Weekly Edition
Return to the Announcements page
| |||||||||||||
The Party of Gno (Linux Magazine)
Over at Linux Magazine, Joe "Zonker" Brockmeier suggests a different course for the free software movement. He says that users aren't responding well to being told "No" (or "Gno" as he puts it in a little dig at the FSF), so those who want to push freedom need to provide realistic alternatives to things like iPads, DRM-encrusted media, proprietary "cloud" services, and so on.
"The free software movement, though, seems to be shrinking. It still has its adherents, of course. But, when I look around at Linux events I see a sea of Mac OS X. Most contributors I know see no problem with proprietary services like Dropbox and Ubuntu One. With very few exceptions, most companies that work in the community have settled on some mixture of proprietary and open source services to try to find a working revenue model. In short, the free software philosophy seems to have gone out the window for most users and contributors. And I'll freely admit, I've advocated the pragmatic approach — because after more than 10 years of working in the community, it's clear that getting things done with a purist approach isnt working."
(Log in to post comments)
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 19:30 UTC (Thu) by DOT (subscriber, #58786) [Link]
This is a very interesting article. It avoids all the usual drama and simply analyzes why some FSF campaigns haven't been working lately. Hopefully, this article will be well-received by the FSF, because it would be a shame to see this decades-old movement slowly die out.
Replacements for the iPad and DRM are not difficult to find. Every hardware company in town is producing tablets now, and DRM has been rejected by the general public.
Web services are more difficult to replace. You can't really expect anyone to buy a server that would provide such services, even if the software was available in an easy bundle. The appeal of these services is that they are easy, cheap, and reliable. A personal server is more expensive, difficult to set up and usually has no reliable backup system in place.
Something like Ubuntu One may be a nice "middle ground" approach. You have it easy, cheap and reliable, but you can also trust it, because everything is also stored locally. Being a fairly limited and proprietary service, Ubuntu One has its problems. But some ideas are pointing to the future of proper web services.
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 20:20 UTC (Thu) by dlang (✭ supporter ✭, #313) [Link]
the biggest problem with running your own server is availability.
your server probably isn't available from the Internet (at least not on a static IP) and even if it does, it probably doesn't have redundant Internet connectivity, it's not HA, the upstream bandwidth when accessing your server from outside the house is probably poor, it probably doesn't have redundant power (it may have a UPS, but that doesn't last very long and it probably doesn't have a generator), it probably doesn't have a high availablity backup server in place, it probably does not have offsite backups (if it has backups at all), it probably doesn't have redundant hard drives. It requires administration, patching, upgrading.
is there any wonder why most people prefer to just use a cloud service?
I run my own servers at home in spite of all this (and no, I don't claim to have addressed all of the issues above) but I definitely understand why people wouldn't want to do it themselves.
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 21:45 UTC (Thu) by emk (guest, #1128) [Link]
Excellent points. However, I think the temptations are even more subtle than you describe. There's a huge wealth of cloud applications out there, and some of them are very well-designed indeed. And for now, at least, it's a very vibrant market with hundreds of major players and thousands of startups.
So if you choose to use best-of-breed cloud applications, you'll get some very nice software. And if you choose to administer your own server, you'll have a much smaller set of applications to choose from. And some of the administration tasks will be non-trivial: Even if you could run your own GitHub, it would be a full-time job. Now multiply that by 20.
And of course, many of the hackers who _could_ write free, AGPLed cloud applications are busy with startups of their own. (Full disclosure: I am.) Free software thrives when there are many disgruntled hackers with spare cycles to burn, often because the vendors have locked up entire software ecosystems.
OTOH, I've spent much of my adult life disagreeing with Richard Stallman about various issues, only to conclude 5 years later that he was prescient, and I was wrong. Intellectual honestly forces me to consider the possibility that he's warning us against serious dangers.
Maybe we need something like a free Heroku clone, with totally standardized application deployment and administration, and a robust grid storage layer? You could, perhaps, administer 20 or 30 apps for the cost of one server. But that would require an unrealistic agreement on which toolsets and database technologies to use.
Let me know if you find answers. :-)
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 0:53 UTC (Fri) by davide.del.vento (guest, #59196) [Link]
Yes I agree with you (and I do use non-free cloud computing stuff, which is almost the only non-free thing that I use).
BUT
this is the same situation that we had at the beginning of gnu/linux. Very good and appealing non-free stuff, against poor free alternatives (if any!)
Recall the history: people go for "free beer" not for "free speach"... Posted Jun 18, 2010 12:32 UTC (Fri) by khim (subscriber, #9252) [Link] Linus started Linux development because Unix workstation was unaffordable. Cox joined because BSD required math-coprocessor. People only start thinking when they are burned. As long as closed-source clouds will not require monthly payments people will not think about free alternatives. That's the main problem here.
The Party of Gno (Linux Magazine) Posted Jun 24, 2010 4:50 UTC (Thu) by jmorris42 (subscriber, #2203) [Link]
> Intellectual honestly forces me to consider the possibility that
> he's warning us against serious dangers.
Oh I guess about 90% of the people reading here agree the danger is real. The argument is about a) what can be done about it and b) tactics.
Personally I think the problem is bigger. Good luck convincing people to care about Freedom when it comes to their data on Facebook or their web browser's source code when they only pay lip service to freedom in general. We have spent a century in the west making people who care about liberty of any sort into wierdos. You aren't supposed to want to be Free, you are supposed to want to be taken care of. And Apple, Facebook, Google, Microsoft, etc. are all too willing to 'take care of you.' as long as you don't inquire too closely at the true pricetag. And nobody is in the habit of looking gift horses in the mouth these days.
And for all the fervor for Free Software here in nerdville, most aren't willing to see beyond this one case to examine the bigger picture. Any society that wants to be taken care of by someone else from cradle to grave doesn't really give a darn about Freedom. Raise yer hand if you oppose the rest of the nanny state too? You gave up the rest of your liberties for supposed security, why are you expecting to be able to pick and choose this one to keep?
Think of it this way. To borrow the example from the article of PETA, imagine how much harder their job would be if nobody even realized animals could suffer or had even bothered to think much about animals at all. And worse treated fellow humans cruelly and violently on a regular basis. (Imagine the world as it existed a thousand or so years ago.) They would have to start with the basic concept that cruelty and needless suffering in general was a bad thing before they could have a hope of bringing up the issue of cruelty to animals.
The Party of Gno (Linux Magazine) Posted Jun 24, 2010 13:38 UTC (Thu) by pboddie (subscriber, #50784) [Link] We have spent a century in the west making people who care about liberty of any sort into wierdos. You aren't supposed to want to be Free, you are supposed to want to be taken care of. And Apple, Facebook, Google, Microsoft, etc. are all too willing to 'take care of you.' as long as you don't inquire too closely at the true pricetag. And nobody is in the habit of looking gift horses in the mouth these days. So true. It is, of course, shocking that people will criticise others for campaigning for various freedoms, effectively campaigning themselves to uphold the "freedom" of the corporate world (and others) to treat people in any way they see fit as long as the goodies keep coming. Indeed, people who are willing to surrender their statutory rights often appear to want to pressure everyone else to surrender theirs in order to keep everyone "in line" and not risk incurring the wrath of those supposedly generous and benevolent corporate masters. Thus, when discussing things like product lifetimes and manufacturer obligations towards the customer, some people will actively criticise others for favouring things like a reasonable warranty period because "it might mean we don't get the new stuff as quickly" and "manufacturers might not want to sell stuff here any more". Suddenly, the threat of being cut off by their corporate masters drives them into action like nothing else. But even if some people cannot be persuaded to reshape their relationships with the companies they do business with, the situation is not as bad as it could be. In many countries there are fairly strong "consumer rights" movements, and even if people have been conditioned into not thinking about freedoms, they have been taught about demanding a fair deal when putting down their money. All the FSF and others need to do is to bridge the gap between the work going on in such areas and the arguably less "mercantile" nature of their existing campaigns.
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 22:10 UTC (Thu) by wmf (guest, #33791) [Link]
Of course, your own server could be in the cloud. Then you'd get freedom, good network connectivity, reliability, availability, etc.
