why didn't he price Apple Computers more competitively with the Wintel stuff?
Because it wasn't in Apple's interests. Mac users are willing to pay Apple prices, so Apple has enormous profit margins.
Think about what cutting prices would actually have done. It would have placed Apple in direct competition with Dell, with Compaq, with Micron, with HP, with a host of large manufacturing companies that are very, very good at shaving down manufacturing costs and operating with tiny profits. There's so little profit in the desktop market today that companies have been exiting market for some time -- focusing on the higher-profit laptops and servers.
That would have been a difficult-to-compete arena for Apple. Apple made a decision that has kept them a successful business -- it was probably the right one from a business standpoint.
Of course, I agree with you WRT to use of Apple products. I gave up on Apple when they revoked clone makers' licenses. People that choose Apple are choosing to work within a niche market, pay significantly higher prices, and have less software and hardware choice. That makes sense for many people (you get a black-box solution that works out of box, which anyone, even the tech illiterate, can comfortably use). It was not a product that I was particularly interested in, but that doesn't make it an invalid business -- Apple's done pretty well for themselves.
And even if it's unlikely that any of this ever influences Microsoft (directly or indirectly), I would not pour scorn on somebody's attempts to do so.
If you've read some of my other posts, you'll see why I don't consider the BIOS situation to be an issue. I don't see any reason to hold back from a debate that might injure someone's efforts to prevent something that really isn't a problem.
Please excuse my "complaining on Slashdot," but I find having freedom extremely important. I do really hope I am not alone in that belief.
The only reason that criticism is there is because you just complained that other people are doing the same.
Of course. TCPA will *never* refuse to obey our commands, should we choose to become second class citizens having absolutely no access to mainstream media whatsoever. One only has to wonder if that is a bad thing indeed...
It all depends upon your definition of "second class citizens". If the concern is that you won't be able to run Trusted Windows software -- no, you certainly won't. I don't use *any* Windows software, nor do Mac OS users. I've also chosen not to own an HDTV, which excludes me from viewing high-definition content. I view the drawbacks as too severe. Everyone will, of course have to make such a choice.
The argument that media in the form of music or movies will become unavailable is not true -- it's quite easy to obtain and watch DivXes of protected DVDs, or MP3s of "copy-protected" CDs, despite the fact that those two are deliberately not manufacturer-supported on my Linux-running computer. De-protecting music and video is too easy -- it will always be done. The issue with respect to software is true -- you will have to make a decision as to whether you want to be able to play "Slag Smasher 2000" or not.
But jokes aside. Do you think University which publishes online papers as Windows Word today is not going to use "Trusted" Word tomorrow?
Perhaps some fields use Word for papers. I majored in computer science and philosophy, and always saw PS or PDF papers online, not Word documents. They made the choice to avoid Word partly for the same reasons you're citing here.
Do you think a bank which has IE-only website will change its mind when there's finally a way to make sure people indeed connect only with "Trusted" IE?
No. But then again, I don't use IE, and have never had any trouble with the two online banking sites I use. I'm unaware of problems with IE-only banking sites *today*. Furthermore, banks have significant benefit in not limiting their audience.
What about government websites? Are they going to allow "untrusted hacker software" to connect to their servers?
At this point, I find your argument to be entering "slippery slope" area. The US government has been exceedingly good at supporting all users using their websites in the past (including ones using very old browsers or slow computers), and I see no reason that would change.
Let me guess what OS are you using right now... Oh, right, I thought so. So you're lucky. But are you going to give a Blinux using person few thousands dollars for software she will need soon? I hope so.
I'm using regular old Linux. I'm not sure why a Blinux user would purchase a Windows copy of Microsoft Access, or how doing so relates to TCPA at all.
Right! Are you really so naive to believe we'll be able to run our own kernals as part of our "trusted" OSen? Please...
Of course not -- that would spoil the nature of a trusted kernel. I'm again curious as to what your beef is. You apparently would find a custom-compiled trusted Linux kernel acceptable. This means that you find not using Windows software also acceptable. I've pointed out that media will hardly be protected by any attempts to copy protect it. Going through the categories, the only type of content that you could possibly still be concerned about (where you would lose support) is software that is currently supported under Linux, but would *not* be supported after a TCPA implementation, and would require a trusted system. I cannot imagine what software you are thinking of.
But all is well, and any concerns are just overblown stupidity in/.
Yes, I'd say that they're exaggerated.
MS is found guilty of abusing its monopoly power.
Netscape is gone thanks to anticompetitve practices.
Yup. Microsoft is big and nasty. Not *as* big and nasty as IBM was twenty years ago, of course -- things have improved quite a bit since then. Not nearly as nasty as Apple's historically been within its own little domain. But, yes, nasty.
We are approaching the IT monoculture were one company alone has the power to decide where the whole industry is going.
No, we've left the PC monoculture where at one point, one company *did* have that power, gone through a series of dips and turns. What people are complaining about in the current story -- DRM in x86 machines -- is a product of a vast number of meetings and agreements, where an industry that *dwarfs* Microsoft decided that DRM would be worthwhile to try. Same thing goes for the world of power-saving. Microsoft doesn't even begin to have the power to implement something like this themselves. They are certainly an influential member of the tech world, but all they can do is coax. And no, they can't "buy out" the industry. MS is big, but they're a drop in the tech industry bucket.
But no, that must be idiots like me, that when faced with a XP machine for the first time could not copy tracks of my CD unless it was done in MS's propietary format.
Yes, if you're using *Windows*, you are subject to what Microsoft decides. That's the sole area where Microsoft *does* have absolute control. Same goes for Mac OS and Apple. If you dislike a vendor's software enough, you'll end up switching -- to Mac OS or Linux. I own a single hybrid CD, and I've happily ripped that to my computer, which is running Linux. There doesn't seem to be any sort of move to prevent Linux for doing whatever it wants to. Microsoft has a fair amount of cash and a fair amount of market share. It can spend it to implement whatever it wants. Keep in mind that unpopular decisions will cost even Microsoft money and market share.
My point was that *Slashdot* doesn't significantly influence Microsoft, not that Microsoft can do whatever the hell it wants to.
hey, guess what, the only commercial software I use on a daily basis is the Windows kernel itself!
I also use cygwin when I use a Windows machine. Frankly, cygwin is nice, but it's effectively a slow, poorly-supported (I mean, I don't mind a bit of a challenge, but the last time I tried building mplayer under Windows, it was a bitch and a half) subset of the functionality I have under Linux. Everyone I know that uses cygwin on their machines has since migrated their primary machine to Linux or BSD (or started using cygwin after doing so). Any cygwin software is also subject to not only the constraints you mention below (your problem with the UNIX filesystem layout), but the frusterations of layering a system on top of Windows...damn/cygdrive/c prefixes.
