use artificial-limitation-free documented open system).
And on a more general level, promote the use of such across the entire computing industry so such programs become available.
Anyway, just give the man some respect, he is somewhat an extremist, but if it weren't for this guy you wouldn't be starting all these startups on shoestring budgets and I know that everyone in their heart of hearts knows that the world is a better place because of RMS.
We'd be using OS X and FreeBSD instead of Linux. Who's laughing now?
I don't think you can really argue that at all. If he hadn't done some useful stuff, someone else would have.
Now imagine for a second if everybody thought this way and left it to somebody else...
I say, if it is in your power to have a positive impact, then it becomes your responsibility.
No, several people happened to invent them at pretty much the same time. It's just luck as to which got the credit for 'inventing'.
Sure, if no one innovated, then we wouldn't progress, but that's not what I was saying.
I was just saying that the OP "if it weren't for this guy you wouldn't be starting all these startups on shoestring budgets" was wrong.
That said, I think he's totally out of his league on the software as a service model. At the very least he needs to spend a lot more time mulling it over. As far as I see it, freeness of javascript more or less meaningless in a web application. The fact is that the barriers to making web applications free are much higher, and there are a lot more legitimate arguments as to why it shouldn't be free. Of course Stallman would argue for freedom under all circumstances, and he's right that web applications do present a risk to freedom of software in the long run, but I don't think it's as easy to argue from his idealogical standpoint when it comes to web apps. Too much baby flying out with the bathwater.
But you'll keep running them on those servers powered by free software running on a free operating system stating how open solutions are better, wont you? How is that for bigotry? It's like you are deliberately trying to stir up a troll-storm.
But please: Do give me some insights on why making web-based applications free poses any more challenges than freakin' kernel-code.
What good is free javascript in a web application where a significant portion of the code lives on the server?
The fundamental issue of software freedom is trivial when you are running software locally. I'm running this software locally, and I want to be free to do with it as I will.
However web application software has the local component (JS) and the server component. The software that is running on the server is clearly not being run by the end user and therefore the notion of freedom is not so clear. Sure, we can make it "free" by releasing the source of the server code, but setting up a server may be non-trivial. Or it may be useless without the data from the service. Maybe the user's data could be released, or maybe it's intertwined with other user's data.
The point is, I care very much about freedom of software in general, but having free javascript by itself is essentially useless to me in the vast majority of cases.
There are a lot of free javascript libraries out there, and quite a few which doesn't directly involve server-callbacks, and leaves you with the power to hook things together as you please. In that case, free javascript libraries still makes a lot of sense.
For situations where the bulk of the solution is actually server-side code, I wouldn't really call that a "javascript library", but some regular server-side solution with added javascript/AJAX functionality. And in those cases I can be willing to agree with your statement.
In my opinion, all software should be free, when all groceries are free.
I assume you didn't mean free in some other way, right?
If I have to release my code under GPL, I can't make money charging thousands for it (low volume type of app). If I can't make money from it, I won't spend years developing it. Not having the app would make it harder for some pharma scientist somewhere to cure cancer. Everyone loses. Communism doesn't work.
However if you choose to release under the GPL you will be standing on the shoulders of giants. You will have whole new avenues of development opportunity opened up to you as a result.
"Communism" does work in the sense of the Linux ecosystem that you see around you every day. Not only that, but companies like Microsoft are forced to innovate because open source eventually catched up to them. Without free software all software would be tied up in a morass of patents and licensing fees. Progress would have ground to a halt, and we would wait with baited breath to pay $500 for the latest version of Windows crafted upon their latest business model to coerce upgrades from all their regular customers.
If you are adding 1000 man hours on top of 50,000 man hours of development that you are getting for free, is it moral for you to give nothing back? The GPL is simply a way for the contributors of the 50,000 hours to say that you can use their code if you share in return. It's not clear at all why you think you should be entitled to that code.
I never said I was. In the GNU utopia, Linux killed MS and all libraries are GPL.
My options are:
1. Spend years writing apps that help cure cancer, GPL them and starve. (not exactly hypothetical)
2. Develop some Web apps that are not bound by GPL. Make money.
What do you think I will choose in your GNU communist utopia?
Legally, sure (modulo plant patents). Practically, you can only do it as long as you don't mind it being a crab apple tree. (Apples tend not to breed true; if you want a tree that is sure to produce edible apples you have to get it by grafting.)
