V4L/V4L2 (Video 4 Linux, Video 4 Linux 2). Ask google about it.
See also, ALSA and OSS.
Regarding your Win32 bit, I dunno what the deal is on your machine. I only see win32 in conjunction with Mozilla source, Java, and various wines on my machine. That, and my distro doesn't know anything about 'mplayer'. Maybe this is a wine-wrapped Windows Media Player (which would definitely explain the reliance on DLLs). I use Xine, mpg123, and others.
heh. I bought a Cisco Aironet 352 PCMCIA card, too.:)
Seriously, though, it's better than a lot of manufacturers (although, Cisco, you still have a ways to go!). The annoying part, though, is that I keep getting several "weird status" notices, such when rebooting or unloading the drivers. (particularly annoying is the one I get on our wireless network here, which floods the virtual consoles with "got weird status 1000" (0x1000) messages. Luckily, I could edit the driver and make it stop annoying me (actually, I named 0x1000 EV_ANNOY:)
Unfortunately, nobody knew what status 0x1000 was, and Cisco wasn't about to tell me. Crappy "proprietary" information. Sheesh. Any cisco employees (or execs) want to "accidentally" leak the info, so that I can at least enjoy about as much use of the hardware I purchased as the Windows people get? Better yet, make it a company policy to completely open the specs! Sheesh.
While y'all are arguing about the GPL versus BSD versus other licenses, or the evillity of Its Billness, here's another, more important point to ponder:
This speech was given at a Microsoft convention for Governments. There, Your Representatives get hob-nobbed, pampered, as well as get their ears bent by Microsoft. That is, Microsoft gets a wonderful chance to come off as Great People® and gets to put some of their thoughts and opinions in the heads of Your Representatives.
So, when are we GPL, BSD, and Other Software Libre/Open Source Software people going to create our own conference for hobnobbing Your Representatives?
Indeed. Someone was hassling me on a newsgroup I read, and I just pointed out that Debian has over 6-7 thousand packages (note: not programs and libraries; there'd be vastly more) to support, so the failure rate is less than.3% so far this year. Compared with however they count MS vulnerabilities, which apparently didn't include the 8 new IIS ones. Sheesh. Silly people.
You have more programs, you increase your chance of any one program in the system being gronky. Including the 8-12 web servers, 15 or so FTP servers, etc. [numbers firmly pulled from my rear; anyone want to actually count?]
Oh, yeah. The guy hasn't responded to my post yet. Maybe shamed him into actually thinking about it?
m. so those coders use what? paper? oh, that would explain the 14 million lines of code (or whatever) for win2k.
No, I have a simpler and more probable explanation for the millions of lines of code, given said code's performance: Millions and millions of monkeys, each with a keyboard. That and some smattering of probability theory.
Just think of it: they only need a cage and a banana every once in a while! The actual human Microsoft employees merely try to compile the output of the monkeys' typing, and save what compiles and works halfway. Easy!
That's why you either do IMAP with pre-authentication over an SSH connection, simply require users to switch from telnet to ssh, or requring a secured IMAP connection.
Does Exchange do IMAP pre-authentication and secured IMAP connections? I know the University of Washington's IMAP server does.
Cheapest solution: keep your existing hardware and install Linux (or BSD, but I'm assuming Linux here) over your Windows stuff. The general process then follows:
Get a test machine to image from
On test machine:
Set up a base Linux distribution (I recommend Debian myself), with only the essentials installed, that is, with only X and the base system installed. No other servers or anything.
Lock it down, hard
Set up any graphics-intensive applications that should be run locally.
Lock it down, hard
Set up a thin-client updating system (is pretty easy actually; use ssh, shells, and possibly NFS) to push out client updates.
Lock it down, hard
Set up wrapper scripts on the Server machine to connect to the client side machine, and run the applications locally (basically, set DISPLAY properly and exec the application).
Lock it down, hard
Test, test, test. You should already be thinking of ways to secure the system. For instance, you should already have realized that your boot loader must be locked down (password + no delay before booting) and non-hard drive booting options disabled in BIOS.
