Basically Cringely is arguing that the court system, whose timetables are based on pre-industrial information flows
Something I'd like to comment about this is that even in pre-industrial times, injunctions used to be imposed to stop the continuing behavior. When you think about it, the same way in which not injuncting allows a generation of computer time to pass, injuncting would allow a generation of computer time to pass *without* the offending product(s). The amount of actual cure that would conceivably provide would be pretty amazing. I'd think the best thing any case being issued now could do would be to place 90% of the effort into just injunctioning and dragging the case out. Touche.
Realize that this is a *distribution* license. So, the best way to take the above is that if you distribute a GPLed program to someone and that someone never distributes the program under the GPL, but they try to sue you, you can't punt the problem up to the person who gave you the program.
The GPL, at each link, prevents handing over liability to the next level. So, generally, each company who distributes a GPLed program is liable. This, nicely, also fits well if companies become the main provider of GPLed software since they're likely selling it to you. Works pretty nice, eh?
.. as the rest is mostly lame apologies for the state of Unix GUI.
They're not Unix GUI. Would you please stop trolling it as Unix. Unix is a trademark. *nix is a vague generalization for all Unix-like OSs. X is the gui of most *nixes, but there are non-*nixes that use X. And for X, I make no apology for some of the weaknesses it has which need correcting (visual modes and clipboards).
>I'm sure you hate WMP's theming then. Or Winamp's.
Yes, I do. However, look at something like Windows Mozilla. It has themes, yet I set my menu color to purple, it respects that change. This is because Windows' basic settings are independant of any particular software monolith.
Windows is the software monolith. If you used Mozilla under gtk+, you'd notice it follows GTK+ themes. Amazing, huh? Software can be designed to follow its toolkit. Magically make it follow other toolkits, though, requires..magic (or a good bit of work).
>It's not a problem that it was "punted to the top". It's called choice.
So you can choose a broken clipboard, or another broken clipboard. Not a choice I want.
Nice reordering of what I said. I addressed the clipboard problem as a real issue. It's an issue of the X layer (having multiple clipboards) and the toolkit layer (choosing to use one or the other clipboard, using their own non-standard encoding format (the X clipboard is rather raw), and just plain forgoing the X clipboard completely in some cases).
I strongly suspect that you are thinking far too much inside the box. Just because Unix has a layered model doesn't mean it's ideal or even good.
Compared to thinking outside the box which causes severe code mingling which causes all sorts of security/stability problems (like in Windows)? Yea, stupid me.
The reason the clipboard is broken in Unix is because nobody implemented a usable low-level layer to perform this service -- largely because that would involve "policy" and would be "hard".
No, because everytime someone comes up with a new, low level service to provide clipboarding it's just *another* clipboard which only works with a few programs. Various toolkits then don't necessarily use them in the "right" way. So, the best method of resolving all of the above would be to block the old clipboard and set up a standard for a new clipboard.
The real question then is, should clipboarding be in a) the kernel, b) a library interface, c) part of a single program (say the X server) which provides it to other programs.
(A) is bad not only because of security/stability issues, but also updating the kernel every time there's another clipboard enhancement becomes a huge hassle.
(B) is better because it provides system wide access, but that also means that every program that wants to use the clipboard has to sanely handle anything you might throw in the clipboard, from simple text up to complex object structures.
(C) is possibly better because it has a higher base standard for programs so requiring them to handle images, audio, complex objects, etc isn't as big of a deal. At the same time, (C) excludes all simple programs which means those not in the (C) group need their own clipboard..and then we've just started our problem all over again.
So, (B) is the best answer and it means that clipboarding will be a headache for most programs. Maybe then we'd have a toolkit/library to handle image->text conversions for the commandline progmras? Whatever, there's serious issues crossing between commandline and gui programs (or really even two gui programs of different paradigms). There's something called "shared memory" which provides just the sort of structure to do the underlying job. There's also the problem that we have to settle on a data storage standard. It has to support images, text, formatting, etc. It also has to properly convert data to the host bit format (big or little endian) and possibly upgrade/d
But one thing Microsoft got right (office aside) is that since Windows 95 there has been a style guide. Not everyone uses it of course, Adobe being the immediate culprit that springs to mind,...
the packages that don't use them look odd and feel odd and eventually got folding in because people expect applications to work in certain ways.
Well, Adobe seems to be going strong regardless of your claim. I guess end users care more about applications than the style guide.
UI consistancy is not down to the OS manufacturer, it's done to developers who want to stick with guidelines, as opposed to either not caring or wanting to do something their way.
Well, if an OEM isn't allowed to include rival products... In seriousness, I'm not sure I can agree to your point about it being the developers job to stick to some guidelines. It might make sense in Windows or Mac environments where there's only one respective company which each have their own one respective style guide, but in *nix there are various toolkits and each can have their own style guide, each of which can clearly be different. So, it's not a simple matter to just throw the blame on developers because even now Windows and Mac developers can't just recompile their programs and they magically fit some other style guide.
Competing UIs are arguably not a good thing for end users, frankly they don't care, they just want consistancy on their computers, something Linux just doesn't offer.
What do you think Mac and Windows are? Or do you think we should all switch to using Macs?
And would you please stop blaming for Linux this? Linux is a kernel. Linux isn't an OS. Yes, I've personally never sees a distro of Linux that has a consistent UI throughout. I think Lycoris and Lindows might come really close, though (from screenshots I've seen). Please realize, though, that Lycoris or Lindows or Redhat are a lot like Mac OS X. The majority of the underlying system wasn't written by Apple. It's the GUI of Mac OS X that stands out as Apple's work, and there's nothing stopping a Linux Distro *cough*Sun's Distro*cough* from trying to be a complete end user experience.
