I'm a free marketer, but yes, sometimes regulation is not evil.
I am always surprised that so many people fall into this false dichotomy. A free market is not a market without regulation any more than a free society is a society without laws.
It's not that KDE 4.0 had a frozen API. If you check the release schedule[0] you'll see that they soft-froze the KDE4 API in May '07, before the KDE4 Alpha 1 release (because API bugs can still get fixed) and hard-froze the libs in July, before Beta 1, well before the 4.0 release in Jan '08.
But even before the first alpha, KDE4 had been in development for more than a year, and the API, while in flux a certain amount, should have been stable enough for early adopter 3rd-party developers to start developing apps based on the platform.
However, before the 4.0 release, no 3rd-party developers were developing their apps to the KDE4 platform. They were still writing to KDE3. Why? Firstly, the theory was (as I understand it) that many developers would not believe in the API freeze until a 4.0 release. Secondly, until users were actually using KDE4, really, what was the point? KDE3 apps would run fine under KDE3 and KDE4, KDE4 apps would only run under KDE4. By developing against an unreleased KDE4, you're severely restricting your user base.
If you want to argue otherwise, please prove me wrong and name 5 apps that started to be developed against the KDE4 platform before the KDE 4.0 release. If you're right, there should be a few given that the API/ABI had been frozen for the best part of 6 months at that point. On my debian "testing" box the "kdebase-runtime" package has 246 reverse dependencies. Most of those apps are part of the kde platform, but a fair few (like "amarok" and "audex") are not. Or, there are likely a few scattered around the net that aren't in Debian. So, if there were many in development before 4.0, it shouldn't be too hard to find them.
By having alpha's be so late in the process What do you call developer releases where you don't want end user bug reports at all?
Uh, well, they're not "releases" then. They're just people (either cow-orkers, or other serious hackers) grabbing the latest sources from revision control and playing with them themselves. If you can do that, you're pretty much by definition a bit more than just an "end user". If you report a bug, you do so against a revision id, or commit id, or branch/date, or similar.
Out of curiosity what industry are you in?
What, you mean aside from "software development"?:-) I've been on teams developing project management apps, web/db middleware, computer aided translation software (think poedit), websites, doing weird things with javascript to integrate with other people's websites, and submitted patches (and even had one or two of them accepted!) to some free software projects.
A transaction draws money from your account immediately, good luck trying to reverse that later, I mean it is YOUR money gone, not the bank's money.
Do you have trouble reading?
There is no money in the account, ever, except when I'm just about to buy something, and then only enough to pay for the thing I'm buying. I never have to worry about reversals, as no unauthorized payment can be taken to be reversed.
First of all you will have the bounce fee.
Yes, I am aware of that. I mentioned it. I am willing to accept that in return for no unauthorized payments being made. I mentioned that too.
You really are having trouble with this whole "reading" thing, aren't you. Heck, if you hadn't managed to grasp the concept that I'm using a debit card at all, I'd be forced to infer that you were completely illiterate.
there have been many instances where banks go ahead and honor the overdrawing
Good job I talked to my bank and got them to remove all overdraft facilities and to give me written confirmation of that. Sorry that I've not lived down to your expectation of how completely fucking dumb you assumed I was.
I use a debit card for online transactions. It has its own separate bank account, with no overdraft or other type of negative balance facility. When I want to buy something online, I get to the checkout page, see how much the total is, use online banking to transfer just enough money into the account from my main account to cover the cost, and then proceed with the purchase. If a retailer tries to take too much, or tries to take payment twice, or if the card number is compromised and is used fraudulently, payment requests just bounce with an "insufficient funds" error.
I think I might get charged for payment bounces, but however much that is is probably going to be less than the value of the invalid payments. And the people messing about get nothing from it.
Actually that's a beta version that is unacceptably buggy but feature complete. alpha versions are not feature complete beta versions are.
Hmmm...and Wikipedia agrees with you.
Strange. In the 5 different companies I've worked for, and for most pre-release software I've tried, alpha and beta versions have been both been feature-complete, and the main difference has been that alpha versions were only distributed to other "programmers"; they were likely to eat your data, had debugging info included, and the developers only wanted bug reports if you were capable of supplementing them with backtraces, able to help out with running the app under a debugger, or accomplish similar technical stuff. Meanwhile beta versions were distributed to (select) end-users, were as free of dataloss bugs as possible, and the devs were happy to get bug reports that the user would not be able to supplement with any more info than what they actually did, but were not yet good enough to be RCs.
Ah well. I guess those terms just get used differently by different people/in different places. *shrug*
Where do you write software? Everywhere else in the world I'm aware of calls something alpha if it's feature-complete but still really buggy. KDE4.0 had all the features they needed for 4.0 - it had just enough user-facing features to be usable as a very basic desktop, and had an API they could declare "frozen" for third-party app developers to start working with. Which third-party app developers generally don't do if you haven't made a point-zero release. If it was KDE alpha-4.19, no-one else would be writing apps for KDE4 yet, they'd still be developing for KDE3.
Anyway, define "feature-complete". There are new features coming in KDE 4.6, and there will be more in 4.7. Is KDE4 "feature-complete" yet? Maybe they should still not have released 4.0.
it'll only work until our malicious plugin installer patches the browser binary
Um, that's why as a user you don't have write access to installed binaries, and binary images are loaded by the OS into write-protected memory regions, which is enforced by the CPU.
Have you ever filled up your Windows so much that it can't find space for swap and hence crashes, and on reboot can't start up, requiring command-line removal of the overgrown files in order to make the system work again?
That's not an app error, that's Windows being shit and not learning from what everyone else has been doing for 30 years. Like reserving 5% of the disk for root/Administrator, returning "disk full" to non-root processes if they try to write, but giving root breathing space to fix the problem. And having a (semi-)sane OOM killer which at least tries to kill processes in "worst offending", and probably least critical, order (even if that heuristic still isn't perfect on the 18th rewrite:-) when memory gets low. And reserving all the swap space it's going to try to use up front.
games should NOT be crashing so hard with filesystem handles open without at least attempting to clean up first, but I've seen programs do exactly that (and thus render those files undeletable - even if they are causing problems - without a reboot).
