Consistently ever since I installed morpheus a few months ago, morpheus has about 550000 users at any time (indicated on the status bar). The largest amount of users I can recall being indicated there is about 750000, about half of what is claimed here. The lowest amount of users I can recall was about 350000.
500000 users is still quite nice since with napster I never had more than around 10000 users to connect to (it wasn't a very scalable network).
I would like to elaborate a bit on the LOC thing. There's some empirical evidence (forgive me that I omit the proper refs here) for the following two claims:
- regardless of the language, maintainance cost is proportional to the LOC
- regardless of the language, programmers have a more or less fixed productivity measured in LOC/timeunit.
In addition I've seen similar evidence as well as had it confirmed by senior software developers in large software companies that the average production of code throughout the development period of a large software project (>>100KLOC) is less than 1 LOC/day/developer.
If you accept this and use common sense you will realize that using a higher level language will allow you to deliver software both quicker and with fewer bugs. Of course, since time is the limiting factor on most sw. projects, the gained time is used to make more complex systems so new systems are generally as buggy but more feature rich.
BTW. I agree with you that the use of C for most real world projects is misguided these days since much better, equally well performing languages are available. It will be interesting to see how quick e.g. kmail (c++, I think) and mozilla (also C++ + scripting languages) will catch up with evolution.
"Actually, I don't know if you ever tried running windows 95"
Yes I did and I agree it wasn't pretty but it worked and for joe average it was good enough to launch into word 6.0 do his thing. I'd say the state of integration then is what Unix is going through now. You have all these apps which either are commandline based or written for stuff like motif or mwm. You can forget about copying pasting anything more elaborate then ascii to and from such apps. It isn't pretty but you can learn to live with it.
700.000 loc is what it takes to write such applications. People sometimes claim there are smaller alternatives but invariably those alternatives are less feature rich.
People have long blamed MS for delivering bloated systems. But it is quite ironic to see that as linux is maturing it is also gaining weight. The hardware requirements for running a full KDE or Gnome desktop are getting awfully close to the hardware requirements of an average MS windows machine. If you consider that MS managed to deliver windows 95 in 1995 on the hardware of that time (pentium/486, 8-16MB) you might actually come to the conclusion they did a better job than Gnome or KDE since in terms of features (not stability of course) it still compares rather well.
No doubt people will reply with references to all sorts of windowmanagers which run rather nice on slow machines claiming they do everything you need. However, they don't fully duplicate the feature set of windows 95 so see above.
I'm sure you are an expert sys admin and obviously I'm not. However, debian doesn't work as advertised. The dogma spread by debian proponents is that you install a base system, point to testing/woody, do a dist-upgrade and happily live ever after. On all occasions I tried this it didn't work even with only very basic stuff like X selected. As long as you stick with potato you are fine but as soon as you move to testing you soon find out why it is marked testing.
"Real corporate environments don't let users install whatever they want. Admins purchase software and then install it on machines."
Yes, they don' go and assemble their own linux distributions from a zillion apt sources. Instead they'll get something that is known to work reliably (i.e. definately not debian unstable)
So you are saying the only way to use debian is to do the testing yourself because debian has such high standards? I hope you see the contradiction here. Obviously the reason there is no stable version of debian you can actually use (without going through some extensive testing) is that debian is so large and has to be tested for all platforms.
So, all I'm saying is that debian is not a viable option unless you want to invest time in testing/tinkering with the packages. In other words it is not a general purpose distribution.
Right now I have no use for debian on my desktop because it is too unstable/unusable. Dll hell is nothing compared to debian unstable. Red Hat may not be ideal and may not be available on all platforms and may not include everything and the kitchen sink (even though it does a great attempt at doing so) but it does offer a set of recent packages that have been tested to such an extent that it will work under most cicumstances (unlike debian unstable).
I suggest that debian splits their packages in essential (kernel, base system, X, Gnome, KDE and essential development stuff) and optional (anything that is not needed on the average server or desktop system -> 80% of the packages qualifies as such IMHO) and evolve them separately. That means that if you need optional you probably have to either use an old core release or use the latest stable core release and invest some time in copiling the optional stuff you need. IMHO this is much better than to have to scavenge stable versions of essential stuff just to be able to install debs of some not so frequently used stuff.
If the need arises, there could be more optional releases. It is simply divide & conquer that is being applied here, nothing fancy. Obviously the intgration effort explodes if you increase the amount of packages. Doing integration testing on 5000 packages is rediculous and has shown to be unfeasible.
Ordinary users shouldn't have to wait two years for stuff like KDE 2.x being integrated into debian because some very obscure packages need to be tested with it. The KDE people do a pretty decent job of testing their stuff and integrating their release code shouldn't take much longer than a few weeks at best. If it takes longer you are doing something wrong because in this period the KDE people will find & fix more issues than the debian people.
So basically that requires an administrator that composes his own distribution (I'd like the debian people to do that sort of thing). In the real world administrators don't have time for that sort of hobbyism. They will just deploy a red hat/suse/whatever image that may or may not be tweaked a bit prior to distribution. Doing so they deploy reasonably well tested software that contains recent, stable versions of all packages.
Debian would be ideal for workstation environments if they'd bother to keep their distribution up to date. What's currently in testing will be obsolete by the time it is finally marked stable (i.e. irrelevant for most users). I consider this a fatal flaw in the way the debian project works. It just takes too long to get stable, production quality software integrated into debian.