There's plenty of open-source stuff to use under Windows.
I find cygwin to hardly be a replacement for using an *IX box. There are a few native open-source Windows projects. PuTTY is an excellent example. There are also a few good free but not OSS tools, like the sysinternals utilities. However, they aren't strace and ltrace, that's for sure.
And it's easier to _find_ a reliable Windows driver for $HARDWARE than to find a Linux one of any sort
It's certainly possible to find hardware for which there is no Linux support. Probably the worst offenders are among cheap, non-HID-compliant USB devices, and some highly specialized PCI boards. However, I have never owned a piece of hardware that lacked a Linux driver -- though I will admit, I have had times that not all *features* were supported. Up until ALSA came along, hardware audio mixing under Linux on all but a few cards required the purchase of a $25 driver suite. There's driver support for most printers out there -- but no ink level utilities. Not actually drivers, but I suppose that related utilities could be considered relevant. It could be the fact that I've always avoided software modems, which were a sticking point for a while (this dates back to the days when I used Mac OS and Windows, where the CPU hit was a significant issue), and webcams (never wanted to broadcast myself).
As a matter of fact, the only time I've run into a n unsupported hardware issue relating to Linux when Windows 2000 didn't support a sound card that had been used under Windows NT. The card vendor and the chipset vendor had both gone out of business, but there was a Linux driver still being maintained because it was GPLed, and my Linux box at home could use the thing.
Oh, do you mean how in Linux you get to choose whether to put something in/bin,/usr/bin,/usr/local/bin,/usr/X11R6/bin,/opt/$whatever/bin, ~/bin, etc., instead of being forced to put it in Program Files? I don't understand why my single-user desktop machine should need to use a file system layout that might concievably have made sense on a mainframe thirty years ago.
No, I was talking about symlinks. I mean that I can drop a data file in one place and then aim programs that look in a particular directory for their data file at it. Shortcuts are an ugly hack.
Finally, I don't consider your claims that you're worried about where to put binaries valid. All these software packages are easily installed via package manager, which handles everything for you. The existence of ~/bin is a solution that Windows does not have -- it lacks a standard location for user-local software installation (and frequently, because of assumptions about registery permissions, software cannot be installed user-locally under Windows).
As a matter of fact, most *Windows* installers prompt you to specify where the software is to be installed (though there's a sensible default in Program Files in almost all cases). On Linux, there is no asking if you simply are installing a package.
I eat up all my ram all the time. I do renderings that max out the RAM. I don't lose control of my computer over it. Now, I will say that over a period of 2 weeks, I end up needing to reboot in one way or another. Is that what you meant maybe?
No. Try the following:
while (1) {
char * foo = malloc(1024);
if (foo)
memset(foo, 'c', 1024); }
That will eat up all the memory on the system, including the pagefile, but not constantly hammer it, so it won't be forcing paging. Then use Explorer for a bit.
Think it'd make a difference if I disabled paging the kernel? That's the first thing I do when I get a new Windows system. I wonder if that's why my machines always tend to stay up...? Can you provide any insight into this?
Because you're running a nonstandard setup, the details of which do not affect even a tenth of a percent of the Windows market. It certainly won't hurt stability to disable paging...
I don't know if I just have damn good luck with hardware or what.
You do. I've yet to bluescreen a 2k+ stable box on harder that wasn't in some way suspect without any third-party drivers on it. I've seen all kinds of blue screens over the years (mostly on friends' systems -- I left for Linux well before 2k was released) that go away when changing driver version or installing a new driver.
I've yet to have any stability problems with Linux drivers -- about four years ago, I managed to hard-freeze a Linux system when hammering it with multicast code, but aside from that, I can't remember ever having seen hard-freezes or kernel oopses when playing with devel kernels. I use ALSA (which *will* be part of the Linux kernel, admittedly), mga_vid (mplayer's included Matrox-specific accelerated driver), and have at various times used lufs (Linux Userspace File System), and the virtual archive file system that Rox Filer used for a while, among others.
Now, I *have* seen what I've considered aberrant behavior during that time -- way back with RH 5.2 and the 2.0 kernel, which would be ge, I remember the latency being so high on the Sound Blaster's MIDI port driver that the serial port driver would occasionally not empty the UART's FIFO fast enough -- so when I used my joystick, my downloads would slow to a crawl due to dropped characters and subsequent retries. I'm sure I've run into other oddities in that time, though I don't remember any off the top of my head.
Windows drivers are decidedly more flaky than Linux ones.
True, and the act of downloading such software infringes their copyrights. However, there would be very little benefit to Valve in attempting to pursue the thousands of people that obtained the code through Slashdot alone, much less through secondary and alternate distribution. They're trying to keep websites from posting the code, but obviously it will never be eliminated.
Seriously, I considered downloading it since I like to toy with 3d programming, and I'd love to see how the "pros" do it, but then I realized that if I were to be caught, anything I might ever want to release myself could be considered infringement in some skewed sense.
No, this is not correct. This is a popular misconception that dates back to the reverse engineering of IBM's PC BIOS. To be *absolutely certain* that there would be no legal challenges (not because it was legally necessary, but because it made it obviously impossible for any engineers to have copied code), Phoenix did not allow any of their engineers that had seen IBM source to be involved in the reverse engineering of the BIOS. While this so-called "clean-room" effort has been cited many times over, it does not have ground in legal *requirements*. Perfectly legal non-clean-room reverse engineering occurs daily in industry -- and you're not even talking about reverse engineering, but simply implementation of a project in a similar category.
I've looked at the quake source code, but it didn't help me much, and by now it's horribly outdated, so seeing some real, working and current, source code to something like this would be beyond incredible for me to pick apart and analyze.
You do not have to worry about legal taint (IANAL). However, if you want a modern 3d engine to look at and are still concerned, you may want to check out the LGPLed Crystal Space.
Because such a port would not be legal in the United States and most other nations. As such, few people would work to maintain it, and few would be willing to distribute it.
Secondly, the game content -- audio, etc, would still be commercial. While I might be willing to spend time porting code of some game to Linux, if I made the extensive kind of time investment involved, I would expect content-creating members of the project to do the same -- to donate their work, so that it may be distributed freely. Obviously, that will not happen in this case.
Morale is: Pirates will manage to pirate it. Do not piss off your customers in the futile process.