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 23:11 UTC (Thu) by dlang (✭ supporter ✭, #313) [Link]
yes and no, you don't get the redundancy without the added administration overhead.
also you don't get the same security as running your own server as the hosting company has control of your data.
you can deal with these issues, but it costs more, and requires more administrative effort.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 8:59 UTC (Fri) by dgm (subscriber, #49227) [Link]
Not really. Redundancy and high availability _may_ be key for some production-critical services, but for the vast majority of private people (and some services in business) it doesn't matter that much. You cannot access the document with the list of your friends' birthdays, or that picture of the last holidays? How inconvenient! but not that important. You will tomorrow once you get home and fix the server.
I would argue that 90% availability is enough for quite a lot of uses (that's just one nine).
I don't know of anybody who prefers the cloud for availability, anyway. They usually prefer it because it's more convenient to have access to your data from anywhere, without having to carry it with you, and also as a means of communication and sharing.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 10:03 UTC (Fri) by nix (subscriber, #2304) [Link]
I want high availability when I'm away from home and can't get back to fix the server, or if the server is inaccessible in some hosting centre. But the thing to do there is to have panic_on_oops and reboot-on-panic turned on, and something else watching it and big-red-switching it (perhaps with X10) if it goes nonresponsive.
For servers running from home we have a bigger problem in the UK anyway: BT's new glittering '21st-century' domestic ADSL network appears to be rife with single points of failure which can take down all customers of one ISP or all ADSL customers in huge geographic regions without warning, and they do this sort of thing routinely for hours on end in the middle of the night, often without bothering to give any kind of warning first. So if we want a network link that stays up, we need to use a hosted server, and *that* then has the physical-accessibility problems mentioned above.
It's a bit of a sod really.
The Party of Gno (Linux Magazine) Posted Jun 19, 2010 21:45 UTC (Sat) by allquixotic (subscriber, #61671) [Link]
I don't think Free Software is incompatible with SaaS or cloud computing / cloud storage. In the case of Ubuntu One, it makes perfect sense to release the software (all of it, not just the core) that runs the system. They already did it for Launchpad; they can (and should) do it for Ubuntu One. Here's why.
All software that is written to require Internet connectivity to a remote server is of a special class of goods, because its supply/demand model is much closer to the traditional model of selling physical goods. The *effective use of* the software requires a non-replicable physical good, where the production costs of each unit (one unit being a server computer residing in a datacenter, with physical connectivity to the Internet) are linear. So you can sell portions of the resources of your company's server assets running your cloud software. Giving away your software is not the same as giving away resources on your server. People will still use your service, because the alternative is for them to purchase their own, non-replicable physical assets needed to run the software (i.e., their own computing assets). Unless they have the same level of redundancy, professional IT technology, server-grade hardware, and best practices that your company maintains, they will be at a distinct disadvantage, putting their data at risk. Customers will recognize this -- even consumers -- when data starts getting lost due to disk failure, or outages start occurring because some idiot hit a telephone pole and knocked out power.
Sure, you will be allowing competitors to also buy physical hardware and undercut you using your own software; but the fact remains that they actually have to *do* that. They'll need a big investment, maybe venture capital, to get the required resources to compete with you at scale. You are then challenged to go to vendors like Intel, AMD, Nvidia and IBM to acquire hardware that is cheaper or more efficient (more power per watt; more resources per dollar) than your competitors, so you can lower your prices and still make a good profit.
And then there's the fact that, if their data is already stored in *your* system, and their friends are all connected to *your* system, and their client programs are configured to connect to *your* system, it's quite a bit of effort for someone to eject from their social, financial and time investment in your system, and migrate to someone else's. And since you wrote the software, you are necessarily the first business to offer the cloud-based service surrounding it, which means the first users have to use your system, due to lack of alternatives. If you want to keep someone from competing with you while you're still developing the software, just don't distribute your free software -- to anyone -- until you have your own service online and available to the general public, at which point you can then distribute the source.
This makes the SaaS / cloud computing market look like almost any other commercial physical goods service. But the benefit to the public of releasing the software that runs your system under a Free Software license is obvious to anyone who reads LWN.
Other than these points, I like the direction the original article points in. I am a card-carrying associate member of the FSF, but I agree with the author that moving towards positive campaigns is the way to go. I am still a bit undecided about the author's friendliness towards commercialization, but the general concept of moving away from negative campaigns in favor of positive campaigns is a novel concept to me, that I think would propel the FSF into a much more positive (and popular) public image.
The Party of Gno (Linux Magazine) Posted Jun 20, 2010 20:38 UTC (Sun) by yuhong (guest, #57183) [Link]
Yea, I was going to say something similar about artificial scarcity too.
Please give me fewer articles, and more diamonds Posted Jun 17, 2010 21:44 UTC (Thu) by coriordan (guest, #7544) [Link]
Reads to me like "Here's a list of extra (really expensive) things I wish you would do."
Board members of FSF have done a GNU-like call for free web services:
Publishing an essay about the problems of cloud computing costs about $100. Replacing a load of web services would cost tens or hundreds of thousands. "Let's have less of the former and more of the latter" doesn't necessarily work :-/
Please give me fewer articles, and more diamonds Posted Jun 17, 2010 22:17 UTC (Thu) by wmf (guest, #33791) [Link]
But the FSF has no choice. The only path that leads to software freedom in the cloud happens to be really expensive, so the FSF should either find a way to raise the money or admit defeat and shut down.
Please give me fewer articles, and more diamonds Posted Jun 18, 2010 12:24 UTC (Fri) by coriordan (guest, #7544) [Link]
(I guess you're joking, but just in case...)
The point was, they're working on it. There's no time limit.
GNU was launched 27 years ago. They didn't give up, and the progress is now pretty impressive.
Please give me fewer articles, and more diamonds Posted Jun 18, 2010 20:34 UTC (Fri) by wmf (guest, #33791) [Link]
I wasn't joking, and I disagree with you again. I think at some point Google and Facebook will become so entrenched that there will be nothing us freedom fighters can do about them. GNU has made great progress on a Unix clone, but I'm not so optimistic about social networking; it looks like the proprietary walled gardens are improving faster than the open stuff so the gap is widening, not closing.
Please give me fewer articles, and more diamonds Posted Jun 18, 2010 16:50 UTC (Fri) by wingo (subscriber, #26929) [Link] There is another way: users can contribute their own compute power to form a "cloud" that respects privacy, freedom, and autonomy.I wrote about this issue a couple months ago. This is the topic of the upcoming GNU Hackers Meeting in the Hague next month. We are working on it.
Please give me fewer articles, and more diamonds Posted Jun 18, 2010 20:28 UTC (Fri) by wmf (guest, #33791) [Link]
Rewriting all the social Web 2.0 stuff that people want as AGPL (as Zonker proposes) will be expensive. Re-architecting all the social Web 2.0 stuff that people want into some encrypted P2P system powered by Guile will be really really expensive. It looks like you're biting off way too much IMO.
Please give me fewer articles, and more diamonds Posted Jun 19, 2010 11:12 UTC (Sat) by zonker (subscriber, #7867) [Link]
I'm sure people said the same thing about UNIX. And databases. And desktop software...
It's funny that FSF-proponents are willing to try the impossible (convincing the general public to stop using things like Facebook because they're non-free) but unwilling to try the possible but fairly difficult (replacing said services with FaiF services).
I guess it is much easier to proselytize than to plan and program...
Please give me fewer articles, and more diamonds Posted Jun 19, 2010 16:47 UTC (Sat) by sebas (subscriber, #51660) [Link]
But those groups are only different in focus and mission, not necessarily in personnel. That is, many "FSF" people *also* write code.
It's really two-fold, you need to tell people why Free services matter, just like Free software does, but you also need to write code. The FSF is about the "tell" side, telling them to write code instead of talking about it is beyond the point.
