| |||||||||||||
Free is too expensive (Economist)Free is too expensive (Economist)Posted Mar 31, 2012 13:16 UTC (Sat) by Del- (guest, #72641)In reply to: Free is too expensive (Economist) by jcm Parent article: Free is too expensive (Economist)
I am afraid you are not being very helpful. I sense and share the notion that desktop linux is progressing too slowly, and that we somehow need to improve. That said, simply asking for desktop linux to be more like windows or osx is not very helpful. To be helpful, you really need to get down and dirty with the details.
First thing up is packages and dependencies. I think many readers will interpret you to advocate that we should get rid of most dependencies, letting each application be it's own island like it often is for proprietary software. While a tempting thought, the dependencies is exactly the mechanism needed for free software to thrive. We build on each others work. Yes it is painful, but there really is no other way. We have to find ways to do that effectively, and the 20.000 packages is currently the answer. It is so effective that Android, ios, osx and win8 now are all copying it, one central management for handling applications and dependencies. Your comment of apt-get install is stupid, you know very well that there have been gui's available many years now, exactly like Android market and the likes. If you want to be helpful, you should avoid such flame baits.
Flash has been distributed for quite some time on the most popular distributions, with little to no hassle for the users. Again, if you want to be helpful then please make the effort of coming up with some of the real issues.
You may have a point on the subject of making distribution of proprietary software on linux easier. I believe there is still a lot of ground we need to cover here, but it is certainly not that bad either. The LSB is nothing to sneeze at, it provides a framework for handling dependencies of the most common libraries across all relevant linux distributions. For other libraries you are simply encouraged to compile statically. The biggest problem is actually that many developers don't even know it exists. Is it perfect? No, it isn't. There are still all the API's that some vendors (like Adobe) complain about, with fragmentation. In this department I believe GNU/Linux have made great strides to improve and avoid unfortunate fragmentation. On sound I would say that it is already settled. Gstreamer and Pulseaudio are now supported by everybody, and can safely be expected by any developer. Looking into the history with the OSS licensing mess, you will see that many of the problems were created by destructive forces exactly because somebody had your pragmatic "holier than thou" attitude. Red Hat's success is precisely because the have been crisp clear on their stand on licensing.
On video, things are progressing so slowly that it makes me want to scream. Intel singlehandedly ruined the desktop push with Ubuntu 10.04 by forcing users over to a GPU driver framework that was nowhere close to ready for prime time. Meanwhile the graphic stack (I am not talking about the proprietary drivers, they are like peeing in your pants to keep warm, in other words no long term solution), is slowly getting better. Intel finally seems to shape up and will have useful drivers for the first time in two years when 12.04 comes out. OpenGL3.0 support seems to come together on all three open drivers too. Wayland will probably be disruptive, just as pulseaudio was, but I *really* *really* *really* hope that the transition will be handled better this time around (i.e., don't throw everybody into the mess and deprecate xorg before things are reasonably stable for wayland).
As goes for KDE, the 3-series provided a good desktop experience, and KDE4 has taken too long (maybe it was too ambitious, but that is in the past now). I finally see signs that KDE is maturing again, and I hope 12.04 will bring it to a point where many non-savvy users can use it. There are already features in there that are simply fantastic, miles ahead of the proprietary offerings. We just need to make the last rough edges go away. As for Gnome and Unity, I don't really follow that development anymore.