No, that's still Windows being shit. Once a process exits, no matter how it exits - cleanly, uncleanly, or spectacular crash - the OS should free all the resources that process was using, be that memory, open file handles, file locks, network sockets, etc... If Windows says a file is undeletable because it thinks a process which no longer exists has it open, that's a major OS issue which needs to be cleaned up. That's without even getting into how some other OSs have always allowed you to unlink in-use files...
Still, though, there's no way for you to know for sure.
OK, just out of curiosity, could you name 3 significant classes/types of work which you think are likely to be on the internet illegally, but are not available for download anywhere legally? Films are available for download from Netflix, music is available from iTunes and numerous internet radio stations, books are available from Amazon, mostly via the kindle, but still from the internet. Much software is now downloadable, but I guess the Windows OS, maybe MS Office, and some computer games, are not. But that's, what, a few hundred unique bits of content, tops? Anything significant?
but it is nearly always legal to download something from the internet.
How do you know?
Because I download hundreds of things from the internet every day. HTML and RSS feed files - all legal. Each and every javascript file, CSS file, image, Flash applet, etc... that make up part of that page - all legal. Updates to my OS in ".deb" format - all legal. ISOs of Linux images - all legal. Youtube videos - nearly all are legal. Flash files - all legal. Mailing list archives, emails from my mail server, usenet postings and archives, calendar subscriptions, instant messages that I've been sent, every single individual patch hosted in any public version control system repo (e.g. cvs, svn, git), etc., etc., etc...
There a billions of bits of content on the internet, mostly held and served up by the copyright holder, or by an authorized redistributor - like these comments, each of which belongs to us, their respective authors, who own the copyright, but which we implicity authorize slashdot to redistribute for us by posting them to their server. I don't see how it is possible for the few bits of content that are hosted illegally can outnumber those, especially when that content is *also* hosted legally, e.g. by Netflix.
OK, I don't know in that I've not actually counted all the bits of information out there. There is a possibility I'm wrong. But I think that that's incredibly unlikely.
And since a typical server is basically automated, the only human left in the loop is the downloader.
That's disingenuous. The sysadmin in charge of a server is responsible for setting up accounts and permissions on it, and ensuring that people can only see what they are allowed to see. Unless you think that servers should honour requests for "/etc/shadow". And the person who uploaded a file and putting it in a place to be downloaded is responsible for the fact that content is available.
Whether or not admins/uploaders personally OK every download is beside the point. They have a responsibility to set up rules and permissions ahead of time so that only those allowed to download a file have the ability to do so.
Files do not magically appear on servers in locations that are automatically world-readable.
Consider an analogy: Alice sets up and aims a sniper rifle at someone.
Your analogy is awful because it is nearly always illegal to shoot people, but it is nearly always legal to download something from the internet.
Everything (bar actual public domain works, whose number is negligible compared to the total number of works on the internet) on the internet is copyrighted by someone, and whenever you download any of them, by, e.g. visiting a web page, and downloading the main HTML file, any associated CSS and JS files, plus all the images inline in the page, you're downloading copyrighted material.
In nearly all cases you are allowed to, because you are requesting the copy from either the copyright holder (e.g. a new article from the news organisation's website) or from someone the copyright holder has authorised to redistribute their work (e.g. videos from youtube).
He can't. The law doesn't care whether he can tell, and holds him liable for infringements even when he had no intent to infringe,
Ah, I was concentrating more on what makes sense, not what the law is. Obviously, because some are being successfully sued for it, people are legally liable for copyright infringment when they download content from a distributor who has no right to send them a copy. I'm just pointing out that I think this is absurd.
If I download an episode of "The Daily Show" from Comedy Central's website, that is fine.
If I download the same episode from some other dodgy Russian website, that is not fine.
If I request a file called "termsandconditions.html" from a dodgy Russian website, and get an episode of "The Daily Show", that is also not fine.
However, if I request a copy of The Daily Show from a dodgy Russian website and they refuse to send me a copy for any reason (e.g. they don't have it, or they're operating a legal paid service and I haven't paid them), or send me something else (e.g. a public domain copy of one of Shakespear's plays) then that is fine.
It is not the downloader who *authorises* or *creates* the copy. It is the distributor. A downloader only ever *requests* a copy of a resource - they can never force a distributor to give them one. It is the distributor's decision how to, or even whether to, respond to a request.
How is a downloader to know whether a distributor has the right to distribute a particular piece of content or not? On first glance, Comedy Central would seem to be the only place to get The Daily Show, but they may have signed an agreement with another website to give that other website exclusive distribution rights. In which case downloading from Comedy Central is not OK, but download from the other website, which might be Russian, is.
However, a distributor must be aware of what rights they have to distribute each piece of content they control. Either they own the copyright, in which case they have the right to distribute unless they have signed that right over to someone else, or they do not, and do not have the right to distribute unless they have an explicit license which says that they can.
Whichever way you look at it, a downloader has no way to know, in general, whether a distributor has the right to distribute any content, or even to know what content is *being* distrubuted until they've actually downloaded it, while a distributor always knows what content they have, and what rights they have to it.
No, it makes no sense at all to make downloaders responsible for the things they download, and absolute sense to make distributors perfectly responsible. Any other way simply cannot work.
Except, I've got a half-dozen server apps on my desktop, some for testing purposes (e.g. private mysql, apache), some because they're useful (sshd, ntpd, ipp) and some because a program decided it wants to listen.
I could figure out, for each of those apps, how to configure them to listen only on the right interface(s) (e.g. only loopback, or maybe only eth0, or maybe loopback + eth0), or maybe to only accept connections from certain sources or source subnets, or how to prevent them from opening a port at all. But that's a lot of different config files, each with their own syntax, some of which are literally impossible to configure how I want - the syntax just isn't flexible enough. So, instead, I could just all apps open whatever ports they like on all interfaces, and manage what ports can be connected to by whom in a single place - my firewall.