Testing/unstable is nice to play with if you have the time to kill to fix the inevitable problems that will pop up. Patching together a usable stable system however is way too difficult for any real life usage of debian. Besides, using unstable packages sort of counters any security policy you might have.
If the debian project wants to matter on the desktop front they will have to address these issues. I'm sure there are a lot of people like me who would like to use a debian desktop but can't be bothered too much with testing or unstable stuff. If a system administrator can compose a stable, up to date desktop environment (which is what you claim), surely the debian project can do a better job and release it to their users.
Using debian in a production environment requires that you use something stable. So consequently since there is no stable debian desktop environment with reasonably modern packages, debian is no option for somebody who wants to use a linux desktop in a production environment.
I tried installing a debian desktop a few weeks ago (I've maintained a debian server in the past so I'm not exactly ignorant). I installed the potato base system, did an upgrade, pointed the sources file to testing, did a dist-upgrade and then launched tasksel and just selected the X environment (nothing else). Then hit enter and tada some conflict and it didn't install. So far I had done nothing special so you can't blame me for doing anything wrong.
Surely it is possible to fix it. An option would be to skip potato entirely and go with the woody boot floppies or something but I didn't bother (too lazy, I know). If I had really wanted I would have probably gotten it in a working condition at some point. The point however is that apt-get rarely works as advertised due to the fact that the packages are untested and immature. For stable it is great, maintaining a potato server is a piece of cake (been there done that).
However on each of the occasions I tried testing or even unstable I ran into stuff that was non trivial to fix. I keep reading about how painless a dist-upgrade to woody is but I have so far been unable to verify this claim on any of my systems. Each time I try I run into issues with some of the packages which either prevent me from completing the dist-upgrade or leave me with an unusable system.
I'm sure support is great on the mailinglists. But unless you have a serious amount of time to waste, mailing lists are no option if you just want to get a system running. Just scanning through archives for your particular problem easily takes up hours. And waiting for a sensible reply generally also takes some time (even though some people are really quick in replying).
The whole point of using debian is apt-get as long as it works it is great as soon as it doesn't work you are on your own.
IBM sells hardware and services. The software is essential for both but in terms of license fees it is peanuts compared to the revenue from the hardware and services.
Therefore stimulating the adoption of their software makes sense. The more developers, the more software, the more software the more hardware to run it is sold as well as services to maintain, deploy and support it. Of the shelf software is rapidly disappearing as a viable businessmodel. IBM still has some packages that profitable but they are increasingly niche markets. DB2 for instance is no doubt profitable. Also whatever license fees IBM harvest probably pales in comparison to the revenue from the associated services that come with it. The long term market is that the DB software is free (maybe even as in speech). MS is already planning to integrate SQL server into their OS. When they do, selling DB engines will no longer be a sound business model.
Nice side effect of all this is that in principle you can have good quality stuff for free. Linux has already benefitted from this enormously since many of the improvements in the 2.4 kernel are contributions from various large UNIX vendors. File systems are no core business to anyone any more so you might as well give them away -> linux has dozens of high quality filesystems to choose from now.
The contribution model is beginning to be somewhat mandatory. Very few vendors can actually afford to continue to support their propietary software. If it doesn't integrate with linux it is dead in the water. IBM knows this, SUN hasn't quite figured it out yet.
Well on the few occasions I tried this, it barfed out halfway with some unresolved dependencies. This kind of stuff shouldn't happen with stable packages but that's just the thing you have to use unstable packages or even testing packages and you may be confronted with this kind of stuff.
I want software that is known to be in a working state for over a year to install without much trouble. Apt-get can technically deliver this experience provided that there are stable packages. And there are none available.
It's not just the kernel. To get a reasonably up to date debian based desktop environment you have no other option then to install unstable packages of software that has had stable releases for over a year in some cases.
And yes unstable actually means unstable. I've played around with debian on several occasions. The apt-get system is brilliant. Unfortunately its usefulness is totally countered by the complete lack of stable packages to install. I've actually managed to get KDE 2.0 up and running once. I just pointed the sources file to some vague http site of some guy who had bothered to create some packages. On several other occasions though I was less fortunate with installing e.g. xfree 4.0, gnome and several other mainstream stuff. I'm sure it is possible to fix if you know what you are doing but the whole point of apt-get is that it is supposed to take care of that instead.
Same for me. I had win2k on my machine earlier. I upgraded to XP and I must say it is slightly faster and just as stable. The UI is actually not that bad once you turn of the pretty colors (theres a silver color scheme which is pretty nice IMHO) I left all the graphical effects on because they really don't get in my way. I noticed that on a slower machine XP automatically turns off some options. And even on this slower machine performance was still adequate (for the record it was a PII233 with 64MB) although I prefer win 98 on it because that is a little lighter on the system resources.
Most stuff is actually in the same place it used to be. Some stuff was moved but there are no radical changes I'm aware of. The new start menu is actually quite nice once you remove the stuff you don't need from it. Easy access to frequently used apps is quite handy.
The new task oriented control panel is the most confusing to people used to the old control panel. But it is organized pretty well and most stuff can be found where you'd expect to find it. If it bothers you there's an option to reverse to the old style UI.