Rely on the ever-more-valid assumption that customer will have machine on always-connected broadband. Implement crucial chunks of the application on a remote server, so that you're essentially an ASP. Try pirating that...
I'm pretty sure he did the obvious thing, though -- took the traditional phrase "800 pound gorilla", comitted the not not-entirely-uncommon error of subsituting "500" for "800" and decided to substitute "elephant" for "gorilla" -- because, after all, an elephant is bigger than a gorilla, and *surely* his point would be better made with a bigger creatre.
The thing is, AFAIK, OF provides minimal benefits to anyone but Apple and Sun -- hardware vendors support x86 as a matter of course.
FWIW, I've seen a bunch of Apple folks on Slashdot already latching on to something they can promote Apple on, claiming that OF isn't a Trusted BIOS (gee, like every other freaking x86 BIOS aside from Phoenix's efforts?)
If Microsoft did somehow find a cure for cancer, it would not be distributed to those who most needed it, but to those who could afford it.
If $LARGE_CORPORATE_ENTITY did somehow find a cure for cancer, it would not be distributed to those who most needed it, but to those who could afford it. This is not a Microsoft-specific issue.
The only reason Microsoft hasn't done stuff like this is because somebody has taken the time to raise public awareness about what freedoms we were about to lose.
No, Microsoft has regularly ignored PR issues WRT to making decisions (at least once it has committed to a decision).
Believe me, the execs at Microsoft (note: not engineers there) care very little about what Slashdot has to say. Slashdot is full of after-the-fact complaining and rumors, and is constantly full of "sky-is-falling" stories. The few that I actually have inside knowledge on have let me realize that Slashdot is a wildly alarmist publication. Generally, people submitting stories have a political agenda, and want to get people upset about something -- and providing misconceptions is an effective way to do something.
Slashdot is fun to read, but it's not a Microsoft leash.
Things where there are *wildly* overblown claims and theories include: United States governmental repression, TCPA, Palladium, most stories on corporate business relationships with either SCO or Microsoft, Gentoo, and Apple products. Articles like these should be treated about the same way you'd treat something from the National Enquirer or the Young Republicans -- with a very, very large block of salt.
I think it is time to renew my EFF membership. I believe everyone should do that instead of just complaining on Slashdot.
And yet, here you are complaining on Slashdot and trying to convince *other* people to donate to the EFF instead of doing so yourself.
TCPA mandates that the system be user-disableable. It's part of the spec. TCPA will *never* refuse to load Linux, though it will *always* refuse to load as a trusted OS (since the kernel isn't signed...though I believe there is a project to get a TCPA-enabled copy of Linux). You're on quite a slippery slope if you're arguing that the current state of affairs would lead to the wildly unlikely state of affairs where all the x86 non-Windows server operators are convinced to use Windows instead. Nope, not going to happen, and vendors are uninterested in being locked in to Microsoft.
There are things that people may not like about TCPA, but they're generally not the ones that the EFF is insinuating (which tend to be improbable rabble-rousers -- they sound nasty, but they aren't the sort of thing that's actually going to be an issue in the foreseeable future.)
* It's incredibly easy to script and build new applications by tying together existing ones via pipes. The results are fast, reliable, and professional -- unlike AppleScript or VB-produced results. This is only relevant to tech users, but it's a big one.
* It's free. Okay, for a professional with a decent salary, the cost of Windows vs Linux itself -- the base package -- really isn't significant. A hundred or two hundred bucks is not a big deal. However, to purchase commercial equivalents of all the Linux apps I use would be extremely expensive. Compilers (think Visual Studio), editors (think Visual SlickEdit), mail clients (think Eudora), system monitors (think all manner of shareware apps), sound editors (think Cakewalk), image editors (think Photoshop), web servers (think IIS), code checkers (think Gimpel Lint), graphing programs (think Visio), math/statistics packages (think MATLAB), and all the rest, there is a *lot* of money involved. Sure, you can pirate it, but that's not an option at work, and pirating software is less and less trivial with the surging prevalance of phone-home features.
* It's secure. Traditionally UNIX (and its apps) have had tighter security design than Windows, especially WRT local security. A couple of Microsoft apps are phenomonally insecure (MSIE, Outlook), and most Windows apps don't have the same emphasis on avoiding attacks.
* It gives better performance. My workstation runs a large set of servers in the background. I don't notice. I have a friend that runs a Windows FTP server that he kills off when he wants to take all the CPU time on his system.
* I can fix bugs that piss me off. If I have an issue, I happen to be a coder, so I can run out and fix it without just complaining to a company's forums and hoping that something happens. I can add features that I want. Obviously, this benefit isn't nearly as good if you aren't a coder, but it's something to consider.
* I can actually see what's going on. Linux has a strong tradition of talking about and letting you see what's *actually* happening on your system. The startup system is just a bunch of scripts that are quite readable. In contrast, if you pick up a book designed for a Microsoft administrator, you'll get a bunch of Microsoft-invented terms ("Enable a service"...am I starting a process listening on a port or what? What the hell is happening?) This also makes troubleshooting much better.
* A richer toolkit. For at least coders, network admins, and security types, good tools exist that have no Windows equivalent. (The reverse tends to be true when it comes to office workers.)
* Choice. If I use Windows, I also must use Explorer, like it or not (and I don't). I can't use the kernel or Windows software without also using the expected file manager (yes, there have been a few hacks to try "replacing" Explorer, such as LiteStep, but they're flaky...more neat toys than pratical tools). On Linux, I have more window managers available than I have fingers. I have a whole collection of file managers. I have docks galore. I can choose my favorite from each category and use that.
* Better design. The fact that Linux uses better file-locking semantics, the fact that Linux uses symlinks instead of shortcuts, the fact that it's easier to write a reliable Linux driver than a reliable Windows driver, all have strong trickle-down effects to the user in the form of fewer reboots, more flexibility in file system layout and control, and a more reliable system.
You decide to upgrade your old Windows 98 to Windows 2000 because it's so much better than Windows 98 was. If you are on a telephone line, dialing up to your AOL account, how are you supposed to get your system patched (in the THREE HOURS it takes to download and install ONLY THE CRITICAL UPDATES) before some cheeze-head hits your system with one of the umpteen remote exploits that you are vulnerable to?
Much as I prefer Linux to Windows, you aren't being entirely fair. The main problem is that Windows runs a horribly complicated and wildly insecure protocol -- NetBIOS -- that was never designed for non-LAN use (hence Microsoft's lame suggestions to "firewall" the service and blaming sysadmins for not doing so instead of fixing the damn thing.)