I've tried it your way, and it got too complex. I find this much easier.
Copyright on the internet does not work the way you think it works.
If comedy central hosts a video of The Daily Show, and I (or Google) links to it, that is fine.
If I copy the video onto my own server, and host it from there, and then I (or Google) link to the video that I am hosting and distributing, then I am committing copyright infringement by distributing the video. I am not committing copyright with my link to my hosted copy of the video, I am committing infringement with the actual copy I am distributing. Google still isn't infringing copyright themselves, but could be thought of as contributing to or aiding copyright infringement, by linking to my infringing copy of TDS. The law you point to is specifically meant to limit Google's (or other search engine's) liability in the case that they link to already-infringing works. They are not infringing by linking.
Read them. Carefully. Please, while allowing for the faint possibility that you might be mistaken, or have been misled by others in the past.
Yes, people have sued over deep linking. But, in the US and Europe, have not won. After all, you can sue someone for anything, but it's winning a suit that lets you know if you were right. No-one in the US or Europe has ever been convicted of copyright infringement for linking to content.
All cases have either been found in favour of the linker, or been settled out-of-court, or the sued party has been convicted of something else. That something else might be aiding or contributing to copyright infringement, or for actual copyright infringement, such as when hosting and serving cached copies of a document (or derivative works, such as thumbnail images). Note that, in this last case, it is because the defendant is hosting and distributing their own copy of the work that they are committing copyright infringement, not because they are linking to the original.
Looking at the cases you are citing: Napster and Limewire (I have the Limewire case right, right?), you'll note that they are both convicted of "contributory infringement" - i.e. helping others commit infringement by linking to infringing content, or "vicarious infringement" - failing to supervise the infringing conduct or others despite having the ability to do so.
They were not convicted of copyright infringment, as they were not hosting any infringing content, or breaking copyright themselves.
They were convicted of contributory and vicarious infringement, not because they were linking to content provided by their users, but rather because the content they were linking to, hosted and distributed by those users, was in breach of copyright, and on a massive scale. The users were hosting and distributing their own copies of other people's files in breach of copyright. That was the issue.
Linking to content is legal.
I wasn't trying to muddy any points. A link is a link is a link is a link. All links are equal. And links could reside anywhere, and point to any content. Or to no content at all. Or to content that has changed since the link was created. It doesn't matter. It's the content itself which is important. And content is hosted and distributed from a particular place. It may be linked to from many places, or none. But it's the actual storage and distribution that is what copyright law applies to.
*sigh*
Even if, after all that, you still convinced I'm wrong, and you're right, be aware that most of the other people on the internet, and plenty of courts by the looks of things, appear to disagre
If you link to their download page and allow the user to select the download, it's the third party distributing it.
So, Google distributes most of the internet? Because Google links to plenty of HTML pages directly. And people click on links from google, and their browser downloads those web pages directly. Are you really saying that Google distributes other people's HTML? All of those end-point HTML pages are copyright by someone other than Google, and I'm pretty sure that Google does not have a license to redistribute those pages. Google links to some of my web pages, allowing people to download them, and I didn't give Google permission to distribute my copyrighted material. Is Google liable for copyright infringement of my content? Is anyone who links to any content on the web without first getting the copyright holders permission guilty of copyright infringement?
And yes, several court cases concerning copyright infringement have backed that statement up. Lime wire and the pirate bay are some of the larger ones.
I'm pretty sure you'll find that in all of those cases, none of the parties were convicted of copyright infringement themselves, because, TBP in particular, were not distributing content. They were merely linking to content that 3rd parties distribute, much like Google does. They may have been charged for promoting the copyright infringement of others, but I'm aware of no cases where a court decided that linking to copyrighted material was equivalent to distributing it.
I'll happily look over any actual evidence to the contrary.
And if this download page has a link to a GPL'd binary file, for it to count as distributing both source and object code, there must also be a link to the source.
A "download" web page does not distribute anything. The thing a download page links to is distributed from some server or other, yes, (and it might even be the same server), but it's the server the file exists on which is where it's distributed from, and the directory in which the file is actually located which is the place that it's distributed from.
How can you not tell the difference between linking to something which can be distributed, and actually distributing it? The whole World Wide Web is based on the fact that person A can link to files (whether those files are.HTML,.PNG,.ISO, or other) that person B hosts and distributes, without needing to host or distribute those files themselves, and also without person A needing to get permission from the copyright holder to link to that file, because person A is not distributing the content!
I know you were able to jump through hoops and find the source
You think that's jumping through hoops? Clicking the ".." link twice and looking for the word "source" is "jumping through hoops"? On the same server where they host and distribute the binaries, the source is only a couple of directories away, in a directory called "source", at the same level as the architecture-specific binaries, and finding that is jumping through hoops? Just how clear do the instructions need to be for you?
[If] clicking your link that takes them directly to the binary, you are technically distributing it.
Uh, really? You think that telling someone where something is, e.g. writing a hyperlink, is the same thing as providing them with that thing yourself, e.g. by owning a computer which hosts and actually *distributes* it? By analogy, would you say that telling people the name of the street where the dope dealers hang out is the same thing as selling dope?
It's certainly an... individual way of looking at things. I can't say I think it's correct.
The clause allowing the downloads to be equated to "distribution with the source code" requires both the binary and source to be linked from the same place.
No, the clause requires that both the binary and source be equivalently accessible from the same place - i.e. from the place where they actually are. Not necessarily from each and any link that happens to point to them, but from where the files really exist.
If it worked your way round, I could actually keep the binary and source files on completely different servers, provided I always maintained links to the binary and source files in pairs.
But your binary server probably doesn't just host GPLd binaries. It might host your media files, or a mailing list archive, or something. So a search spider comes along and indexes your binary server, traversing all the directories on it, finds the GPLd binaries, and provides links to them. Now someone's breaking the law, because the search engine now points to a binary, but someone who finds it has no hope of finding the equivalent source file. Is it the spider? It's just found and indexed publically available content. Is it you? You've maintained your pairs of links scrupulously.