So, the idea of a "free apple" is a pipe dream.
Not cost-nothing, but without-legal-or-other-arbitrarily-imposed-limitation-on-use.
FOSS can't compete in this space because they can't afford the servers - you have to have revenue to make it in SAAS.
I think RMS's positions are extreme. I don't share most of his views, including his latest concern about javascript. But I admire the man for knowing his own mind and speaking for his beliefs directly and honestly. You don't get mealy-mouthed bullshit from RMS. No evasions, no fog-of-words, no equivocations. I love him for that.
Also, he is more right than I once thought. When I first read "Right To Read" I thought it was ridiculous, extremist, absurd. Well, after seeing what "they" have tried to do with DRM, I was wrong. I think we are still a long way from that dystopia, and the momentum (right now at least) is in the right direction. But he saw something I didn't, and I try not to forget it when reading other things of his I don't agree with.
However, I'm not sure that it wouldn't be easily done if there was a demand for it. Most web frameworks come with easy ways to run the environment locally, so you should be able to release a RoR app open source without many problems. It just doesn't seem to be in high demand.
GreaseMonkey can do that in Chromium [http://dev.chromium.org/developers/design-documents/user-scr...]. Stallman mentions Greasefire later on, but it sounds like he might not clearly understand that Greasefire finds applicable GreaseMonkey scripts, while GreaseMonkey enables user scripting. He may not realize it, but I think the infrastructure is already mostly available for building the detect-and-substitute feature he wants.
what, I think, Stallman envisions is being able to use your local copy of Gmail instead of the one Google's serving you without any loss of functionality.
Chromium has this "early injection" feature for GreaseMonkey scripts, so that's one free browser for which it would be very easy to build Stallman's replacement feature. See the link I gave earlier in this thread and also http://www.mail-archive.com/chromium-discuss@googlegroups.co... for some background.
You have to know how many lines from the header file you are "distributing" (in the GNU lingo). If you are using C++ templates, you are probably screwed. You have to make sure you link dynamically (or else). LGPL seems to require advertising the library. And if you want do static linking, talk to a lawyer. Linus and RMS disagree on what constitutes "derived work", btw.
Edit: I'd like to know what motivated people to upmod parent, when it's factually incorrect and can cost developers dearly if they take the message at face value.
* a) Give prominent notice with each copy of the object code that the Library is
used in it and that the Library and its use are covered by this License.
* b) Accompany the object code with a copy of the GNU GPL and this license document.
I understand that you want to be able to charge money for the software you write, but that you seem to perceive the LGPL as a threat to this ability seems excessively fear-mongery. I agree that anyone concerned about licensing terms should consult an attorney, but because you want to understand things better, not because you are fearful of software development turning socialist.
2. I'd assume as lines are normally counted. The number of line-feeds and/or carriage returns in a file. I don't see what the big deal is anyway; if you are at all concerned, just include the license text. Is doing so that big a deal?
Ten lines per function, per header, per library, per program?
Edit: I think you didn't understand LGPL yourself either. I think in the quoted lines, "object code" refers to your code, and accompanying them with a copy of GPL refers to having to release your code under GPL. Or do you think you only have to say you release it under GPL and not actually release it?
If you have just your object code, you are not bound by the license of the library, then you don't need to include it.
If you have just the library header code, then you are not incorporating it into anything else, and have no obligations by the license.
If you have both the library header code incorporated into your object code , then you are bound by the license to provide attribution along with that incorporated object code.
As for you second question, I'd guess lines per file. But it is a bit ambiguous, as almost all legalese is. Again, you could just err toward caution, and include the license text regardless. Are you averse to this for some reason?
Oh, and as per the wiki link, object code is the compiled binary of your source code. At this point, whatever headers you included would already be integrated.
The fact that your understanding of the license is different shows just how sneaky and obfuscated LGPL is.
> Lesser GPL permits use of the library in proprietary programs
That means "in some cases", not "in all cases". Logical fallacy on your part.
> ordinary GPL for a library makes it available only for free programs
Here, GPL'ers contradict themselves (for both senses of "free"). GPL code can, technically, be non-gratis. And GPL code can be used in non-GPL programs (you just can't distribute them easily).
Those terms to which you object mean, I think, the following: If your object code includes bits of the library header files, and if those bits are substantial, then you have to (1) say that you're using the library, (2) say what licence the library is covered by, and (3) include a copy of the documents that define that licence.
