Except for games that use audio or graphics, Linux has pretty much spot-on compatibility with any IF game, as do most operating systems. IF games are extremely portable, written to one of a number of portable VMs (and all this years before Java...and with better compatibility than Java).
TADS (IMHO the most advanced engine, though Inform is very close) just plain runs on Linux. You want this to play.gam files. There is Frotz to run Inform (.z5 files...I believe a couple other.zX formats, but I've only played.z5). There is an ADRIFT implementation called SCARE for Linux. It has a less-than-perfect parser. To be honest, ADRIFT is a much simpler engine, and I generally fine TADS or Inform games to be much more fun and impressive.
Note that other classic adventure game VMs -- the ones for commercial graphical adventures -- like the Sierra (King's Quest, among others) and Lucasarts (Day of the Tentacle, Sam and Max, Secret of Monkey Island, among others) VMs have been ported to Linux in the form of Sarien, FreeSCI, and ScummVM. I don't believe there have been any new AGI/SCI/SCUMM adventures made -- the engines are static and no improved games will be made for them, but they're still neat projects to have fun playing the originals on.
While you're there, try Babel (for the TADS engine) and Bliss also for TADS.
Babel, a sci-fi horror game, was the first IF game I beat without needing any hints (though it's still significantly more difficult than many IF games out there), and Bliss...probably the best example I've ever seen of using the IF format. Describing it would probably spoil it, so I can only recommend playing it. When you've finished it, you'll understand.
Note that both of these games are dark and potentially disturbing. If you like fluffy or fantastic or slice-of-life or something else, look elsewhere...
The main problem is that an AI capable of entertaining and interacting with a human would have to be very complex. The complexity of an IF world is very small compared to what it would take to train an AI.
You know those stories about people that get stranded somewhere in a foreign country, and have to perform an amazing act to learn and understand what's going on? A human baby learns extremely quickly. Each human manages to do this, to go from zero knowledge of any language to speaking and comprehending knowledge with a few years of work. The problem is that it takes *four years* of constant interaction with new people and languages, plus visual input, to learn something like this. If you consider how much sheer *stuff* there is in four years of a human life, the task becomes staggering in scope. And we'd still be working with only a text-based interface...given the lower data bandwidth, and the fact that visual knowledge from the outside world is incredibly important to understand what is going on in a game for us, it would probably take far longer with just text to work with.
So maybe it could be done, but all the IF writers in the world have not written enough content (and some of it is surrealistic or misleading). All of it offers a much less powerful world than the one humans are in.
I've thought about this -- I like playing with both speech synth and IF -- but decided that you really have to think intensely about IF. I'm uncomfortable even listening to audio books, since I can't stop and reflect when I need to. IF would be much more difficult to play without being able to stop.
While I'm posting, I wanted to say that it's wonderful to see an IF story put on Slashdot. IF fell off a lot of people's radar with Infocom. The games have advanced a lot and, IMHO, gotten much better. Problems like "hunt the verb" have been largely eliminated, and many clever tricks and elements introduced.
You know what would be a great idea? X. Too bad it would cost a bunch of money to develop X into something usefull...if only there were some way to help try to ensure that I could recover the money I spend developing it...
There are a couple of problems with this logic. First, you're worried that someone will rip off your idea. Patents were designed in a day and age when the things being patented were *simple*. An industrial process, or a simple machine. The thing is, it's *really hard* to just duplicate the functionality of a worthwhile piece of software. If you can't just take the software (and copyright takes care of that), it's generally not going to be cheap or quick for you to reimplement the idea...and in that time, the original person has moved beyond where he was. Software needs patents much less than old processes once did.
Second, you're giving an example of an exceptional idea, something really amazing. The problem is that software development is so complex compared to earlier systems that you could find something to patent in almost every new system made. This is, frankly, not how the patent system is intended to operate at all.
Third, you talk about "expense" of developing the new idea. It really *was* expensive to develop some older things -- if you want to build a new machine and figure out how to make it work well, it could take many years and lots of expensive and painstaking ironwork -- and the simple result could be copied. However, software is (comparably ) incredibly cheap to work with. You think, write a hundred lines of code, and you have an implementation to test out and work with. You don't write up a blueprint and then have an implementation to test two months later.
Fourth, older devices were much more static. A plow is a plow is a plow. Maybe someone comes up with a way to hollow out part of the thing and make it lighter...then no improvements for a while. In the software field, there are constantly surging improvements. The whole goal of an engineer is to improve on existing systems...rather unlike the masses of plow companies, that might just produce different plows of roughly the same design. Patents are *much* more onerous in software.