(Log in to post comments)
Free is too expensive (Economist) Posted Mar 31, 2012 17:06 UTC (Sat) by khim (subscriber, #9252) [Link] It is so effective that Android, ios, osx and win8 now are all copying it, one central management for handling applications and dependencies. FAIL. Android iOS, OSX and Win8 are built on top on NeXTStep bundles idea. They quite explicitly don't track dependencies. At all. The whole store is flat: you have libraries and capabilities in the OS and application use them to talk to each other. They don't use capabilities from other bundles directly. This is what makes it possible to create a store with 500'000 applications faster then Linux could build one with 50'000 applications. The LSB is nothing to sneeze at, it provides a framework for handling dependencies of the most common libraries across all relevant linux distributions. FAIL. LSB is nice piece of technology but it's totally unsuitable for real-world applications. For example it still does not offer a way to play sound or video - not even in latest incarnation. Come on: today is 2012, not 1982! For other libraries you are simply encouraged to compile statically. Yeah, nice idea. The only problem: ALSA or PulseAudio don't work this way. You need to somehow find the configuration and/or daemons before they'll work as user expects. On sound I would say that it is already settled. Gstreamer and Pulseaudio are now supported by everybody, and can safely be expected by any developer. Huh? This is nice to hear, but I still can not find them in LSB. Perhaps I'm missing some announcement which moved LSB site or something? There are already features in there that are simply fantastic, miles ahead of the proprietary offerings. That's cool, but again: when these features will be available for LSB users? Or how can I use them if I link all the libraries statically? How can I create applet for a KDE panel which is usable on all LSB-based distributions? You are kind of right: most of the pieces needed to build stable Linux desktop ABI are there. But few small, yet important pieces are missing. glick2 looks like a nice try to finally create cohesive whole, but it does not yet look ready for prime time.
Free is too expensive (Economist) Posted Apr 1, 2012 6:38 UTC (Sun) by Del- (guest, #72641) [Link]
"Android iOS, OSX and Win8 are built on top on NeXTStep bundles idea. They quite explicitly don't track dependencies. At all. The whole store is flat: you have libraries and capabilities in the OS and application use them to talk to each other. They don't use capabilities from other bundles directly."
I should have re-phrased my sentence. The app stores do not handle dependencies, they leave that up to the individual apps. Nevertheless, it is through them that apps have access to their dependencies, and then depend on them being present in the app store. If you build an Qt app for Android you will experience how this is used to pull in the Qt libraries. Problem is that this model is much more fragile, since you leave it up to packagers to handle all dependencies. That model simply does not scale. I repeat, the devil is in the details, you are mixing the dependencies which is a back-bone of open development with ABI breakage. For meaningful discussion we will need to separate between the two.
"This is what makes it possible to create a store with 500'000 applications faster then Linux could build one with 50'000 applications."
I beg to differ. There are far more linux applications than 50.000. I haven't checked, but Debian shouldn't be far away from 50.000 in the official repos by now. If you are familiar with what it takes for a pacage to enter Debian, you would now that you are badly mixing apples and oranges here. Bring in all the ppa's of ubuntu, and you have something more comparable. Then take into account the open development model, and you see how we avoid the 499.000 utterly useless apps. Even with a tiny tiny fraction of the attention, the widgets available for KDE already rivals the ones available on Android (my personal opinion of course). The point though is that proprietary applications are not tailored for in Debian. Ubuntu has adapted their software center for proprietary applications, so it is finally offered, but admittedly still in it's infancy.
"LSB is nice piece of technology but it's totally unsuitable for real-world applications. For example it still does not offer a way to play sound or video"
No, it doesn't, but it pretty much takes care of the ABI issues for most other standard libraries. Maybe I was not clear on this, but addressing sound and video in my post directly was exactly because LSB is not very helpful there. I agree though that media dependencies should have a place in LSB too.
"ALSA or PulseAudio don't work this way. You need to somehow find the configuration and/or daemons before they'll work as user expects."
ALSA has done hoops to cater to compatibility issues, supporting OSS based applications for instance. Still programming against ALSA is probably a bad idea, which is why I did not list it. I did mention Gstreamer on the other hand. Pulseaudio is a server, so I am not sure exactly what grievance you have developing against it. Along with Gstreamer I can also add SDL. Both should give you sane ways to provide sound in your applications that by now should be quite robust between linux distributions, and even across operatingsystems. Which is more than I can say for the alternatives on other platforms.
"How can I create applet for a KDE panel which is usable on all LSB-based distributions?"