We have the right to improve GPL'd programs with horrible documentation.
The implementation of JavaScript is most likely open source, and the script's code is obviously available. Am I missing something? Is he saying that it's not "Open Source" if it doesn't have enough comments?
Trying to prevent people from releasing source code without comments because it's not easily hackable is a bit much. Almost everything about RMS is "a bit much" though, and that's why so many hackers like him ;)
The suggestion in the article that you put a license delimiter and then parse the text of the license seems error prone; not because the license is particular difficult text to parse, but because text parsing in general is a rather fussy and fragile enterprise. And someone could put after the licensing tag "Just kidding, here's the real license..." which the parser wouldn't catch, but someone litigating a case of copyright infringement would.
Maybe I'm not being imaginative enough, but I can't think of a way this is feasible with just the client attempting to glean licensing on its own.
Having a global registry of free JavaScripts wouldn't solve the problem, either. Assuring correct URL parsing in every case is a risky thing to rely on for such restrictive behavior, and nothing stops the web site owner from renaming the free script to something else, or renaming one of his non-free scripts to the same name as the free script. Verifying file hashes wouldn't really help, since the free script is allowed to be modified.
Regardless of the political ramifications, the feasibility of this seems really questionable. Are there ways of doing this I am not realizing?
His position is, in my mind, completely untenable. He sees everything through his 'free software' glasses, the same way a dyed-in-the-wool communist sees everything through his Breznospecs. It's not enough for him that a Google Docs javascript file be free (gratis) – he objects that the file is not easy to read and has no comments. The fact that removing unnecessary white space is a commonly-used practice in speeding-up page loading is not mentioned, even in passing. What is this guy on?
I have in mind a special device for Stallman, Raymond and the other gratis software nutters. It consists of a big plastic bag that envelops their heads and into which is pumped the aroma of roasted coffee beans. It is required that they use this apparatus until such a time as their eyes light up and they utter the required phrase.
Today, more than ever, we live in a world of economic reality. In the recent past it was easy to find some chump to lend you all the money required to buy that house/car/tv you really couldn't afford, or finance your startup based on a 10 page deck and the words “web 2.0”. But things have changed. And perhaps for the better.
The open software movement cares more about its users than it does its developers. Hacker kudos doesn't pay the fucking rent. I want to know whether all the people writing iPhone or Facebook apps would have tried so hard if they knew their expected return would be exactly $0? Somehow I don't think so.
We've got to move away from this notion that software should be free (gratis). If you use it and can afford to pay for it then why should I give it to you for free? Note: I'm not arguing against open-source, I'm arguing against working for free. If I spend a great deal of time writing a funky database application and big_multi_national dumps Oracle in favor of my code and saves itself $20million in the process, why shouldn't I get my sniff? There's nothing wrong with making an honest dime...
If the same code is used by some charity or someone's personal website then I'm happy to tip my hat and say 'glad to be of service', but not if they're Coca Cola, or Hertz, or... (you name it).
If there existed a fair and balanced way of rewarding open source developers then I believe the whole sector would explode with a level of commitment and energy that would dwarf the already impressive achievements seen by the community.
That time has come.
It's time for revolution's founders to retire gracefully, and for a more realistic (though still ethical) guard to take its place.
David Semeria Milan, Italy
You can still provide a link to the full, unobfuscated source code. That isn't difficult to do. Just put the link in an HTML comment where the source is loaded, or at the top of the source file.
You're another one of those so-called pragmatists who only thinks about the short-term, ja?
Just think, 20+ years ago there wasn't very much free software (nor very much open source software). Now look how much there is! It is now expected that developers release the source code of their projects. True that they are mostly side-projects but there are many serious businesses that are built using Free Software.
>It's time for revolution's founders to retire gracefully, and for a more realistic (though still ethical) guard to take its place.
That was tried a number of years ago with the Open Source movement. Their philosophy is to never really talk about freedom, just talk about the benefits, mainly economic. Has it succeeded? Yes, but only as a marketing slogan, with companies co-opting it and the consumer never really knowing exactly what it means for software to be free.
>We've got to move away from this notion that software should be free (gratis).
You're allowed to charge money for any Free Software. The only catch is that you can't keep the source code closed or forbid the user from re-distributing it. Which basically means that when you charge money for Free Software, you're a redistribution fee, not a license-to-use fee.