Please give me fewer articles, and more diamonds Posted Jun 20, 2010 2:05 UTC (Sun) by coriordan (guest, #7544) [Link]
I think we all (here) agree that it would be great if the success of GNU could be replicated (ASAP) for web services. But GNU wasn't an overnight success, and it still has to fight constantly to even get recognised. Here's a 1986 speech about why we need software freedom:
http://www.gnu.org/philosophy/stallman-kth.html
This would fall under your "proselytising", which you seem to view negatively. In 1986, there was no GCC, GDB, glibc, etc. etc. Richard worked on it, and published essays, and built support, and he coded along with those supporters, and 27 years later here we are.
He's doing the same thing again (although this time seeking programmers to help from the outset), and we're in the early years, and there's no big success yet. You can either help out, or you can be like the people of the early 80s who said that RMS's campaign was utopian.
(As for Windows7Sins - success is judged on how much effort it cost to get the results it got. Maybe FSF found someone who'd do it really cheap. Maybe it cost next to nothing, and whatever result it had is a bonus. Not every campaign has to change the world in order to be a positive force.)
The xkcd guy has drawn RMS a few times. What makes you think that RMS's proselytising wasn't the inspiration for that comic you loved?
Please give me fewer articles, and more diamonds Posted Jun 18, 2010 12:50 UTC (Fri) by andrel (subscriber, #5166) [Link] RMS didn't start a social movement by giving people articles, he did it by giving away diamonds. Lofty ideals backed by working code is the winning formula.
Please give me fewer articles, and more diamonds Posted Jun 18, 2010 15:29 UTC (Fri) by coriordan (guest, #7544) [Link]
And are we going to leave all the work to him again this time?
And in 2037, when there's a new threat and he's retired, will we call him out then too?
There's a moment where *we* have to start doing enough to defend our own freedom.
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 22:59 UTC (Thu) by BrucePerens (guest, #2510) [Link] The cloud is not so attractive that folks who believe in free anything, including their own freedom, have to use it. It's still incredibly cheap to rent an entire server with great reliability and 24 hour hands and eyes, and use that. And it's easier to manage than a cloud.And I knew there was a reason that I didn't put more stuff on facebook :-)
The Party of Gno (Linux Magazine) Posted Jun 19, 2010 11:15 UTC (Sat) by zonker (subscriber, #7867) [Link]
Yeah, Bruce - sure. It's easy for you.
If the only folks you care about are other people like you, then ... great. Enjoy your freedom while the general public keeps using non-free software and becomes further entrenched in non-free software. I'd like to see free software for everybody, not just the people who happen to have the technical chops to maintain their own server + services.
Not to mention the financial ability to do so. Lots of folks can't afford $120 and up per year to rent a server of their own.
An alternative to web services (was: The Party of Gno (Linux Magazine)) Posted Jun 18, 2010 1:03 UTC (Fri) by cesarb (subscriber, #6266) [Link]
An alternative to web services could be distributed services.
For instance, for file storage, you could have several people each store a small piece of an encrypted copy of your file, with several extra pieces for redundancy (using FEC). This is what Freenet does, but if you do not need the heavy anonymity layer it uses, you could use simpler algorithms and gain some extra performance.
They would not all be in low-powered desktops with metered network connections; some people would set them up on real servers (much like the use of seedboxes on BitTorrent).
An alternative to web services (was: The Party of Gno (Linux Magazine)) Posted Jun 18, 2010 7:56 UTC (Fri) by michaeljt (subscriber, #39183) [Link]
> An alternative to web services could be distributed services.
This has been on my mind since the articles on free Facebook replacements: in this case a peer-to-peer based Facebook alternative. You push the bits of your profile you wish to friends who are online and whom you trust with those parts, and they can push them further to other people they think you will trust with them. (Makes for an interesting trust issue, but in the end this and more is possible with Facebook as it is.) You loose the guaranteed always-on-ness of course. Perhaps this could even be marketed as a feature though - social profile browsing instead of anonymous browsing!
An alternative to web services (was: The Party of Gno (Linux Magazine)) Posted Jun 18, 2010 8:49 UTC (Fri) by dgm (subscriber, #49227) [Link]
I think this is really the way to go. Far more useful than the "semantic" stuff some people have been pushing (I'm looking at you, KDE).
An alternative to web services (was: The Party of Gno (Linux Magazine)) Posted Jun 18, 2010 16:53 UTC (Fri) by wingo (subscriber, #26929) [Link]
A triple-store, RDF or no, probably still has a place in such a system. It's the only way I know of to have an extensible relational database (for some meaning of those words).
An alternative to web services (was: The Party of Gno (Linux Magazine)) Posted Jun 18, 2010 13:52 UTC (Fri) by mattdm (subscriber, #18) [Link]
http://www.foaf-project.org/, started in 2000.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 14:38 UTC (Fri) by sepreece (subscriber, #19270) [Link]
"and DRM has been rejected by the general public."
Really? That must be why manufacturers have found it impossible to sell HD TVs, cable and satellite companies have been unable to sell decoder boxes, and nobody ever bought those DVD things. Not.
The general public has rejected DRM in a couple of specific niches, but broadly accepted many other applications of DRM as a cost of access to content.
That's not to say I wouldn't prefer to have DRM-free content (and most of the general public probably would, too), but it doesn't rise to the level of a barrier to purchase for most people, in most applications.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 15:35 UTC (Fri) by davecb (subscriber, #1574) [Link]
It's a slight overstatement, but it's correct in the main. One of my old employers researched it heavily in the Windows 2.0 days, and decide it sounded attractive, but was actually snake-oil and a large customer support cost whenever it failed, as the default was to deny the customer the use of his purchased product.
We found:
--dave
Dongles live. Posted Jun 18, 2010 15:58 UTC (Fri) by dmarti (subscriber, #11625) [Link] Dongles live on in vertical applications, from health care to retail to CAD. These applications typically don't get installed by end users, but by partner companies. Looks like the software companies don't trust their partners not to install 20 seats, bill the client for 20, and report 10 to the vendor.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 16:39 UTC (Fri) by sepreece (subscriber, #19270) [Link]
Yes, software is one of the niches I referred to. DCopy protection didn't work out so well for video cassette tapes, either.
On the other hand, DRM protection associated with content is pervasive and growing. I don't claim people *like* it, just that they're willing to accept it, especially when it largely doesn't get in the way of what they want to do with the content.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 17:56 UTC (Fri) by lmb (subscriber, #39048) [Link]
There's a huge difference between having DRM on your content or DRM on your software licenses, too.
The Party of Gno (Linux Magazine) Posted Jun 18, 2010 18:38 UTC (Fri) by cesarb (subscriber, #6266) [Link]
Even if it is not a barrier to purchase, all else being equal, DRM-free content has better fitness than DRM-restricted content.
For instance, while with DRM-restricted content you might have to carry several large physical round objects with you to wherever you might want to use the content, with DRM-free content you could just duplicate it everywhere you might want to use it, carry a copy of it with you within a few very small rectangular objects, and even get a quick copy from a friend if you forgot yours. DRM-free is clearly superior here, and it becomes even more superior when new uses which were not tought of by the DRM makers are found (suppose for instance you want to classify your movie collection by the average color of each movie; the DRM will probably not let you).
Not to mention that, in the particular case of free (as in freedom) content, the amount of available content tends to only rise. If I have $50, I can have all the free content ever created plus $50 of non-free content*. Given time, the amount of available free content (which is of course DRM-free) will be large enough that the annoyance of having to put up with DRM starts making less and less sense ("You mean I cannot watch this movie I just bought in my high-quality computer monitor, which is what I use to watch every single one of the other movies I have?").
The extra flexibility of DRM-free content means it gets used more, which means it brings more value to the user, who then will seek more DRM-free content (since after all it is better value). (But remember that this is "all else being equal".)
* Simplifying out the costs of network access, disk space, and so on. If you include them, it looks even more unbalanced; it is either $50 of non-free content or $50 to pay for enough bandwitdh and disk space for a lot more content, so you might decide not to buy the non-free content after all and use the money instead to pay for more bandwidth and disk space.