I worked in a research lab for a while, and I think that I can safely claim that software patents are minimally useful to society. It's fairly rare that a really good, reasonable, legitimate software patent exists -- the type of research encouraged by software patents is of the "lock people out" variety, rather than the "make something better" variety. I do not think that research would be signifiantly impacted by a lack of software patents, and I *do* think that software engineering would be much easier.
The problem, I think, is that developers will happily use something if it makes their life easier, but not if it makes it more difficult.
If there's a good, standard parser out there, developers will use it.
GConf sounds vaguely like what I want, but does many things wrong. It implements a registry, instead of letting each program have its own file. It requires a daemon to be running (and I have had a *ton* of problems with said daemon crashing or malfunctioning in various ways in the past). It's really much more complicated than a simple "here's my config file, please parse".
I used an integrated Mac for years, and I was very happy to get out of that mess. I *hate* integrated computers. If you're making a laptop, great. Otherwise, they're a pain in the ass -- a pain to work on, a pain to upgrade, a pain to replace a broken part.
This is great for the advanced user but it floods the general user with too many choices.
I've never agreed with this philosophy.
Do what ESR did with fetchmailconf. Make all the options available. Just have the ability to autodetect everything you can. You don't need to hide options (this is a hugely irritating factor of Windows wizards), you just need to have intelligent defaults and good autodetection, so that things can work pretty much out of box.
When you're reading a mailing list and someone emails "how do I do X", the usual response, as ESR has pointed out, is to simply answer them. If the problem comes up a couple of times, perhaps it goes into a FAQ. There is one missing thing, though, that doesn't happen. The author should ask "what could I change in the *software*, not the documentation or FAQ or whatnot, that would keep users from coming to ask me about this again."
ESR's "bugger me with a chainsaw" appears to be a reference to the "fuck me gently with a chainsaw" phrase in Heathers. I got a kick out of that one...
Look. I love the UNIX printing system. It's a real pain in the ass to configure things, but it's also terribly powerful.
I've used LPR and then LPRng. I get network transparency, ability to batch-print easily (this is *not* trivial...try printing out hard copies of 200 source files pretty printed in Windows). I can manage print queues.
The thing is, UNIX printing systems are usually lots of little parts cobbled together with some scripts that vary from distro to distro, and then a config GUI that the distro maintainer puts out. The UNIX printing world is terribly disorganized compared to most other things in the UNIX world.
Can you identify the function of and tell the difference between all of these? LPR, LPRng, CUPS, gimp-print, foomatic, Omni, gnome-print, printman, printtool, desktop-printing, enscript, a2ps, ghostscript, pna2ppa, samba, hpoj, gsview, gv, ps2ps, ggv, redhat-config-printer, printconf, and mpage. All of these printing-related utilities and more have been on my system in the last few years. Keep in mind that I don't even use KDE, and that most distros vary the choice of what to use and you have an interesting set of knowledge to amass. The different print spoolers have different auth systems and config formats.
On the other hand, I have an old Apple LaserWriter without enough memory to print much of anything. I salvaged it when my old university threw it out. I hooked it up, and started cobbling together bits into a print filter. Sure, it took some doing and learning, but when I was done, all pages on the printer were rendered on the computer (where all the RAM in the world was available), converted to a bitmap and compressed, and sent in an embedded postscript file to the printer. On Windows I would have been simply SOL.
So, I'm not sure that an all-in-one system would be great. I *do* think that the printing situation could be cleaned up a lot, that the distros *really* need to get together and standardize on an interface (if you want to differentiate yourselves, please don't do it on something as basic as printing, which is a huge impediment to office use everywhere), and that it'd be nice to have some degree of autodetection of intelligent defaults (After a click on "add printer", "You have a Model Foobar attached. The proper driver is being selected.")
ESR also did fetchmail. Fetchmail has an *excellent* configuration interface.
* First, the config file is simple and small. A typical configuration should be simple and small. Take a look at the difference between the size of a basic sendmail and a basic postfix installation, and you'll notice an astonishing size difference -- thousands and thousands of lines.
* Second, fetchmail enjoys good defaults. If you enter the minimal set of options in the config file, it generally works properly.