As far as I can tell, that's just absurd.
The only thing that makes sense to me is to either provide the binary and source in the same place, or to provide instructions on how to get the source in the same place as the binary. That's the actual place where the binary is. Not on any and each page that happens to link to the binary, but where the binary is.
Because, once the source, or the instructions to get it, are in the same place as the binary, you don't need to duplicate links any more. Everyone knows the source is in the same place as the binary. It doesn't matter what the page that linked to the binary says, or doesn't say, or if it was auto-generated by a spider, because the source is right there.
Well, seeing how the GPL doesn't apply unless you are distributing the code, I assumed it was given that all this talk was about distributing the code.
Sorry, in your post that I replied to, you wrote:
There are restrictions to what code you can link to or use with the programs/*pl'd code, how it can be used and so on. It's not like Lgpl code can link to GPL code and Mycode licenses software at the same time.
I really thought you were talking about "use", not "distribution". Or the actual act of linking (i.e. linking bits of compiled code yourself), not the distribution of linked binaries.
Apologies for misunderstanding you.
Well, the GPL doens't apply to getting copies or using them so I'm not sure how you are equating [the (L)GPL] to end use license agreements in good faith.
I'm trying to get my head around the statement (not yours though) that:
So is GPL more or less free than a commercial license? Usually less free
I was trying to point out that for >99.9% of commercial software that exists, whatever license or permissions you get with it, as far as I can tell, is "less free" than the GPL. This includes commercial redistributable libraries and frameworks, as most of these will not give you the freedom to modify the libraries.
Did I mention debian specifically?
No, you mentioned "some distro". I had to pick one to test, so I picked Debian because I that's the distro I happen to use. *shrug*
Why was it that you didn't want me to select minimal install?
Because the minimal install is specifically for people who want to download as little as possible. That's what "minimal" means. If you want a minimal install, you are almost certainly not interested in the source.
But, still, let's have a look. The minimal install for amd64 points to:
Oh yes, there's a "source" directory. But this brings us to:
Is the source for the minimal install in the same place
This is the real point of contention, isn't it.
Well, there I happen to think you're being a bit picky w.r.t. "the same place". It's on the same server (and therefore doesn't activate the "on a different server" subclause of the GPL), as part of a well-organised directory tree, and trivial to find given the filenames involved.
For a binary file called "[base]/amd64/foo.tgz", is it really substantially different to call the corresponding source file "[base]/source/foo.tgz" instead of "[base]/amd64/foo.source.tgz"?
If you're going to be that strict about how you interpret "in the same place", surely you could interpret it to mean that the source must be present in the same file (iso, tar, deb, etc...) as the binary. After all, by your reasoning, if it's in a different file, is that not "a different place"?
By your interpretation, the source has to be in the same directory as the amd64 isos, and again in the same directory as the x86 isos, and in the same directory as the sparc isos, ad nauseum. Do you really think that's the intent of the GPL? Do you think that's practical? (I know, I know; whether something is practical has no bearing on whether it actually adheres to the license. I just balk at impracticality.)
Hmmm....I think we have a real, legitimate difference of interpretation here. The distros happen to agree with me. That doesn't mean I'm right of course (that would be argumentum ad populum) but I will point out that the distros do have lawyers who have gone over this.
There are restrictions to what code you can link to or use with the programs/*pl'd code,
No, only if you *redistribute* that code. You can take GPL'd code, and link it with any code you like, no matter what license, for your own use. It's only when you redistribute GPL'd code to another user that you have to give them the same freedoms the GPL gave you.
Im think he is talking about the license to distribute the software and not the EULA
But most commercial software does not even have one of these licenses. Try getting a license for Windows that allows you do distribute copies you've made. Or Photoshop. Or any other commercial app. (Note - not pass on existing copies, which you can also do with GPL'd software, but distribute new copies) Or even if talking about libraries or frameworks (a very small subset of commercial software) try getting a license for a commercial library that allows you to redistribute modified versions at all.
I don't see how, comparing like with like, that for any reasonable definition of "free", the GPL is "less free" than 99.9% of commercial software licenses.
You see, if you offer the binary for download, you need to offer the source on the same server in just as conspicuous of a way. Try finding an ISO with all the source on it on some Distro's download page.
Well, I'm a Debian user, so let's look, shall we?
Go to the Debian homepage, click "CD ISO Images", and then select any "download" option other than the "download a minimal bootable CD image". Hey, look at that. Right there, with the links to the binaries for each architecture, there's a link labelled "source". I wonder what that does.:-)
Hell, just try to find the source code to the updates the package manager automagically installs when you set it to update.
Hmmm...right next to the arch specific directory for my downloads, again there's a "source" directory. Wow!
The GPL says 3 years after the last offering,
No, it's a bit more subtle than that.
GPL2 3 states: If distribution of executable or object code is made by offering access to copy from a designated place, then offering equivalent access to copy the source code from the same place counts as distribution of the source code, even though third parties are not compelled to copy the source along with the object code.
GPL3 6d states: Convey the object code by offering access from a designated place (gratis or for a charge), and offer equivalent access to the Corresponding Source in the same way through the same place at no further charge. You need not require recipients to copy the Corresponding Source along with the object code. If the place to copy the object code is a network server, the Corresponding Source may be on a different server (operated by you or a third party) that supports equivalent copying facilities, provided you maintain clear directions next to the object code saying where to find the Corresponding Source. Regardless of what server hosts the Corresponding Source, you remain obligated to ensure that it is available for as long as needed to satisfy these requirements.
So, if you distribute over the internet, you can fulfil your obligations under the GPL by making the source accessible alongside the binaries, for only as long you distribute the binaries. If a client chooses not to get the source from you at the time they get the binaries, you need not offer them the source at a later date. The "3 year offer" is
GPL is very restrictive as to what the licensee can or cannot do with the work. BSD/MIT both allow nearly any use/modification/extension, but requires the licensee to retain copyright notices.
Um, (L)GPL allows for any use/modification/extension. Those are freedoms 0 and 1.