The Party of Gno (Linux Magazine) Posted Jun 23, 2010 14:32 UTC (Wed) by kirschner (subscriber, #62102) [Link] (I also posted this comment on my blog.) [...] I fully agree with Joe's point that it is important to do positive campaigning. This is especially true in Europe. That is why we -- FSF's European sister organisation -- always tried to formulate our message in a positive way. Only with positive campaigning FSFE is able to be recognised by European politicians and in return get their recognition, e.g. with this year's Theodor Heuss Medal. But also the FSF already had positive campaigns. Here some recent examples from the FSFs:
There are more examples of positive activies on FSFE's activity page, and there are also more examples on FSF's campaign page. Although there are a lot of "Gyes" campaigns out there, the problem is to get media coverage with positiv campagins. Beside that sometimes you have to critise bad things. For example Digitial Restriction Management (DRM) is bad for software users. The FSFs work that users have to control over their computing. So we have to defend them and say "No" to DRM, even if we do not have a solution how to make money with music and videos in the future. It is important for the FSFs to keep a firm stand on users software freedom. --Matthias Kirschner
Gno Thanks Posted Jun 17, 2010 20:15 UTC (Thu) by dowdle (subscriber, #659) [Link]
Thanks for the advice, but gno thanks. Why? Because people have been trying to tell RMS to compromise for years and years and years. The Free Software movement didn't get to where it is by compromising.
Nothing is wrong with negative campaigns. They are more about education than gaining converts.
Provide alternatives? That's what the FSF has been trying to do since day one and continues to do. Their abilities are certainly limited and they can't do it all... and they do encourage others to make alternatives.
So what if some people have Macs at Linux / Free Software / Open Source events? Yeah, I do find them mildly annoying myself but what are you going to do about it?
Free Software is in the best shape it has ever been even though more people seem to refer to it as Open Source. The Free Software Foundation's work will never be done.
There is absolutely NO WAY the FSF or anyone else is going to compete with Apple on things like the iPod and iPad. Why do I say that? I'm referring to the advertising dollars the spend with their media saturation. We can't compete with that. So, having a goal of completely eliminating proprietary hardware, software and services is unrealistic to begin with, and it'll never happen, but it is still a good goal.
I myself am not 100% pure but I do make an effort where there are alternatives... and I do promote Free Software... and I don't encourage the use of proprietary things... and I contribute as best as I can to the free software projects I care about. Here's to RMS! Keep up the good work.
Gno Thanks Posted Jun 17, 2010 21:50 UTC (Thu) by trasz (guest, #45786) [Link]
It's not about advertising - Open Source has lots of it, and it had even more in the past - but about creating something that is actually useful. It succeeded on servers and in embedded, but completely failed when it comes to desktops.
Gno Thanks Posted Jun 17, 2010 22:29 UTC (Thu) by NAR (subscriber, #1313) [Link]
On servers and in embedded the users are tech-savy people (server administrators, integrators, application developers) while on the desktop the users tend to be less tech-savy. The first group appreciates the hackability/extendibility that free software provides, while the second group couldn't care less.
Gno Thanks Posted Jun 17, 2010 22:38 UTC (Thu) by dowdle (subscriber, #659) [Link]
I'm just saying that the advertising that Apple does, and the unpaid placements they get (like on The Colbert Report for example)... really helps create a want and drives sales. I'm not seriously suggesting that free software should spend the tens/hundreds of millions on advertising that Apple does.
So, you say that free software already advertises? Really. Where might that be? I have yet to see a single commercial on television, radio or in the newspaper. The only ones I remember are the IBM Linux commercials from years ago. Oh, and there was that full-page ad the Mozilla folks took out some time ago. Other than that, nada.
I guess I have seen some ads for some free software on the tech savy websites I visit, but I wouldn't really count that.
Gno Thanks Posted Jun 19, 2010 13:58 UTC (Sat) by mmcgrath (subscriber, #44906) [Link]
> I'm just saying that the advertising that Apple does, and the unpaid placements they get (like on The Colbert Report for example)... really helps create a want and drives sales.
So does having a highly functional operating system. If OSX was free, or the Linux desktop was paid only, we'd see an even more major shift to OSX then we do now. It's because we have all the people that care about freedom over function and unfortunately for us, it's a very tiny minority of users. The larger majority prefers function and the Linux desktop offerings just don't compare to the non-free experiences.
Gno Thanks Posted Jun 17, 2010 22:52 UTC (Thu) by Zack (guest, #37335) [Link]
>It's not about advertising - Open Source has lots of it, and it had even more in the past - but about creating something that is actually useful. It succeeded on servers and in embedded, but completely failed when it comes to desktops.
If I read the above correctly, you are saying that in spite of adequate advertising, Open Source has not succeeded on the desktop because it failed to produce useful software.
Is that how you intended it to be read ?
Gno Thanks Posted Jun 18, 2010 7:41 UTC (Fri) by trasz (guest, #45786) [Link]
It succeeded in creating various useful pieces, but failed to create something useful as a whole desktop for the average user. Actually, not just for the average user - just look at the number of technical people that moved to Macs.
Gno Thanks Posted Jun 18, 2010 9:24 UTC (Fri) by seyman (subscriber, #1172) [Link]
Useful or superior to OS X ?
I'll argue the Open Source desktop has reached the former (it does everything I want it to do and, for most people, reading proprietary formats looks to be the only problem).
Gno Thanks Posted Jun 18, 2010 10:06 UTC (Fri) by Zack (guest, #37335) [Link]
>Actually, not just for the average user - just look at the number of technical people that moved to Macs.
Well, it's been a while. A lot of "us" are now in their 30s or pushing 40. A looming midlife-crisis makes one buy things that are hip, easy to understand and "connects" you to younger people, whilst giving the semblance you've been successful in the meantime.
Gno Thanks Posted Jun 18, 2010 10:47 UTC (Fri) by boudewijn (subscriber, #14185) [Link] Grin... Well, the one thing that connects me (40 years old and busy in free sofware since I started writing a GPL mail/usenet client using Visual Basic when I was 24) to young people is working in a free software project. Working together with students, mentoring them in the Google Summer of Code, drinking beer during sprints. It's fun, rewarding and definitely beats being stuck with a laptop with one mouse button and no delete key! And being the maintainer of a project that gains new contributors and enables artists to do cool stuff definitely makes me feel succesful!
Gno Thanks Posted Jun 18, 2010 0:22 UTC (Fri) by agajan (subscriber, #7859) [Link]
> Nothing is wrong with negative campaigns. They are more about education than gaining converts.
Yes, the FSF campaigns are crucial for educating the public about IP freedom.
I work at a medium sized college. I posted some Windows7Sins fliers around campus. Afterwards, I got the impression that the fliers didn't go over well with the IT staff that support Windows. I don't know if it is possible for the campaigns to be effective without offending some people.
Gno Thanks Posted Jun 18, 2010 7:56 UTC (Fri) by trasz (guest, #45786) [Link]
Except that Windows7Sins is FUD comparable only to "Get The Facts" campaign, which makes the "educating" part a little dubious.
Gno Thanks Posted Jun 18, 2010 7:59 UTC (Fri) by michaeljt (subscriber, #39183) [Link]
>> Nothing is wrong with negative campaigns. They are more about education than gaining converts.
> I don't know if it is possible for the campaigns to be effective without offending some people.
Perhaps the answer is to campaign positively, focussing on what free software does better (the FSF do this too of course, but I always feel they focus too much on the benefits for software developers, preaching to the converted so to speak). This is probably also cheeper than trying to match proprietary feature for feature.
Gno Thanks Posted Jun 18, 2010 0:51 UTC (Fri) by davide.del.vento (guest, #59196) [Link]
> Because people have been trying to tell RMS to compromise for
Disclosure. I am a FSF member and I strongly support it, although I don't have any decision making role with them.
They do compromise: take LGPL and Affero. They decided to create both as a compromise, which was supposed to do a better job in achieving the final goal. In fact, a no-compromise would have been no LGPL at all and the inclusion of the Affero clause directly into GPLv3. The latter didn't happen because FSF listened to companies that wanted to do cloud computing and not be forced to disclose the source code. Unfortunately this is what we are now are crying about.
