Slashdot Mirror

Did it with BattleTech years ago.

What would normally be a 3-4 hour game became a 7 hour game.

The problem is making certain everybody's on the same page (and not cheating).

Now tabletop simulations like Megamek outstrip tabletop over IRC by orders of magnitude.

And there's a Powertool from Microsoft (as well as third-party applications) which can be used to make use of them.

Frankly, ZoneAlarm is goofy.

AppArmor is vastly superior, in that it also can be used to regulate resource privelidges on a per-application basis, but instead of only controlling network access, AppArmor works on a system-wide basis. Furthermore, AppArmor can isolate applications from one another. The GUI isn't so bad.

Oddly enough, we have "Ask Slashdot:" articles discussing the very technology that underlies AppArmor, LSM, and how one might be able to find a similar thing on Windows.

Furthermore, the types of issues that cause you to use ZoneAlarm aren't nearly as prevalent on Linux. You don't get malware, and OpenSource and/or high-end pro software tend not to phone home randomly.

If you're really, really, really determined to have lots of really, really annoying popsups (remember that things like keyloggers are resolved by AppArmor), you can use either Program Guard or SysTrace for Linux. Program Guard annoys you about TCP/IP access on a per-application basis, while SysTrace annoys you about everything.

TuxGuardian is apparently another app like this

NetLimiter: I do not understand the point of this application. Why would you ever want to do per-application bandwidth shaping when you can do global L7 QoS? Furthermore, it seems to me that you can use a combination trickle for hard "per-application" limits (which, IMHO, don't _ever_ make sense_, and global QoS to acheive any combination of features you could potentially acheive with NetLimiter.

This is a list of GUI iproute2 QoS configurators, but I think you're pretty much fine running Wondershaper, and perhaps watching pretty graphics go by with MasterShaper.

As it is, I run 6 desktops, 3 vonage lines, and 3 laptops over a Comcast 8Mbps/768kbps connection. I use one firewall on the router, running linux, with QoS enabled and global L7 traffic shaping. We have no problems when simultaneously running Limewire, Bittorrent, Vonage, and generalized web access (everything remains responsive).

The real problem with pointing at these sorts of applications is that this kind of functionality is just not needed on Linux. Proper application isolation, lack of malware, high quality global QoS, and decent packet filtering means that these kinds of annoying GUIs that are really nothing other than system maintenace and mundane micromanaging are not needed. I don't need to rate limit my downloads or uploads in order to preserve network responsiveness; I don't need to watch my applications to see if they are phoning home or not. I don't need to worry about whether or not my financial data is being read by malware; I don't need to worry about whether compromised user-apps on my system are affecting admin-level system services.

If you really, really, really, really want, the tools are out there, in proper Java, QT, and/or GTK form. But the reason they aren't widely deployed is because you really shouldn't be using them; a computer is a tool for work or entertainment, not an adventure game on its own. We don't live in the Tron world; and much like you don't need to have pressure gauges and per-pump control over your automobiles fluidic systems, you don't need to have direct control over this stuff on Unixy systems. It just works, and that's good enough for 99.9999% of non-super-geeks out there. For the remaining .0001% of us, we write our own GUIs, hunt out little known programs, or use the commandline. But the vast majority of computer users out there shouldn't need to be familiar with a tool like ZoneAlarm, and shouldn't have to worry about all those bloody popups. For the m

Frankly, ZoneAlarm is goofy.

AppArmor is vastly superior, in that it also can be used to regulate resource privelidges on a per-application basis, but instead of only controlling network access, AppArmor works on a system-wide basis. Furthermore, AppArmor can isolate applications from one another. The GUI isn't so bad.

Oddly enough, we have "Ask Slashdot:" articles discussing the very technology that underlies AppArmor, LSM, and how one might be able to find a similar thing on Windows.

Furthermore, the types of issues that cause you to use ZoneAlarm aren't nearly as prevalent on Linux. You don't get malware, and OpenSource and/or high-end pro software tend not to phone home randomly.