The.Net shit, pricing and legal issues aside I think it is safe to say that windows XP is MS' best product so far. If you intend to use it, make sure to put lots of memory in your machine. It will run with 64 MB (i tried it with office XP on the slow machine) but you will have a much smoother experience with 256MB or more (I have 512MB). Also make sure your hardware is supported since that seems to be the biggest issue with XP. I have a aureal vortex 2 soundcard and a voodoo 3 card. Both are supported by XP. However, the voodoo driver is worthless for gaming. Luckily there's an unsupported hacked together driver available on the internet. The soundcard works with the default driver but since the manufacturer no longer exists there will be no new drivers.
I mostly agree with you actually. I just don't agree that the solution should aim for a minimal set of features that is common to all platforms since I strongly suspect that just like with AWT it will be an ugly mess. AWT provided just the basic features. You could instantiate a button, put it in a window and so on. However, even this simple functionality was too different from platform to platform which resulted in applications that had to be tested on each intended target platform to deal with all the platform specific quirks.
I agree however that there is a need for a crossplatform GUI kit. I only think that it should be a full featured one rather than some arbitrary subset. More importantly, I think it should be object oriented and language independent. Otherwise we'll see duplicated efforts for perl, ruby, whatever. Most of the existing GUI kits are language dependent or at least assume a great deal about the language used to work with them. This limits their usefulness. A good example is the GTK - QT debate. C programmers will likely favour GTK whereas C++ programmers will likely prefer QT. Swing is great when you use Java, but using it from another language is usually hard (unless it has an implementation ontop of the JVM).
SUN has been working for years to create a cross platform application framework. It is called swing. Swing is very complex and heavy on resources but it does the job. It supports printing, drag & drop, integration with the native clipboard, key bindings, skinning, all sorts of graphical stuff and lots of other stuff.
While you might disagree that Swing is a good solution, the fact remains that Sun realized that all of the above is needed if you want to create competitive GUI apps.
The solution suggested for python seems to make the same design mistakes sun made early on (AWT) and seems to be based on the same naive view on what comprises a good GUI. In addition it seems to ignore a whole lot of other perfectly good solutions (qt, gtk, XUL, Kylix, Swing,...). Swing integration is easy if you use jython. I think there are also python bindings for Mozilla so you should be able to create XUL applications in python. Presumably integration with GTK or QT is also easy.
Wrapping is no good solution for anything but the most trivial applications. As soon as you make things more complex, you will have more and more trouble keeping things crossplatform since each platform works slightly different, has its own bugs to work around and may or may not support what you need.
The trouble with this fix is that it will cause the server to serve up IE specific code rather than w3c compliant code at whatever site you visit. Any browser detection code will incorrectly assume that you use IE and provide you with the appropriate HTML. Quite a bad thing if you have a browser that can do a good job of displaying w3c compliant stuff but is not bug for bug compatible with whatever MS chose to implement.
Well that is the problem of the linux community in general. They seam to be out of touch with reality. Right now linux gives me the experience the earlier versions of win95 (i.e. before Billy boy realized it was worth investing in the Internet). I mean come on, I like it as a server OS but as a desktop OS it is downright pathethic. Its cheap explorer rip offs, the ie 2.0 look alikes competing for the web browser market, the total lack of a omnipresent component model, the lack of clipboard functionality that actually works for most applications. I mean we could deny that these problems exist, convince users they really don't need such functionality but the rest of the world has had this functionality since windows 3.0 or even earlier if they were apple users.
Well to paraphrase a popular phrase by Ghandi
"first they deny you, then they fight you and then you win"
When it comes to usability linux is still in the denial phase, occasional even fighting the "a GUI is actually useful" notion. I must admit some KDE developers do have a grasp of reality and I admire their work. But the majority of the Unix community actually is incapable of looking beyond the command line. I perceive that as a weakness.
And yes, I had a few beers too many so don't be hard on me:-)
You are not tech savvy and yet you want to have tight control over your computer?? Sorry, even MS can't deliver you that and linux requires you to be tech savvy to do just about anything windows does out of the box. Don't point at mandrake please, I've seen it and it is way too buggy for end users who don't know what they are doing. If you care to know, I thought win9x was actually more reliable than that piece of shit. And that particular OS caused me to swear in very nasty ways.
In any case, if you get me a linux version that is as easy to use as win XP, as secure as linux and as feature rich as win xp, I'm your man but I think we both know such an OS does not exist today (well maybe OS X but that has some hardware requirements and stability issues). I'll stick with win XP for the moment and by all means, use linux on your desktop if you have to. I just wanted to let people hear a different sound on this site. All this shit about linux being ready for the desktop is kind of premature from where I'm standing.
Privacy issues aside (I'm not denying there are any and I'm annoyed by them as well), windows XP is a very nice product. Compared to win2k it is a nice, incremental update. It is reasonably stable, it's slightly faster then win2k, it has some nice features, a polished UI that you can configure if you are offended by the use of colours. I've used it for a month now and I think a tech savvy user can eliminate most of the privacy issues: remove activation (or install corporate edition); don't sign up for passport; get rid of messenger; install a third party firewall (e.g. sygate); install virus scanner (e.g. norton) and you are all set for a reasonable safe eXPerience:-).
Install a firewall, that's what I did. But if you are willing to trade features for security it is a free country and I could very well imagine moving to linux for this reason one day. Just don't claim that it is the same thing as windows because clearly it is not. If you use linux you will have to use the commandline because most file managers are crap, you get less integration between apps and so on.