Gno Thanks Posted Jun 18, 2010 2:45 UTC (Fri) by yuhong (guest, #57183) [Link]
The first GPLv3 drafts allowed a Affero-like provision to be added as an *optional* clause.
Gno Thanks Posted Jun 18, 2010 20:32 UTC (Fri) by davide.del.vento (guest, #59196) [Link]
> The first GPLv3 drafts allowed a Affero-like provision
I know and that was for flexibility and compromise too. The no-compromise would have been GPLv3 with mandatory Affero provision, always on.
Gno Thanks Posted Jun 18, 2010 7:51 UTC (Fri) by michaeljt (subscriber, #39183) [Link]
> They do compromise: take LGPL and Affero.
On that note, I stumbled across this - http://www.gnu.org/philosophy/selling-exceptions.html - yesterday
LGPL-like license Posted Jun 25, 2010 23:38 UTC (Fri) by spitzak (guest, #4593) [Link]
Can you tell me if the FSF recommends or has defined a license that is basically "what people think the LGPL is". I would dearly love to be able to stick a TLA on my software and have everybody know what it is, rather than the current kludge of saying "GPL + linking exception".
Just to be clear, what I am looking for is that somebody can make a program using the licensed software in any way, as long as they don't modify the software itself, and do anything with that result, such as sell it without source. However if they *modify* the software, they must include the source code of their modified version (they can then make a closed-source program using the modified version). There may be a few other rules to avoid tricks: mostly that the modifications must be stand-alone and result in a piece of software of equal reusability as the original. In addition you must be able to combine the software with GPL/LGPL licensed software, in which case the result is GPL/LGPL, this requirement makes the CDDL unuseable.
LGPL-like license Posted Jun 26, 2010 4:09 UTC (Sat) by foom (subscriber, #14868) [Link] Huh, I always thought the LGPL was to protect the users, by allowing them to fix bugs in the software.It thus requires that you allow the replacement of the LGPL-licensed work in your application with a newly compiled version. Makes sense to me... Now, a name for the license you want which abandons the users would be nice, but I would've omitted the "what people think the LGPL is" bit of your comment. :)
LGPL-like license Posted Jun 29, 2010 1:18 UTC (Tue) by spitzak (guest, #4593) [Link]
I have found the LGPL to be a pain for developing open-source libraries.
On the assumption that I actually want to encourage the use of my library by allowing closed-source programs to use it, the only way to do this with the LGPL is to make a shared library (so that the end user can, supposedly, replace it with a new version). Not only does this make it a pain to distribute the application, a much more odious problem is now I am forced to be binary-compatible, which I certainly do not want to do for an under-development library (in reality I don't bother with binary compatability but this means that nobody can actually replace it in a program anyway, thus defeating the entire purpose of this LGPL requirement).
The FSF does support licenses like I want, they typically call it "GPL with a linking exception". What I would like is a SHORT name for this, and official legalese code. The lack of this is probably the main reason for OSS license proliferation because everybody writes their own.
Unlike the CDDL and some others, I want to allow my code to be used by anybody including GPL projects. And I like things like the reverse-engineering clause, etc, from the GPL, which are often deleted as well. I just want end users to be free to write any program using my code and distribute it any way they want, while being unable to make a closed-source modification to my code.
LGPL-like license Posted Jun 29, 2010 4:43 UTC (Tue) by nix (subscriber, #2304) [Link] Not only does this make it a pain to distribute the application, a much more odious problem is now I am forced to be binary-compatibleAs you pointed out in the next line, no you aren't. There's nothing stopping you setting the soname to .0, or to the program version, and just dumping binary compatibility (although if a library interface is changing *that* fast once it's in active use and it's not a compatibility library like gnulib, I'd have to wonder 'why'?) This is not exactly rocket science. Hell, it's how half the shared libraries on your average Linux system are 'versioned' (the ones that are associated with only one package). Distributing static libraries is bad for a vast number of reasons beside the lack of easy LGPL support; notably, you lose all backward-compatibility guarantees for glibc, so you need to ship a new version for every glibc minor version. Do you really do that? It seems like a hell of a lot of work for very little gain.
LGPL-like license Posted Jun 29, 2010 18:05 UTC (Tue) by spitzak (guest, #4593) [Link] Distributing static libraries is bad for a vast number of reasons beside the lack of easy LGPL support; notably, you lose all backward-compatibility guarantees for glibc, so you need to ship a new version for every glibc minor version.Huh? This would also apply to programs. You are saying that libraries must be recompiled for every minor change to glibc, but for some reason programs don't? I think possibly the confusion is that you think I am talking *about* glibc. I am not. I am talking about MINOR libraries of useful funcitions, used by approximately ONE program on a system. A shared library is a waste of time, and is counter productive to me. And I find it hard to believe you think I should be required to finalize the API before I let anybody test my software!
LGPL-like license Posted Jun 29, 2010 21:02 UTC (Tue) by nix (subscriber, #2304) [Link]
No. glibc makes no guarantee of binary compatibility except on the level of ELF executable objects (that means executables and shared libraries, which are very nearly identical beasts in many ways with different things conventionally done to them). Specifically, there is no guarantee at all for .o files, and that means for .a files too. The only .o files you normally see on a Linux system are a few startup/shutdown files in /usr/lib which come with glibc (and so the lack-of-guarantee is irrelevant) and GCC (and these contain no glibc references).
That is, you cannot compile against one version of glibc and link against another, ever. That's not what the interface guarantee allows. What you can do is compile and link against one version and then *dynamically* link (i.e. *run*) against a later version, safely, even if functions change in ways that might disturb your program or structures like 'struct stat' change size.
There are two reasons for this.
Firstly, .a files cannot make reference to specific versions of versioned symbols: that work is done by the linker. As most of the glibc interface-compatibility maintenance scheme depends on versioned symbols, this means that scheme cannot operate for .a files at all. Also, there are a good few functions in glibc where the old version depends on the header files the code was compiled with doing one thing, and the new version depends on the header files doing some other thing. Compile against one version of glibc and link against another, and pain results.
Secondly, glibc depends (because it must) on determining the version of (potentially versioned) structures bound into a program by means of linking a .a file -- libc_nonshared.a -- into that program and having your program pass a function in that file a version number tag defined in the headers at compile time. Compile against one version and link against another, and the tag may be wrong (hell, the structure may have different sizes in different parts of your program: obvious disaster). (See http://sourceware.org/glibc/wiki/Development/Versioning_A... for much more information.) (This one will not have bitten you so far, because none of these versioned structures have ever changed definition yet. But they might, and if they do, little things like stat() will rely on them.)
So you can ship .a files if you like, as long as they never use any functions in glibc that take any structures as parameters whose definitions might ever possibly change, and never call any functions from glibc which might ever be versioned (that would be all of them). I hope your useful functions never use glibc. ;)
Note that you don't have to finalize the API. You just have to *make a shared library*. You can give it a useless soname -- increment it whenever you release your program, soname it after the date, anything -- and change the API whenever you like, and laugh at anyone else who DT_NEEDs it, but you cannot ship .a files and expect them to work, at all, across glibc upgrades. The X people used to do as you do -- experimental extensions got .a files only, so they didn't have to give the shared library an soname -- and glibc upgrades broke them several times as I recall. Eventually X.org shifted to using shared libraries for everything as part of the modular rework, and just using an soname of {blah}.0 until the interface stabilized. It was a good decision.
LGPL-like license Posted Jun 30, 2010 0:08 UTC (Wed) by spitzak (guest, #4593) [Link]
I think I might have been misleading as to what I wanted to do. I am not shipping .a files or anything else. I want to ship the *source* for my software, but allow somebody to use my source in a closed-source program, provided they do not modify the source files themselves, but only add other files to them. Most likely they will compile my stuff to a .a file but they will have no trouble making a new .a file if they need to, and what they distribute will be an executable, not a library.
What I am looking for is a short name and legal description, blessed by the FSF, for what I have been calling "GPL with linking exception".