If you're really, really, really determined to have lots of really, really annoying popsups (remember that things like keyloggers are resolved by AppArmor), you can use either Program Guard or SysTrace for Linux. Program Guard annoys you about TCP/IP access on a per-application basis, while SysTrace annoys you about everything.

TuxGuardian is apparently another app like this

NetLimiter: I do not understand the point of this application. Why would you ever want to do per-application bandwidth shaping when you can do global L7 QoS? Furthermore, it seems to me that you can use a combination trickle for hard "per-application" limits (which, IMHO, don't _ever_ make sense_, and global QoS to acheive any combination of features you could potentially acheive with NetLimiter.

This is a list of GUI iproute2 QoS configurators, but I think you're pretty much fine running Wondershaper, and perhaps watching pretty graphics go by with MasterShaper.

As it is, I run 6 desktops, 3 vonage lines, and 3 laptops over a Comcast 8Mbps/768kbps connection. I use one firewall on the router, running linux, with QoS enabled and global L7 traffic shaping. We have no problems when simultaneously running Limewire, Bittorrent, Vonage, and generalized web access (everything remains responsive).

The real problem with pointing at these sorts of applications is that this kind of functionality is just not needed on Linux. Proper application isolation, lack of malware, high quality global QoS, and decent packet filtering means that these kinds of annoying GUIs that are really nothing other than system maintenace and mundane micromanaging are not needed. I don't need to rate limit my downloads or uploads in order to preserve network responsiveness; I don't need to watch my applications to see if they are phoning home or not. I don't need to worry about whether or not my financial data is being read by malware; I don't need to worry about whether compromised user-apps on my system are affecting admin-level system services.

If you really, really, really, really want, the tools are out there, in proper Java, QT, and/or GTK form. But the reason they aren't widely deployed is because you really shouldn't be using them; a computer is a tool for work or entertainment, not an adventure game on its own. We don't live in the Tron world; and much like you don't need to have pressure gauges and per-pump control over your automobiles fluidic systems, you don't need to have direct control over this stuff on Unixy systems. It just works, and that's good enough for 99.9999% of non-super-geeks out there. For the remaining .0001% of us, we write our own GUIs, hunt out little known programs, or use the commandline. But the vast majority of computer users out there shouldn't need to be familiar with a tool like ZoneAlarm, and shouldn't have to worry about all those bloody popups. For the m

http://sourceforge.net/

rgds

http://popfile.sourceforge.net/

... the ones which have worked best (for me) are Bayesian Spam Filters (A Plan for Spam, SpamBayes - a free filter) and CRM114 The Controllable Regex Mutilator (Paul Graham mentions it here). I've always had a very high success rate with these.

... the ones which have worked best (for me) are Bayesian Spam Filters (A Plan for Spam, SpamBayes - a free filter) and CRM114 The Controllable Regex Mutilator (Paul Graham mentions it here). I've always had a very high success rate with these.

Theres always links!

Mozilla got so much money. But do they invest in development? I fear they don't. Just look at Sunbird, Lightening or whatever the calendar is called these days. Or NVU's son KompoZer?

Mozilla has the ressources to cross-finance development of other tools, to bootstrap open source. But it seems they don't want to.

I mean we have a successful tool called Firefox every company likes to play with, including a fanatic user community. We have a a wonderful mail client which lacks a calendar tool.

But what about other tools of the community? Chatzilla - wouldn't it be nice to get a standalone version? Or Fireftp stabdalone? A preconfigured Bugzilla server distribution. KompoZer. Better spellchecking tools and dictionaries. Tools for Internet Cafés, I think of a kind Browser-Only plugandplay Linux distribution. Brushed Theme for Thunderbird. An ODF view plugin. A usable pdf viewer. Developer Conferences coorganised by their mozilla-hungry AJAX-fanatic bigbusiness friends.

Real Networks, oh well.

Take me back to the days of BlackBox

Go for it:

http://blackboxwm.sourceforge.net/

http://www.bb4win.org/news.php

KFG

I used to be a big RH fan until my personal subscription was cancelled early around the end of RH9 days. I struggled with Fedora for a while, but it was always flaky and updates seemed to break more than they fixed. I moved to Slackware... which I liked for the same reasons I really liked Gentoo -- it was easy to install packages from source; however neither of those distros is as easy to manage in a commercial setting as other more popular distros.