* Third, and this is the biggie, fetchmailconf is an excellent GUI config tool. It can autodetect most of the configuration, and if there are multiple supported protocols/auth methods, it uses the "best", which is really better than most commercial email clients can do. Note that one *still* has full access to the simple, readable output that it produces. It doesn't hide anything from you at all, so it doesn't hurt power users that know exactly what they want the software to do, but it makes things much easier for new users.
I think it is called being caught between a rock and a hard place. Now if only somone would start an opensource configuration project. But that is the problem isn't it. Those who could build that don't need it and those who need it can't build it.
That's really not a bad idea.
The problem is that data would need to be stored in a more structured manner -- right now, often only the program that reads a config file has a parser that can understand it.
If more UNIX software used XML (which would, unfortunately, probably decrease the hand-editability of the config files) and shipped with DTDs/schemas, there could be a Single Configuration UI that could work in GUI or text mode. It would be possible to easily compare configurations across systems. It might even be possible to automate some testing.
However, this would require getting a *lot* of people onboard. It would need someone like the FSF (or, come to think of it, ESR) backing it.
Right now, every piece of UNIX software, instead of using a standard "libconfig" to read and write an XML-based config file, has its own config file parser, which isn't that great of a system.
I'd like to point out that the person bitching is ESR, who is decidedly a geek.
However, ESR's fetchmail does a really good job of both having a config utility capable of autodetecting almost everything and of having the software guess good defaults if the config file doesn't have a piece of information it might want.
It really is true that very little open source software does this -- includes an intelligent autoconfiguration utility, though most software (most *cough* sendmail) has good defaults as well.
What about "hold down control" to have guesses made about proper settings...or even better, have a button on each dialog for "suggest settings".
Actually, neither is probably the right way to do it from a UI standpoint, but I'm not sure that requiring people to constantly try holding down control in different dialogs is a great solution either.
Why the hell would they make a liveCD? The install cd is a rescue disk as well
Because a real OS is a much nicer recovery environment than a highly limited shell. I've had uncomfortable moments where it's really difficult to fix something with the extremely stunted rescue environment -- but if I could just have booted into Windows, I'd be fine. I've done low-level Windows work of the sort that can easily render Windows unbootable and eventually gave up trying to use the lousy rescue environments, and just installed Windows on two partitions on each machine I was working on, so that I had a second copy of Windows around to do live boot recovery. It's technically possible to do this with Windows. There have been hacks from c't for 9X and a somewhat more broken hack for NT, Bart provides a hack to pull it off with certain versions of already installed Windows, and Microsoft even uses live-boot CDs for special (non-resuce purposes), but they don't provide a troubleshooting CD. My guess is that they're trying to maintain the illusion that the Windows NT/2k/XP password prompt provides some kind of security to the user, and hence want to make bypassing it marginally more difficult.
, and the OS is designed to boot from the MBR to avoid boot-partition messups.
What are you talking about? That doesn't make any sense at all. The MBR is synonymous with the boot blocks of a drive.
Auto update?? THE DON'T HAVE ANY RIGHTS TO MOST OF THE SOFTWARE ON THE SYSTEM. And the ones that they do own, and that are free you can update through windows update, so don't be dumb.
[sigh] Look. They can quite easily provide a service that lets people download updated software over the network automatically, as all current major Linux distros do. That doesn't mean that people can just download Corel's suite, it means they have a tool that scans all the software on your system and looks for patches to that software, and can download and patch that suite.
It means that your system maintenance is automated, as opposed to a slow task of looking for updates and manually checking what version you have. Heck, I'm not even sure where to *do* this for Windows -- at least the Mac has versiontracker.
I'm not sure about this, but does Linux have an equivelant to directX, where a game can bypass the videocard drivers to increase speed?
Yes. First of all, DirectX does not bypass drivers, not in the sense that you're talking about. Bypassing drivers would be a really awful idea, since you'd lose all your hardware acceleration. What it does is bypass the Windows GDI API, which is a slow imaging toolkit used to draw most things in the Windows interface. The equivalent to this on Linux is bypassing X11 (and preferably SHM as well), and using DGA. Since it's a pain to have to deal directly with the crap beneath the GDI, Microsoft put out a nice interface called DirectX. SDL (and associated libraries, like SDL_image and so on) is a front end that provides this kind of fast graphics functionality.
DirectX can be handy because it's a single library that does sound, 3d graphics, 2d graphics, etc. Normally on Linux, one uses individual libraries for each component one is working with. Most of the popular ones lie on top of SDL.