Politician speak alert!
If there existed a fair and balanced way of rewarding open source developer
Buy their software. And with it, get the source code and the freedom to do whatever you like with it.
Just as a dyed-in-the-wool capitalist sees everything through their Smith-o-specs.
We have always lived in a world of complete economic reality, by obvious tautology. It is that economic reality that gave rise to the GPL and the Free Software Movement. These were market instruments made to route around other market instruments perceived as damaging or inefficient. That there is so much under the GPL right now suggests that there was real market damage or inefficiency being routed around. Just because there isn't money involved in this doesn't mean there isn't a market involved. This market just happens to trade in something other than money; you seem to see something wrong with that.
I think I see why, as well. These people, who will work on software for free -- for hacker kudos -- are perceived as a threat to those who would work on that same software for money. They are not forced, they chose to, but it still means market forces threaten to drive the salary of a developer down. You think this will drive hackers into economic poverty. If anything, it has had the opposite effect. The software people release for free earns them a reputation that allows them to later release other software for a very handsome profit.
The open software movement cares more about its users than it does its developers.
The movement is very little apart from the developers, so this reduces down to the developers caring more about the users than they do about themselves. Are they being too altruistic? I don't know. If they are, they'll soon find themselves being less altruistic out of need. If not, I don't see why we need to cast aspersions their way. They choose to work this way.
As for threatening the livelihood of programming work-for-pay, it isn't. There are still problems out there that you would need to pay people to work on, partly because they haven't forgotten about their rent in all of this, and partly because there just isn't that intrinsic motivation for some of the software people need written.
The problem with that belief is that every piece of free software makes it possible to build more valuable new things on top of it for cheaper. Software is additive in a way that no other prior asset has been.
As such, the value of free software is tremendous, almost incalculable. At every step somebody bemoans that the GPL is preventing them from grabbing free software and turning it into an incredibly profitable proprietary product. They quickly point to the lost potential revenue, and completely ignore the value that would be lost by the free software being supplanted by a proprietary product.
What if Red Hat had forked Linux to a proprietary model 10 years ago and done enough to become the standard. Do you think they could have delivered as much as the entire community has done since then?
What I don't love is the notion that either all software should be free (gratis) à la Stallman, or the less extreme view that you can make money from it as long as you do so in a very inefficient and non-scalable manner (basically charging for support).
Also, I find it quite ironic that some people describe (with a straight face) the supply of proprietary extensions to open source code as an 'open source business model'. That's akin to funding a peace march with the proceeds from arms sales.
And whilst I have lots of respect for what the community has achieved, and the people who helped make it happen (including Stallman and Raymond) I would argue that there is no inherent conflict between open source ideals and the notion that some users should pay for the code. If a developer wants to give his code away for free, then that's his or her choice and I respect that. But if another developer decided that his code is not free for commercial use, then I respect that position as well.
I hold the view that a low-friction and equitable licensing system for open source software would have a further galvanizing effect on this already vibrant sector.
DS
(Also, note that the guy's opinions and world view have been getting more and more popular.)
As leader of our national Ukrainian Linux Users Group, I involved in fixing of various problems with support for Ukrainain language in foreign programs.
Here is my experience: MS: some problems still around for more than 20 years (from MS DOS era till today); SUN: almost 10 years to fix trivial problem; Apple: good support; IBM: good support of official standards, not so good support for net standards; Google: some problems are still around for years, nobody response to my letters, no way to fix. Linux/*BSD/GNU: very easy to fix (or create) problem in any supported program, much harder when program in no longer in development (need to maintain our own version of program in our own distribution).
All these problems are trivial (charset support, proper charset translation, national keyboard support, message translations, etc.), so they are very easy to fix. Currently, then only OS'es that have no problems with Ukrainian language is GNU/Linux. Even Non-GNU/Linux systems are easy to fix in most cases.
I agree with Stallman: web services are closed, so they introduce high risk for our freedom.
In Stallman's dream world, you would be free to charge for your service, or put ads up, or whatever you like. But your service would use standardized file formats, and the user would be able to modify the application (both server-side and client-side) to suit his needs.
I have zero experience with Firefox extension development, but everything I _have_ seen about it points at it being very difficult to mirror in a closed source environment.