The only restrictions the (L)GPL places on works licensed under it are when it comes to redistribution of the code - you have to give the person you redistribute someone else's GPL'd code to the same freedoms that you got when you received it.
So is GPL more or less free than a commercial license? Usually less free, but it depends on what you want.
Really? Last I checked, the licences for commercial software usually do not allow you to redistribute copies of the code at all, in any form (binary or source, oh, and good luck with getting the source), under any circumstances.
The GPL is more restrictive in that source code must be provided, no charge, to anyone who gets a binary copy.
Yes, but you're allowed to include, in the amount you charge people for the binary version, costs to cover providing them with the source in the future, which you'll then have whether or not they ever ask for the source. Or you could always ship the source alongside the binaries.
I'd say that smaller companies tend to be less restrictive in their licenses than the GPL
Isn't it a pretty negligible inconvenience? Really, how hard is it to just not have one drink? Is the mood really improved that much for you that you think it's worth putting other people's lives at increased risk, however negligible that increase might be? Do you think those other people in your community would agree with that, if they had a choice in the matter?
Slashdot Mirror
I am always surprised that so many people fall into this false dichotomy. A free market is not a market without regulation any more than a free society is a society without laws.
Hmmm....
Ah, yes. The thinly-veiled, plausibly-deniable version of "I think you're dumb". Well, excuse me for reading between the lines.
Or, maybe that's just the attitude you inspire in people (well, it worked for me) and you just suck at helping them.
Sorry, I've not been completely clear.
It's not that KDE 4.0 had a frozen API. If you check the release schedule[0] you'll see that they soft-froze the KDE4 API in May '07, before the KDE4 Alpha 1 release (because API bugs can still get fixed) and hard-froze the libs in July, before Beta 1, well before the 4.0 release in Jan '08.
But even before the first alpha, KDE4 had been in development for more than a year, and the API, while in flux a certain amount, should have been stable enough for early adopter 3rd-party developers to start developing apps based on the platform.
However, before the 4.0 release, no 3rd-party developers were developing their apps to the KDE4 platform. They were still writing to KDE3. Why? Firstly, the theory was (as I understand it) that many developers would not believe in the API freeze until a 4.0 release. Secondly, until users were actually using KDE4, really, what was the point? KDE3 apps would run fine under KDE3 and KDE4, KDE4 apps would only run under KDE4. By developing against an unreleased KDE4, you're severely restricting your user base.
If you want to argue otherwise, please prove me wrong and name 5 apps that started to be developed against the KDE4 platform before the KDE 4.0 release. If you're right, there should be a few given that the API/ABI had been frozen for the best part of 6 months at that point. On my debian "testing" box the "kdebase-runtime" package has 246 reverse dependencies. Most of those apps are part of the kde platform, but a fair few (like "amarok" and "audex") are not. Or, there are likely a few scattered around the net that aren't in Debian. So, if there were many in development before 4.0, it shouldn't be too hard to find them.
Uh, well, they're not "releases" then. They're just people (either cow-orkers, or other serious hackers) grabbing the latest sources from revision control and playing with them themselves. If you can do that, you're pretty much by definition a bit more than just an "end user". If you report a bug, you do so against a revision id, or commit id, or branch/date, or similar.
What, you mean aside from "software development"? :-) I've been on teams developing project management apps, web/db middleware, computer aided translation software (think poedit), websites, doing weird things with javascript to integrate with other people's websites, and submitted patches (and even had one or two of them accepted!) to some free software projects.
[0] http://en.wikipedia.org/wiki/KDE4#Release_schedule
Do you have trouble reading?
There is no money in the account, ever, except when I'm just about to buy something, and then only enough to pay for the thing I'm buying. I never have to worry about reversals, as no unauthorized payment can be taken to be reversed.
Yes, I am aware of that. I mentioned it. I am willing to accept that in return for no unauthorized payments being made. I mentioned that too.
You really are having trouble with this whole "reading" thing, aren't you. Heck, if you hadn't managed to grasp the concept that I'm using a debit card at all, I'd be forced to infer that you were completely illiterate.
Good job I talked to my bank and got them to remove all overdraft facilities and to give me written confirmation of that. Sorry that I've not lived down to your expectation of how completely fucking dumb you assumed I was.
I use a debit card for online transactions. It has its own separate bank account, with no overdraft or other type of negative balance facility. When I want to buy something online, I get to the checkout page, see how much the total is, use online banking to transfer just enough money into the account from my main account to cover the cost, and then proceed with the purchase. If a retailer tries to take too much, or tries to take payment twice, or if the card number is compromised and is used fraudulently, payment requests just bounce with an "insufficient funds" error.
I think I might get charged for payment bounces, but however much that is is probably going to be less than the value of the invalid payments. And the people messing about get nothing from it.
Hmmm...and Wikipedia agrees with you.
Strange. In the 5 different companies I've worked for, and for most pre-release software I've tried, alpha and beta versions have been both been feature-complete, and the main difference has been that alpha versions were only distributed to other "programmers"; they were likely to eat your data, had debugging info included, and the developers only wanted bug reports if you were capable of supplementing them with backtraces, able to help out with running the app under a debugger, or accomplish similar technical stuff. Meanwhile beta versions were distributed to (select) end-users, were as free of dataloss bugs as possible, and the devs were happy to get bug reports that the user would not be able to supplement with any more info than what they actually did, but were not yet good enough to be RCs.
Ah well. I guess those terms just get used differently by different people/in different places. *shrug*
Where do you write software? Everywhere else in the world I'm aware of calls something alpha if it's feature-complete but still really buggy. KDE4.0 had all the features they needed for 4.0 - it had just enough user-facing features to be usable as a very basic desktop, and had an API they could declare "frozen" for third-party app developers to start working with. Which third-party app developers generally don't do if you haven't made a point-zero release. If it was KDE alpha-4.19, no-one else would be writing apps for KDE4 yet, they'd still be developing for KDE3.