Free Software is free, true, but it doesn't just fall from the sky. It takes a lot of work from a lot of people, so hopefully businessmen such as yourself see the responsibility to replenish the pool, not just a resource to exploit.
It's not being exploited.
I'm *glad* to have more users of my code, even if they aren't contributing anything, as long as they aren't rudely complaining and demanding free support. It's no skin off my back if someone wants to and can make money with the service.
I do think that good bug reports (which are not as easy to produce as one might think) *are* a reasonable expectation. If you don't want to hassle with Microsoft over an Office bug, that's pretty reasonable. Heck, why do their work for them, especially since it takes *fighting* to get anything fixed? But if Mozilla crashes, it's pretty easy to sit down with Bugzilla and report the problem, and I'd consider it a social obligation.
Having a userbase familiar with your project is not to be scoffed at. Microsoft enjoys tremendous benefits from having a large userbase familiar with their software.
Heh. Well, I think Windows is less good than you're making it out to be (though current Windows machines are more suited to serving than they once were -- it's just that Linux is *really* *really* great for serving).
It's a nice GUI, yes.
I dunno. The GUI on Windows is definitely grating in many ways to Mac folks. I normally use Linux, and when I use a Windows box, I badly miss my pager, a nice clean background without icons on it, powerful keyboard control (launching programs, killing windows, zipping about workspaces), and the ability to use more mouse buttons usefully (resizing/moving windows).
It's also a decent standalone desktop system,
I don't agree. It could be, but Microsoft made three really big mistakes with Windows.
1) They don't provide a live-boot Windows CD (yes, there's the Bart hack, which while very cool, is definitely not a Microsoft-endorsed approach) for troubleshooting. This sort of thing is terribly useful for desktops.
2) They don't provide a comprehensive auto-update service for all the software on the system, or a mechanism for someone *else* to provide such a service. This is huge for home users.
3) They run several network services out-of-box. This is bad from a security standpoint (there have *always* been remote holes found in copies of Windows), and not a good idea for desktop machines.
and (at least comparably) wonderful for games.
Actually, I've been underwhelmed by Windows' gaming prowess. There are a lot of good games for Windows, but that's because the game authors wrote them for Windows, not because of really good gaming capabilities. I will grant that DirectX is very nice, and not to be overlooked. However, Microsoft:
* Does not provide native support for joysticks in their application software. This sort of stuff is important to encourage adoption of gaming peripherals. Mplayer can be controlled with a joystick, but Windows Media Player cannot.
* Has an API that requires applications to do work to deal with application-switching or being fullscreened/de-fullscreened. A *lot* of game vendors didn't do the extra work required, and hence all the nastiness with games that die when minimzied or don't minimize, or games that cannot be alt-tabbed out of. On Linux, folks generally use SDL, and have a single API call to toggle fullscreen. There are no problems on Linux with new windows popping up.
* Linux currently provides excellent scheduler and context-switching latency. This is important for soft-real-time tasks like games.
It integrates on networks poorly.
I'm not sure what you mean by that. I *do* think that it's awfully easy to set up a plug-and-play network of Windows machines. "I can see your computer in Network Neighborhood!" I'd say Windows is actually pretty easy to set up for basic networking stuff by an end user. Note that Apple's had a tradition of doing same.
Locking a machine down for the desktop(let's say comparable to a -default- RedHat desktop install) takes a LOT of work
I agree. However, I think some of this may not be due to the OS so much as the users -- Linux security issues are better widely understood and there is better security information out there for Linux.
I'm disturbed about how *close* to a police state the United States is getting -- I think that the line beyond which I am not happy has been crossed.
What is being done in Iraq is not pretty, and is on par with other hostile occupations (and is rather different from situations where a leader is simply ousted, which puts the lie to Bush's current claim "we just wanted to get rid of Hussein".)
However, despite me wanting it to move more towards individual freedoms,the United States is most definitely not a police state, and is in fact freer than most countries in the world.
Well, the one obvious notable factor is that all of you live in Ontario, which isn't the warmest of climes. It could just be that the components Apple chose for the iPod don't do all that well with constant cold weather.
Slashdot Mirror
Except for games that use audio or graphics, Linux has pretty much spot-on compatibility with any IF game, as do most operating systems. IF games are extremely portable, written to one of a number of portable VMs (and all this years before Java...and with better compatibility than Java).
.gam files. .zX formats, but I've only played .z5).