- firefox - open, free, gets money from google
- openmoko - open mobile, you can even download their pcb design and CAD files, but they do sell the product
- specialised apps - company I work for releases some code as open-source, but that doesn't mean that our users have resources to provide the same service themselves, so we don't lose anything
edit: Sure, downvote me if you like. If you like to think adblock etc actually block all adverts I guess it's up to you. It blocks a small class of 'obvious adverts'. Several websites will pass outgoing links through a jump script which may or may not then go through affiliate links. Adblock is useless against such things. There are several such examples.
I don't do it to stop site owners 'making' money from me. If you can get non-annoying ads to me, and make money from me, that's brilliant. If that's all anyone did, I'd get rid of *block altogether.
Although I think what Stallman would probably prefer is the source to the site / web application being available so that people can run their own instances.
HN and reddit have their source available. I don't see anyone using the reddit code in a way that's any threat to reddit, mainly since the size of the community is an asset that's hard to compete with.
Potentially off-topic, since I'm not familiar with the prevalent licensing, but the various django applications look pretty nice for this kind of thing.
Actually, no. He is for totally abrogating copyright on works which have "practical use" and for maintaining certain forms of exclusive economic control on entertainment and advocacy.
Essentially, he thinks he should be able to redistribute my software for free, for any purpose whatsoever, but that I should not be able to commercially exploit his speeches or modify them, because that "misrepresents" him.
See:
http://en.wikisource.org/wiki/Free_Software_and_Beyond:_Huma...
But let's turn to the next category, works that state the views of certain parties. Now here my answer is different, I don't think modified versions of these works contribute to society, all they do is misrepresent the authors. So I propose a compromised copyright system which says that everybody is free non-commercially to redistribute exact copies. But modifications require permission and commercial use require permission. So this compromised copyright system would provide revenues more or less as the current system does.
ad hominem.
> I should allow you to access my servers and services for > free?
You can charge for Free software. If you're using any Free software in your web application stack, if you're using any Free documentation, and Free data sources, then isn't it a little hypocritical of you to lock-down what basically amounts to a component assembly of Free software?
> I should allow you to control how you use my resources in > a way that guarantees I loose money?
Make all the money you want, companies are making billions out of Free software. And nobody wants to use your server, bandwidth or material resources. However, the _software_ should be given to the people who USE it. Giving the public a freebie web service for now, until you figure out how to sell their information and useage stats to advertisers, does not make your service free.
People are typing their twitter logins to little apps that are poping up everywhere. Wouldn't you rather have these web apps open, somewhere in github, with a transparent database backend that you can see is not hoarding any personal infromation about you?
> Stallman is a digital socialist.
And that's a label many would self-attribute, with pride.
> The idea of services (or anything other than direct > labor) that people charge for offends him. That offends me.
Fine. Don't use Free software then :-)
If I had the front-end javascript source code to gmail or salesforce or basecamp, that would neither give me access to other user's data, nor give me free access. Proof: I do have access to their complete front-end source code, just minified. Having it more readable would just be convenient.
Minification / Obfuscation != Security. Don't put anything in your front end code that you would not want an adversary to read (true of the backend too, if perhaps less urgently).
I think web developers are over estimating the sophistication of their work :-)
You can't argue based on a 'network effect' website that has obvious lock-in.
Look at some non-social webapp. Say an online image editor. If the source was available, copy cats competitors could spring up and take away market share. There's no upside to releasing the source code, and obvious downside (Assuming you're in it to make money and not just spread love).
Image-editing algorithms are published in text-books. Graphics API for the web (flash, java, canvas, etc.) are published in API documentation. Nothing is secret. If someone copies your GPLed Free software and improves on it you can go and copy the improved version if you want it.
If a competitor is blindly copying you, you probably have nothing to worry about (should I elaborate?.) And if your competitive edge can be taken away with cut-and-paste, then you have ALLOT to worry about.
But allow me to add that the great majority of web startups are running on the value of a "social network". Remove the users and the service is worth nothing. You can't say that about shrink-wrap software; even if the great majority of the people don't use it, there is that one or two lucrative corporate or government contract to make it all worthwhile.
So in that light, shouldn't the users of web apps be entitled to something a little more than "free use"? Say, a little assurance that the platform they trust so much of their private lives is transparent, extensible, and should the company sink tomorrow, available for replication?
What happens when you strike it rich, sell the startup to a Mega Corp and move to Tahiti? The users will be at the mercy of greedy suits who will fill data warehouses with every click, persona and profile of your users, the very same people who made you rich.