Anyway, define "feature-complete". There are new features coming in KDE 4.6, and there will be more in 4.7. Is KDE4 "feature-complete" yet? Maybe they should still not have released 4.0.
Um, that's why as a user you don't have write access to installed binaries, and binary images are loaded by the OS into write-protected memory regions, which is enforced by the CPU.
You want nethack.
(Or, if you value your free time, you don't)
That's not an app error, that's Windows being shit and not learning from what everyone else has been doing for 30 years. Like reserving 5% of the disk for root/Administrator, returning "disk full" to non-root processes if they try to write, but giving root breathing space to fix the problem. And having a (semi-)sane OOM killer which at least tries to kill processes in "worst offending", and probably least critical, order (even if that heuristic still isn't perfect on the 18th rewrite :-) when memory gets low. And reserving all the swap space it's going to try to use up front.
No, that's still Windows being shit. Once a process exits, no matter how it exits - cleanly, uncleanly, or spectacular crash - the OS should free all the resources that process was using, be that memory, open file handles, file locks, network sockets, etc... If Windows says a file is undeletable because it thinks a process which no longer exists has it open, that's a major OS issue which needs to be cleaned up. That's without even getting into how some other OSs have always allowed you to unlink in-use files...
OK, just out of curiosity, could you name 3 significant classes/types of work which you think are likely to be on the internet illegally, but are not available for download anywhere legally? Films are available for download from Netflix, music is available from iTunes and numerous internet radio stations, books are available from Amazon, mostly via the kindle, but still from the internet. Much software is now downloadable, but I guess the Windows OS, maybe MS Office, and some computer games, are not. But that's, what, a few hundred unique bits of content, tops? Anything significant?
Because I download hundreds of things from the internet every day. HTML and RSS feed files - all legal. Each and every javascript file, CSS file, image, Flash applet, etc... that make up part of that page - all legal. Updates to my OS in ".deb" format - all legal. ISOs of Linux images - all legal. Youtube videos - nearly all are legal. Flash files - all legal. Mailing list archives, emails from my mail server, usenet postings and archives, calendar subscriptions, instant messages that I've been sent, every single individual patch hosted in any public version control system repo (e.g. cvs, svn, git), etc., etc., etc...
There a billions of bits of content on the internet, mostly held and served up by the copyright holder, or by an authorized redistributor - like these comments, each of which belongs to us, their respective authors, who own the copyright, but which we implicity authorize slashdot to redistribute for us by posting them to their server. I don't see how it is possible for the few bits of content that are hosted illegally can outnumber those, especially when that content is *also* hosted legally, e.g. by Netflix.
OK, I don't know in that I've not actually counted all the bits of information out there. There is a possibility I'm wrong. But I think that that's incredibly unlikely.
That's disingenuous. The sysadmin in charge of a server is responsible for setting up accounts and permissions on it, and ensuring that people can only see what they are allowed to see. Unless you think that servers should honour requests for "/etc/shadow". And the person who uploaded a file and putting it in a place to be downloaded is responsible for the fact that content is available.
Whether or not admins/uploaders personally OK every download is beside the point. They have a responsibility to set up rules and permissions ahead of time so that only those allowed to download a file have the ability to do so.
Files do not magically appear on servers in locations that are automatically world-readable.
Your analogy is awful because it is nearly always illegal to shoot people, but it is nearly always legal to download something from the internet.
Everything (bar actual public domain works, whose number is negligible compared to the total number of works on the internet) on the internet is copyrighted by someone, and whenever you download any of them, by, e.g. visiting a web page, and downloading the main HTML file, any associated CSS and JS files, plus all the images inline in the page, you're downloading copyrighted material.
In nearly all cases you are allowed to, because you are requesting the copy from either the copyright holder (e.g. a new article from the news organisation's website) or from someone the copyright holder has authorised to redistribute their work (e.g. videos from youtube).
Ah, I was concentrating more on what makes sense, not what the law is. Obviously, because some are being successfully sued for it, people are legally liable for copyright infringment when they download content from a distributor who has no right to send them a copy. I'm just pointing out that I think this is absurd.
If I download an episode of "The Daily Show" from Comedy Central's website, that is fine.
If I download the same episode from some other dodgy Russian website, that is not fine.
If I request a file called "termsandconditions.html" from a dodgy Russian website, and get an episode of "The Daily Show", that is also not fine.
However, if I request a copy of The Daily Show from a dodgy Russian website and they refuse to send me a copy for any reason (e.g. they don't have it, or they're operating a legal paid service and I haven't paid them), or send me something else (e.g. a public domain copy of one of Shakespear's plays) then that is fine.
It is not the downloader who *authorises* or *creates* the copy. It is the distributor. A downloader only ever *requests* a copy of a resource - they can never force a distributor to give them one. It is the distributor's decision how to, or even whether to, respond to a request.
How is a downloader to know whether a distributor has the right to distribute a particular piece of content or not? On first glance, Comedy Central would seem to be the only place to get The Daily Show, but they may have signed an agreement with another website to give that other website exclusive distribution rights. In which case downloading from Comedy Central is not OK, but download from the other website, which might be Russian, is.
However, a distributor must be aware of what rights they have to distribute each piece of content they control. Either they own the copyright, in which case they have the right to distribute unless they have signed that right over to someone else, or they do not, and do not have the right to distribute unless they have an explicit license which says that they can.
Whichever way you look at it, a downloader has no way to know, in general, whether a distributor has the right to distribute any content, or even to know what content is *being* distrubuted until they've actually downloaded it, while a distributor always knows what content they have, and what rights they have to it.
No, it makes no sense at all to make downloaders responsible for the things they download, and absolute sense to make distributors perfectly responsible. Any other way simply cannot work.
Except, I've got a half-dozen server apps on my desktop, some for testing purposes (e.g. private mysql, apache), some because they're useful (sshd, ntpd, ipp) and some because a program decided it wants to listen.