An old copy of Linux will run into the same problem. Red Hat Linux 5.x-era stuff used to run plenty of daemons out of box, since the technical crowd knew what to disable. If you didn't axe said daemons before hooking up to the 'Net to get updates, you could be equally screwed.
So, with Win2k, you need to do the same thing. Disable the Server service, along with any other services listening for connections. Then plug your computer into the network and start updating it. Aside from TCP/IP-level exploits, you should be pretty safe.
Modern Linux distros run few daemons out of box (though I still wish they were tighter). Win 2003 Server does the same thing. Microsoft learns, abeit slowly and after a sufficient number of users have been badly burned.
Granted, there are still some incredibly bad technical decisions that Microsoft made for political and business reasons. Making Internet Explorer (big, incredibly complex, fragile, full of security issues) the client to use for automatic updates is pretty horrific.
Yes, Windows 2000 is better than Windows 95. Big deal. My wristwatch is better than Windows 95.
If you're talking about reliability, that's actually a pretty high standard. Your wristwatch doesn't have any nonvolatile state to muck up, so worst case you can remove the battery. It's fairly simple -- I don't believe I've ever seen a buggy wristwatch (though I have seen other buggy embedded devices, like network printers.) Your wristwatch is a pretty reliable realtime devices that can take a physical pounding, is probably waterproof, and can run for a year and a half without a battery change.
A stock, vanilla Windows install seems to be fairly reliable for basic tasks. (Not necessarily for remaining usable or stable under heavy load or resource usage -- try eating up all free memory and watch Explorer die the first time it hits an unchecked malloc()) However, if you get new hardware, you're probably going to be using third-party drivers. These aren't necessarily as solid. While some of this is just the fact that there are more companies making lousy hardware for Windows employing incompetent device driver authors, there is also the issue that Windows device drivers are far, far harder to write correctly than Linux device drivers. Windows has pageable kernel memory, requires all sorts of special work WRT power management, etc. Windows is a much harder environment to write a reliable device driver in, and driver quality reflects this.
OpenFirmware is absolutely INCREDIBLE, and if more companies were on-board it would get even better. On a Macintosh (O.F.) you just hold 'option' at boot and you get a menu of all bootable drives connected to the machine, be they FireWire, IDE, SCSI, or USB (actually USB is disabled out of sanity). You can get a device list even better than most Operating Systems can provide from OF.
You know, I hate to say it, but I just plain haven't booted off a non-ATA volume in forever. Haven't had any reason to, and haven't even used USB-boot features of an x86 BIOS.
From what I can tell, the benefit of OF is essentially to help Apple and Sun. Both have fairly marginalized hardware platforms (SPARC and PPC), and worry about lack of architecture support from hardware vendors. So they come up with a solution that provides architecture-independent support, hoping that they won't be ignored by hardware vendors. The x86 world doesn't have this problem, and Intel has not, to the best of my knowledge, been particularly interested in making Open Firmware universal, as it provides little benefit to them (as a matter of fact, I believe that they have their own, different solution to handle IA64/IA32 support).
And yet, despite all this, I don't think that spam is going to go away because of legal efforts. It will take technical ones.
I think that email is a doomed beast, though I'm sure much effort will be spent trying to save it. The ability for anyone to freely send you email into your inbox is a relic of the 70s and 80s, much as anonymous ftp sites with incoming directories for handing files around were. The email system obviously is great if it can operate successfully -- but I think that it's ill-adapted to today's social environment. Reminds me a bit of communism -- sure, it'd be great if folks could get it working, but you have to assume that people will willingly cooperate and not abuse the system. These assumptions have been found to be false, and modifications made to all other major in-use protocols to prevent abuse. (I've been involved in adding anti-abuse modifications to P2P clients, and have run into plenty of this myself.)
Email in its current form (where almost anyone can send almost anyone email freely) will eventually (and probably painfully) die, and alternate systems, where most communications are whitelisted, will come to the fore.
You know, I really hate to say it, but I'm sorry to see the source go down (Valve says that they're asking websites to take it down -- not sure if that translates to asking or sending legal threats). I'm skimming through a copy of it that I downloaded with some interest. It's not often that you get such insight into game development (post-mortems are interesting as well, but source hasn't been neatly edited). While I doubt a commercial gaming company would ever swipe code from Valve (too much potential damage -- if there's even a 10% risk of exposure, they're better off just licensing it.) Cheating will obviously be a problem...if I were Valve, I think I'd consider significant protocol revisions.
Some people have said that Valve has included GPLed code in the Half Life 2 code. Dunno as to whether this is true, but I'd like to point out that while this is technically not kosher, I suspect that a lot of places do it -- as long as it's out by release time, I very much doubt that anyone will complain. (On the other hand, if it *isn't* out...)
This is a good example of why internal security is very important. I don't use any computers at work that don't talk to each other through encrypted connections. I maintain a single trust relationship (pubkey based, not IP based), from a machine that has a superset of the information on a second machine, so there's little point in exploiting trust relationships (plus, if superset machine A were compromised, a keygrabber could easily allow compromise of machine B anyway). I don't use Windows filesharing. These are all very easy to overlook, especially during crunch time, but as Valve has discovered, while the chances of things going sour may be low, the potential damages are enormous. I would urge folks who are working with *any* kind of important IP to do the same -- do *not* rely on Windows filesharing, do *not* use trust relationships, and do not use unencrypted connections, even on your local network. SFTP exists and there are free clients all over -- you do not have a good excuse for using FTP.
This is also another example of why it may be worthwhile to have a network admin that does regular security audits. It takes additional time, and the vast majority of time that cost is overhead, but Valve is certainly regretting not doing so at the moment. (We have irregular security audits, which is better than nothing, but obviously not ideal.)
Finally, I'd like to say "chin up" to the folks at Valve. This sort of thing can be very frusterating, and I'm sure it hasn't helped morale at Valve much, but it's not a game-killer, even if it necessitates changes in the protocol or game engine, and a release delay. Good luck -- I probably won't buy your game, since it's unlikely that there'll be a Linux client, but I expect you'll have healthy sales.
As for other folks -- remember crack.com, remember Valve -- secure your damn networks already.
Thank you. This is indeed a good, spot-on resource that I hadn't seen before. I was quite satisfied by the fact that his points seemed to be similar to my own.:-)
People need to stop trashing Earthstation 5. It's a fantastic program, and does exactly as advertised. Plus, it seems to have built-in compression software -- my free disk space has been steadily increasing ever since I installed it!