TADS (IMHO the most advanced engine, though Inform is very close) just plain runs on Linux. You want this to play
There is Frotz to run Inform (.z5 files...I believe a couple other
There is an ADRIFT implementation called SCARE for Linux. It has a less-than-perfect parser. To be honest, ADRIFT is a much simpler engine, and I generally fine TADS or Inform games to be much more fun and impressive.
Note that other classic adventure game VMs -- the ones for commercial graphical adventures -- like the Sierra (King's Quest, among others) and Lucasarts (Day of the Tentacle, Sam and Max, Secret of Monkey Island, among others) VMs have been ported to Linux in the form of Sarien, FreeSCI, and ScummVM. I don't believe there have been any new AGI/SCI/SCUMM adventures made -- the engines are static and no improved games will be made for them, but they're still neat projects to have fun playing the originals on.
While you're there, try Babel (for the TADS engine) and Bliss also for TADS.
Babel, a sci-fi horror game, was the first IF game I beat without needing any hints (though it's still significantly more difficult than many IF games out there), and Bliss...probably the best example I've ever seen of using the IF format. Describing it would probably spoil it, so I can only recommend playing it. When you've finished it, you'll understand.
Note that both of these games are dark and potentially disturbing. If you like fluffy or fantastic or slice-of-life or something else, look elsewhere...
The main problem is that an AI capable of entertaining and interacting with a human would have to be very complex. The complexity of an IF world is very small compared to what it would take to train an AI.
You know those stories about people that get stranded somewhere in a foreign country, and have to perform an amazing act to learn and understand what's going on? A human baby learns extremely quickly. Each human manages to do this, to go from zero knowledge of any language to speaking and comprehending knowledge with a few years of work. The problem is that it takes *four years* of constant interaction with new people and languages, plus visual input, to learn something like this. If you consider how much sheer *stuff* there is in four years of a human life, the task becomes staggering in scope. And we'd still be working with only a text-based interface...given the lower data bandwidth, and the fact that visual knowledge from the outside world is incredibly important to understand what is going on in a game for us, it would probably take far longer with just text to work with.
So maybe it could be done, but all the IF writers in the world have not written enough content (and some of it is surrealistic or misleading). All of it offers a much less powerful world than the one humans are in.
I've thought about this -- I like playing with both speech synth and IF -- but decided that you really have to think intensely about IF. I'm uncomfortable even listening to audio books, since I can't stop and reflect when I need to. IF would be much more difficult to play without being able to stop.
While I'm posting, I wanted to say that it's wonderful to see an IF story put on Slashdot. IF fell off a lot of people's radar with Infocom. The games have advanced a lot and, IMHO, gotten much better. Problems like "hunt the verb" have been largely eliminated, and many clever tricks and elements introduced.
Write a script in bash or perl -- it can do *anything*.
Then have xbindkeys, a simple program that runs commands when you tap specified key combinations, set up to trigger it.
Many window operations can be performed from the keyboard in powerful window managers like sawfish.
Making your environment dance at the touch of a key is what Linux does best.
No. All it means is that he *came up with* the idea of being stupid and then decided that it was a marketable concept...oh. Yes, I see.
You know what would be a great idea? X. Too bad it would cost a bunch of money to develop X into something usefull...if only there were some way to help try to ensure that I could recover the money I spend developing it...
There are a couple of problems with this logic. First, you're worried that someone will rip off your idea. Patents were designed in a day and age when the things being patented were *simple*. An industrial process, or a simple machine. The thing is, it's *really hard* to just duplicate the functionality of a worthwhile piece of software. If you can't just take the software (and copyright takes care of that), it's generally not going to be cheap or quick for you to reimplement the idea...and in that time, the original person has moved beyond where he was. Software needs patents much less than old processes once did.
Second, you're giving an example of an exceptional idea, something really amazing. The problem is that software development is so complex compared to earlier systems that you could find something to patent in almost every new system made. This is, frankly, not how the patent system is intended to operate at all.
Third, you talk about "expense" of developing the new idea. It really *was* expensive to develop some older things -- if you want to build a new machine and figure out how to make it work well, it could take many years and lots of expensive and painstaking ironwork -- and the simple result could be copied. However, software is (comparably ) incredibly cheap to work with. You think, write a hundred lines of code, and you have an implementation to test out and work with. You don't write up a blueprint and then have an implementation to test two months later.