I could figure out, for each of those apps, how to configure them to listen only on the right interface(s) (e.g. only loopback, or maybe only eth0, or maybe loopback + eth0), or maybe to only accept connections from certain sources or source subnets, or how to prevent them from opening a port at all. But that's a lot of different config files, each with their own syntax, some of which are literally impossible to configure how I want - the syntax just isn't flexible enough. So, instead, I could just all apps open whatever ports they like on all interfaces, and manage what ports can be connected to by whom in a single place - my firewall.
I've tried it your way, and it got too complex. I find this much easier.
s/beyond/irrelevant to/
Copyright on the internet does not work the way you think it works.
If comedy central hosts a video of The Daily Show, and I (or Google) links to it, that is fine.
If I copy the video onto my own server, and host it from there, and then I (or Google) link to the video that I am hosting and distributing, then I am committing copyright infringement by distributing the video. I am not committing copyright with my link to my hosted copy of the video, I am committing infringement with the actual copy I am distributing. Google still isn't infringing copyright themselves, but could be thought of as contributing to or aiding copyright infringement, by linking to my infringing copy of TDS. The law you point to is specifically meant to limit Google's (or other search engine's) liability in the case that they link to already-infringing works. They are not infringing by linking.
Seriously.
Read the wikipedia articles on Deep linking and Copyright_aspects_of_hyperlinking_and_framing.
Read them. Carefully. Please, while allowing for the faint possibility that you might be mistaken, or have been misled by others in the past.
Yes, people have sued over deep linking. But, in the US and Europe, have not won. After all, you can sue someone for anything, but it's winning a suit that lets you know if you were right. No-one in the US or Europe has ever been convicted of copyright infringement for linking to content.
All cases have either been found in favour of the linker, or been settled out-of-court, or the sued party has been convicted of something else. That something else might be aiding or contributing to copyright infringement, or for actual copyright infringement, such as when hosting and serving cached copies of a document (or derivative works, such as thumbnail images). Note that, in this last case, it is because the defendant is hosting and distributing their own copy of the work that they are committing copyright infringement, not because they are linking to the original.
Looking at the cases you are citing: Napster and Limewire (I have the Limewire case right, right?), you'll note that they are both convicted of "contributory infringement" - i.e. helping others commit infringement by linking to infringing content, or "vicarious infringement" - failing to supervise the infringing conduct or others despite having the ability to do so.
They were not convicted of copyright infringment, as they were not hosting any infringing content, or breaking copyright themselves.
They were convicted of contributory and vicarious infringement, not because they were linking to content provided by their users, but rather because the content they were linking to, hosted and distributed by those users, was in breach of copyright, and on a massive scale. The users were hosting and distributing their own copies of other people's files in breach of copyright. That was the issue.
Linking to content is legal.
I wasn't trying to muddy any points. A link is a link is a link is a link. All links are equal. And links could reside anywhere, and point to any content. Or to no content at all. Or to content that has changed since the link was created. It doesn't matter. It's the content itself which is important. And content is hosted and distributed from a particular place. It may be linked to from many places, or none. But it's the actual storage and distribution that is what copyright law applies to.
*sigh*
Even if, after all that, you still convinced I'm wrong, and you're right, be aware that most of the other people on the internet, and plenty of courts by the looks of things, appear to disagre
So, Google distributes most of the internet? Because Google links to plenty of HTML pages directly. And people click on links from google, and their browser downloads those web pages directly. Are you really saying that Google distributes other people's HTML? All of those end-point HTML pages are copyright by someone other than Google, and I'm pretty sure that Google does not have a license to redistribute those pages. Google links to some of my web pages, allowing people to download them, and I didn't give Google permission to distribute my copyrighted material. Is Google liable for copyright infringement of my content? Is anyone who links to any content on the web without first getting the copyright holders permission guilty of copyright infringement?
I'm pretty sure you'll find that in all of those cases, none of the parties were convicted of copyright infringement themselves, because, TBP in particular, were not distributing content. They were merely linking to content that 3rd parties distribute, much like Google does. They may have been charged for promoting the copyright infringement of others, but I'm aware of no cases where a court decided that linking to copyrighted material was equivalent to distributing it.
I'll happily look over any actual evidence to the contrary.
A "download" web page does not distribute anything. The thing a download page links to is distributed from some server or other, yes, (and it might even be the same server), but it's the server the file exists on which is where it's distributed from, and the directory in which the file is actually located which is the place that it's distributed from.
How can you not tell the difference between linking to something which can be distributed, and actually distributing it? The whole World Wide Web is based on the fact that person A can link to files (whether those files are .HTML, .PNG, .ISO, or other) that person B hosts and distributes, without needing to host or distribute those files themselves, and also without person A needing to get permission from the copyright holder to link to that file, because person A is not distributing the content!
You think that's jumping through hoops? Clicking the ".." link twice and looking for the word "source" is "jumping through hoops"? On the same server where they host and distribute the binaries, the source is only a couple of directories away, in a directory called "source", at the same level as the architecture-specific binaries, and finding that is jumping through hoops? Just how clear do the instructions need to be for you?
Uh, really? You think that telling someone where something is, e.g. writing a hyperlink, is the same thing as providing them with that thing yourself, e.g. by owning a computer which hosts and actually *distributes* it? By analogy, would you say that telling people the name of the street where the dope dealers hang out is the same thing as selling dope?
It's certainly an ... individual way of looking at things. I can't say I think it's correct.
No, the clause requires that both the binary and source be equivalently accessible from the same place - i.e. from the place where they actually are. Not necessarily from each and any link that happens to point to them, but from where the files really exist.
If it worked your way round, I could actually keep the binary and source files on completely different servers, provided I always maintained links to the binary and source files in pairs.
But your binary server probably doesn't just host GPLd binaries. It might host your media files, or a mailing list archive, or something. So a search spider comes along and indexes your binary server, traversing all the directories on it, finds the GPLd binaries, and provides links to them. Now someone's breaking the law, because the search engine now points to a binary, but someone who finds it has no hope of finding the equivalent source file. Is it the spider? It's just found and indexed publically available content. Is it you? You've maintained your pairs of links scrupulously.
As far as I can tell, that's just absurd.