But, the above just leads back to the idea of an OS and a toolkit being inseperable (something you seem to prefer). So, maybe that's a good idea for some business to work on making a whole environment distro. That seems to be the plan of various companies which use Linux, anyways. And the resulting distros can compete against each other which might give some idea of what users want. Or it might help none. I think it's silly, though, to place all of this on Linux's shoulders though, since FreeBSD or NetBSD or just about any *nix-like OS could be an easy replacement for the kernel. And if you want to say there are no *nixes that have a good GUI, I guess you tell Apple. The problem obviously is more a question of work by people trying to sell a product than developers just trying to reasonably make an interface that does something.
Layering is a good thing, but I seriously disagree that it was "designed properly" for Unix.
It's not good design to stick drawing routines in the kernel to make the system faster at the risk of stability, for starters. Layering provides not only discrete stability (you know clearly what breaks, if you update a thing at a time), but it also provides discrete security. Not to mention that networked gfx is pretty useful if you ever actually use machines on a LAN and would like to actually not have to copy over everything you're working on all the time.
Sure you can run 5 different toolkits at the same time... but the clipboard doesn't work properly between them
Yes, X has a few too many clipboards and a clear standard does need to be written so everything works sanely together. Of course, I'm still waiting for every app in Windows to properly handle copying bold, colored, etc text + images.
, they all have different theme defintions and look different,
I'm sure you hate WMP's theming then. Or Winamp's. Or how IE6 looks different than IE5, but only in Win XP. And all that theme stuff added to Win XP which doesn't work with all apps..yea, that's horrible too.
and you have to recompile apps to get aliased fonts.
Most distros have this magically thing we call packages. In them, there are these things called binaries. And those have what we call "features". If you want to bitch that not all programs support aliased fonts (just gtk+, qt, and possibly other toolkits), go right ahead since there wasn't a standard lib to display anti-aliased fonts until recently. You can also bitch that it took so long. But, claiming that it requires recompiling is silly for any good distro you're using (exception to all the distros that are source base, but then compilation is your choice, not a forced situation).
This is all because stuff that should be in a mid-layer was punted up to the top. Note that Windows/Mac have had none of these problems -- a very strong indicator that Unix was not well designed in this regard.
It's not a problem that it was "punted to the top". It's called choice. Mac and Windows have one way of doing things, so it's obviously easy to modify one library to change all programs. With X, you can choose your toolkit (or use straight X). Are you going to complain that in-game text in Windows/Mac isn't anti-aliased? Or should they have punted to the mid-layer? The fact is, X gives developers a lot more choice and that means that users have to live with it to some extent (for any good distro, that just means it makes the distro's job harder). If you don't want choice, then just use apps written in one toolkit with all the eye candy you want. There's even distros that specialize in it (Lindows and Lycoris spring to mind).
Bitching that Windows is only 90% consistent is rather silly when Unix is *at best* 50% consistent.
Linux isn't Unix. Mac OS X isn't Linux. Mac OS X isn't Unix. Etc. If you want to say most Linux distros are *at best* 50% consistent, I'd agree with you. My advice would be that if you want consistency to pick a distro that is 90%+ consistent, since they do exist.
Well, first off I have to say I don't know much about commercial Unixes as I've only used Solaris. However, while you attribute the blame to Unix and Dennis Richie, I don't believe that the source of the problem is layering, however.
Layering in general is a great way to develop a whole system in stable parts so that whole layers can be easily supersede without breaking the whole system. Because layers are designed properly, it's possible to run gtk+, qt, motif, etc applications side-by-side. As a result, having a consistent UI policy in one toolkit does nothing to guarantee the overall system will look proper (unless, of course, you only use programs from one toolkit).
But, having a consistent UI policy within each toolkit isn't always guaranteed, and that's hardly a *nix-centric problem. Upon releasing Microsoft Windows 95, the Find dialog violated UI policy while MS Office 95 came with its own file dialogs. Of course, Microsoft has cleared up most the UI policy issues (by stuffing everything into Internet Explorer, apparently) so you'd think it'd be a moot point. But people are going to Mac OS X over Windows precisely *because* Mac OS X has a better interface.
Since Mac OS X is built off of BSD, this really tells me there's something to it other than kernel philosophy. The root issue of UI development is not only usability but also *consistency*, and Windows (and Linux, really) is more about making a new flashy UI to attract attention than about usability which screws with the consistency. The obvious answer then is to have one company who makes the UI consistent (Apple for Mac OS X, Redhat/Sun/Novell/ for Linux, Microsoft for Windows). To me, having 3+ companies working with Linux each to make their own consistent UI is a good thing (and before you start flaming about how this just splinters Linux, this is why Distros are OSs all their own and Linux isn't (besides, competing UIs is a good thing..and 98%+ (a guess) compatible packaging is good too)).
Yes, lets put small understaffed local companies in charge of the emergency phone system.
Compared to a small, understaffed local government in charge of handling the actual calls? Why, we should give have Verizon take over that too and see how many people die as a result. Or maybe hand it over to a small company.
George Bush is this thing we call the President. That means, that when he says, "I'll create more jobs", you ask yourself, "Why isn't he doing that right now?"
If Howard Dean said it, we could bitch about how that it would mean more taxes regardless of whether he made mention of it. With Bush, this is the nth package he's talked about which would involve a good deal of spending without raising taxes. Given that eventually we can't load ourselves enough money to allow for all the programs required with the current tax level, there seems no indication that taxes will go up, and no indication that current programs will be cutback, all of the above either leads to George Bush being a huge liar about really supporting all the programs he talks about or he's setting up for rampant inflation/a recession.