Fourth, older devices were much more static. A plow is a plow is a plow. Maybe someone comes up with a way to hollow out part of the thing and make it lighter...then no improvements for a while. In the software field, there are constantly surging improvements. The whole goal of an engineer is to improve on existing systems...rather unlike the masses of plow companies, that might just produce different plows of roughly the same design. Patents are *much* more onerous in software.
I worked in a research lab for a while, and I think that I can safely claim that software patents are minimally useful to society. It's fairly rare that a really good, reasonable, legitimate software patent exists -- the type of research encouraged by software patents is of the "lock people out" variety, rather than the "make something better" variety. I do not think that research would be signifiantly impacted by a lack of software patents, and I *do* think that software engineering would be much easier.
The problem, I think, is that developers will happily use something if it makes their life easier, but not if it makes it more difficult.
If there's a good, standard parser out there, developers will use it.
GConf sounds vaguely like what I want, but does many things wrong. It implements a registry, instead of letting each program have its own file. It requires a daemon to be running (and I have had a *ton* of problems with said daemon crashing or malfunctioning in various ways in the past). It's really much more complicated than a simple "here's my config file, please parse".
I used an integrated Mac for years, and I was very happy to get out of that mess. I *hate* integrated computers. If you're making a laptop, great. Otherwise, they're a pain in the ass -- a pain to work on, a pain to upgrade, a pain to replace a broken part.
This is great for the advanced user but it floods the general user with too many choices.
I've never agreed with this philosophy.
Do what ESR did with fetchmailconf. Make all the options available. Just have the ability to autodetect everything you can. You don't need to hide options (this is a hugely irritating factor of Windows wizards), you just need to have intelligent defaults and good autodetection, so that things can work pretty much out of box.
What feature is it that you need to change from the default, and what is inappropriate about the Web config GUI?
When you're reading a mailing list and someone emails "how do I do X", the usual response, as ESR has pointed out, is to simply answer them. If the problem comes up a couple of times, perhaps it goes into a FAQ. There is one missing thing, though, that doesn't happen. The author should ask "what could I change in the *software*, not the documentation or FAQ or whatnot, that would keep users from coming to ask me about this again."
ESR's "bugger me with a chainsaw" appears to be a reference to the "fuck me gently with a chainsaw" phrase in Heathers. I got a kick out of that one...
Look. I love the UNIX printing system. It's a real pain in the ass to configure things, but it's also terribly powerful.
I've used LPR and then LPRng. I get network transparency, ability to batch-print easily (this is *not* trivial...try printing out hard copies of 200 source files pretty printed in Windows). I can manage print queues.
The thing is, UNIX printing systems are usually lots of little parts cobbled together with some scripts that vary from distro to distro, and then a config GUI that the distro maintainer puts out. The UNIX printing world is terribly disorganized compared to most other things in the UNIX world.
Can you identify the function of and tell the difference between all of these? LPR, LPRng, CUPS, gimp-print, foomatic, Omni, gnome-print, printman, printtool, desktop-printing, enscript, a2ps, ghostscript, pna2ppa, samba, hpoj, gsview, gv, ps2ps, ggv, redhat-config-printer, printconf, and mpage. All of these printing-related utilities and more have been on my system in the last few years. Keep in mind that I don't even use KDE, and that most distros vary the choice of what to use and you have an interesting set of knowledge to amass. The different print spoolers have different auth systems and config formats.
On the other hand, I have an old Apple LaserWriter without enough memory to print much of anything. I salvaged it when my old university threw it out. I hooked it up, and started cobbling together bits into a print filter. Sure, it took some doing and learning, but when I was done, all pages on the printer were rendered on the computer (where all the RAM in the world was available), converted to a bitmap and compressed, and sent in an embedded postscript file to the printer. On Windows I would have been simply SOL.
So, I'm not sure that an all-in-one system would be great. I *do* think that the printing situation could be cleaned up a lot, that the distros *really* need to get together and standardize on an interface (if you want to differentiate yourselves, please don't do it on something as basic as printing, which is a huge impediment to office use everywhere), and that it'd be nice to have some degree of autodetection of intelligent defaults (After a click on "add printer", "You have a Model Foobar attached. The proper driver is being selected.")
ESR also did fetchmail. Fetchmail has an *excellent* configuration interface.
* First, the config file is simple and small. A typical configuration should be simple and small. Take a look at the difference between the size of a basic sendmail and a basic postfix installation, and you'll notice an astonishing size difference -- thousands and thousands of lines.
* Second, fetchmail enjoys good defaults. If you enter the minimal set of options in the config file, it generally works properly.