Then a colleague mentioned CentOS and I have not looked back. CentOS3 is still 2.4 kernel based and would be a good replacement for the older RH7x/8x/9x versions. CentOS4 is 2.6 kernel based and a great distro. I have moved everything I had from RH to CentOS and not looked back.

I even use CentOS4 on my work desktop with the KDE-RedHat repository and the Dag Wieers repository supplying packages for software that RH/CentOS do not.

I think I am going to revisit Kubuntu in the near future though, because I do like apt, I simply cannot stand Gnome (no offense - I simply prefer KDE).

• Primer arranque (First boot): Allows graphics booting using gfxboot, something quite nice for "normal" users, as I remember my flatmate got scared at the Ubuntu screen boot, with the list of the [OK] and [FAILED] services status (background here).
  
• Instalación (Installing): Just a graphical installer with graphical partition resizing, I saw this already when installing Ubuntu and Mandriva.
  
• Más comodidad (More confortability [is that a word?]): Just the old root/user password option with automatic login.
  
• Un Escritorio más vivo (more alive desktop),Mantente a la última (stay at the edge), El nuevo Actualizar LinEx (New LinEx Update): Some desktop backgrounds, system update and package installer. Nothing too fancy IMHO.
  
• Aptéalo con APTZILLA ("Aptate it" with APTZILLA) : This is something which I believe is worth to mention, I have never seen something simillar in any other distribution. It seems to be a Firefox extension that enables to install software from an internet page. It would be very interesting to try it because from what it seems it would be a way to achieve the "click+download+click^x+install" behaviour in Windows for the end user (my father for example wont be able to install Repast framework in Ubuntu because it is not in the repositores, whereas to install it on Windows he just have to download the installer and run it).
  
• El Panel de Control de gnuLinEx: A control panel similar to what a lot of other distributions have. HOWEVER I find quite relevant that they embed the WINE emulator (which btw I.N.an E.), I imagine they try to make as easy as possible to enable Windows applications to run in Linux. That is the other property worth to note, as I have not seen any distribution that gives so much importance to it (well, besides the commecial distros like xandros, lindows [ya ya I like to call it Lindows], etc).
  
• El wiki personal (The Personal Wiki [see, spanish is not that hard]): That is the other interesting application, which a wiki like note taker (the application seems to be Tomboy.
  

Have you tried Lincity? It's not a clone of any of the SimCity genre, but it's still quite fun.

I've never been interested in the Sims genre, so I haven't looked up any Linux equivalents.

A very nifty tool, not very well known: http://hugin.sourceforge.net/

It's not the best, but it's open source and it does work to read UPC and EAN codes: http://sourceforge.net/projects/barcr-reader/

hugin does photo-stitching pretty well, I find, and is open source.

I use Hugin Panorama Tools. I think it works really well, though I don't have a lot of experience with other stitchers. It can do panorama stitching, and also correct for barrel distortion and such.

Quite right, but there's also X-Moto!

PySol, Fish Fillets and FloboPuyo are also nice.

http://fillets.sourceforge.net/

http://freshmeat.net/projects/flobopuyo/

I wouldn't be surprised if they targeted C# after that.

See my comment in 2000 to Doug Engelbart's Bootstrap List at:
    http://www.bootstrap.org/dkr/discussion/0754.html
From there [with some outdated links removed]:

I'd love to make a souped up version of this for OHS/DKR use: (Read about in May 2000 Popular Mechanics) "Cybiko Introduces First Handheld Internet Wireless Entertainment System At Toy Fair 2000"

US $149.00 The Cybiko system combines instant messaging, interactive gaming, email and personal information manager (PIM) capabilities in an all-in-one device. ... Available in four translucent colors, Cybiko has a full QWERTY keyboard to compose messages, LCD display, .5 MB memory (expandable to 16MB), a high frequency transmitter and Vibration Alert feature. The unit measures 4.8 x 2.8-inches and weighs under four ounces making it light, thin and small enough to carry in a book bag, purse or shirt pocket. ... With Cybiko, kids and teens can communicate instantly with others within a radius of 150 to 300 feet, depending on the environment, creating their very own virtual community.