Anyone suggesting linux as an alternative to windows XP for end users does not understand a thing about usability. I was reading a review about mandrake earlier today it boiled down to nice look and feel, pitty it's so buggy (this was in the review and is also my opinion entirely).
Windows XP is a nice product. It's a pitty the kernel tries to contact Redmond through the internet on a regular basis, but that doesn't change the fact that it is also setting some standards from the point of view of usability.
I notice the linux community is acting very conservative and even scared towards win xp. Jokes about clippy, blue screens and so on are made a lot whereas proposals to include new useful features into e.g. KDE or Gnome are a lot less frequent. Windows XP means a lot of new work for those projects because they are way behind delivering the same out of the box experience as windows XP.
I have windows XP at home. After some fiddling I managed to get it reasonably secure and prevent it from contacting MS too often. I love it. A lot of the new interface makes sense. Of course you want to at least change the colour scheme but a lot of things work really nice.
At this point, linux is a downgrade for end users. You lose features. You may say you don't like the features, you don't need the features or don't want the features but the bottom line is you lose them. And if you really need a unix environment, most of the more useful UNIX tools have win32 ports available. And yes windows xp is stable and won't bsod unless you install crappy drivers. If you load a buggy kernel module you have the same experience in linux.
Just for clarity, I have used linux for years and maintain a debian server. I know my shit and have hands on experience with most mainstream distributions.
The question is not whether it will run but whether it will integrate. Can I drag and drop between konqueror and evolution? Can I drag an url to a new message? What about Koffice (cut and paste without losing layout)? What about themes? The better it integrates the more useful it is.
It's not about right and wrong. What you find wrong (i.e. copying & distributing music presumably) is simply a technical reality, deal with it and get over it. Money is not protected by copyright but by additional laws which basically make it criminal to print your own money so that's a bogus argument on your side.
"And if you've no money coming in from record sales, then how are you going to buy equipment and studio time? Or even afford to eat? Naw, you just want to justify your greedy hording of MP3's"
Well I guess you could get out and get a job or something. Sorry, it's a tough world. All I'm saying is that it no longer makes sense to create and distribute physical copies of music since it can all be done using networks at virtually no cost. If your business model is selling plastic discs with music on them you obviously have a revenue problem. Basically all techniques for preventing illegal downloads of music (and movies) appear to be fundamentally flawed (see my previous post) so you can't replace your obsolete business model by simply asking money for the downloads (again sorry, there's no moral judgement here just a technical argument).
My earlier suggestions will certainly not replace all of the lost revenue. It just serves as an example of how you can still make money after the collapse of the recording industry. And after all, creativity does not require revenue at all. Some people just love to make good music because it pleases them. Basically aside from the past 75 years, throughout history there was no way to distribute copies of music. Still people like Mozart and Beethoven delivered some very fine music so it is not the end of the world.
Software requires an exact digital copy and duplication is somewhat more complex due to that. An encoded mp3 you can record by playing it and holding a microphone next to the speaker. Similarly you can record a movie. Admittedly the quality will suck but there are ways of improving on quality by using e.g. better equipment.
You can't do this with software. That doesn't mean pirating software is impossible. It is just a bit harder.
The whole point is that copyright law no longer is working. The law was built with the idea of physical copies in mind. Currently there is a whole industry around distributing physical copies of music to clients. With fat internet pipes readily available, zero cost duplication of content is possible (exluding the price of using the connection of course).
The music industry is still charging as if there were a physical copy but that is no longer true.
Instead of argueing about what is theft and what is not I suggest a more pragmatic course of action. Pragmatism starts with realizing that:
- if it can be played it can be recorded and vice versa so recording enables other people recording again.
- if it is available digitally it can be distributed at no cost. With a peer to peer setup this basically means you don't even need an expensive server setup.
- if it can be done, it will be done whether you stamp your feet on the ground loudly or not.
This applies to movies and music. These are not things that are open to debate, these are facts of life. Once you realize that, you also realize that the cd producing industry as we know it today is doomed to die eventually. The factories that create the machines to create the cds are no longer necessary, the factories that create the cases for the cds are no longer necessary, the shops that sell the cds are no longer necessary, etc.
The things that are necessary are artists to create content and supporting staff and equipment to help them record the content and optionally marketing people to market the content. Most artists consider albums to be marketing material for their live shows. Generally they don't make a lot of money from these cds since most of the profits go to the record industry. Of the 20$ you pay for a cd, only a fraction of that actually ends up in the artist's wallet Really only the very big artists can make a living out of cd sales.
So how can you make money of content creation? We have already established that the distribution has no meaningful cost associated with it so realistically it is the content creation that should generate the revenue and not the distribution.
Suggestions:
- Live performances. People love live performances and are generally willing to pay for it.
- Commercial activities. If you're famous, you can help promote stuff for money. You could for instance get a sponsor. Many sports people for instance wear clothes from their sponsor and get paid for it.
- Video clips are broadcasted on tv channels who make money by receiving revenue from advertisements.
These are only a few examples. All of these activities actually benefit from free distribution of content. And more importantly, for many artists these are already the primary source of income.
Reality is that I have a/mp3 directory which is huge and it will be a long time before I rm -rf it.