Personally, though, I wouldn't believe any presidential candidate who was offering such things, nor do I think it's the government's business to fund such. Ie, I'd be just as much against Howard Dean if he supported it. (The only way I can take exception to that is if there was good proof that the telecommunication conglomerates were unfairly holding back broadband to cause intentional overpricing in which case there might be a basis for an anti-trust case which *might* eventually lead to ubiquitous broadband, and the would-be President could push towards such a case.)
I can understand modding me -1 Flaimbait, -1 Troll, or -1 Overrated (if you don't think Bush is a tyrant). But -1 Offtopic when the topic is about tyranny and this forum is online? Hell, the post is a prime example of tyranny talk plus the web. At least bump me down for a good reason.
> Bush has dared to tell the truth about Iraq over and over again.
Yes, the whole WMD thing. We're just rolling in those. Yes, I know, he had "credible sources" that Saddam had access to Uranium, yet now he says he's wrong and so does everyone else. So, he was at minimal untruthful. Now, there seems to be various people saying he was interested in attacking Iraq prior to said "credible sources" giving him any reason to. This indicates he had a predisposition towards accepting "credible sources" even if he couldn't back them up solidly with other supporting information. So, either Bush was untruthful (invalid information) and deceptive (had an ulterior motive) or Bush was untruthful and utterly gullible and ineffectual (given that there were *no* other sources supporting the one he was going on and several that specific stated the opposite).
>You forget that Cheney quit Halliburton before taking office.
So? Cheney worked for Halliburton, quit and is hiring Halliburton, and Halliburton is (and very likely will) donate money to Rebuplicans. Even more:
Cheney earned forty-four million dollars during his tenure at Halliburton. Although he has said that he ?severed all my ties with the company,? he continues to collect deferred compensation worth approximately a hundred and fifty thousand dollars a year, and he retains stock options worth more than eighteen million dollars. He has announced that he will donate proceeds from the stock options to charity. -- http://www.newyorker.com/fact/content/?040216fa_fa ct
But, maybe I should branch out with how the President has often ignored the Constitution (and so has the Senate). Conveniently pretending that the Constitution says "citizen" and not "person" to justify indefinitely holding people on US soil. Whether Bush deserves to be President (considering the whole Florida fiasco) could be challenged as well, though that has a lot less standing (we're still, unfortunately, under the reign of our Representatives hiring our President which obviously creates issue on the ability to have a sanely balanced tri-branch government).
All in all, I would say that Bush is at minimal a borderline tyrant who while without absolute power has done many questionably lawful things and has been tyrannical in other countries. Removing Saddam might have been a good thing, but the reason wasn't. And if he couldn't come out at first with the balls to tell the reason for his move, then I and all other people in this country will be left questioning his real motives for his actions. That, my friend, is the trait of a liar.
It certainly hasn't worked to oust Bush. But then I guess we're too civilized to overthrow him. Of course, we could tell our Representatives and Senators to impeach and eject him and the vice president for very apparently lying (George Bush for his motives over Iraq (which would be lying to the American people--something that the Republicans wanted to Impeach Bill Clinton over) and Dick Cheney's apparent conspiracy before and after related to Halenburten (sp?) (which would obviously make him worthy of questioning of a crime)). Do realize that impeach means to charge with a crime, not find guilty of a crime. I very well question the ethics of the Republican (and Democratic, though the currently don't have much of a voice in the Senate) parties for not at least doing a serious investigation. It's nice to know we can trust our representatives to think of the people first and their party second.
> Its amusing how many of your cry about how Linux is still lacking desktop penetration, yet you are quick to run anybody down using it to make a product for PROFIT
The question wasn't about Linux, it was about MythTV. Linux is a kernel. MythTV is a user program. And someone else already made a proper comment about it being important to do proper research on GPL compliance instead of asking a half-assed questioned. The Slashdot editors do not well represent the Linux community any more than any other single person. Editors, also, aren't journalists and really shouldn't be making journalistic questions they aren't prepared to answer.
PS: MythTV can be run on FreeBSD (at minimal) if not a large variety of other OSs (though, admittedly, MythTV still assumes you're using Linux so using a Linux API emulation layer might be necessary (for those who do know more aobut this, please comment)). So, the whole Linux/Windows thing isn't relevant.
May I ask why you assume that Windows will be more capable of providing good installation defaults over any distribution of Linux that exists? Or are you stating that there will be at least one good version of Windows (comparable to how 2003 Server was a huge improvement, installation defaults wise) and that the fractured nature of Linux ensures that a smaller percentage of distributions will have as good of results?
I would guess that if anything, both could have very similar sane results given their underlying security models are near identical. I would also guess that Linux will definitely have such a version because the NSA (at minimal) has shown a very serious interest in using Linux as a secure core of an operating system which to me means they will extended it to an entire set of tools. I do not, however, know if Microsoft will ever have a strong enough interest to make such a version of Windows for a very small subset of the populace. It is likely, IMHO, that there will be a version of Windows and a distribution of Linux which will provide an *adequate* amount of security which will be less secure but more practical. I do not believe that such a distribution of Linux or version of Windows exists currently, though, that even rates as adequate. I don't know enough about OpenBSD to begin to comment on it..
Of course, you can't stop "soft money" expenditures by companies (as that really is an encroachment on freedom of speech)
Companies and corporations aren't people. Corporations are only "people" to the extent of limited liability. The Constitution protects the right of a company president, acting as a person, to give money to someone else.