The KDE desktop has been a rather fast moving target lately, I would even say partly due to proprietary development of Qt. If you are happy with the approach taken by other platforms, i.e., using a fixed low performing high level language, you can do it in Python quite easily and expect it to provide your precious stable ABI to kingdom come. If you do it in QML, chances are you are in luck. Qt is moving to an open governance model, and with it I expect a relatively stable ABI. In my book, linux has had a stable ABI. It is the drivers that don't have a stable binary interface, and I think we should be happy about that.
"most of the pieces needed to build stable Linux desktop ABI are there."
Indeed, let's get those last pieces nailed too.
Free is too expensive (Economist) Posted Apr 1, 2012 6:47 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]
There are 38579 entries in my apt cache. Of them, 21536 are without the 'lib' prefix. So 20000 apps is a fairly good estimation.
It's a far cry from 500000 apps in iPhone/Android store.
And about uselessness, a lot of Debian packages are not particularly useful for end-users. So I won't be the one casting stones if I were you.
Free is too expensive (Economist) Posted Apr 1, 2012 9:10 UTC (Sun) by Del- (guest, #72641) [Link]
You must be joking. There are practically no useful apps on ios or android. I can count the number I use of them on one hand, and of them a couple is bundled on any linuxbox. Granted I see my family members enjoy two or three more apps. To be generous I will estimate the number of useful apps at about twenty. Do you really want all that garbage on linux?
For your reference libs count like apps on android and ios too. It is not particularly positive to have few libs available. If you check out KDE, you will see that distribution of widgets and other stuff (like wallpapers, icons, plugins, etc) are outside the package system (in the package system you will typically find bundles of them, but only for some). So your estimate is disingenuous at best. The right answer in an apples to apples comparison is quite favourable for a debian based distribution.
Talk about throwing stones, that goes both ways.
Free is too expensive (Economist) Posted Apr 1, 2012 17:21 UTC (Sun) by pboddie (subscriber, #50784) [Link] You must be joking. There are practically no useful apps on ios or android. Indeed. It's like arguing with someone who claims to have 500000 bookmarks, because the threshold for "app" status is more or less at that level.
Free is too expensive (Economist) Posted Apr 1, 2012 21:24 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]
Sure. And all those 300000000 people with Android have thrown away their smartphones and switched to featurephones.
Right now I have on my phone: Sygic navigation (there's no alternative on Linux at all), Yandex Metro (subway navigation for the Moscow subway), Yelp, barcode scanner, etc. Care to provide similar packages for OpenMoko?
Besides, let's look at Debian. I've chosen few random packages:
Am I supposed to be excited by them more than I'm excited by a wet toilet paper?
Free is too expensive (Economist) Posted Apr 1, 2012 21:54 UTC (Sun) by Del- (guest, #72641) [Link]
I am having problems taking you seriously. Maybe you are bored, and just want to argue for the sake of arguing for no particular reason?
Navigation is provided by Google, so it hardly counts in this setting. Moreover, it is not exactly an interesting application for a desktop now is it?
Openmoko was a nice project, just as Qtopia was. Nokia thought the latter was good enough to bet their future on, and I think the N9 proved the point. You are of course free to bash the N9, but in my book Microsoft did what they could to defeat it and Qt, the product itself was excellent. We simply got ass-raped by big money just to make sure it never got momentum. Just like everybody comes after Android these days.
As goes for Debian, there are so many packages I enjoy there the picking random ones is just stupid. Good luck with that.
Free is too expensive (Economist) Posted Apr 1, 2012 22:45 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]
Google has no turn-by-turn navigation for almost all countries other than the US. It doesn't show current speed limit, there's no voice functionality and it doesn't work offline.
So which package out of 50000 Debian packages should I install on my phone to fix this? My phone actually runs Ubuntu in a chroot, btw!
No navigation? When what about barcodes and QR codes? Is there a simple scanner app with integrated search functionality for this in one of the 50000 packages?
>As goes for Debian, there are so many packages I enjoy there the picking random ones is just stupid. Good luck with that.
That goes for the 500000 apps in the Google Play. People enjoy their favorite games and tools even though they may seem stupid to you.
And yes, you have to provide comparable number of applications to attract users and vendors.