Slashdot Mirror
Consistently ever since I installed morpheus a few months ago, morpheus has about 550000 users at any time (indicated on the status bar). The largest amount of users I can recall being indicated there is about 750000, about half of what is claimed here. The lowest amount of users I can recall was about 350000.
500000 users is still quite nice since with napster I never had more than around 10000 users to connect to (it wasn't a very scalable network).
I would like to elaborate a bit on the LOC thing. There's some empirical evidence (forgive me that I omit the proper refs here) for the following two claims:
- regardless of the language, maintainance cost is proportional to the LOC
- regardless of the language, programmers have a more or less fixed productivity measured in LOC/timeunit.
In addition I've seen similar evidence as well as had it confirmed by senior software developers in large software companies that the average production of code throughout the development period of a large software project (>>100KLOC) is less than 1 LOC/day/developer.
If you accept this and use common sense you will realize that using a higher level language will allow you to deliver software both quicker and with fewer bugs. Of course, since time is the limiting factor on most sw. projects, the gained time is used to make more complex systems so new systems are generally as buggy but more feature rich.
BTW. I agree with you that the use of C for most real world projects is misguided these days since much better, equally well performing languages are available. It will be interesting to see how quick e.g. kmail (c++, I think) and mozilla (also C++ + scripting languages) will catch up with evolution.
"Actually, I don't know if you ever tried running windows 95"
Yes I did and I agree it wasn't pretty but it worked and for joe average it was good enough to launch into word 6.0 do his thing. I'd say the state of integration then is what Unix is going through now. You have all these apps which either are commandline based or written for stuff like motif or mwm. You can forget about copying pasting anything more elaborate then ascii to and from such apps. It isn't pretty but you can learn to live with it.
700.000 loc is what it takes to write such applications. People sometimes claim there are smaller alternatives but invariably those alternatives are less feature rich.
People have long blamed MS for delivering bloated systems. But it is quite ironic to see that as linux is maturing it is also gaining weight. The hardware requirements for running a full KDE or Gnome desktop are getting awfully close to the hardware requirements of an average MS windows machine. If you consider that MS managed to deliver windows 95 in 1995 on the hardware of that time (pentium/486, 8-16MB) you might actually come to the conclusion they did a better job than Gnome or KDE since in terms of features (not stability of course) it still compares rather well.
No doubt people will reply with references to all sorts of windowmanagers which run rather nice on slow machines claiming they do everything you need. However, they don't fully duplicate the feature set of windows 95 so see above.
I'm sure you are an expert sys admin and obviously I'm not. However, debian doesn't work as advertised. The dogma spread by debian proponents is that you install a base system, point to testing/woody, do a dist-upgrade and happily live ever after. On all occasions I tried this it didn't work even with only very basic stuff like X selected. As long as you stick with potato you are fine but as soon as you move to testing you soon find out why it is marked testing.
"Real corporate environments don't let users install whatever they want. Admins purchase software and then install it on machines."
Yes, they don' go and assemble their own linux distributions from a zillion apt sources. Instead they'll get something that is known to work reliably (i.e. definately not debian unstable)
So you are saying the only way to use debian is to do the testing yourself because debian has such high standards? I hope you see the contradiction here. Obviously the reason there is no stable version of debian you can actually use (without going through some extensive testing) is that debian is so large and has to be tested for all platforms.
So, all I'm saying is that debian is not a viable option unless you want to invest time in testing/tinkering with the packages. In other words it is not a general purpose distribution.
Right now I have no use for debian on my desktop because it is too unstable/unusable. Dll hell is nothing compared to debian unstable. Red Hat may not be ideal and may not be available on all platforms and may not include everything and the kitchen sink (even though it does a great attempt at doing so) but it does offer a set of recent packages that have been tested to such an extent that it will work under most cicumstances (unlike debian unstable).
I suggest that debian splits their packages in essential (kernel, base system, X, Gnome, KDE and essential development stuff) and optional (anything that is not needed on the average server or desktop system -> 80% of the packages qualifies as such IMHO) and evolve them separately. That means that if you need optional you probably have to either use an old core release or use the latest stable core release and invest some time in copiling the optional stuff you need. IMHO this is much better than to have to scavenge stable versions of essential stuff just to be able to install debs of some not so frequently used stuff.
If the need arises, there could be more optional releases. It is simply divide & conquer that is being applied here, nothing fancy. Obviously the intgration effort explodes if you increase the amount of packages. Doing integration testing on 5000 packages is rediculous and has shown to be unfeasible.
Ordinary users shouldn't have to wait two years for stuff like KDE 2.x being integrated into debian because some very obscure packages need to be tested with it. The KDE people do a pretty decent job of testing their stuff and integrating their release code shouldn't take much longer than a few weeks at best. If it takes longer you are doing something wrong because in this period the KDE people will find & fix more issues than the debian people.
or use one of the p2p solutions. I'm currently downloading it using morpheus. Probably it is also available on the gnutella network.
So basically that requires an administrator that composes his own distribution (I'd like the debian people to do that sort of thing). In the real world administrators don't have time for that sort of hobbyism. They will just deploy a red hat/suse/whatever image that may or may not be tweaked a bit prior to distribution. Doing so they deploy reasonably well tested software that contains recent, stable versions of all packages.
Debian would be ideal for workstation environments if they'd bother to keep their distribution up to date. What's currently in testing will be obsolete by the time it is finally marked stable (i.e. irrelevant for most users). I consider this a fatal flaw in the way the debian project works. It just takes too long to get stable, production quality software integrated into debian.