There's nothing in the Constitution that'd stop there being a law blocking companies (ie, people acting as something other than people (like a large company)) from donating soft money. And there's nothing stopping an owner of a company paying himself a large sum of money from a company and donating to politicians (beyond, of course, whatever happens now as punishment for giving away lots of money). Do realize, though, that such legal transfers mean more taxes (paying yourself == income tax) which would minorly decrease the resulting soft money received by a small percent. Then we'd get to find out the legality of forcing people to contribute money they get in their pay check towards political groups (or do you really think they'd gives a single CEO a couple million dollars which he's "encouraged" to hand over to some lobbying group?).
Well, Linux 2.6 contains labels for security as well (SE Linux had it first, but Win 2003 obviously beat Linux to "official" support of it). I'm not sure if SE Linux had MAC too, but I'd assume so given it was worked on by the NSA. Personally, I hope that MAC for Windows doesn't rely on people not spoofing IP addresses (as apparently some people believe you can't fake an IP address...).
The security mechanisms in the NT derivatives are considerably more sophisticated than those in traditional UNIX systems.
I'd really like to know what those security mechanisms are. Far as I knew, the only major difference between NT and *nix, security wise, is it's much more trivial to manage giving rights of files to groups of people than it is in *nix (in general, since not all *nixes have more than the 3 octal permission setup) and you can have more combinations.
Extended ACLs in ext3 might eventually change that for Linux, SGI has had XFS for some time, and I have no idea about what JFS supports. UFS, as far as I know, doesn't have any such extensions. Beyond this, there's only group and user run ids to concern about, and as far as I'm aware NT isn't any different with those.
It has come to light recently that Linux is eatting at the very fabric of our future society. In a disturbing trend documented at a popular Linux enthusiast website, Slashdot, documentation of sexual relations between device programmers and Linux have surfaced. Many such programmers have died, leaving a parent or sometimes a girlfriend emotionally devastated without a significant source of income to provide in future years.
Even more devastating are the programmers who, aftering mating with Linux, form a lifelong relationship with Linux thereby inhibiting the production of future generations of programmers. Even worse, Linux being open source carries the real possibility of future Software EXchange diseases which could cripple said developers before their time. Various comments on Slashdot imply that such a coupling with Linux is healthy for the current economics, yet there exist no document involving developers mating with Windows.
It is clear, the future is using Windows on embedded devices where the clear relationship lines are drawn. Being proprietary, Windows can offer a safe and company backed relationship that can be broken off if there's ever trouble. And because people pay for Windows or competing embedded services the free market is at work allowing for no single competitor to somehow strong arm developers into using one product. For me, the future looks bright with Microsoft.
The fact is, this idea of trying to segregate off the "bad businesses" makes sense in the real world to some extent because factories produce a lot of noise/pollution, generally. In the virtual world, www.whitehouse.gov isn't just by www.whitehousf.gov necessarily. Sure, mistyping the name might get you to the wrong place, but unless whitehousf.gov is trying to defraud you into believing its whitehousf.gov, you know you took a wrong step.
That's life, though, and I don't think it's at all something that should be under the power of the government to enforce who can use what name for whatever reason. If the US gets pissy enough, maybe the EU will start up their own DNS service to provide less strict names. Then they'll only default to asking the US.com server whenever "necessary". The US shouldn't unilaterally try to force control over the internet when it could very well fracture the internet in a very bad way. If the US wants to offer more TLDs, fine. But, forcing the use of one TLD over another is likely to just move people out of the US and possibly into their own DNS system which isn't so anally retentive about protecting the poor stupid Americans.
I agree with your point that it's not high broadband. Most worms are small enough to only take 8 seconds are less to send. Even being always on is hardly anything, since most dial-up users stay on for hours at a time. I don't think that the morris worm really woke up people as much as you say it did. Neither has Windows worms.
The major difference is until recently, there weren't that many machines online relatively and in the interim *nix in all its forms (though not really Linux, specifically) had the chance to fix most bugs. *BSD have too because they got to hear the same problems. Because Linux includes *BSD tools and GNU tools (based in idea off of Unix tools), it's at about the same level of security checks as *nixes.
Now, that that's covered, you should realize that spam receiving isn't something that only Windows users suffer. And most spam relays spoof their address.
Windows machines are ideal candidates more because their admin is less likely to properly admin (aka, security patch) things which are inately on. More of the core system hasn't been tested, while on *nix security flaws tend to be in auxillary programs (apache has become more core and hasn't had the same level of testing as say finger...and the Linux kernel is also relatively new) where less testing has occurred.
Overall, this means more Windows systems (by percentage, not by populace) are infected and that coupled with populace and broadband means more spam, not more worms. After all, you can get/send 1,000 spam messages a day, but worms tend to be incompatible with each other meaning stable systems tend to only stay infected with two or three worms. More broadband just makes you a better spam relay, which means more clogging of the internet. I don't think it'll cause any more of a collapse than P2P has. At the same time, I wonder if at some point ISPs will start trying to regulate e-mail more to save costs from all the junk being sent through them. The most ironic part to me is not only how little people seem to care about their machines, but that there's been no citizen outcry trying to sue for violating of their machines.
Slashdot Mirror
Basically Cringely is arguing that the court system, whose timetables are based on pre-industrial information flows
Something I'd like to comment about this is that even in pre-industrial times, injunctions used to be imposed to stop the continuing behavior. When you think about it, the same way in which not injuncting allows a generation of computer time to pass, injuncting would allow a generation of computer time to pass *without* the offending product(s). The amount of actual cure that would conceivably provide would be pretty amazing. I'd think the best thing any case being issued now could do would be to place 90% of the effort into just injunctioning and dragging the case out. Touche.