The only thing that makes sense to me is to either provide the binary and source in the same place, or to provide instructions on how to get the source in the same place as the binary. That's the actual place where the binary is. Not on any and each page that happens to link to the binary, but where the binary is.
Because, once the source, or the instructions to get it, are in the same place as the binary, you don't need to duplicate links any more. Everyone knows the source is in the same place as the binary. It doesn't matter what the page that linked to the binary says, or doesn't say, or if it was auto-generated by a spider, because the source is right there.
Crap, it looks like I replied to my comment instead of yours. See my other reply below. Sorry.
Sorry, in your post that I replied to, you wrote:
I really thought you were talking about "use", not "distribution". Or the actual act of linking (i.e. linking bits of compiled code yourself), not the distribution of linked binaries.
Apologies for misunderstanding you.
I'm trying to get my head around the statement (not yours though) that:
I was trying to point out that for >99.9% of commercial software that exists, whatever license or permissions you get with it, as far as I can tell, is "less free" than the GPL. This includes commercial redistributable libraries and frameworks, as most of these will not give you the freedom to modify the libraries.
No, you mentioned "some distro". I had to pick one to test, so I picked Debian because I that's the distro I happen to use. *shrug*
Because the minimal install is specifically for people who want to download as little as possible. That's what "minimal" means. If you want a minimal install, you are almost certainly not interested in the source.
But, still, let's have a look. The minimal install for amd64 points to:
http://cdimage.debian.org/debian-cd/5.0.5/amd64/iso-cd/debian-505-amd64-netinst.iso
So, let's look at:
http://cdimage.debian.org/debian-cd/5.0.5/
Oh yes, there's a "source" directory. But this brings us to:
This is the real point of contention, isn't it.
Well, there I happen to think you're being a bit picky w.r.t. "the same place". It's on the same server (and therefore doesn't activate the "on a different server" subclause of the GPL), as part of a well-organised directory tree, and trivial to find given the filenames involved.
For a binary file called "[base]/amd64/foo.tgz", is it really substantially different to call the corresponding source file "[base]/source/foo.tgz" instead of "[base]/amd64/foo.source.tgz"?
If you're going to be that strict about how you interpret "in the same place", surely you could interpret it to mean that the source must be present in the same file (iso, tar, deb, etc...) as the binary. After all, by your reasoning, if it's in a different file, is that not "a different place"?
By your interpretation, the source has to be in the same directory as the amd64 isos, and again in the same directory as the x86 isos, and in the same directory as the sparc isos, ad nauseum. Do you really think that's the intent of the GPL? Do you think that's practical? (I know, I know; whether something is practical has no bearing on whether it actually adheres to the license. I just balk at impracticality.)
Hmmm....I think we have a real, legitimate difference of interpretation here. The distros happen to agree with me. That doesn't mean I'm right of course (that would be argumentum ad populum) but I will point out that the distros do have lawyers who have gone over this.
No, only if you *redistribute* that code. You can take GPL'd code, and link it with any code you like, no matter what license, for your own use. It's only when you redistribute GPL'd code to another user that you have to give them the same freedoms the GPL gave you.
But most commercial software does not even have one of these licenses. Try getting a license for Windows that allows you do distribute copies you've made. Or Photoshop. Or any other commercial app. (Note - not pass on existing copies, which you can also do with GPL'd software, but distribute new copies) Or even if talking about libraries or frameworks (a very small subset of commercial software) try getting a license for a commercial library that allows you to redistribute modified versions at all.
I don't see how, comparing like with like, that for any reasonable definition of "free", the GPL is "less free" than 99.9% of commercial software licenses.
Well, I'm a Debian user, so let's look, shall we?
Go to the Debian homepage, click "CD ISO Images", and then select any "download" option other than the "download a minimal bootable CD image". Hey, look at that. Right there, with the links to the binaries for each architecture, there's a link labelled "source". I wonder what that does. :-)
Well, my package manager downloads binaries from:
http://ftp.debian.org/debian/dists/stable/main/
Hmmm...right next to the arch specific directory for my downloads, again there's a "source" directory. Wow!
No, it's a bit more subtle than that.
GPL2 3 states: If distribution of executable or object code is made by offering access to copy from a designated place, then offering equivalent access to copy the source code from the same place counts as distribution of the source code, even though third parties are not compelled to copy the source along with the object code.
GPL3 6d states: Convey the object code by offering access from a designated place (gratis or for a charge), and offer equivalent access to the Corresponding Source in the same way through the same place at no further charge. You need not require recipients to copy the Corresponding Source along with the object code. If the place to copy the object code is a network server, the Corresponding Source may be on a different server (operated by you or a third party) that supports equivalent copying facilities, provided you maintain clear directions next to the object code saying where to find the Corresponding Source. Regardless of what server hosts the Corresponding Source, you remain obligated to ensure that it is available for as long as needed to satisfy these requirements.
So, if you distribute over the internet, you can fulfil your obligations under the GPL by making the source accessible alongside the binaries, for only as long you distribute the binaries. If a client chooses not to get the source from you at the time they get the binaries, you need not offer them the source at a later date. The "3 year offer" is
Um, (L)GPL allows for any use/modification/extension. Those are freedoms 0 and 1.
The only restrictions the (L)GPL places on works licensed under it are when it comes to redistribution of the code - you have to give the person you redistribute someone else's GPL'd code to the same freedoms that you got when you received it.
Really? Last I checked, the licences for commercial software usually do not allow you to redistribute copies of the code at all, in any form (binary or source, oh, and good luck with getting the source), under any circumstances.
Yes, but you're allowed to include, in the amount you charge people for the binary version, costs to cover providing them with the source in the future, which you'll then have whether or not they ever ask for the source. Or you could always ship the source alongside the binaries.
Can you give an example?
SMBs? Huh? Does that even make sense?
Isn't it a pretty negligible inconvenience? Really, how hard is it to just not have one drink? Is the mood really improved that much for you that you think it's worth putting other people's lives at increased risk, however negligible that increase might be? Do you think those other people in your community would agree with that, if they had a choice in the matter?