Free is too expensive (Economist) Posted Apr 2, 2012 6:51 UTC (Mon) by Del- (guest, #72641) [Link]
Well, the only device using debian packages where navigation makes sense is N9, you will find all the features you mention there. I leave it up to you to search for a debian packaged barcode/qr-code scanner there, see if you can find it yourself.
As for desktop debian you should check out Marble, it is already competitive with Google maps where I live, and I believe it is ported to N9/N900 and can give you off line maps.
As for Qr-code, did you know that generating qr-codes is built into the clip board tool in KDE, so that you can generate a qr-code out of everything on the clip board with a click? One of hundreds of useful features I miss on any proprietary desktop.
Keep an eye on plasma active. With little funding, no devices, it still already is quite capable. If devices with it comes with GPS, you can expect work on free navigation software to catch on. I do believe devices will come with camera, so I am willing to bet that barcode/qr-code reader will follow at some point if it is not already there. And yes plasma active is already distributed by kubuntu as debs.
Free is too expensive (Economist) Posted Apr 2, 2012 7:23 UTC (Mon) by Cyberax (✭ supporter ✭, #52523) [Link]
>Well, the only device using debian packages where navigation makes sense is N9, you will find all the features you mention there. I leave it up to you to search for a debian packaged barcode/qr-code scanner there, see if you can find it yourself.
Nope. N9 uses common app-store approach with solid packages, N9 does NOT use .deb-files-with-dependencies approach for software packaging.
>As for Qr-code, did you know that generating qr-codes is built into the clip board tool in KDE, so that you can generate a qr-code out of everything on the clip board with a click? One of hundreds of useful features I miss on any proprietary desktop.
So I want to scan QR/bar codes with my camera, recognize them and search information based on it. There are tons of 'useless' Android apps that do this.
Certainly, there should be at least one mega-useful Debian QR-scanner among these 50000 packages, right? (Well, no, I've just rechecked it)
Note, I'm thinking as a typical user.
>If devices with it comes with GPS, you can expect work on free navigation software to catch on.
So far, the only freely available map is OpenStreetMap which is woefully inadequate.
Free is too expensive (Economist) Posted Apr 2, 2012 7:57 UTC (Mon) by Del- (guest, #72641) [Link]
Hm, so you challenge me to find phone oriented debs in repo's made for non-existent devices. Exactly what are you trying to prove? That open development typically avoid apps that are of absolutely no use? Frequently termed bloat. On this we agree.
I see you are not satisfied with openstreetmap. I beg to differ. With the progress it has had, I already find it useful, and within reasonable time it is set to rival any proprietary map. Then finally we may go over to a more productive model where just about anybody can develop a navigation app. But for free navigation apps to catch on, we need devices in peoples hands. The Vivaldi is the first device, openmoko was a nice attempt but never got functional enough to count.
Free is too expensive (Economist) Posted Apr 2, 2012 22:06 UTC (Mon) by Del- (guest, #72641) [Link]
I had a look at openstreetmap and accidentally found a navigation application. It was of course conveniently available in Debian and Ubuntu repos too. Check out gpsdrive. I have no idea how good it is, but without relevant devices, even navigation is already in place on Debian. I think it is time you eat some of your own words.
Free is too expensive (Economist) Posted Apr 2, 2012 22:31 UTC (Mon) by Cyberax (✭ supporter ✭, #52523) [Link]
Yup. That's EXACTLY why Linux is failing.
I might use a free program for the sake of freeness, but normal users would cry and run away in horror from something like gpsdrive. Have you actually checked it? It can't do even a quarter of what a good navigation system should be able to do.
But let's go further. Why would something like gpsdrive need full Debian package management system?
Free is too expensive (Economist) Posted Apr 3, 2012 8:13 UTC (Tue) by bronson (subscriber, #4806) [Link]
You might want to actually try using gpsdrive before presenting it as an example of a good software experience. (oh, the memories...) While they had a release in 2010, I don't think much has changed since 2004. It was certainly best-in-class in 2004!