Realize that this is a *distribution* license. So, the best way to take the above is that if you distribute a GPLed program to someone and that someone never distributes the program under the GPL, but they try to sue you, you can't punt the problem up to the person who gave you the program.
The GPL, at each link, prevents handing over liability to the next level. So, generally, each company who distributes a GPLed program is liable. This, nicely, also fits well if companies become the main provider of GPLed software since they're likely selling it to you. Works pretty nice, eh?
.. as the rest is mostly lame apologies for the state of Unix GUI.
They're not Unix GUI. Would you please stop trolling it as Unix. Unix is a trademark. *nix is a vague generalization for all Unix-like OSs. X is the gui of most *nixes, but there are non-*nixes that use X. And for X, I make no apology for some of the weaknesses it has which need correcting (visual modes and clipboards).
>I'm sure you hate WMP's theming then. Or Winamp's.
Yes, I do. However, look at something like Windows Mozilla. It has themes, yet I set my menu color to purple, it respects that change. This is because Windows' basic settings are independant of any particular software monolith.
Windows is the software monolith. If you used Mozilla under gtk+, you'd notice it follows GTK+ themes. Amazing, huh? Software can be designed to follow its toolkit. Magically make it follow other toolkits, though, requires..magic (or a good bit of work).
>It's not a problem that it was "punted to the top". It's called choice.
So you can choose a broken clipboard, or another broken clipboard. Not a choice I want.
Nice reordering of what I said. I addressed the clipboard problem as a real issue. It's an issue of the X layer (having multiple clipboards) and the toolkit layer (choosing to use one or the other clipboard, using their own non-standard encoding format (the X clipboard is rather raw), and just plain forgoing the X clipboard completely in some cases).
I strongly suspect that you are thinking far too much inside the box. Just because Unix has a layered model doesn't mean it's ideal or even good.
Compared to thinking outside the box which causes severe code mingling which causes all sorts of security/stability problems (like in Windows)? Yea, stupid me.
The reason the clipboard is broken in Unix is because nobody implemented a usable low-level layer to perform this service -- largely because that would involve "policy" and would be "hard".
No, because everytime someone comes up with a new, low level service to provide clipboarding it's just *another* clipboard which only works with a few programs. Various toolkits then don't necessarily use them in the "right" way. So, the best method of resolving all of the above would be to block the old clipboard and set up a standard for a new clipboard.
The real question then is, should clipboarding be in a) the kernel, b) a library interface, c) part of a single program (say the X server) which provides it to other programs.
(A) is bad not only because of security/stability issues, but also updating the kernel every time there's another clipboard enhancement becomes a huge hassle.
(B) is better because it provides system wide access, but that also means that every program that wants to use the clipboard has to sanely handle anything you might throw in the clipboard, from simple text up to complex object structures.
(C) is possibly better because it has a higher base standard for programs so requiring them to handle images, audio, complex objects, etc isn't as big of a deal. At the same time, (C) excludes all simple programs which means those not in the (C) group need their own clipboard..and then we've just started our problem all over again.
So, (B) is the best answer and it means that clipboarding will be a headache for most programs. Maybe then we'd have a toolkit/library to handle image->text conversions for the commandline progmras? Whatever, there's serious issues crossing between commandline and gui programs (or really even two gui programs of different paradigms). There's something called "shared memory" which provides just the sort of structure to do the underlying job. There's also the problem that we have to settle on a data storage standard. It has to support images, text, formatting, etc. It also has to properly convert data to the host bit format (big or little endian) and possibly upgrade/d
But one thing Microsoft got right (office aside) is that since Windows 95 there has been a style guide. Not everyone uses it of course, Adobe being the immediate culprit that springs to mind, ...
the packages that don't use them look odd and feel odd and eventually got folding in because people expect applications to work in certain ways.
Well, Adobe seems to be going strong regardless of your claim. I guess end users care more about applications than the style guide.
UI consistancy is not down to the OS manufacturer, it's done to developers who want to stick with guidelines, as opposed to either not caring or wanting to do something their way.
Well, if an OEM isn't allowed to include rival products... In seriousness, I'm not sure I can agree to your point about it being the developers job to stick to some guidelines. It might make sense in Windows or Mac environments where there's only one respective company which each have their own one respective style guide, but in *nix there are various toolkits and each can have their own style guide, each of which can clearly be different. So, it's not a simple matter to just throw the blame on developers because even now Windows and Mac developers can't just recompile their programs and they magically fit some other style guide.
Competing UIs are arguably not a good thing for end users, frankly they don't care, they just want consistancy on their computers, something Linux just doesn't offer.
What do you think Mac and Windows are? Or do you think we should all switch to using Macs?
And would you please stop blaming for Linux this? Linux is a kernel. Linux isn't an OS. Yes, I've personally never sees a distro of Linux that has a consistent UI throughout. I think Lycoris and Lindows might come really close, though (from screenshots I've seen). Please realize, though, that Lycoris or Lindows or Redhat are a lot like Mac OS X. The majority of the underlying system wasn't written by Apple. It's the GUI of Mac OS X that stands out as Apple's work, and there's nothing stopping a Linux Distro *cough*Sun's Distro*cough* from trying to be a complete end user experience.