Slashdot Mirror
why didn't he price Apple Computers more competitively with the Wintel stuff?
Because it wasn't in Apple's interests. Mac users are willing to pay Apple prices, so Apple has enormous profit margins.
Think about what cutting prices would actually have done. It would have placed Apple in direct competition with Dell, with Compaq, with Micron, with HP, with a host of large manufacturing companies that are very, very good at shaving down manufacturing costs and operating with tiny profits. There's so little profit in the desktop market today that companies have been exiting market for some time -- focusing on the higher-profit laptops and servers.
That would have been a difficult-to-compete arena for Apple. Apple made a decision that has kept them a successful business -- it was probably the right one from a business standpoint.
Of course, I agree with you WRT to use of Apple products. I gave up on Apple when they revoked clone makers' licenses. People that choose Apple are choosing to work within a niche market, pay significantly higher prices, and have less software and hardware choice. That makes sense for many people (you get a black-box solution that works out of box, which anyone, even the tech illiterate, can comfortably use). It was not a product that I was particularly interested in, but that doesn't make it an invalid business -- Apple's done pretty well for themselves.
And even if it's unlikely that any of this ever influences Microsoft (directly or indirectly), I would not pour scorn on somebody's attempts to do so.
If you've read some of my other posts, you'll see why I don't consider the BIOS situation to be an issue. I don't see any reason to hold back from a debate that might injure someone's efforts to prevent something that really isn't a problem.
Please excuse my "complaining on Slashdot," but I find having freedom extremely important. I do really hope I am not alone in that belief.
The only reason that criticism is there is because you just complained that other people are doing the same.
Of course. TCPA will *never* refuse to obey our commands, should we choose to become second class citizens having absolutely no access to mainstream media whatsoever. One only has to wonder if that is a bad thing indeed...
It all depends upon your definition of "second class citizens". If the concern is that you won't be able to run Trusted Windows software -- no, you certainly won't. I don't use *any* Windows software, nor do Mac OS users. I've also chosen not to own an HDTV, which excludes me from viewing high-definition content. I view the drawbacks as too severe. Everyone will, of course have to make such a choice.
The argument that media in the form of music or movies will become unavailable is not true -- it's quite easy to obtain and watch DivXes of protected DVDs, or MP3s of "copy-protected" CDs, despite the fact that those two are deliberately not manufacturer-supported on my Linux-running computer. De-protecting music and video is too easy -- it will always be done. The issue with respect to software is true -- you will have to make a decision as to whether you want to be able to play "Slag Smasher 2000" or not.
But jokes aside. Do you think University which publishes online papers as Windows Word today is not going to use "Trusted" Word tomorrow?
Perhaps some fields use Word for papers. I majored in computer science and philosophy, and always saw PS or PDF papers online, not Word documents. They made the choice to avoid Word partly for the same reasons you're citing here.
Do you think a bank which has IE-only website will change its mind when there's finally a way to make sure people indeed connect only with "Trusted" IE?
No. But then again, I don't use IE, and have never had any trouble with the two online banking sites I use. I'm unaware of problems with IE-only banking sites *today*. Furthermore, banks have significant benefit in not limiting their audience.
What about government websites? Are they going to allow "untrusted hacker software" to connect to their servers?
At this point, I find your argument to be entering "slippery slope" area. The US government has been exceedingly good at supporting all users using their websites in the past (including ones using very old browsers or slow computers), and I see no reason that would change.
Let me guess what OS are you using right now... Oh, right, I thought so. So you're lucky. But are you going to give a Blinux using person few thousands dollars for software she will need soon? I hope so.
I'm using regular old Linux. I'm not sure why a Blinux user would purchase a Windows copy of Microsoft Access, or how doing so relates to TCPA at all.
Right! Are you really so naive to believe we'll be able to run our own kernals as part of our "trusted" OSen? Please...
Of course not -- that would spoil the nature of a trusted kernel. I'm again curious as to what your beef is. You apparently would find a custom-compiled trusted Linux kernel acceptable. This means that you find not using Windows software also acceptable. I've pointed out that media will hardly be protected by any attempts to copy protect it. Going through the categories, the only type of content that you could possibly still be concerned about (where you would lose support) is software that is currently supported under Linux, but would *not* be supported after a TCPA implementation, and would require a trusted system. I cannot imagine what software you are thinking of.
But all is well, and any concerns are just overblown stupidity in /.
Yes, I'd say that they're exaggerated.
MS is found guilty of abusing its monopoly power.
Netscape is gone thanks to anticompetitve practices.
Yup. Microsoft is big and nasty. Not *as* big and nasty as IBM was twenty years ago, of course -- things have improved quite a bit since then. Not nearly as nasty as Apple's historically been within its own little domain. But, yes, nasty.
We are approaching the IT monoculture were one company alone has the power to decide where the whole industry is going.
No, we've left the PC monoculture where at one point, one company *did* have that power, gone through a series of dips and turns. What people are complaining about in the current story -- DRM in x86 machines -- is a product of a vast number of meetings and agreements, where an industry that *dwarfs* Microsoft decided that DRM would be worthwhile to try. Same thing goes for the world of power-saving. Microsoft doesn't even begin to have the power to implement something like this themselves. They are certainly an influential member of the tech world, but all they can do is coax. And no, they can't "buy out" the industry. MS is big, but they're a drop in the tech industry bucket.
But no, that must be idiots like me, that when faced with a XP machine for the first time could not copy tracks of my CD unless it was done in MS's propietary format.
Yes, if you're using *Windows*, you are subject to what Microsoft decides. That's the sole area where Microsoft *does* have absolute control. Same goes for Mac OS and Apple. If you dislike a vendor's software enough, you'll end up switching -- to Mac OS or Linux. I own a single hybrid CD, and I've happily ripped that to my computer, which is running Linux. There doesn't seem to be any sort of move to prevent Linux for doing whatever it wants to. Microsoft has a fair amount of cash and a fair amount of market share. It can spend it to implement whatever it wants. Keep in mind that unpopular decisions will cost even Microsoft money and market share.
My point was that *Slashdot* doesn't significantly influence Microsoft, not that Microsoft can do whatever the hell it wants to.
hey, guess what, the only commercial software I use on a daily basis is the Windows kernel itself!
/cygdrive/c prefixes.
/bin, /usr/bin, /usr/local/bin, /usr/X11R6/bin, /opt/$whatever/bin, ~/bin, etc., instead of being forced to put it in Program Files? I don't understand why my single-user desktop machine should need to use a file system layout that might concievably have made sense on a mainframe thirty years ago.