The best OSM navigation tool I've found is called OSMAnd but it's still quite painful. Terrible address/city lookup, poor route selection (sometimes abysmal), difficult UI. I wish I could rely upon it but I just can't. I'm using Google Maps all the time.
For the past decade (since Bruce Perens was working with the Tiger data in 1999?), the year of solid Linux/OSM navigation has seemed pretty close. Just like the year of the Linux desktop, it's only a couple of years out! Too bad it seems to be constant time to completion.
Free is too expensive (Economist) Posted Apr 2, 2012 23:08 UTC (Mon) by wookey (subscriber, #5501) [Link]
Not sure why you thinkg OSM is 'woefully inadequate'. OSM has been better than Google maps for at least 5 years where I live (Cambridge, UK). It's bang uptodate (when a bridge that had been closed for 6 months reopened, OSM reflected that within 24 hours. It took google maps another 6 months. Every footpath, bike route, pub and college is on it. Google maps just doesn't have this detail.
OSM has a cyclemap view and all sort of related tools and sites because it can be re-sused in much more versatile ways than google maps. e.g http://cyclestreets.net for cycle route planning anywhere in the UK.
Google's draggable routing tool is still very cool, but it is very car-centric, and OSM is a much better map for most purposes I need.
I have a tomtom device which illustrates the advantages of OSM. The maps that came with it 8 years ago are now _very_ out of date. I can pay a small fortune to get newer ones, or I can switch to an OSM-based nav tool. Not quite as slick interface but totally uptodate maps forever more at no cost. That's a big deal.
Free is too expensive (Economist) Posted Apr 3, 2012 0:58 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]
Cambridge probably has a lot of students with lots of free time. So it's no wonder that the coverage is great there. It's not so great in other areas, though.
Commercial map providers often use data from official sources, so it's consistent everywhere. Quite often that requires obtaining various government licenses to work with it and/or converting complex formats.
This is not cheap, so applications have to be sponsored somehow. Through ads or maybe through direct sales.
Free is too expensive (Economist) Posted Apr 3, 2012 7:25 UTC (Tue) by anselm (subscriber, #2796) [Link] It's not so great in other areas, though. The OpenStreetMap maps for Germany are, by now, widely considered superior to anything available on the commercial map market. Some loose ends need to be tied up but on the whole things are looking not at all bad.
Free is too expensive (Economist) Posted Apr 3, 2012 16:41 UTC (Tue) by wookey (subscriber, #5501) [Link]
Cambridge is just an example, and in fact most of the initial mapping was done by a man with a job. Most of the UK now has excellent OSM coverage.
Mapping is the canonical example of something that is better crowd-sourced than done commercially. Getting data from official sources is about governments providing data in open formats.
I sometimes wonder why you post here - you don't really seem to believe in the idea of collaborative effort or free software at all. You just keep telling how much better commercial OSes and commercial map providers and massive phone companies are, when the point is that those entities can use and contribute to open data and free software too if they want and we all get better stuff.
Where are these places where the OSM mapping data is poor? I'd expect such places to be getting quite thin on the ground now. Clearly there is more data to add to OSM so that it is good for all purposes (such as motorway lane details for turn-by-turn style nav, speed restrictions) but no doubt that is all ongoing at a rate of knots.
For the last Debconf (in Bosnia and Herzegovina) it was notable that so far as Google or Deutsche Bahn were concerned the place was empty and devoid of roads or trains. OSM had excellent detailed mapping. That seems to be true for many 'lesser' countries where the big providers can't be bothered.
Free is too expensive (Economist) Posted Apr 3, 2012 17:44 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]
Well, for example, Google Maps show the hotel where I'm staying right now: http://g.co/maps/3xm25 and OSM doesn't. In my home city OSM gives me incorrect address for the US embassy: http://www.openstreetmap.org/?lat=50.460314&lon=30.51...
And so on. I've been burned a couple of times by the OSM, so I won't use it for anything important (like getting from point A to point B in time).
So no, maps are definitely not something I'd trust to Wikipedia-style development. Now, maps based on official geodata with wiki-style corrections would be great. And we actually have it in Russia ("folks' map" by Yandex).