Wow!

Imagine what we could have for $1000 by the end of this year by integrating technology that already exists:

Develop a beefed up version supporting a distributed file system like Freenet...
    http://freenet.sourceforge.net/

Using technology like this 6GB in 14 ounces $500 portable audio player/recorder: [nomad Jukebox]:

And a two mile radio range: [Motorola walky talky]

Maybe with a next generation StrongARM 600Mhz processor:
    http://www.intel.com/pressroom/archive/releases/em 050399.htm
Like a faster version of: [BossaNova mobile processor]

Running Squeak (and maybe Linux) as an open source OS/Development environment:
    http://www.squeak.org/

Using Bootstrap OHS/DKR type ideas for the interface...

Powered by solar energy and/or Baygen radio windup technology and/or fuel cells.

And with a digital camera for fun and creation of educational how-to tutorials... (And on the spot news reporting...)

And remember that in five years this entire thing will cost US$100 each.

As an alternative, this could be a set of HandSpring modules instead: [Springboard]

Consider a couple of these souped up devices given to each village in Africa. Anyone with $1 billion for true development aid to 500,000 African villages? (This is just the cost of one unfinished dam or one shut down nuclear plant.)

Consider millions of these devices airdropped into Iraq and Yugoslavia -- instead of more expensive cruise missiles! Anybody got $1 billion to spend on ensuring democracy with a true defense against tyranny in those places? (This is probably what the U.S. military's spends on gas/oil for a month cruising the area...)

This is like a system I wanted to develop and deploy pre-Y2K just in case... But it still has much value in preparing for any potential (natural, political, economic, biological) disaster, as well as aiding the development of democracy.

It's somewhat like the wearable crystals described in The Skills of Xanadu" by Theodore Sturgeon (available in his book The Golden Helix), although the one thing it lacks is easy self-repliaction...

Developing and then deploying this sort of device is the sort of thing the UN or a major foundation should fund (if they were on the ball). But luckily, there is hope from toymakers!

====

Anyway, glad to see six years later this is going ahead at that $100 price point (and developed by other than toymakers). My hat goes off to the dedicated people making this happen.

Oh my oh my, why isn't wesnoth listed here ? It's a kickass game for being opensource portable etc.

http://www.wesnoth.org/

I like it, only wish that the online games would last a bit longer ;)

As a close second for me, come torcs http://torcs.sourceforge.net/ and danger from the deep http://dangerdeep.sourceforge.net/

Linux is quite playable today, if you're just into fps shooters, go for q3 or enemy territory :)

Oh my oh my, why isn't wesnoth listed here ? It's a kickass game for being opensource portable etc.

http://www.wesnoth.org/

I like it, only wish that the online games would last a bit longer ;)

As a close second for me, come torcs http://torcs.sourceforge.net/ and danger from the deep http://dangerdeep.sourceforge.net/

Linux is quite playable today, if you're just into fps shooters, go for q3 or enemy territory :)

You can build a proof of concept for under $500. There is an application out there called vblade[1], which is a virtual EtherDrive blade. Basically, it allows you to export a local block device as an ATA over Ethernet device. Turning your junker PC in the corner with a couple of extra disks in it into a ATAoE testbed.

It's GPLed code, works with the new native kernel AoE support, and was even written by one of the Coraid guys.

[1] One of the aoetools, available on sourceforge here.

I've been playing with this on a spare machine in the test rack in our data center, and I have to admit, it is pretty slick. There's potential here for use in a number of areas.

Not a few cores, just one: Core 5

yum install xen kernel-xen0 kernel-xenU
on FC 4 works fine, but is not very cross version compatible with FC 5.
FC 5 has a nice auto build script for creating new domU's.
People have done mixed FC 4/FC 5 Xen configs, some info is in the FC/Xen mailing list.