But, the above just leads back to the idea of an OS and a toolkit being inseperable (something you seem to prefer). So, maybe that's a good idea for some business to work on making a whole environment distro. That seems to be the plan of various companies which use Linux, anyways. And the resulting distros can compete against each other which might give some idea of what users want. Or it might help none. I think it's silly, though, to place all of this on Linux's shoulders though, since FreeBSD or NetBSD or just about any *nix-like OS could be an easy replacement for the kernel. And if you want to say there are no *nixes that have a good GUI, I guess you tell Apple. The problem obviously is more a question of work by people trying to sell a product than developers just trying to reasonably make an interface that does something.
Layering is a good thing, but I seriously disagree that it was "designed properly" for Unix.
... but the clipboard doesn't work properly between them
It's not good design to stick drawing routines in the kernel to make the system faster at the risk of stability, for starters. Layering provides not only discrete stability (you know clearly what breaks, if you update a thing at a time), but it also provides discrete security. Not to mention that networked gfx is pretty useful if you ever actually use machines on a LAN and would like to actually not have to copy over everything you're working on all the time.
Sure you can run 5 different toolkits at the same time
Yes, X has a few too many clipboards and a clear standard does need to be written so everything works sanely together. Of course, I'm still waiting for every app in Windows to properly handle copying bold, colored, etc text + images.
, they all have different theme defintions and look different,
I'm sure you hate WMP's theming then. Or Winamp's. Or how IE6 looks different than IE5, but only in Win XP. And all that theme stuff added to Win XP which doesn't work with all apps..yea, that's horrible too.
and you have to recompile apps to get aliased fonts.
Most distros have this magically thing we call packages. In them, there are these things called binaries. And those have what we call "features". If you want to bitch that not all programs support aliased fonts (just gtk+, qt, and possibly other toolkits), go right ahead since there wasn't a standard lib to display anti-aliased fonts until recently. You can also bitch that it took so long. But, claiming that it requires recompiling is silly for any good distro you're using (exception to all the distros that are source base, but then compilation is your choice, not a forced situation).
This is all because stuff that should be in a mid-layer was punted up to the top. Note that Windows/Mac have had none of these problems -- a very strong indicator that Unix was not well designed in this regard.
It's not a problem that it was "punted to the top". It's called choice. Mac and Windows have one way of doing things, so it's obviously easy to modify one library to change all programs. With X, you can choose your toolkit (or use straight X). Are you going to complain that in-game text in Windows/Mac isn't anti-aliased? Or should they have punted to the mid-layer? The fact is, X gives developers a lot more choice and that means that users have to live with it to some extent (for any good distro, that just means it makes the distro's job harder). If you don't want choice, then just use apps written in one toolkit with all the eye candy you want. There's even distros that specialize in it (Lindows and Lycoris spring to mind).
Bitching that Windows is only 90% consistent is rather silly when Unix is *at best* 50% consistent.
Linux isn't Unix. Mac OS X isn't Linux. Mac OS X isn't Unix. Etc. If you want to say most Linux distros are *at best* 50% consistent, I'd agree with you. My advice would be that if you want consistency to pick a distro that is 90%+ consistent, since they do exist.
Well, first off I have to say I don't know much about commercial Unixes as I've only used Solaris. However, while you attribute the blame to Unix and Dennis Richie, I don't believe that the source of the problem is layering, however.
Layering in general is a great way to develop a whole system in stable parts so that whole layers can be easily supersede without breaking the whole system. Because layers are designed properly, it's possible to run gtk+, qt, motif, etc applications side-by-side. As a result, having a consistent UI policy in one toolkit does nothing to guarantee the overall system will look proper (unless, of course, you only use programs from one toolkit).
But, having a consistent UI policy within each toolkit isn't always guaranteed, and that's hardly a *nix-centric problem. Upon releasing Microsoft Windows 95, the Find dialog violated UI policy while MS Office 95 came with its own file dialogs. Of course, Microsoft has cleared up most the UI policy issues (by stuffing everything into Internet Explorer, apparently) so you'd think it'd be a moot point. But people are going to Mac OS X over Windows precisely *because* Mac OS X has a better interface.
Since Mac OS X is built off of BSD, this really tells me there's something to it other than kernel philosophy. The root issue of UI development is not only usability but also *consistency*, and Windows (and Linux, really) is more about making a new flashy UI to attract attention than about usability which screws with the consistency. The obvious answer then is to have one company who makes the UI consistent (Apple for Mac OS X, Redhat/Sun/Novell/ for Linux, Microsoft for Windows). To me, having 3+ companies working with Linux each to make their own consistent UI is a good thing (and before you start flaming about how this just splinters Linux, this is why Distros are OSs all their own and Linux isn't (besides, competing UIs is a good thing..and 98%+ (a guess) compatible packaging is good too)).
Don't you mean killall?
Yes, lets put small understaffed local companies in charge of the emergency phone system.
Compared to a small, understaffed local government in charge of handling the actual calls? Why, we should give have Verizon take over that too and see how many people die as a result. Or maybe hand it over to a small company.
Both sound ridiculous to me.
George Bush is this thing we call the President. That means, that when he says, "I'll create more jobs", you ask yourself, "Why isn't he doing that right now?"
If Howard Dean said it, we could bitch about how that it would mean more taxes regardless of whether he made mention of it. With Bush, this is the nth package he's talked about which would involve a good deal of spending without raising taxes. Given that eventually we can't load ourselves enough money to allow for all the programs required with the current tax level, there seems no indication that taxes will go up, and no indication that current programs will be cutback, all of the above either leads to George Bush being a huge liar about really supporting all the programs he talks about or he's setting up for rampant inflation/a recession.