>I sometimes wonder why you post here - you don't really seem to believe in the idea of collaborative effort or free software at all.
I certainly do. Only an idiot would deny the success of the Linux kernel (btw, I hate the combination 'Linux kernel', sounds too much like 'ATM machine').
However, I'm not blind and I can see where the _non-commercial_ community development model fails. It usually fails in tasks that require a lot of drudgery and/or interaction with real users.
Linux kernel by now is not non-commercial, it's developed by for-profit companies as a way to avoid developing their own completely new OS. Besides, Linux is hardly boring at all.
OSM might actually be picked up by companies which need reliable mapping data but which don't want to pay to develop it from scratch and/or license it. That's arguably already happening (Apple is using OSM in one of their products).
But no company bets on Linux desktop right now. And unfortunately, a lot of desktop-related development is very boring stuff. Like keeping compat wrappers for old API or making sure you don't break anything with new updates. So we see the result - a lot of wonderful new development with no regards for backward compatibility and regressions in functionality.
A similar problem - there are no good tax/accounting packages for Linux. Just our favorite grumpy editor.
Free is too expensive (Economist) Posted Apr 2, 2012 22:55 UTC (Mon) by wookey (subscriber, #5501) [Link]
Navit or tangoGPS?
No idea about QR codes - there is something called zbarcam that decodes qr codes from a video device. Is that what you mean?
It is true that a lot of android and iphone 'apps' are little more than a URL. It does turn out that those are popular though, and we probably should do more to make such things easy to come by on linux desktops. It seems tome that this is actually a problem of organisation - how do you find what you want in half-a-million bits of software?
Free is too expensive (Economist) Posted Apr 2, 2012 23:18 UTC (Mon) by khim (subscriber, #9252) [Link] Navit or tangoGPS? Still no good. The first thing you need to have good navigation is to have raw data. Not just maps (OpenStreetMap slowly becomes better here), but traffic data (including historical traffic data), photos of complex interchanges, etc. This data is rarely available for free: either it needs to be sold for money or it may be supported via ads. Which basically excludes distros out of the equation immediately. As Cyberax said: I might use a free program for the sake of freeness, but normal users would cry and run away in horror from something like gpsdrive. Not because the program is not any good, but because the lack of data makes it sub-standard. The same with QR-codes, NFC, etc: it's trivial to write program which parses QR code, read NFC or does something like this. And, again, I might use a free program for the sake of freeness, but normal users would cry and run away in horror because they don't need to just read the QR code. They expect that program will do “something sensible” with data (if it's URL it must be opened, if it's a vCard it must be added to address book, etc). If I touch phone with my metro ticket Simple? Yes. Boring? Sure. But this makes it actually useful for Joe Average! Similarly with NFC: yes, I can probably find program which will read data from card, but will it show number of trips you have left on subway ticket like “useless” Android's app?
Free is too expensive (Economist) Posted Apr 3, 2012 15:29 UTC (Tue) by mathstuf (subscriber, #69389) [Link]
> OpenStreetMap slowly becomes better here
I hope Apple's OSM usage[1] will accelerate OSM's coverage.
Free is too expensive (Economist) Posted Apr 1, 2012 12:21 UTC (Sun) by anselm (subscriber, #2796) [Link] There may be 500.000 apps in the iOS or Android app store, but it is fairly safe to say that, for most users, at least 495.000 of them are not really worth looking at, let alone downloading, installing, or running. I'm pretty sure that my phone won't like me installing more than a few hundred apps at a time in the first place, so, as a conservative estimate, out of the 500.000, 499.500 or more are useless to me in any case. If you separate out the »apps« that are really glorified web bookmarks or exist strictly because somebody wants to sell you something, I don't think the difference to what is in the Debian repository is all that significant any longer in practical terms. Whether there are 999 or only 9 useless apps to every single one that is actually worth my trouble doesn't really bother me a lot.