As far as FC 6 delays due to Xen, because of the rapid pace of dev on both the FC 6 Linux kernel and Xen 3.X.Y there is fairly frequent breakage of some type found every time one or the other is updated. Lots of people, some associated with RH/the Fedora Project as well as individual users, work daily to debug and tune the FC/Xen releases.

https://sourceforge.net/projects/xenman/

I just tested this last week. vblade installation was a snap. From there it's a matter of creating an empty file, associating a loopback device with it, and having vblade "export" it to the LAN. On another machine, load the aoe kernel module, install the aoe-tools, and run aoe-stat to see that the device is exported. You can then mount it and stick a filesystem on it from the remote machine. As a test, I started with ext3. Worked fine. Then I put OCFS2 on it to see if I could mount it read-write on 2 hosts. That worked too. All in all, I'm impressed. This should work well for backup, cheap HA, or Xen environments.

The AoE tools, including vblade, can be found here.

:)

The big issue now is symbian signed. With S60 version 3 onward, they've seriously locked down the platform. If your code isn't signed, it won't run on most devices, and even where it will, it won't be allowed to do interesting things (write to filesystem, talk to network etc). If you want to get your code signed, you have to have an expensive verisign certificate, and pay a bunch of cash to have your app reviewed.

I was able to install Putty for Symbian OS and other self signed software on my Nokia E series phone running S60V3. I had to turn off the signature checking in App. Manager to enable installation of self signed apps. This is set to 'Signed only' by default which does keep out malware but is still kind of annoying but Putty works as close to what was advertised as one can expect from Beta software.

In the context of using this in low-cost environments with Linux I can hardly see how this could kill iSCSI. Last week I implemented an iSCSI setup for about 500 euros (target serves out 500GB disk space for non-critical backup) using standard components, FC5, iSCSI Enterprise Target and Microsoft iSCSI Initiator.

Works great and is a lot (>10x) faster than the about similarly priced NAS device that was used for the same task before.

Hopefully, the engineer who designed this hybrid drive has, at a minimum, integrated an LCD counter and a tiny speaker into the drive. The counter shall display the running total of the number of writes to the flash memory. The tiny speaker shall beep like crazy when the total exceeds 99900.

More likely, they'd use SMART to let you monitor the flash's health programmatically, just as they use for the rest of the drive. This is far from the only designed limitation of modern hard drives, and SMART is a nice system for tracking them. Makes sense to keep using it.

As someone else mentioned, it'd be need to be more than a simple counter - the 100,000 writes lifetime estimate is per-sector. I don't know if it's practical or necessary to maintain such a large counter array. Decent wear leveling may be enough to ensure this limit is not reached.

I see lots of misconception of how FATFS deletion works, therefore I'm posting here what I hope is a comprehensive explanation of how it works.

First, you can go to my site and find FATFS specification from the developer's documention section.

You will also find a description of a directory entry.

Please note I'm differenciating between FATFS and FAT, FAT being a component of FATFS.

FATFS is divided in System Zone, FAT, and allocatable blocks (clusters). Before FAT32FS, root directory was also treated specially (a remain from DOS 1.0 when directories were unavailable).

FAT is a singly linked string list of blocks, and the first block is stored in the directory entry. As a practical example, a directory entry says that file first block is at b1. You look at FAT[b1]=b2 to get next block or an EOF mark. Then you read FAT[b2]=b3 to get next block and so on. It's quite inefficient with big fragmented FS because you need to seek only to get the block list. It was OK at the time of DOS 1.0 because the FAT size of a 360KiB floppy would only use one KiB and could be stored permanently in RAM.

When deleting a file, it's correct that only the first char is changed (to 0xE5) in the directory entry, but the allocation string in the FAT is also erased - because it's the only mean for FATFS to know that those blocks are available.

First block of erased file is still retrievable because it is stored in the directory entry, but other blocks can't, you can only guess, because if a file isn't fragmented, FAT[i]=i+1, and file size is stored in directory entry so you know when to stop.

On most other filesystems, a directory entry points to an inode, which in turn contains all the allocation string (maybe using indirection blocks). Free blocks are written in another structure, an allocation bitmap. So when you're erasing a file on those FS, you only need to erase the directory entry (filename is lost), free blocks in the allocation bitmap(s), free inode in the inode bitmap. You don't need to touch at the inode itself. Therefore, if you have the inode number (instead of the file name) it's in fact easier and more reliable to undelete files on those filesystems than with a FATFS. Except ext3fs zeroes inode block list (but it works on ext2fs).