I also use cygwin when I use a Windows machine. Frankly, cygwin is nice, but it's effectively a slow, poorly-supported (I mean, I don't mind a bit of a challenge, but the last time I tried building mplayer under Windows, it was a bitch and a half) subset of the functionality I have under Linux. Everyone I know that uses cygwin on their machines has since migrated their primary machine to Linux or BSD (or started using cygwin after doing so). Any cygwin software is also subject to not only the constraints you mention below (your problem with the UNIX filesystem layout), but the frusterations of layering a system on top of Windows...damn
There's plenty of open-source stuff to use under Windows.
I find cygwin to hardly be a replacement for using an *IX box. There are a few native open-source Windows projects. PuTTY is an excellent example. There are also a few good free but not OSS tools, like the sysinternals utilities. However, they aren't strace and ltrace, that's for sure.
And it's easier to _find_ a reliable Windows driver for $HARDWARE than to find a Linux one of any sort
It's certainly possible to find hardware for which there is no Linux support. Probably the worst offenders are among cheap, non-HID-compliant USB devices, and some highly specialized PCI boards. However, I have never owned a piece of hardware that lacked a Linux driver -- though I will admit, I have had times that not all *features* were supported. Up until ALSA came along, hardware audio mixing under Linux on all but a few cards required the purchase of a $25 driver suite. There's driver support for most printers out there -- but no ink level utilities. Not actually drivers, but I suppose that related utilities could be considered relevant. It could be the fact that I've always avoided software modems, which were a sticking point for a while (this dates back to the days when I used Mac OS and Windows, where the CPU hit was a significant issue), and webcams (never wanted to broadcast myself).
As a matter of fact, the only time I've run into a n unsupported hardware issue relating to Linux when Windows 2000 didn't support a sound card that had been used under Windows NT. The card vendor and the chipset vendor had both gone out of business, but there was a Linux driver still being maintained because it was GPLed, and my Linux box at home could use the thing.
Oh, do you mean how in Linux you get to choose whether to put something in
No, I was talking about symlinks. I mean that I can drop a data file in one place and then aim programs that look in a particular directory for their data file at it. Shortcuts are an ugly hack.
Finally, I don't consider your claims that you're worried about where to put binaries valid. All these software packages are easily installed via package manager, which handles everything for you. The existence of ~/bin is a solution that Windows does not have -- it lacks a standard location for user-local software installation (and frequently, because of assumptions about registery permissions, software cannot be installed user-locally under Windows).
As a matter of fact, most *Windows* installers prompt you to specify where the software is to be installed (though there's a sensible default in Program Files in almost all cases). On Linux, there is no asking if you simply are installing a package.
No. Try the following:
That will eat up all the memory on the system, including the pagefile, but not constantly hammer it, so it won't be forcing paging. Then use Explorer for a bit.
Think it'd make a difference if I disabled paging the kernel? That's the first thing I do when I get a new Windows system. I wonder if that's why my machines always tend to stay up
Because you're running a nonstandard setup, the details of which do not affect even a tenth of a percent of the Windows market. It certainly won't hurt stability to disable paging...
I don't know if I just have damn good luck with hardware or what.
You do. I've yet to bluescreen a 2k+ stable box on harder that wasn't in some way suspect without any third-party drivers on it. I've seen all kinds of blue screens over the years (mostly on friends' systems -- I left for Linux well before 2k was released) that go away when changing driver version or installing a new driver.
I've yet to have any stability problems with Linux drivers -- about four years ago, I managed to hard-freeze a Linux system when hammering it with multicast code, but aside from that, I can't remember ever having seen hard-freezes or kernel oopses when playing with devel kernels. I use ALSA (which *will* be part of the Linux kernel, admittedly), mga_vid (mplayer's included Matrox-specific accelerated driver), and have at various times used lufs (Linux Userspace File System), and the virtual archive file system that Rox Filer used for a while, among others.
Now, I *have* seen what I've considered aberrant behavior during that time -- way back with RH 5.2 and the 2.0 kernel, which would be ge, I remember the latency being so high on the Sound Blaster's MIDI port driver that the serial port driver would occasionally not empty the UART's FIFO fast enough -- so when I used my joystick, my downloads would slow to a crawl due to dropped characters and subsequent retries. I'm sure I've run into other oddities in that time, though I don't remember any off the top of my head.
Windows drivers are decidedly more flaky than Linux ones.
VALVe officially "0wnZ j00," should've posted anonymously...
True, and the act of downloading such software infringes their copyrights. However, there would be very little benefit to Valve in attempting to pursue the thousands of people that obtained the code through Slashdot alone, much less through secondary and alternate distribution. They're trying to keep websites from posting the code, but obviously it will never be eliminated.
Seriously, I considered downloading it since I like to toy with 3d programming, and I'd love to see how the "pros" do it, but then I realized that if I were to be caught, anything I might ever want to release myself could be considered infringement in some skewed sense.
No, this is not correct. This is a popular misconception that dates back to the reverse engineering of IBM's PC BIOS. To be *absolutely certain* that there would be no legal challenges (not because it was legally necessary, but because it made it obviously impossible for any engineers to have copied code), Phoenix did not allow any of their engineers that had seen IBM source to be involved in the reverse engineering of the BIOS. While this so-called "clean-room" effort has been cited many times over, it does not have ground in legal *requirements*. Perfectly legal non-clean-room reverse engineering occurs daily in industry -- and you're not even talking about reverse engineering, but simply implementation of a project in a similar category.
I've looked at the quake source code, but it didn't help me much, and by now it's horribly outdated, so seeing some real, working and current, source code to something like this would be beyond incredible for me to pick apart and analyze.
You do not have to worry about legal taint (IANAL). However, if you want a modern 3d engine to look at and are still concerned, you may want to check out the LGPLed Crystal Space.
Because such a port would not be legal in the United States and most other nations. As such, few people would work to maintain it, and few would be willing to distribute it.
Secondly, the game content -- audio, etc, would still be commercial. While I might be willing to spend time porting code of some game to Linux, if I made the extensive kind of time investment involved, I would expect content-creating members of the project to do the same -- to donate their work, so that it may be distributed freely. Obviously, that will not happen in this case.
I don't get it.
How is gunpower supposed to go "unstable" to the point where water will make it detonate?
Morale is: Pirates will manage to pirate it. Do not piss off your customers in the futile process.
Rely on the ever-more-valid assumption that customer will have machine on always-connected broadband. Implement crucial chunks of the application on a remote server, so that you're essentially an ASP. Try pirating that...