Free is too expensive (Economist) Posted Apr 1, 2012 9:26 UTC (Sun) by khim (subscriber, #9252) [Link] Nevertheless, it is through them that apps have access to their dependencies, and then depend on them being present in the app store. Have you ever actually used any appstore? Of course if someone will try to create app store which works this way it'll not work. Applications are not supposed to depend on app store. At all. Yes, some of them include instructions which say “install package XXX before you can use this package”, but these are rare: developers know to the very bone that when you add any such dependency you risk alienating users. For the most part application is standalone thing and it can be installed in isolation provided you have required version of Android. Some allow you to further install plugins - but this is handled by application itself, not by app store. App store reliably solves dependencies problem - by refusing to support them at all. This approach is tested and it does scale. It uses more resources then distro packaging approach, but frankly in a world where entry-level phones have gigabytes of flash and entry-level laptops have hundreds of gigabytes of space on HDD this is not an issue. If you build an Qt app for Android you will experience how this is used to pull in the Qt libraries. Yes, I know. Qt abuses the system slightly. Time will tell if it'll work Ok for them or if people will complain too much. Problem is that this model is much more fragile, since you leave it up to packagers to handle all dependencies. That model simply does not scale. Somehow it scales pretty well on Web: all these Google Maps APIs, Facebook APIs and so on are supported by developers without distribution-appointed police officers and so far Web works. I repeat, the devil is in the details, you are mixing the dependencies which is a back-bone of open development with ABI breakage. For meaningful discussion we will need to separate between the two. I don't care about any back-bones. I care about the ability to build and deploy the program. "This is what makes it possible to create a store with 500'000 applications faster then Linux could build one with 50'000 applications." Yes. Show me one repo with at least this much - and you'll have a point. Bring in all the ppa's of ubuntu, and you have something more comparable. Nope. If you'll bring all the ppa's of ubuntu on one system then most likely result will be something crazy - I'll be surprised if the resulting system will even boot. PPAs are dangerous beasts. They conflict with each other, some of them can break your system even if they are used in isolation, etc. IOW: they push the burden on user. And the last thing user want is to play “private distro packager”. Ubuntu has adapted their software center for proprietary applications, so it is finally offered, but admittedly still in it's infancy. It's worse: it does not even address API stability seriously. The only thing it says: Short, simple set of requirements, right? Well, let's start. Be in one, self-contained directory when installed. How can I do this when Ubuntu does not provide stable enough foundation for the development? No, it doesn't, but it pretty much takes care of the ABI issues for most other standard libraries. Rilly? SDL is not there, crypto (neither openssl nor gnutls) is not there, desktop integration (dock, notifications, etc) is not there, HTML widget is not included, etc. Just what kind of desktop application can I build using this foundation? I did mention Gstreamer on the other hand. Which is not in LSB either. Pulseaudio is a server, so I am not sure exactly what grievance you have developing against it. Along with Gstreamer I can also add SDL. My grievance is in the fact that nobody guarantees their availability. And if you'll take a look here you'll see that GStreamer is on the road to incompatible 1.0 version. When this happened with openssl Ubuntu immediately punted old version it to universe which effectively rendered all packages which used it out of “one, self-contained directory” category. SDL in the same boat. It looks like PulseAudio does not have plans to do this any time soon, but still, it's a gamble, not a stable platform with a stable ABI. If you are happy with the approach taken by other platforms, i.e., using a fixed low performing high level language, you can do it in Python quite easily and expect it to provide your precious stable ABI to kingdom come. Rilly? Can you gunrantee that two years from now KDE will not decide to switch to python3, for example? In my book, linux has had a stable ABI. Linux kernel has remarkably stable ABI. Linux desktop is a disaster. That's what we are discussing here.
Free is too expensive (Economist) Posted Apr 1, 2012 16:41 UTC (Sun) by tuna (guest, #44480) [Link]
You will be able to run both GStreamer 0.10 and 1.0 concurrently, just like you can run GTK2 and GTK3 together.
Free is too expensive (Economist) Posted Apr 1, 2012 17:28 UTC (Sun) by khim (subscriber, #9252) [Link] But will GStreamer 0.10 come preinstalled or not? That's the question. Or else it'll me more of the same: I can't even get some Humble games installed on Linux, so I just reboot into Windows.