LGPL-like license Posted Jun 30, 2010 19:25 UTC (Wed) by foom (subscriber, #14868) [Link]
> the only way to do this with the LGPL is to make a shared library (so that the end user can, supposedly, replace it with a new version)
That's not true at all. I don't believe shared libraries existed (or at least, they weren't yet common) when the LGPL was introduced.
LGPL: 1991.
The accepted way to follow the LGPL without using shared libraries is for the closed-source program to distribute the .o files you can pass to a linker, along with the .a library, to create a final executable. I understand that you don't want them to have to do this, but that's different from saying it's impossible or not allowable.
As the LGPL says:
LGPL-like license Posted Jun 30, 2010 20:50 UTC (Wed) by farnz (guest, #17727) [Link] It's also possible for the closed source application to do a partial link; take the bunch of .o files that make up the closed source application, link them together with appropriate options (see GNU ld's --relocatable option, if that's your linker of choice), and strip as much of the symbol table as possible. The resulting .o file gives away very little about the original closed-source code (on a par with a closed source app using the LGPL libraries as shared libraries), but is still good enough for linking against many other .o files (those from a LGPL library, for example) to get a final binary.
Gno Thanks Posted Jun 18, 2010 11:14 UTC (Fri) by zonker (subscriber, #7867) [Link]
dowdle, please point out in the article where I suggest that the FSF compromise. I don't think the FSF should budge on its principles. I think the FSF needs to wake up and understand the real world, and understand how people outside of the "movement" think about their computers.
And, actually, the free software movement *did* get to where it is by compromising. It bootstrapped on proprietary systems. It has compromised the GPL by inventing the LGPL so that proprietary software could be written on top of free systems.
But I'm not asking the FSF to compromise. I'm asking it to look forward and think about how to create free services for cloud computing the same way that it created a free replacement for unix. Even if FSF doesn't do all the work or even act as a focal point, it would be nice if the FSF had a vision for free software that extends to SaaS and other "cloud" computing.
There is a great deal wrong with negative campaigns. In particular, the ones being run by the FSF currently do nothing to help - they're not persuasive to people who aren't already convinced. They're actively harmful outside of that circle, because they make it sound like free software supporters are a bunch of radical clowns who are entirely out of touch with how people want to work with their computers.
They may not be able to do it all, but it seems the organization has stuck a line in the sand and said "ok, we've done the unix replacement bit, now we're going to tell people computing should stop here."
Gno Thanks Posted Jun 20, 2010 2:43 UTC (Sun) by coriordan (guest, #7544) [Link] > I'm asking [FSF] to look forward and think about how to create free services for cloud computing the same way that it created a free replacement for unix. Even if FSF doesn't do all the work or even act as a focal point, it would be nice if the FSF had a vision for free software that extends to SaaS and other "cloud" computing. FSF published the AGPL in 2002, the Franklin Street Statement in 2008, essays on javascript and SaaS recently:
The Party of Gno (Linux Magazine) Posted Jun 17, 2010 20:46 UTC (Thu) by jdave23 (guest, #27160) [Link]
It's an editorial opinion, obviously, and one which I don't agree with.
The author admits that his observations are his perceptions; things "seem" a certain way. There are too many assumptions behind it, well argued or otherwise, and although he has determined the FSF approach "isn't working" it "seems" like a non-issue to me.
He's not getting it Posted Jun 17, 2010 22:45 UTC (Thu) by BrucePerens (guest, #2510) [Link] Nokia n900 seems to have gotten a great lot of attention, even though iphone and droid are built for the mass appeal. It's too bad that the whole phone platform in the n900 is closed, but the folks who are buying it are doing so for what is open.What he isn't getting is: the fact that this is a minority audience isn't such a problem, they're a very high quality audience. They're still the leading edge, and there are more of them now than ever before. The fact that lots of people are running mixed platforms only means that changing what system people run isn't the same as changing the people.
He's not getting it Posted Jun 18, 2010 1:23 UTC (Fri) by drag (subscriber, #31333) [Link]
The solution is just to produce superior software.
If 'Free Software' is to succeed as 'Free software' it needs to produce superior product and out compete.
For example:
Linux distributions are incompatible with one another. Ensuring binary compatibility and a unified package management and so on and so forth is a critical step in growth for Linux, but it has never been done.
Package Management is the solution? NO! Package management is the bandaid and is not the solution! The longer people fool themselves into beleiving that package management is a good substitute for good system design the bigger the problem is getting.
Apple implemented package management system for 'iOS', but unlike Linux they made developer-friendly environment and a developer-friendly OS.
-------------------------------------------
How many useful applications are in Debian's repositories?
I am not talking about pure packages, no *-dev packages or libraries or any crap like that. I am talking about real swear-to-god applications that people can use for different things.
10 thousand? 20 thousand?
How long has Debian been around? 17 years?
Now compare this to iPhone: 3 years and ~140,000 applications.
Android: One year = 40,000+ applications. Probably will hit 100,000 by the end of the year.
Sure sure these are smaller applications and many are nothing more then trivial fart games or whatever, but the pace of development of those systems far far outstrip anything ever seen on Linux prior to Android.
--------------------
What is these platforms doing so well that GNu/Linux is utterly failing at?
One thing for certain is that no end user ever had to recompile a application from scratch to get it to work on their phone...
He's not getting it Posted Jun 18, 2010 5:54 UTC (Fri) by k8to (subscriber, #15413) [Link]
drag, you're smarter than this.
If you're going to claim stuff like "binary compatability is what we need" you need to actually address all the *real* problems that it brings.
He's not getting it Posted Jun 18, 2010 11:02 UTC (Fri) by drag (subscriber, #31333) [Link]
> If you're going to claim stuff like "binary compatability is what we need" you need to actually address all the *real* problems that it brings.
If I understood the problem better then I'd be able to do that.
I know that the oil spilling into the gulf is a problem, but I don't know how to fix it. But it's still a problem, right?
It's the same thing with applications on the Linux desktop. Packages should be done by upsteam, as part of normal build process, and collected by distributions. Those packages should be installable on any Linux desktop I happen to come across. It's probably acceptable that newer packages will have a cut-off for how old the system they will run easily on and there is going to be a cut-off on how new of a system that will be supported.
I know that this is how package systems should really work.
People here tell me that diversity is good, choice is good, and all that. Which is, generally, something that is easy to agree with.
But I just fail to see the advantage that not being able to install Ubuntu packages on Fedora brings me, as a end user. I know that they are using just about the same kernel versions, similar gcc versions, similar libc libraries, simple gtk/qt libraries, upstart, etc etc. I bet that for the vast majority of packages in Ubuntu or Fedora's latest releases everything less then a year difference in terms of upstream releases.
I don't understand everything, obviously. But what is so much better about Fedora's C++ ABI over the one that Ubuntu has? What is the big win in one vs the other? Is it really a problem?
--------------------------------
I know that I can download the Blender 2.5 alpha release in a tarball and it 'just works' in Debian. But if I try to install packages built for Ubuntu it won't. http://www.graphicall.org/builds/
What is the big win in Ubuntu or Debian that requires users to deal with these sort of incompatibilities between the systems?
In this case the package management system is a huge lose for me. In Windows I can download the zip release and it'll 'just work'... And the tarball binary version will work also. I know that for Debian the the packages for Ubuntu won't really, at least not any of the half a dozen I tried. Apt-get itself is very unhelpful in this regard.
-----------------------------------
Here is another example:
Opera browser...
If you go to Opera's website they offer package formats for all sorts of different Linux versions.
http://www.opera.com/download/
And one may think.. Aha! Opera cares about the end users so they are willing to put the work int producing packages for all these different systems.
Well on more careful examination it quickly becomes apparent that except for one system that uses QT4, all the binaries supplied in the packages are _exactly_the_same_. There is some differences in code paths and in package metadata, of course.
But as far as the actual executable goes they are all using the same exact _dynamically_linked_ Linux executable!
And again the tarball works just fine on my system.
--------------------
Ditto for Firefox...
Yeah they include a lot of libs with the release, but if you only use the Firefox-specific libs for things like libxul and use system libs whenever possible it'll still work.