I see lots of misconception of how FATFS deletion works, therefore I'm posting here what I hope is a comprehensive explanation of how it works.

First, you can go to my site and find FATFS specification from the developer's documention section.

You will also find a description of a directory entry.

Please note I'm differenciating between FATFS and FAT, FAT being a component of FATFS.

FATFS is divided in System Zone, FAT, and allocatable blocks (clusters). Before FAT32FS, root directory was also treated specially (a remain from DOS 1.0 when directories were unavailable).

FAT is a singly linked string list of blocks, and the first block is stored in the directory entry. As a practical example, a directory entry says that file first block is at b1. You look at FAT[b1]=b2 to get next block or an EOF mark. Then you read FAT[b2]=b3 to get next block and so on. It's quite inefficient with big fragmented FS because you need to seek only to get the block list. It was OK at the time of DOS 1.0 because the FAT size of a 360KiB floppy would only use one KiB and could be stored permanently in RAM.

When deleting a file, it's correct that only the first char is changed (to 0xE5) in the directory entry, but the allocation string in the FAT is also erased - because it's the only mean for FATFS to know that those blocks are available.

First block of erased file is still retrievable because it is stored in the directory entry, but other blocks can't, you can only guess, because if a file isn't fragmented, FAT[i]=i+1, and file size is stored in directory entry so you know when to stop.

On most other filesystems, a directory entry points to an inode, which in turn contains all the allocation string (maybe using indirection blocks). Free blocks are written in another structure, an allocation bitmap. So when you're erasing a file on those FS, you only need to erase the directory entry (filename is lost), free blocks in the allocation bitmap(s), free inode in the inode bitmap. You don't need to touch at the inode itself. Therefore, if you have the inode number (instead of the file name) it's in fact easier and more reliable to undelete files on those filesystems than with a FATFS. Except ext3fs zeroes inode block list (but it works on ext2fs).

There is the Wayback File System http://wayback.sourceforge.net/

It's a feature, and a pretty cool one. I wouldn't mind this in Linux. This is not a bad thing.

The VMS filesystem (Files 11) was an evolution of earlier DEC filesystems and had versioning buit in from the start. There's also a more user-oriented versioning filesystem which has been in development for Linux for the past few years.
http://sourceforge.net/projects/versionfs/

Cut the crap and the Microsoft bashing, I'm much more concerned about the spate of port 22 scans, and the brute force ssh password attacks going on right now.

Here's a project that implements what you are talking about:
http://dmx.sourceforge.net/

of course, at some point, you'll still have a video card, but..

When Noga and others came up with LegOS, an operating system for the Lego Mindstroms that enabled the writing of sophisticated programs, they were forced to change name, to BrickOS, I guess under legal thread from the Lego company due to misuse of trademark. So much for supporting the community! And the sad irony is that they must have sold lots of Mindstorms due to LegOS - pardon, BrickOS.

So this is a real direction change! I have a lot invested in LegOS code, and I am waiting to see if anyone will port BrickOS to that, or whether there will be any half-decent RTOS that runs on them... I am not holding my breath though.

for most of the geeks :-) Before being bashed by FOSS gurus, let me show my example.

I'm a Java and Ruby (on rails) programmer. I've dropped windows about 2 years ago, and used various flavors of linux in this meantime (debian, ubuntu, gentoo and ubuntu again). Less than one year ago I bought my first mac (mac mini).