Free is too expensive (Economist) Posted Apr 1, 2012 16:55 UTC (Sun) by Del- (guest, #72641) [Link]
"Have you ever actually used any appstore?"
Yes, of course. I have even made an Android app just for fun, and had to use necessitas to get a decent development environment. I am very happy to have most of my time focussed on linux applications, android and ios developers have my sympathy.
"App store reliably solves dependencies problem - by refusing to support them at all. This approach is tested and it does scale"
No, it doesn't. It will create the same disaster we have on the proprietary desktops.
"Somehow it scales pretty well on Web: all these Google Maps APIs, Facebook APIs and so on are supported by developers without distribution-appointed police officers and so far Web works."
No, web doesn't work. You cannot trust these api's. Facebook is already quite busy figuring out how to disable them to have people come over to their web-site instead. Google has changed api's on gmail and on google calendar for no good reason before. IMAP is the only sane way to talk to gmail, and google contacts are not available through any sane api yet. Moreover, you have probably seen the concerns over the internet being tested only on three engines. Html is no silver bullet, it may prove to look more like a thorny road.
"I don't care about any back-bones."
Then I suggest you go and troll elsewhere. The GPL is about building on each others work. If you don't care about that, and want to throw it away, I suggest you stop wasting your time here and go play on windows. Deploying programs on linux is not that difficult, I deploy a number of them across multiple distributions, and it hardly costs me any time. It is not a big issue. Those who want to do it manages. Your link to gstreamer only shows me that it finally is set to mature with version 1.0, hardly bad news for sound on linux.
Free is too expensive (Economist) Posted Apr 1, 2012 17:33 UTC (Sun) by khim (subscriber, #9252) [Link] No, it doesn't. It will create the same disaster we have on the proprietary desktops. What disaster? If you mean “the system where you can play games, easily watch latest videos and do my work”, then yes, I do want such “disaster”. No, web doesn't work. I think is the point after which the discussion becomes pointless. If you think that web does not work and linux distributions do then you deserve what you are getting.
Free is too expensive (Economist) Posted Apr 1, 2012 18:19 UTC (Sun) by Del- (guest, #72641) [Link]
"What disaster?"
The disaster is that practically all software categories end up with one dominating vendor, also known as monopoly. Notably, games still have numerous vendors, but they enforce the monopoly on the desktop. Have a look at win7, it contains practically nothing, and costs more than a $100 in the cheapest stripped down version. It is a tragedy (or disaster if you like).
"I think is the point after which the discussion becomes pointless."
That is up to you, I simply informed you that some of your prime examples have proved messy environments for programmers and users. Changing API's and web-developers who only cater to a subset of browsers. Thanks to the open model you seem to detest the web has improved tremendously over the last years, but it has at least as long to go as the linux desktop before it can be considered a stable environment. You are not alone though. Everything on the web seems to be the new mantra, it will fail as badly as similar foolishness before it.
Free is too expensive (Economist) Posted Apr 1, 2012 21:34 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]
Windows 7 allows to run other programs. And users are more than happy to pay for this.
I've actually tried to migrate a couple of mid-range (50-100 users) organizations to Linux. We've tried pilot projects with just a few users. Invariably, it turned out that some small but critical app they are using since 90-s can't work in Wine. Then they try do arithmetic:
Large companies might decide to go on with the migration. Small companies usually just continue with Windows.
Free is too expensive (Economist) Posted Apr 2, 2012 7:43 UTC (Mon) by Del- (guest, #72641) [Link]
You really do open up new topics as we go along here. Migrating company desktops to linux is indeed challenging. Mostly because of lock-in. Sure, there are applications, some places more than other places. In HPC related industry you frequently have the opposite, applications, yes even proprietary ones, are often only available on linux. Guess what, proprietary 3D viewers are abundant on linux.
Start a new thread and I can join you in this discussion. For this thread we have already expanded the capabilities of LWN.
|
|||||||||||||