* Third, and this is the biggie, fetchmailconf is an excellent GUI config tool. It can autodetect most of the configuration, and if there are multiple supported protocols/auth methods, it uses the "best", which is really better than most commercial email clients can do. Note that one *still* has full access to the simple, readable output that it produces. It doesn't hide anything from you at all, so it doesn't hurt power users that know exactly what they want the software to do, but it makes things much easier for new users.
I think it is called being caught between a rock and a hard place. Now if only somone would start an opensource configuration project. But that is the problem isn't it. Those who could build that don't need it and those who need it can't build it.
That's really not a bad idea.
The problem is that data would need to be stored in a more structured manner -- right now, often only the program that reads a config file has a parser that can understand it.
If more UNIX software used XML (which would, unfortunately, probably decrease the hand-editability of the config files) and shipped with DTDs/schemas, there could be a Single Configuration UI that could work in GUI or text mode. It would be possible to easily compare configurations across systems. It might even be possible to automate some testing.
However, this would require getting a *lot* of people onboard. It would need someone like the FSF (or, come to think of it, ESR) backing it.
Right now, every piece of UNIX software, instead of using a standard "libconfig" to read and write an XML-based config file, has its own config file parser, which isn't that great of a system.
I'd like to point out that the person bitching is ESR, who is decidedly a geek.
However, ESR's fetchmail does a really good job of both having a config utility capable of autodetecting almost everything and of having the software guess good defaults if the config file doesn't have a piece of information it might want.
It really is true that very little open source software does this -- includes an intelligent autoconfiguration utility, though most software (most *cough* sendmail) has good defaults as well.
What about "hold down control" to have guesses made about proper settings...or even better, have a button on each dialog for "suggest settings".
Actually, neither is probably the right way to do it from a UI standpoint, but I'm not sure that requiring people to constantly try holding down control in different dialogs is a great solution either.
Why the hell would they make a liveCD? The install cd is a rescue disk as well
Because a real OS is a much nicer recovery environment than a highly limited shell. I've had uncomfortable moments where it's really difficult to fix something with the extremely stunted rescue environment -- but if I could just have booted into Windows, I'd be fine. I've done low-level Windows work of the sort that can easily render Windows unbootable and eventually gave up trying to use the lousy rescue environments, and just installed Windows on two partitions on each machine I was working on, so that I had a second copy of Windows around to do live boot recovery. It's technically possible to do this with Windows. There have been hacks from c't for 9X and a somewhat more broken hack for NT, Bart provides a hack to pull it off with certain versions of already installed Windows, and Microsoft even uses live-boot CDs for special (non-resuce purposes), but they don't provide a troubleshooting CD. My guess is that they're trying to maintain the illusion that the Windows NT/2k/XP password prompt provides some kind of security to the user, and hence want to make bypassing it marginally more difficult.
, and the OS is designed to boot from the MBR to avoid boot-partition messups.
What are you talking about? That doesn't make any sense at all. The MBR is synonymous with the boot blocks of a drive.
Auto update?? THE DON'T HAVE ANY RIGHTS TO MOST OF THE SOFTWARE ON THE SYSTEM. And the ones that they do own, and that are free you can update through windows update, so don't be dumb.
[sigh] Look. They can quite easily provide a service that lets people download updated software over the network automatically, as all current major Linux distros do. That doesn't mean that people can just download Corel's suite, it means they have a tool that scans all the software on your system and looks for patches to that software, and can download and patch that suite.
It means that your system maintenance is automated, as opposed to a slow task of looking for updates and manually checking what version you have. Heck, I'm not even sure where to *do* this for Windows -- at least the Mac has versiontracker.
I'm not sure about this, but does Linux have an equivelant to directX, where a game can bypass the videocard drivers to increase speed?
Yes. First of all, DirectX does not bypass drivers, not in the sense that you're talking about. Bypassing drivers would be a really awful idea, since you'd lose all your hardware acceleration. What it does is bypass the Windows GDI API, which is a slow imaging toolkit used to draw most things in the Windows interface. The equivalent to this on Linux is bypassing X11 (and preferably SHM as well), and using DGA. Since it's a pain to have to deal directly with the crap beneath the GDI, Microsoft put out a nice interface called DirectX. SDL (and associated libraries, like SDL_image and so on) is a front end that provides this kind of fast graphics functionality.
DirectX can be handy because it's a single library that does sound, 3d graphics, 2d graphics, etc. Normally on Linux, one uses individual libraries for each component one is working with. Most of the popular ones lie on top of SDL.