---------------------
So I know what I saying is certainly possible. Blender, Operah, and Firefox are far from simple programs and yet they can produce binaries that will work across multiple Linux distributions. Sure there is probably extract work that goes into making this happen and it's probably not trivial, but it's certainly possible.
LSB? (was: He's not getting it) Posted Jun 18, 2010 11:28 UTC (Fri) by cesarb (subscriber, #6266) [Link]
Isn't the LSB designed to be such a common packaging format? I have never actually tried it, but from what I read most of the work is already done.
LSB? (was: He's not getting it) Posted Jun 18, 2010 12:27 UTC (Fri) by sorpigal (subscriber, #36106) [Link]
LSB mandates RPM, which is not my idea of how compatibility should be assured.
LSB? (was: He's not getting it) Posted Jun 18, 2010 13:08 UTC (Fri) by seyman (subscriber, #1172) [Link]
> LSB mandates RPM, which is not my idea of how compatibility should be assured.
This isn't exactly true. LSB mandates RPMv3 or an installer which is itself LSB-compliant.
LSB? (was: He's not getting it) Posted Jun 19, 2010 7:41 UTC (Sat) by AndreE (subscriber, #60148) [Link]
Not to mention that the actual format of the package is actually not even the problem
We still have to consider naming schemes and separation guidelines, and other packaging issues (e.g. Fedora's non-bundled system libs).
Just getting everyone to play RPM doesn't help. Just Try installing a fedora rpm on a suse or mandriva box to find out.
LSB? (was: He's not getting it) Posted Jun 18, 2010 13:03 UTC (Fri) by cortana (subscriber, #24596) [Link]
Are there any applications that actually make use of the LSB package format? If I buy a game for Linux, today, I'll get an executable shell script that will perform some arbitrary actions to drop the game on my filesystem, register it with the freedesktop.org menu, etc. Not an RPM.
LSB? (was: He's not getting it) Posted Jun 19, 2010 4:45 UTC (Sat) by da4089 (subscriber, #1195) [Link]
LSB is a set of minimum guarantees for things an LSB-compliant system will provide.
But it's so minimal as to be useless. Last time I tried to package an LSB-compliant application, it didn't provide libxml, OpenSSL, zlib, etc.
So your application has to package all of that stuff internally.
And then, in 6 months (or 18 months for enterprise folks), the whole crop of sexy stuff (think of changes like Xorg, PulseAudio, DBus, etc) that's available for native packages isn't available to your LSB application.
It's a well-meaning, but busted model.
He's not getting it Posted Jun 18, 2010 16:31 UTC (Fri) by jwakely (subscriber, #60262) [Link]
> But what is so much better about Fedora's C++ ABI over the one that Ubuntu has?
They use the same one, and the C++ shared libraries are backward compatible and have been for about 6 years.
I suspect all that's done by Opera and Firefox is to compile and link on a non-bleeding edge OS, so that the shared lib dependencies are suitably old that they can be found on any suitably recent system. It's not rocket science.
iPhone =/= Debian app Posted Jun 18, 2010 9:39 UTC (Fri) by tialaramex (subscriber, #21167) [Link]
The _vast_ majority of iPhone apps are basically analogous to web pages, not Debian applications.
Remember when the web "took off". Every double-glazing sales company, rat catcher and accountant wanted a web site right away. They'd all call up some spotty kid who happened to know a bit of HTML and he'd make them their very own web site for a reasonable price. And then they'd go pay the sign painter to write "Threepalmvalleycentralheating.obscureisp.com" on the side of their vans.
The spotty kid would eventually figure out that he's paying himself minimum wage because he's lousy at business finance and he's no longer excited about making web pages, and the rat catcher would figure out that the web isn't bringing in much business. But meanwhile business was "booming".
That's where the iPhone is now. Countless people have breathlessly told me that they're now an iPhone developer, and hey that means they can buy a new iThing every couple of months and see it as a business expense, and oh, the money? Well they don't have the money part figured out yet. In 18 months most of them will be back working for the Man. But meanwhile they're generating thousands of me too apps, and more importantly, tens of thousands of very niche special purpose apps for just a handful of people, which outside the Reality Distortion Field would have just been web pages (especially since they often don't work without Internet access)
Imagine if there were debian packages like view-lwn-kinda-like-a-web-browser.deb and view-phoronix-sorta-like-a-web-browser.deb would you be more impressed and believe Debian was going to explode in popularity? Or would you think its developers had gone crazy? For Debian thousands of worthless packages is a cost, so they're against it. For Apple every "Hello, world" app is money in the bank.
iPhone =/= Debian app Posted Jun 18, 2010 13:52 UTC (Fri) by pboddie (subscriber, #50784) [Link]
Once again, someone posts a great comment that succinctly dissects the hype around something in a way that no-one else has managed to do! In this case the bubble is iWhatever applications, and Apple, like the various banks and venture capital houses in the dot-bomb era, is the player skimming off the cream while the champagne is still flowing.
iPhone =/= Debian app Posted Jun 19, 2010 4:40 UTC (Sat) by bartszyszka (guest, #67794) [Link]
I don't think your analogy to websites during the dot-com bubble works exactly. During the dot-com bubble, people gave everything away for free without a business model, running on borrowed time until the investors' money ran out. With iPhone apps, the actual customer pays for what the developer makes (with Apple taking a cut, they are a business afterall). I, as a consumer, pay $9.99 for an iPhone app and the developer gets a majority of that money and developers make real profits. That's a big deal. There are very few websites people are willing to pay even a dollar for.
iPhone =/= Debian app Posted Jun 19, 2010 8:11 UTC (Sat) by k8to (subscriber, #15413) [Link]
Sure, but the developer paid several hundred developers this year, and so did you, to Apple. And both of you are going to pay several hundred dollars again, when the next revs comes out in another couple of years. That's a bigger deal.
iPhone =/= Debian app Posted Jun 19, 2010 8:49 UTC (Sat) by asaz989 (subscriber, #67798) [Link]
Spot-on article. On a side note, this is what I think is behind Apple's promotion of web standards - they don't think that the iPhone-app craze is sustainable, if only because people now have non-Apple smartphones and developers will not put in that much effort for a smaller bite of the mobile market. Apple wants the iPhone to still have consumer appeal if and when sites realize that web pages should be designed for web browsers (perhaps with a touch-screen version or stylesheet), and not for any specific platform.
iPhone =/= Debian app Posted Jun 19, 2010 16:21 UTC (Sat) by job (guest, #670) [Link] Hardly. Apple's appreciation of web standards is because they've been on the wrong side Internet Explorer, and that hurt.
Debian is not a good comparison (was: He's not getting it) Posted Jun 18, 2010 11:21 UTC (Fri) by cesarb (subscriber, #6266) [Link]
Debian's repository is not a good comparison target. At least for "main", the packages have to meet a fairly strict set of rules, and need to be (at least in theory) maintained by at least one Debian developer. While for "non-free" (and "contrib") the rules are more relaxed, they (and the need for one Debian developer) are still enough to reduce the amount of programs which are packaged.
A better comparison would be with all packages at http://freshmeat.net/; but even then, it will miss a large number of the non-free software available for Linux (especially the more specialized software).
He's not getting it Posted Jun 18, 2010 16:19 UTC (Fri) by ballombe (subscriber, #9523) [Link]
> What is these platforms doing so well that GNu/Linux is utterly failing at?
So well ? If you remove free software from them, they become useless bricks.