Testing/unstable is nice to play with if you have the time to kill to fix the inevitable problems that will pop up. Patching together a usable stable system however is way too difficult for any real life usage of debian. Besides, using unstable packages sort of counters any security policy you might have.
If the debian project wants to matter on the desktop front they will have to address these issues. I'm sure there are a lot of people like me who would like to use a debian desktop but can't be bothered too much with testing or unstable stuff. If a system administrator can compose a stable, up to date desktop environment (which is what you claim), surely the debian project can do a better job and release it to their users.
Using debian in a production environment requires that you use something stable. So consequently since there is no stable debian desktop environment with reasonably modern packages, debian is no option for somebody who wants to use a linux desktop in a production environment.
I tried installing a debian desktop a few weeks ago (I've maintained a debian server in the past so I'm not exactly ignorant). I installed the potato base system, did an upgrade, pointed the sources file to testing, did a dist-upgrade and then launched tasksel and just selected the X environment (nothing else). Then hit enter and tada some conflict and it didn't install. So far I had done nothing special so you can't blame me for doing anything wrong.
Surely it is possible to fix it. An option would be to skip potato entirely and go with the woody boot floppies or something but I didn't bother (too lazy, I know). If I had really wanted I would have probably gotten it in a working condition at some point. The point however is that apt-get rarely works as advertised due to the fact that the packages are untested and immature. For stable it is great, maintaining a potato server is a piece of cake (been there done that).
However on each of the occasions I tried testing or even unstable I ran into stuff that was non trivial to fix. I keep reading about how painless a dist-upgrade to woody is but I have so far been unable to verify this claim on any of my systems. Each time I try I run into issues with some of the packages which either prevent me from completing the dist-upgrade or leave me with an unusable system.
I'm sure support is great on the mailinglists. But unless you have a serious amount of time to waste, mailing lists are no option if you just want to get a system running. Just scanning through archives for your particular problem easily takes up hours. And waiting for a sensible reply generally also takes some time (even though some people are really quick in replying).
The whole point of using debian is apt-get as long as it works it is great as soon as it doesn't work you are on your own.
IBM sells hardware and services. The software is essential for both but in terms of license fees it is peanuts compared to the revenue from the hardware and services.
Therefore stimulating the adoption of their software makes sense. The more developers, the more software, the more software the more hardware to run it is sold as well as services to maintain, deploy and support it. Of the shelf software is rapidly disappearing as a viable businessmodel. IBM still has some packages that profitable but they are increasingly niche markets. DB2 for instance is no doubt profitable. Also whatever license fees IBM harvest probably pales in comparison to the revenue from the associated services that come with it. The long term market is that the DB software is free (maybe even as in speech). MS is already planning to integrate SQL server into their OS. When they do, selling DB engines will no longer be a sound business model.
Nice side effect of all this is that in principle you can have good quality stuff for free. Linux has already benefitted from this enormously since many of the improvements in the 2.4 kernel are contributions from various large UNIX vendors. File systems are no core business to anyone any more so you might as well give them away -> linux has dozens of high quality filesystems to choose from now.
The contribution model is beginning to be somewhat mandatory. Very few vendors can actually afford to continue to support their propietary software. If it doesn't integrate with linux it is dead in the water. IBM knows this, SUN hasn't quite figured it out yet.
Well on the few occasions I tried this, it barfed out halfway with some unresolved dependencies. This kind of stuff shouldn't happen with stable packages but that's just the thing you have to use unstable packages or even testing packages and you may be confronted with this kind of stuff.
I want software that is known to be in a working state for over a year to install without much trouble. Apt-get can technically deliver this experience provided that there are stable packages. And there are none available.
It's not just the kernel. To get a reasonably up to date debian based desktop environment you have no other option then to install unstable packages of software that has had stable releases for over a year in some cases.
And yes unstable actually means unstable. I've played around with debian on several occasions. The apt-get system is brilliant. Unfortunately its usefulness is totally countered by the complete lack of stable packages to install. I've actually managed to get KDE 2.0 up and running once. I just pointed the sources file to some vague http site of some guy who had bothered to create some packages. On several other occasions though I was less fortunate with installing e.g. xfree 4.0, gnome and several other mainstream stuff. I'm sure it is possible to fix if you know what you are doing but the whole point of apt-get is that it is supposed to take care of that instead.
Same for me. I had win2k on my machine earlier. I upgraded to XP and I must say it is slightly faster and just as stable. The UI is actually not that bad once you turn of the pretty colors (theres a silver color scheme which is pretty nice IMHO) I left all the graphical effects on because they really don't get in my way. I noticed that on a slower machine XP automatically turns off some options. And even on this slower machine performance was still adequate (for the record it was a PII233 with 64MB) although I prefer win 98 on it because that is a little lighter on the system resources.
.Net shit, pricing and legal issues aside I think it is safe to say that windows XP is MS' best product so far. If you intend to use it, make sure to put lots of memory in your machine. It will run with 64 MB (i tried it with office XP on the slow machine) but you will have a much smoother experience with 256MB or more (I have 512MB). Also make sure your hardware is supported since that seems to be the biggest issue with XP. I have a aureal vortex 2 soundcard and a voodoo 3 card. Both are supported by XP. However, the voodoo driver is worthless for gaming. Luckily there's an unsupported hacked together driver available on the internet. The soundcard works with the default driver but since the manufacturer no longer exists there will be no new drivers.