Free Software is free, true, but it doesn't just fall from the sky. It takes a lot of work from a lot of people, so hopefully businessmen such as yourself see the responsibility to replenish the pool, not just a resource to exploit.
It's not being exploited.
I'm *glad* to have more users of my code, even if they aren't contributing anything, as long as they aren't rudely complaining and demanding free support. It's no skin off my back if someone wants to and can make money with the service.
I do think that good bug reports (which are not as easy to produce as one might think) *are* a reasonable expectation. If you don't want to hassle with Microsoft over an Office bug, that's pretty reasonable. Heck, why do their work for them, especially since it takes *fighting* to get anything fixed? But if Mozilla crashes, it's pretty easy to sit down with Bugzilla and report the problem, and I'd consider it a social obligation.
Having a userbase familiar with your project is not to be scoffed at. Microsoft enjoys tremendous benefits from having a large userbase familiar with their software.
Heh. Well, I think Windows is less good than you're making it out to be (though current Windows machines are more suited to serving than they once were -- it's just that Linux is *really* *really* great for serving).
It's a nice GUI, yes.
I dunno. The GUI on Windows is definitely grating in many ways to Mac folks. I normally use Linux, and when I use a Windows box, I badly miss my pager, a nice clean background without icons on it, powerful keyboard control (launching programs, killing windows, zipping about workspaces), and the ability to use more mouse buttons usefully (resizing/moving windows).
It's also a decent standalone desktop system,
I don't agree. It could be, but Microsoft made three really big mistakes with Windows.
1) They don't provide a live-boot Windows CD (yes, there's the Bart hack, which while very cool, is definitely not a Microsoft-endorsed approach) for troubleshooting. This sort of thing is terribly useful for desktops.
2) They don't provide a comprehensive auto-update service for all the software on the system, or a mechanism for someone *else* to provide such a service. This is huge for home users.
3) They run several network services out-of-box. This is bad from a security standpoint (there have *always* been remote holes found in copies of Windows), and not a good idea for desktop machines.
and (at least comparably) wonderful for games.
Actually, I've been underwhelmed by Windows' gaming prowess. There are a lot of good games for Windows, but that's because the game authors wrote them for Windows, not because of really good gaming capabilities. I will grant that DirectX is very nice, and not to be overlooked. However, Microsoft:
* Does not provide native support for joysticks in their application software. This sort of stuff is important to encourage adoption of gaming peripherals. Mplayer can be controlled with a joystick, but Windows Media Player cannot.
* Has an API that requires applications to do work to deal with application-switching or being fullscreened/de-fullscreened. A *lot* of game vendors didn't do the extra work required, and hence all the nastiness with games that die when minimzied or don't minimize, or games that cannot be alt-tabbed out of. On Linux, folks generally use SDL, and have a single API call to toggle fullscreen. There are no problems on Linux with new windows popping up.
* Linux currently provides excellent scheduler and context-switching latency. This is important for soft-real-time tasks like games.
It integrates on networks poorly.
I'm not sure what you mean by that. I *do* think that it's awfully easy to set up a plug-and-play network of Windows machines. "I can see your computer in Network Neighborhood!" I'd say Windows is actually pretty easy to set up for basic networking stuff by an end user. Note that Apple's had a tradition of doing same.
Locking a machine down for the desktop(let's say comparable to a -default- RedHat desktop install) takes a LOT of work
I agree. However, I think some of this may not be due to the OS so much as the users -- Linux security issues are better widely understood and there is better security information out there for Linux.
Fortunately, the phrase "sweaty monkey man" contains neither "l"s nor "r"s.
Wow, I didn't realize that there *was* a Japanese Slashdot. I always wondered why the Japanese seemed underrepresented on Slashdot.
How many different Slashdots are there?
I can't agree.
I'm disturbed about how *close* to a police state the United States is getting -- I think that the line beyond which I am not happy has been crossed.
What is being done in Iraq is not pretty, and is on par with other hostile occupations (and is rather different from situations where a leader is simply ousted, which puts the lie to Bush's current claim "we just wanted to get rid of Hussein".)
However, despite me wanting it to move more towards individual freedoms,the United States is most definitely not a police state, and is in fact freer than most countries in the world.
Well, the one obvious notable factor is that all of you live in Ontario, which isn't the warmest of climes. It could just be that the components Apple chose for the iPod don't do all that well with constant cold weather.