This is an interesting tidbit.
I'm pretty sure he did the obvious thing, though -- took the traditional phrase "800 pound gorilla", comitted the not not-entirely-uncommon error of subsituting "500" for "800" and decided to substitute "elephant" for "gorilla" -- because, after all, an elephant is bigger than a gorilla, and *surely* his point would be better made with a bigger creatre.
The thing is, AFAIK, OF provides minimal benefits to anyone but Apple and Sun -- hardware vendors support x86 as a matter of course.
FWIW, I've seen a bunch of Apple folks on Slashdot already latching on to something they can promote Apple on, claiming that OF isn't a Trusted BIOS (gee, like every other freaking x86 BIOS aside from Phoenix's efforts?)
If Microsoft did somehow find a cure for cancer, it would not be distributed to those who most needed it, but to those who could afford it.
If $LARGE_CORPORATE_ENTITY did somehow find a cure for cancer, it would not be distributed to those who most needed it, but to those who could afford it. This is not a Microsoft-specific issue.
The only reason Microsoft hasn't done stuff like this is because somebody has taken the time to raise public awareness about what freedoms we were about to lose.
No, Microsoft has regularly ignored PR issues WRT to making decisions (at least once it has committed to a decision).
Believe me, the execs at Microsoft (note: not engineers there) care very little about what Slashdot has to say. Slashdot is full of after-the-fact complaining and rumors, and is constantly full of "sky-is-falling" stories. The few that I actually have inside knowledge on have let me realize that Slashdot is a wildly alarmist publication. Generally, people submitting stories have a political agenda, and want to get people upset about something -- and providing misconceptions is an effective way to do something.
Slashdot is fun to read, but it's not a Microsoft leash.
Things where there are *wildly* overblown claims and theories include: United States governmental repression, TCPA, Palladium, most stories on corporate business relationships with either SCO or Microsoft, Gentoo, and Apple products. Articles like these should be treated about the same way you'd treat something from the National Enquirer or the Young Republicans -- with a very, very large block of salt.
I think it is time to renew my EFF membership. I believe everyone should do that instead of just complaining on Slashdot.
And yet, here you are complaining on Slashdot and trying to convince *other* people to donate to the EFF instead of doing so yourself.
TCPA mandates that the system be user-disableable. It's part of the spec. TCPA will *never* refuse to load Linux, though it will *always* refuse to load as a trusted OS (since the kernel isn't signed...though I believe there is a project to get a TCPA-enabled copy of Linux). You're on quite a slippery slope if you're arguing that the current state of affairs would lead to the wildly unlikely state of affairs where all the x86 non-Windows server operators are convinced to use Windows instead. Nope, not going to happen, and vendors are uninterested in being locked in to Microsoft.
There are things that people may not like about TCPA, but they're generally not the ones that the EFF is insinuating (which tend to be improbable rabble-rousers -- they sound nasty, but they aren't the sort of thing that's actually going to be an issue in the foreseeable future.)
Other good reasons to use Linux:
* It's incredibly easy to script and build new applications by tying together existing ones via pipes. The results are fast, reliable, and professional -- unlike AppleScript or VB-produced results. This is only relevant to tech users, but it's a big one.
* It's free. Okay, for a professional with a decent salary, the cost of Windows vs Linux itself -- the base package -- really isn't significant. A hundred or two hundred bucks is not a big deal. However, to purchase commercial equivalents of all the Linux apps I use would be extremely expensive. Compilers (think Visual Studio), editors (think Visual SlickEdit), mail clients (think Eudora), system monitors (think all manner of shareware apps), sound editors (think Cakewalk), image editors (think Photoshop), web servers (think IIS), code checkers (think Gimpel Lint), graphing programs (think Visio), math/statistics packages (think MATLAB), and all the rest, there is a *lot* of money involved. Sure, you can pirate it, but that's not an option at work, and pirating software is less and less trivial with the surging prevalance of phone-home features.
* It's secure. Traditionally UNIX (and its apps) have had tighter security design than Windows, especially WRT local security. A couple of Microsoft apps are phenomonally insecure (MSIE, Outlook), and most Windows apps don't have the same emphasis on avoiding attacks.
* It gives better performance. My workstation runs a large set of servers in the background. I don't notice. I have a friend that runs a Windows FTP server that he kills off when he wants to take all the CPU time on his system.
* I can fix bugs that piss me off. If I have an issue, I happen to be a coder, so I can run out and fix it without just complaining to a company's forums and hoping that something happens. I can add features that I want. Obviously, this benefit isn't nearly as good if you aren't a coder, but it's something to consider.
* I can actually see what's going on. Linux has a strong tradition of talking about and letting you see what's *actually* happening on your system. The startup system is just a bunch of scripts that are quite readable. In contrast, if you pick up a book designed for a Microsoft administrator, you'll get a bunch of Microsoft-invented terms ("Enable a service"...am I starting a process listening on a port or what? What the hell is happening?) This also makes troubleshooting much better.
* A richer toolkit. For at least coders, network admins, and security types, good tools exist that have no Windows equivalent. (The reverse tends to be true when it comes to office workers.)
* Choice. If I use Windows, I also must use Explorer, like it or not (and I don't). I can't use the kernel or Windows software without also using the expected file manager (yes, there have been a few hacks to try "replacing" Explorer, such as LiteStep, but they're flaky...more neat toys than pratical tools). On Linux, I have more window managers available than I have fingers. I have a whole collection of file managers. I have docks galore. I can choose my favorite from each category and use that.
* Better design. The fact that Linux uses better file-locking semantics, the fact that Linux uses symlinks instead of shortcuts, the fact that it's easier to write a reliable Linux driver than a reliable Windows driver, all have strong trickle-down effects to the user in the form of fewer reboots, more flexibility in file system layout and control, and a more reliable system.
You decide to upgrade your old Windows 98 to Windows 2000 because it's so much better than Windows 98 was. If you are on a telephone line, dialing up to your AOL account, how are you supposed to get your system patched (in the THREE HOURS it takes to download and install ONLY THE CRITICAL UPDATES) before some cheeze-head hits your system with one of the umpteen remote exploits that you are vulnerable to?
Much as I prefer Linux to Windows, you aren't being entirely fair. The main problem is that Windows runs a horribly complicated and wildly insecure protocol -- NetBIOS -- that was never designed for non-LAN use (hence Microsoft's lame suggestions to "firewall" the service and blaming sysadmins for not doing so instead of fixing the damn thing.)