Most stuff is actually in the same place it used to be. Some stuff was moved but there are no radical changes I'm aware of. The new start menu is actually quite nice once you remove the stuff you don't need from it. Easy access to frequently used apps is quite handy.
The new task oriented control panel is the most confusing to people used to the old control panel. But it is organized pretty well and most stuff can be found where you'd expect to find it. If it bothers you there's an option to reverse to the old style UI.
The
I mostly agree with you actually. I just don't agree that the solution should aim for a minimal set of features that is common to all platforms since I strongly suspect that just like with AWT it will be an ugly mess. AWT provided just the basic features. You could instantiate a button, put it in a window and so on. However, even this simple functionality was too different from platform to platform which resulted in applications that had to be tested on each intended target platform to deal with all the platform specific quirks.
I agree however that there is a need for a crossplatform GUI kit. I only think that it should be a full featured one rather than some arbitrary subset. More importantly, I think it should be object oriented and language independent. Otherwise we'll see duplicated efforts for perl, ruby, whatever. Most of the existing GUI kits are language dependent or at least assume a great deal about the language used to work with them. This limits their usefulness. A good example is the GTK - QT debate. C programmers will likely favour GTK whereas C++ programmers will likely prefer QT. Swing is great when you use Java, but using it from another language is usually hard (unless it has an implementation ontop of the JVM).
SUN has been working for years to create a cross platform application framework. It is called swing. Swing is very complex and heavy on resources but it does the job. It supports printing, drag & drop, integration with the native clipboard, key bindings, skinning, all sorts of graphical stuff and lots of other stuff.
...). Swing integration is easy if you use jython. I think there are also python bindings for Mozilla so you should be able to create XUL applications in python. Presumably integration with GTK or QT is also easy.
While you might disagree that Swing is a good solution, the fact remains that Sun realized that all of the above is needed if you want to create competitive GUI apps.
The solution suggested for python seems to make the same design mistakes sun made early on (AWT) and seems to be based on the same naive view on what comprises a good GUI. In addition it seems to ignore a whole lot of other perfectly good solutions (qt, gtk, XUL, Kylix, Swing,
Wrapping is no good solution for anything but the most trivial applications. As soon as you make things more complex, you will have more and more trouble keeping things crossplatform since each platform works slightly different, has its own bugs to work around and may or may not support what you need.
The trouble with this fix is that it will cause the server to serve up IE specific code rather than w3c compliant code at whatever site you visit. Any browser detection code will incorrectly assume that you use IE and provide you with the appropriate HTML. Quite a bad thing if you have a browser that can do a good job of displaying w3c compliant stuff but is not bug for bug compatible with whatever MS chose to implement.
>Oddly, I am not aware of any such "features"
:-)
Well that is the problem of the linux community in general. They seam to be out of touch with reality. Right now linux gives me the experience the earlier versions of win95 (i.e. before Billy boy realized it was worth investing in the Internet). I mean come on, I like it as a server OS but as a desktop OS it is downright pathethic. Its cheap explorer rip offs, the ie 2.0 look alikes competing for the web browser market, the total lack of a omnipresent component model, the lack of clipboard functionality that actually works for most applications. I mean we could deny that these problems exist, convince users they really don't need such functionality but the rest of the world has had this functionality since windows 3.0 or even earlier if they were apple users.
Well to paraphrase a popular phrase by Ghandi
"first they deny you, then they fight you and then you win"
When it comes to usability linux is still in the denial phase, occasional even fighting the "a GUI is actually useful" notion. I must admit some KDE developers do have a grasp of reality and I admire their work. But the majority of the Unix community actually is incapable of looking beyond the command line. I perceive that as a weakness.
And yes, I had a few beers too many so don't be hard on me
You are not tech savvy and yet you want to have tight control over your computer?? Sorry, even MS can't deliver you that and linux requires you to be tech savvy to do just about anything windows does out of the box. Don't point at mandrake please, I've seen it and it is way too buggy for end users who don't know what they are doing. If you care to know, I thought win9x was actually more reliable than that piece of shit. And that particular OS caused me to swear in very nasty ways.
:-).
In any case, if you get me a linux version that is as easy to use as win XP, as secure as linux and as feature rich as win xp, I'm your man but I think we both know such an OS does not exist today (well maybe OS X but that has some hardware requirements and stability issues). I'll stick with win XP for the moment and by all means, use linux on your desktop if you have to. I just wanted to let people hear a different sound on this site. All this shit about linux being ready for the desktop is kind of premature from where I'm standing.
Privacy issues aside (I'm not denying there are any and I'm annoyed by them as well), windows XP is a very nice product. Compared to win2k it is a nice, incremental update. It is reasonably stable, it's slightly faster then win2k, it has some nice features, a polished UI that you can configure if you are offended by the use of colours. I've used it for a month now and I think a tech savvy user can eliminate most of the privacy issues: remove activation (or install corporate edition); don't sign up for passport; get rid of messenger; install a third party firewall (e.g. sygate); install virus scanner (e.g. norton) and you are all set for a reasonable safe eXPerience
Install a firewall, that's what I did. But if you are willing to trade features for security it is a free country and I could very well imagine moving to linux for this reason one day. Just don't claim that it is the same thing as windows because clearly it is not. If you use linux you will have to use the commandline because most file managers are crap, you get less integration between apps and so on.