Java development in Windows is "standard", in linux is good and in Mac it is great. You have the same tools as Windows or Linux and, since java is "portable", all other tools (frameworks) works fine. The difference between Windows and Linux/Mac is that Windows restricts you *a lot*. Ex.: I put all my libs in just one place, and make sym links to them in the projects I need. In Windows, its not possible (afaik) (yes, a simple example, but try to keep the libs updated on a windows box...) . And the difference between Linux and Mac is performance. *Usually*, a Java application runs faster on Mac than on Linux, because the Java VM in Mac is done by Apple, meaning that its built by the ones who knows the OS. In Linux, as you certainly know, is a certain pain to install Java (you need to follow one or another howto to get things working), and the performance is *usually* worse than in a Mac, because the VM is done by Sun, which is concerned mainly in getting things working. Yes, they care about performance, but not that much :-) (just remember that IBM and BEA's VM's performs better than Sun's)

When programming with Ruby, Mac is really awesome. Again, Windows looses here. In Windows, you have a set of tools (editors/IDE's) that also exists in other platforms, but its performance is poor (afaik). Also, some Ruby libs requires some sort of compilation (mysql, rmagick, ...), and it can become a pain to get things working. On Linux, things are far better than Windows. You have almost the same tools, but its far easier to get things working: just apt-get / emerge / whatever and you are ready to go. In a Mac, just "port install" what you need, just like linux. The difference between Mac and Linux is in the tools. The same ones + a fantastic editor (and cheap for some, expensive for others). Ok, its not that smart to left an inexpensive OS to go to an "expensive" one just because of an editor. But trust me, it worth.

Besides these work-related details, you also get an OS that just works, with enough applications to do what you usually do on a PC, a good terminal (I definitely cannot use the "cmd" anymore), a more than nice UI and so on... And for people who asks me "why use a mac", I just ask the same: "why use a Windows". There is no reason to use Windows. I can't find something that Windows does better than Mac (ok, I left an space here for some +5 Funny comments).

But yes, there *are* reasons to use Linux instead of Mac. Specially if you want "all the freedom you can get", if you don't want to spend a penny in software or simply don't care about the UI.

Of course, I talked about just the OS itself. The hardware *is* more expensive, specially here in Brazil (macs comes from US, which means they are taxed in *only* 100%). But if you think a bit better, it probably worth. In my case, I spend more than 10 hours/day looking at a computer, so, it certainly worth for me :-)

And I'm sorry, this would be a single-line comment, but it simply grows :-(

Unlike almost every other feld in the world, programming is something anyone who wants to can. There are tons of free books and courses on-line as well as source code, with a whole new source repository opening.

There simply isn't the discrimination you so desperately wish to see in the field of programming, only an overall lack of interest in entering such a field by one sex. As a percentage there are a lot fewer women working as movers. Why not start working on getting the percentage up over there?

http://cmusphinx.sourceforge.net/html/cmusphinx.ph p
http://sourceforge.net/projects/cmusphinx/

It's been around for a while. I think it's pretty good, though quite resource-demanding. The peeps at the tech-report usually benchmark it.

Here for instance:

http://techreport.com/reviews/2006q3/core2/index.x ?pg=12 -core 2 duo

http://techreport.com/reviews/2002q1/athlonxp-2100 /index.x?pg=6 -early Athlon XP vs P4
    (middle of the page)

As you can see, we can do real-time sphinx now.

http://cmusphinx.sourceforge.net/html/cmusphinx.ph p
http://sourceforge.net/projects/cmusphinx/

It's been around for a while. I think it's pretty good, though quite resource-demanding. The peeps at the tech-report usually benchmark it.

Here for instance:

http://techreport.com/reviews/2006q3/core2/index.x ?pg=12 -core 2 duo

http://techreport.com/reviews/2002q1/athlonxp-2100 /index.x?pg=6 -early Athlon XP vs P4
    (middle of the page)

As you can see, we can do real-time sphinx now.

Open Source Speech Recognition?

I believe one such system is called Sphinx.

['[S]ome of the "best" comments (obviously there's some subjectivity to it) are ones that may not be as highly moderated as some decent ones which happen to have been made earlier and therefore had more time to be moderated up.']

"That sounds a lot like "Moderation is broken"."