An old copy of Linux will run into the same problem. Red Hat Linux 5.x-era stuff used to run plenty of daemons out of box, since the technical crowd knew what to disable. If you didn't axe said daemons before hooking up to the 'Net to get updates, you could be equally screwed.
So, with Win2k, you need to do the same thing. Disable the Server service, along with any other services listening for connections. Then plug your computer into the network and start updating it. Aside from TCP/IP-level exploits, you should be pretty safe.
Modern Linux distros run few daemons out of box (though I still wish they were tighter). Win 2003 Server does the same thing. Microsoft learns, abeit slowly and after a sufficient number of users have been badly burned.
Granted, there are still some incredibly bad technical decisions that Microsoft made for political and business reasons. Making Internet Explorer (big, incredibly complex, fragile, full of security issues) the client to use for automatic updates is pretty horrific.
Yes, Windows 2000 is better than Windows 95. Big deal. My wristwatch is better than Windows 95.
If you're talking about reliability, that's actually a pretty high standard. Your wristwatch doesn't have any nonvolatile state to muck up, so worst case you can remove the battery. It's fairly simple -- I don't believe I've ever seen a buggy wristwatch (though I have seen other buggy embedded devices, like network printers.) Your wristwatch is a pretty reliable realtime devices that can take a physical pounding, is probably waterproof, and can run for a year and a half without a battery change.
Not necessarily.
A stock, vanilla Windows install seems to be fairly reliable for basic tasks. (Not necessarily for remaining usable or stable under heavy load or resource usage -- try eating up all free memory and watch Explorer die the first time it hits an unchecked malloc()) However, if you get new hardware, you're probably going to be using third-party drivers. These aren't necessarily as solid. While some of this is just the fact that there are more companies making lousy hardware for Windows employing incompetent device driver authors, there is also the issue that Windows device drivers are far, far harder to write correctly than Linux device drivers. Windows has pageable kernel memory, requires all sorts of special work WRT power management, etc. Windows is a much harder environment to write a reliable device driver in, and driver quality reflects this.
OpenFirmware is absolutely INCREDIBLE, and if more companies were on-board it would get even better. On a Macintosh (O.F.) you just hold 'option' at boot and you get a menu of all bootable drives connected to the machine, be they FireWire, IDE, SCSI, or USB (actually USB is disabled out of sanity). You can get a device list even better than most Operating Systems can provide from OF.
You know, I hate to say it, but I just plain haven't booted off a non-ATA volume in forever. Haven't had any reason to, and haven't even used USB-boot features of an x86 BIOS.
From what I can tell, the benefit of OF is essentially to help Apple and Sun. Both have fairly marginalized hardware platforms (SPARC and PPC), and worry about lack of architecture support from hardware vendors. So they come up with a solution that provides architecture-independent support, hoping that they won't be ignored by hardware vendors. The x86 world doesn't have this problem, and Intel has not, to the best of my knowledge, been particularly interested in making Open Firmware universal, as it provides little benefit to them (as a matter of fact, I believe that they have their own, different solution to handle IA64/IA32 support).
And yet, despite all this, I don't think that spam is going to go away because of legal efforts. It will take technical ones.
I think that email is a doomed beast, though I'm sure much effort will be spent trying to save it. The ability for anyone to freely send you email into your inbox is a relic of the 70s and 80s, much as anonymous ftp sites with incoming directories for handing files around were. The email system obviously is great if it can operate successfully -- but I think that it's ill-adapted to today's social environment. Reminds me a bit of communism -- sure, it'd be great if folks could get it working, but you have to assume that people will willingly cooperate and not abuse the system. These assumptions have been found to be false, and modifications made to all other major in-use protocols to prevent abuse. (I've been involved in adding anti-abuse modifications to P2P clients, and have run into plenty of this myself.)
Email in its current form (where almost anyone can send almost anyone email freely) will eventually (and probably painfully) die, and alternate systems, where most communications are whitelisted, will come to the fore.
You know, I really hate to say it, but I'm sorry to see the source go down (Valve says that they're asking websites to take it down -- not sure if that translates to asking or sending legal threats). I'm skimming through a copy of it that I downloaded with some interest. It's not often that you get such insight into game development (post-mortems are interesting as well, but source hasn't been neatly edited). While I doubt a commercial gaming company would ever swipe code from Valve (too much potential damage -- if there's even a 10% risk of exposure, they're better off just licensing it.) Cheating will obviously be a problem...if I were Valve, I think I'd consider significant protocol revisions.
Some people have said that Valve has included GPLed code in the Half Life 2 code. Dunno as to whether this is true, but I'd like to point out that while this is technically not kosher, I suspect that a lot of places do it -- as long as it's out by release time, I very much doubt that anyone will complain. (On the other hand, if it *isn't* out...)
This is a good example of why internal security is very important. I don't use any computers at work that don't talk to each other through encrypted connections. I maintain a single trust relationship (pubkey based, not IP based), from a machine that has a superset of the information on a second machine, so there's little point in exploiting trust relationships (plus, if superset machine A were compromised, a keygrabber could easily allow compromise of machine B anyway). I don't use Windows filesharing. These are all very easy to overlook, especially during crunch time, but as Valve has discovered, while the chances of things going sour may be low, the potential damages are enormous. I would urge folks who are working with *any* kind of important IP to do the same -- do *not* rely on Windows filesharing, do *not* use trust relationships, and do not use unencrypted connections, even on your local network. SFTP exists and there are free clients all over -- you do not have a good excuse for using FTP.
This is also another example of why it may be worthwhile to have a network admin that does regular security audits. It takes additional time, and the vast majority of time that cost is overhead, but Valve is certainly regretting not doing so at the moment. (We have irregular security audits, which is better than nothing, but obviously not ideal.)
Finally, I'd like to say "chin up" to the folks at Valve. This sort of thing can be very frusterating, and I'm sure it hasn't helped morale at Valve much, but it's not a game-killer, even if it necessitates changes in the protocol or game engine, and a release delay. Good luck -- I probably won't buy your game, since it's unlikely that there'll be a Linux client, but I expect you'll have healthy sales.
As for other folks -- remember crack.com, remember Valve -- secure your damn networks already.
Thank you. This is indeed a good, spot-on resource that I hadn't seen before. I was quite satisfied by the fact that his points seemed to be similar to my own. :-)
NEVER install software from a company you've either never heard...
...or don't trust
Screws me over WRT most open source software.
And that wipes out closed source software.
People need to stop trashing Earthstation 5. It's a fantastic program, and does exactly as advertised. Plus, it seems to have built-in compression software -- my free disk space has been steadily increasing ever since I installed it!