Personally, though, I wouldn't believe any presidential candidate who was offering such things, nor do I think it's the government's business to fund such. Ie, I'd be just as much against Howard Dean if he supported it. (The only way I can take exception to that is if there was good proof that the telecommunication conglomerates were unfairly holding back broadband to cause intentional overpricing in which case there might be a basis for an anti-trust case which *might* eventually lead to ubiquitous broadband, and the would-be President could push towards such a case.)
Read the chapter on oligarchies, which is how a collection of monopolistic (copyrighted works) markets behave.
I can understand modding me -1 Flaimbait, -1 Troll, or -1 Overrated (if you don't think Bush is a tyrant). But -1 Offtopic when the topic is about tyranny and this forum is online? Hell, the post is a prime example of tyranny talk plus the web. At least bump me down for a good reason.
> Bush has dared to tell the truth about Iraq over and over again.
a ct
Yes, the whole WMD thing. We're just rolling in those. Yes, I know, he had "credible sources" that Saddam had access to Uranium, yet now he says he's wrong and so does everyone else. So, he was at minimal untruthful. Now, there seems to be various people saying he was interested in attacking Iraq prior to said "credible sources" giving him any reason to. This indicates he had a predisposition towards accepting "credible sources" even if he couldn't back them up solidly with other supporting information. So, either Bush was untruthful (invalid information) and deceptive (had an ulterior motive) or Bush was untruthful and utterly gullible and ineffectual (given that there were *no* other sources supporting the one he was going on and several that specific stated the opposite).
>You forget that Cheney quit Halliburton before taking office.
So? Cheney worked for Halliburton, quit and is hiring Halliburton, and Halliburton is (and very likely will) donate money to Rebuplicans. Even more:
Cheney earned forty-four million dollars during his tenure at Halliburton. Although he has said that he ?severed all my ties with the company,? he continues to collect deferred compensation worth approximately a hundred and fifty thousand dollars a year, and he retains stock options worth more than eighteen million dollars. He has announced that he will donate proceeds from the stock options to charity. -- http://www.newyorker.com/fact/content/?040216fa_f
But, maybe I should branch out with how the President has often ignored the Constitution (and so has the Senate). Conveniently pretending that the Constitution says "citizen" and not "person" to justify indefinitely holding people on US soil. Whether Bush deserves to be President (considering the whole Florida fiasco) could be challenged as well, though that has a lot less standing (we're still, unfortunately, under the reign of our Representatives hiring our President which obviously creates issue on the ability to have a sanely balanced tri-branch government).
All in all, I would say that Bush is at minimal a borderline tyrant who while without absolute power has done many questionably lawful things and has been tyrannical in other countries. Removing Saddam might have been a good thing, but the reason wasn't. And if he couldn't come out at first with the balls to tell the reason for his move, then I and all other people in this country will be left questioning his real motives for his actions. That, my friend, is the trait of a liar.
It certainly hasn't worked to oust Bush. But then I guess we're too civilized to overthrow him. Of course, we could tell our Representatives and Senators to impeach and eject him and the vice president for very apparently lying (George Bush for his motives over Iraq (which would be lying to the American people--something that the Republicans wanted to Impeach Bill Clinton over) and Dick Cheney's apparent conspiracy before and after related to Halenburten (sp?) (which would obviously make him worthy of questioning of a crime)). Do realize that impeach means to charge with a crime, not find guilty of a crime. I very well question the ethics of the Republican (and Democratic, though the currently don't have much of a voice in the Senate) parties for not at least doing a serious investigation. It's nice to know we can trust our representatives to think of the people first and their party second.
Phirebox?
> Its amusing how many of your cry about how Linux is still lacking desktop penetration, yet you are quick to run anybody down using it to make a product for PROFIT
The question wasn't about Linux, it was about MythTV. Linux is a kernel. MythTV is a user program. And someone else already made a proper comment about it being important to do proper research on GPL compliance instead of asking a half-assed questioned. The Slashdot editors do not well represent the Linux community any more than any other single person. Editors, also, aren't journalists and really shouldn't be making journalistic questions they aren't prepared to answer.
PS: MythTV can be run on FreeBSD (at minimal) if not a large variety of other OSs (though, admittedly, MythTV still assumes you're using Linux so using a Linux API emulation layer might be necessary (for those who do know more aobut this, please comment)). So, the whole Linux/Windows thing isn't relevant.
May I ask why you assume that Windows will be more capable of providing good installation defaults over any distribution of Linux that exists? Or are you stating that there will be at least one good version of Windows (comparable to how 2003 Server was a huge improvement, installation defaults wise) and that the fractured nature of Linux ensures that a smaller percentage of distributions will have as good of results?
I would guess that if anything, both could have very similar sane results given their underlying security models are near identical. I would also guess that Linux will definitely have such a version because the NSA (at minimal) has shown a very serious interest in using Linux as a secure core of an operating system which to me means they will extended it to an entire set of tools. I do not, however, know if Microsoft will ever have a strong enough interest to make such a version of Windows for a very small subset of the populace. It is likely, IMHO, that there will be a version of Windows and a distribution of Linux which will provide an *adequate* amount of security which will be less secure but more practical. I do not believe that such a distribution of Linux or version of Windows exists currently, though, that even rates as adequate. I don't know enough about OpenBSD to begin to comment on it..
Of course, you can't stop "soft money" expenditures by companies (as that really is an encroachment on freedom of speech)
Companies and corporations aren't people. Corporations are only "people" to the extent of limited liability. The Constitution protects the right of a company president, acting as a person, to give money to someone else.