Well, I guess the way I'd put it is more like "Moderation is imperfect." That's one reason it's constantly being tweaked; it's certainly broken compared to Utopia! But until time can be manipulated like play-doh — or cookie dough, or any kind of dough — older (earlier) comments are always going to have been, just by definition, available longer for moderation to take place. Moderation is helpful, but will never be the one true path to enlightenment :)

Another way of saying it: moderation lets Slashdot function as an extended conversation, not just a shouting match; that's a seemingly low goal to shoot for, but it's a trickier thing to achieve than it sounds, and calls for a lot of juggling. I am very glad that there are clever, thoughtful coders who stew over the details. (And if there's a specific bug you think could be fixed in the moderation system or other parts of the code that runs Slashdot, they take requests.)

Cheers,

timothy

..which means you now have to have an insecure file on your computer storing your different made-up answer for each site... I hope to god that's encrypted and password-protected out the wazoo.

KeePass

I spent the whole yesterday refreshing the slashdot frontpage, and somehow managed to miss the story!

Anyway, this technique reminded me (yes I know they're very different) of airpwn, a piece of code which sniffs out the images and replaces them with the ones you specify, the authors had some fun at defcon 12

http://unxutils.sourceforge.net/

dd is included. I use these on my Windows servers at work all the time.

Furthermore, it implies that all one need do is supply their own BIOS and bootloader code that uploads the hashes from the original BIOS and bootloader and the OS underneath will be none the wiser because the message it gets back from the TPM will be exactly what it's expecting to see.

Good question. The way the system works is kind of subtle. This is perhaps why so many people prefer to believe the falsehoods that our out there, about how the system will only run signed code and keep you from changing your own software.

The TPM chip is completely passive in a TC system. It relies entirely on the BIOS, bootloader and OS to feed it data. Yet the system is designed to be secure. Here is how it works.

The first part of the BIOS is non-flashable in a TC system. It is hard-coded to hash the whole rest of the BIOS and feed that data to the TPM, first thing.

Then the BIOS, before transferring control to the boot loader, hashes it and sends the hash to the TPM.

And the boot loader similarly hashes the OS and some config files, sending that data to the TPM, before transferring control. The Linux Trusted Grub project has a patched boot loader that does this.

The TPM can later report these hash values, signed with an on-chip crypto key that can't be spoofed.

People would need to know what these values are supposed to be on an untampered system. Then there's no way to alter a system and let it boot such that the TPM gets the same values and so you could fool people about what you are running. For example, you could try patching Trusted Grub to send a fake OS hash to the TPM so you could lie about what OS version you were booting. But recall that the BIOS hashes the boot loader before executing it. If you've changed Trusted Grub, it will have a different hash, and this will be reported to the TPM. (Keep in mind that the TPM does not check these hashes for "correctness", it just remembers them and can report them later.)

So yes, you can patch your boot loader to lie, but you still won't end up with the same system configuration "fingerprint" because the boot loader got hashed and the TPM told about it before it got control. In this way the system achieves the ability to security report its boot sequence and configuration, even though people can change the software involved.

And how does the TPM chip know that the information that the bios/os loader/os itself sends to it is actually the real information about the bios/os loader/os and not faked information from an unmodified version?

This happens via a staged-boot process.

The BIOS itself in a TPM computer has a non-flashable portion that runs at startup. This takes a hash of the rest of the BIOS and feeds it into the TPM. The BIOS then hashes the boot loader into the TPM (Grub, in the case of Linux) before transferring control to it. The boot loader can hash the OS before switching control to that. The Trusted Grub project on Linux has been enhanced to do this. Then you could have the OS hash application data into the TPM as it loads; so far only a few experimental projects do that, like Enforcer.

Now, the BIOS is hard-wired to tell the truth. But the later components could lie. You could have a patched boot loader (it's open source!) which does not send truthful data about the OS kernel and configuration to the TPM.

But you couldn't get away with this fraud. Recall that the BIOS hashes the boot loader and sends the data to the TPM before running it. If you patch Trusted Grub to lie, it will have a different hash, and the TPM will be told about it. (Note, the TPM doesn't have any "expectations" regarding what these hashes are supposed to be, it just remembers what it was told and can report it later, signed with a crypto key.) So with this concept, if someone knows what the system "fingerprint" is supposed to be of a secure BIOS + Trusted Grub + Enforcer Linux boot sequence, it's impossible for you to patch it and end up with that same exact pattern.