Anyone suggesting linux as an alternative to windows XP for end users does not understand a thing about usability. I was reading a review about mandrake earlier today it boiled down to nice look and feel, pitty it's so buggy (this was in the review and is also my opinion entirely).
Windows XP is a nice product. It's a pitty the kernel tries to contact Redmond through the internet on a regular basis, but that doesn't change the fact that it is also setting some standards from the point of view of usability.
I notice the linux community is acting very conservative and even scared towards win xp. Jokes about clippy, blue screens and so on are made a lot whereas proposals to include new useful features into e.g. KDE or Gnome are a lot less frequent. Windows XP means a lot of new work for those projects because they are way behind delivering the same out of the box experience as windows XP.
I have windows XP at home. After some fiddling I managed to get it reasonably secure and prevent it from contacting MS too often. I love it. A lot of the new interface makes sense. Of course you want to at least change the colour scheme but a lot of things work really nice.
At this point, linux is a downgrade for end users. You lose features. You may say you don't like the features, you don't need the features or don't want the features but the bottom line is you lose them. And if you really need a unix environment, most of the more useful UNIX tools have win32 ports available. And yes windows xp is stable and won't bsod unless you install crappy drivers. If you load a buggy kernel module you have the same experience in linux.
Just for clarity, I have used linux for years and maintain a debian server. I know my shit and have hands on experience with most mainstream distributions.
The question is not whether it will run but whether it will integrate. Can I drag and drop between konqueror and evolution? Can I drag an url to a new message? What about Koffice (cut and paste without losing layout)? What about themes? The better it integrates the more useful it is.
It's not about right and wrong. What you find wrong (i.e. copying & distributing music presumably) is simply a technical reality, deal with it and get over it. Money is not protected by copyright but by additional laws which basically make it criminal to print your own money so that's a bogus argument on your side.
"And if you've no money coming in from record sales, then how are you going to buy equipment and studio time? Or even afford to eat? Naw, you just want to justify your greedy hording of MP3's"
Well I guess you could get out and get a job or something. Sorry, it's a tough world. All I'm saying is that it no longer makes sense to create and distribute physical copies of music since it can all be done using networks at virtually no cost. If your business model is selling plastic discs with music on them you obviously have a revenue problem. Basically all techniques for preventing illegal downloads of music (and movies) appear to be fundamentally flawed (see my previous post) so you can't replace your obsolete business model by simply asking money for the downloads (again sorry, there's no moral judgement here just a technical argument).
My earlier suggestions will certainly not replace all of the lost revenue. It just serves as an example of how you can still make money after the collapse of the recording industry. And after all, creativity does not require revenue at all. Some people just love to make good music because it pleases them. Basically aside from the past 75 years, throughout history there was no way to distribute copies of music. Still people like Mozart and Beethoven delivered some very fine music so it is not the end of the world.
Software requires an exact digital copy and duplication is somewhat more complex due to that. An encoded mp3 you can record by playing it and holding a microphone next to the speaker. Similarly you can record a movie. Admittedly the quality will suck but there are ways of improving on quality by using e.g. better equipment.
You can't do this with software. That doesn't mean pirating software is impossible. It is just a bit harder.
The whole point is that copyright law no longer is working. The law was built with the idea of physical copies in mind. Currently there is a whole industry around distributing physical copies of music to clients. With fat internet pipes readily available, zero cost duplication of content is possible (exluding the price of using the connection of course).
/mp3 directory which is huge and it will be a long time before I rm -rf it.
The music industry is still charging as if there were a physical copy but that is no longer true.
Instead of argueing about what is theft and what is not I suggest a more pragmatic course of action. Pragmatism starts with realizing that:
- if it can be played it can be recorded and vice versa so recording enables other people recording again.
- if it is available digitally it can be distributed at no cost. With a peer to peer setup this basically means you don't even need an expensive server setup.
- if it can be done, it will be done whether you stamp your feet on the ground loudly or not.
This applies to movies and music. These are not things that are open to debate, these are facts of life. Once you realize that, you also realize that the cd producing industry as we know it today is doomed to die eventually. The factories that create the machines to create the cds are no longer necessary, the factories that create the cases for the cds are no longer necessary, the shops that sell the cds are no longer necessary, etc.
The things that are necessary are artists to create content and supporting staff and equipment to help them record the content and optionally marketing people to market the content. Most artists consider albums to be marketing material for their live shows. Generally they don't make a lot of money from these cds since most of the profits go to the record industry. Of the 20$ you pay for a cd, only a fraction of that actually ends up in the artist's wallet Really only the very big artists can make a living out of cd sales.
So how can you make money of content creation? We have already established that the distribution has no meaningful cost associated with it so realistically it is the content creation that should generate the revenue and not the distribution.
Suggestions:
- Live performances. People love live performances and are generally willing to pay for it.
- Commercial activities. If you're famous, you can help promote stuff for money. You could for instance get a sponsor. Many sports people for instance wear clothes from their sponsor and get paid for it.
- Video clips are broadcasted on tv channels who make money by receiving revenue from advertisements.
These are only a few examples. All of these activities actually benefit from free distribution of content. And more importantly, for many artists these are already the primary source of income.
Reality is that I have a