History editing Posted Jun 18, 2010 14:18 UTC (Fri) by pboddie (subscriber, #50784) [Link] I want to see free software, not just open source, succeed. Open source has really already succeeded. [...] Much of the infrastructure we all use on the Internet every day is open source, and it will continue to grow. The free software movement, though, seems to be shrinking. This is a fine example of tweaking definitions in order to make a point. Much of the infrastructure is Free Software, too. One can argue that the people who deployed it didn't care about the philosophical arguments around Free Software, but just wanted code that was available for free, worked on a bunch of platforms, and had licensing terms that they could live with. But at the same time, the classic arguments around open source - stuff like "many eyes make bugs shallow" - might not have been bought into by such people either. As for whether people respond to a negative message, most people will be told umpteen times that their lack of interest in wider issues - like who exercises control over a platform or over a device - will eventually catch up with them, and it's true that most people just won't be told. Of course, they then learn their lessons the hard way: one recent classic example was the iPad developer whose application was pulled from the App Store after having "stuck up" for Apple in the face of criticism of that company's policies. In every generation you'll always have people learning things through bitter experience even though they could have listened to people who've seen it all before. Claiming that the FSF isn't doing enough is just absurd: they're one of the more active organisations, together with other people in the Free Software movement, when it comes to advocating open standards and open platforms, resisting the expansion of the patent system, and publicising alternatives to proprietary cloud services. Instead of belittling and berating the FSF, one could justifiably ask where other organisations, more palatable to those who advocate "pragmatism" (presumably all those people using Macs at conferences), have got to. Is it not pertinent to ask where the Open Source Initiative stands on these matters, or do we hold some organisations to a higher standard than others? Or is this just another attempt to "rebase" the achievements of the Free Software movement in terms of the activities of some other group of people whose visibility and coherency on the wider issues leaves a lot to be desired?
History editing Posted Jun 18, 2010 15:33 UTC (Fri) by jdave23 (guest, #27160) [Link]
I agree -- the article seemed to be projecting the author's interpretation of events and desired outcomes on to the FSF.
Open vs. Free Posted Jun 18, 2010 18:02 UTC (Fri) by dbruce (subscriber, #57948) [Link]
AFAICT these terms refer to almost precisely the same bodies of software. Once you specify that "Free" means "Free as in Four Freedoms", and that "Open" means "Open Source Definition" and not just "the code is visible", Free Software and Open Source Software are practically identical.
Of course, these terms have very different implications with regard to the goals and beliefs of the developers. Nonetheless, it makes little sense to claim that the use of open source software is thriving, but free software is not. The author obviously knows this - I assume he is saying that the devs and users of Free/Open software seem to be caring less and less about freedom as the years pass.
History editing Posted Jun 19, 2010 11:19 UTC (Sat) by zonker (subscriber, #7867) [Link]
"Claiming that the FSF isn't doing enough is just absurd:"
They're not doing the right things. DBD is a complete and utter waste of time. Win7 Sins is a complete waste of time and FSF contributor dollars. RMS spending his time on the GNOME foundation list arguing for "GNU/Linux" instead of spending time constructively is a waste of his time and the time of other people on the lists...
"Is it not pertinent to ask where the Open Source Initiative stands on these matters, or do we hold some organisations to a higher standard than others?"
OSI, last I checked, wasn't seeking funds to squander to run pointless campaigns.
History editing Posted Jun 21, 2010 1:05 UTC (Mon) by coriordan (guest, #7544) [Link] RMS isn't paid by FSF, so how he spends his time is his business. He's been watching this issue develop for 27 years and he's come to the conclusion that terminology is critical. Was Windows7Sins a flop? It wasn't my favourite campaign, but FSF were right to do *something* for the release of Windows 7, and maybe another demographic liked it. Did it consume donations? I see a website, a mail-out of 500 letters, a morning spent throwing Windows boxes into a bin in a nearby park, and a call for donations if people want more such letters sent. Not exactly extravagant. How do you know the bulk of the work wasn't done by volunteers? Or that someone didn't suggest it and foot the bill? If you think scratching off actions like Windows7Sins could free up resources to make a substantial start on web services, I think that running an NGO would deeply disappoint you :-) (and as I pointed out above, FSF is working on web services.)
History editing Posted Jun 21, 2010 11:05 UTC (Mon) by pboddie (subscriber, #50784) [Link] They're not doing the right things. DBD is a complete and utter waste of time. Win7 Sins is a complete waste of time and FSF contributor dollars. How are these campaigns a "waste of time"? They may not target the man in the street, but if they target a demographic that can then influence the man in the street, then they're worthwhile. I notice that at least one commenter in this discussion has referred to Windows 7 Sins as "FUD", but I now take that to mean "I don't like those people, and I can't think of a response" rather than anything which can actually detract from the message of the campaign. Do you regard any kind of "consumer education" campaign as a "waste of time"? Why not just give the consumers something which is just as shiny but less "evil" so that they don't really have to think about any issues? This works up to the point where Apple, say, make a public relations play about how they've made a more "open" product, and then you're back to square one. OSI, last I checked, wasn't seeking funds to squander to run pointless campaigns. Last I checked, OSI had no direction whatsoever, and probably no inclination to execute any campaigns that might be deemed "ideological", in the fear that it may offend all those people who have worked so hard over the years to not have a position on anything other than maybe how great the resulting code is: a position which is actually easily undermined and largely tangential to the matter of openness and its principal, blatantly obvious benefits. Indeed, due to the nature of many of the other organisations around "open source" - that they eschew what they regard as "ideology" in favour of not rocking anyone's boat - there are actually few of them who are inclined to pursue the causes that the FSF pursues. If one dislikes the nature of the FSF's campaigns, one would be better advised to improve the FSF's campaigning instead of lining up some "pragmatic" superheroes, only to see them issue a collective "meh" on whichever issue one would have them confront.
History editing Posted Jun 19, 2010 16:24 UTC (Sat) by bronson (subscriber, #4806) [Link]
Here's my oversimplified view:
Open Source: MIT, BSD, LGPL, or Apache licensed. The developer just wants to get her code in the open, see others use it, and start receiving patches. If someone else tries to close the code back down, that's unfortunate but it sure doesn't keep her up at night.
Free Software: GPLV2orLater, GPLv3, AGPLv3: The developer can recite the Four Freedoms, cares deeply about future developers, and is willing to expend a lot of effort to lock those freedoms in stone. If it means a license will be huge and complex, or GCC will not have a plugin interface because it could be misused by proprietary developers, then that's a price worth paying.
By this, I think Zonker is right. In my experience, and as evidenced from the astounding percentage of Apple laptops I've seen at every developer conference I've gone to in the past few years, the more pragmatic Open Source movement is thriving and the more idealistic Free Software movement is shrinking.
Personally, I know the FSF has done a really good job of pushing me toward the Open Source camp with GPLv3. I'd hate to use a license so complex for any code that I release! So maybe I'm not the most impartial party to be attempting this observation. :)
History editing Posted Jun 20, 2010 19:32 UTC (Sun) by dlang (✭ supporter ✭, #313) [Link]
you forgot to add GPLv2 to the open source group
Reinventing The Party of Gno Posted Jun 18, 2010 22:51 UTC (Fri) by crosstan (guest, #67792) [Link]
I think you are trying to reinvent the wheel. We have any number of attempts at the party of Gno...
The party of Gnovell
The party of Sco.
All trying to compromise the principles of free software and doing such a good job (at compromising anyway).
The Party of Gno (Linux Magazine) Posted Jun 19, 2010 1:46 UTC (Sat) by dps (subscriber, #5725) [Link]
I am not sure all of the companies that contribute to the free software pool ever where free software purists. The cheapness and flexibility of free software appeals to commercial users. My current work also appreciates the good 64 bit support.
You could push the flexibility, support for your target environment and impact on your cost base as reasons to prefer free software. Using free software avoids the need to track the royalties owed to lots people under dozens of different licenses, which is likely to be expensive.
You can also argue for contributing your changes back to the community on pure cost grounds.
I know my work, which writes commercial software that runs only on linux x86_64, values the qualities above and can find commercial reasons for feeding patches they require upstream. The GPL is not regarded as a problem.
conflation? Posted Jun 19, 2010 13:12 UTC (Sat) by wingo (subscriber, #26929) [Link] From the quoted snippet:And I'll freely admit, I've advocated the pragmatic approach because after more than 10 years of working in the community, it's clear that getting things done with a purist approach isnt working. Not sure what you're getting at here, zonker; might you be conflating "negative" with "purist"? Surely there are ways to be an idealist and to be positive.
|
|||||||||||||