There's nothing in the Constitution that'd stop there being a law blocking companies (ie, people acting as something other than people (like a large company)) from donating soft money. And there's nothing stopping an owner of a company paying himself a large sum of money from a company and donating to politicians (beyond, of course, whatever happens now as punishment for giving away lots of money). Do realize, though, that such legal transfers mean more taxes (paying yourself == income tax) which would minorly decrease the resulting soft money received by a small percent. Then we'd get to find out the legality of forcing people to contribute money they get in their pay check towards political groups (or do you really think they'd gives a single CEO a couple million dollars which he's "encouraged" to hand over to some lobbying group?).
I prefer:
for i in *.jpeg; do mv "${i}" "${i/%.jpeg/.jpg}"; done
Well, Linux 2.6 contains labels for security as well (SE Linux had it first, but Win 2003 obviously beat Linux to "official" support of it). I'm not sure if SE Linux had MAC too, but I'd assume so given it was worked on by the NSA. Personally, I hope that MAC for Windows doesn't rely on people not spoofing IP addresses (as apparently some people believe you can't fake an IP address...).
The security mechanisms in the NT derivatives are considerably more sophisticated than those in traditional UNIX systems.
I'd really like to know what those security mechanisms are. Far as I knew, the only major difference between NT and *nix, security wise, is it's much more trivial to manage giving rights of files to groups of people than it is in *nix (in general, since not all *nixes have more than the 3 octal permission setup) and you can have more combinations.
Extended ACLs in ext3 might eventually change that for Linux, SGI has had XFS for some time, and I have no idea about what JFS supports. UFS, as far as I know, doesn't have any such extensions. Beyond this, there's only group and user run ids to concern about, and as far as I'm aware NT isn't any different with those.
Pluto is a dog. Sedna is a plane. What's all this talk about planets?
Dissociative Press - March 20, 2004
It has come to light recently that Linux is eatting at the very fabric of our future society. In a disturbing trend documented at a popular Linux enthusiast website, Slashdot, documentation of sexual relations between device programmers and Linux have surfaced. Many such programmers have died, leaving a parent or sometimes a girlfriend emotionally devastated without a significant source of income to provide in future years.
Even more devastating are the programmers who, aftering mating with Linux, form a lifelong relationship with Linux thereby inhibiting the production of future generations of programmers. Even worse, Linux being open source carries the real possibility of future Software EXchange diseases which could cripple said developers before their time. Various comments on Slashdot imply that such a coupling with Linux is healthy for the current economics, yet there exist no document involving developers mating with Windows.
It is clear, the future is using Windows on embedded devices where the clear relationship lines are drawn. Being proprietary, Windows can offer a safe and company backed relationship that can be broken off if there's ever trouble. And because people pay for Windows or competing embedded services the free market is at work allowing for no single competitor to somehow strong arm developers into using one product. For me, the future looks bright with Microsoft.
Stupid question, but then would
.com server whenever "necessary". The US shouldn't unilaterally try to force control over the internet when it could very well fracture the internet in a very bad way. If the US wants to offer more TLDs, fine. But, forcing the use of one TLD over another is likely to just move people out of the US and possibly into their own DNS system which isn't so anally retentive about protecting the poor stupid Americans.
s/black/gay/g
or
s/porn/gay/g
be valid? That is obviously off-topic, btw.
The fact is, this idea of trying to segregate off the "bad businesses" makes sense in the real world to some extent because factories produce a lot of noise/pollution, generally. In the virtual world, www.whitehouse.gov isn't just by www.whitehousf.gov necessarily. Sure, mistyping the name might get you to the wrong place, but unless whitehousf.gov is trying to defraud you into believing its whitehousf.gov, you know you took a wrong step.
That's life, though, and I don't think it's at all something that should be under the power of the government to enforce who can use what name for whatever reason. If the US gets pissy enough, maybe the EU will start up their own DNS service to provide less strict names. Then they'll only default to asking the US
I agree with your point that it's not high broadband. Most worms are small enough to only take 8 seconds are less to send. Even being always on is hardly anything, since most dial-up users stay on for hours at a time. I don't think that the morris worm really woke up people as much as you say it did. Neither has Windows worms.
The major difference is until recently, there weren't that many machines online relatively and in the interim *nix in all its forms (though not really Linux, specifically) had the chance to fix most bugs. *BSD have too because they got to hear the same problems. Because Linux includes *BSD tools and GNU tools (based in idea off of Unix tools), it's at about the same level of security checks as *nixes.
Now, that that's covered, you should realize that spam receiving isn't something that only Windows users suffer. And most spam relays spoof their address.
Windows machines are ideal candidates more because their admin is less likely to properly admin (aka, security patch) things which are inately on. More of the core system hasn't been tested, while on *nix security flaws tend to be in auxillary programs (apache has become more core and hasn't had the same level of testing as say finger...and the Linux kernel is also relatively new) where less testing has occurred.
Overall, this means more Windows systems (by percentage, not by populace) are infected and that coupled with populace and broadband means more spam, not more worms. After all, you can get/send 1,000 spam messages a day, but worms tend to be incompatible with each other meaning stable systems tend to only stay infected with two or three worms. More broadband just makes you a better spam relay, which means more clogging of the internet. I don't think it'll cause any more of a collapse than P2P has. At the same time, I wonder if at some point ISPs will start trying to regulate e-mail more to save costs from all the junk being sent through them. The most ironic part to me is not only how little people seem to care about their machines, but that there's been no citizen outcry trying to sue for violating of their machines.
So...Microsoft's going to kill Hebrew?