Finish setting up and securing your server system.
Lock it down, hard
After sufficient testing (any new system must be thoroughly checked (but you probably already know that), deploy your install via a hard drive imaging tool, another, or a homebrew system (I installed a lab once with NFS and a boot zip disk).
Running analysis...
* Site is discussing punishment of Microsoft, due to monopolistic practices.
* URLs end in ".htm".
* The following text appears at the beginning of the page http://www.usdoj.gov/atr/cases/ms-major.htm:
<meta content="text/html; charset=windows-1252" http-equiv=Content-Type><!-- META instructs browser to insure latest version of this page is loaded: -->
<meta content="Mon, 01 Jan 1996 23:59:59 GMT" http-equiv=Expires>
<meta content="MSHTML 5.00.2014.210" name=GENERATOR></head>
Preliminary Analysis returns POSITIVE check for ironic content.
re-run ironyanalysis -d <level> to run a deeper irony analysis.
operator overloading can be critical to some applications. It's annoying to have to use my_complex_number.multiply(another_complex_number) instead of simply doing my_complex_number*another_complex_number. Same with vectors (although you need to make sure you know what * does, is it vector or scalar product), usw. Operator overloading, when done correctly, can vastly simplify programming, especially for scientific application.
Not bad. Not as good as it could be. The weak point in your scheme is that you're still compiling and installing the programs as the user you walk around as.
Thus, I present:
Trelane's Installation Scheme for the Fairly Paranoid
Have two additional accounts: compile user and install user. They are not allowed to log in except via su (maybe give them a bad shell and make their home directories minimal and mod 400/500) Install user must be a member of group {staff,wheel}, but compile-user should be exactly as your regular user, i.e. the same as every other ordinary user on the system, i.e., the least trusted user on the system.:)
su -c bash compileuser
compile package
make install prefix='some package directory'
exit
su -c bash installuser
chmod -R u-w $package-dir
chmod -R go=rX $package-dir
chown -R bin:bin $package-dir
install the package
exit
Naturally, there are still holes in this scheme, but I'll leave that to Trelane's Installation Scheme for the More Paranoid
Heh. If you see the string "Linux engineers are weenies" or "seineew era sreenigne xuniL" in it, then let slip the dogs of conspiracy theories.;)
Re:maybe the best server OS is not the best PDA OS
on
Linux PDA Part Deux
·
· Score: 2, Insightful
How is the Linux kernel dead last at usability? After all Linux only refers to the kernel. The "usability" you probably are referring to is the GUI that the user sees.
My friend, that is not Linux. Linux is the kernel. The fluff that you use to interact with the kernel is the operating environment, to borrow from Sun's terminology. It's the distribution that the user sees. While you can whine about the "usability" of various distros, it is really a matter of taste. I like the command line. Many do not. For me, Debian is the distribution. For a Windows user looking around, I'd recommend Red Hat or especially Mandrake, since they are very much GUI-oriented. Seeing the RH7 installer made me very happy, since I could finally show the Windows users something they could relate to. For some reason, block graphics turn people off faster than, well, some very fast things.
Linux is not just for servers. I've run it successfully on several desktops and my laptop. It's just what I could ask for. It can be point-and-drool if you want. It can be all command-line if you want. I'd say that it has better usability than some *ahem* other Operating Environments out there, since it gives you a choice and makes both of them just as good as the other. You're not herded into anything.
Anywho, the point is wake up. Linux isn't just for geeks anymore. Anyone who would tell you otherwise is probably trying to sell you something (e.g. XP)
[on the subject of making things work right]
Now go use them and stop whining about stuff you don't know is IE specific (*unlike* the Moz standard of adding "-moz-" in front of things to alert you to the fact they're not standard)
If your page doesn't validate, you're wrong. Fix it.
I read up on the other comments and knew that it was a troll. I just had an interesting idea [the posting in order to poison the data] and wanted to see what other people thought.
Actually, this does bring up an interesting question. I'll betcha Microsoft is out there, posting FUD under Anonymous Cowards and fake identities. Anyone have concrete evidence?
Why waste time working on conquering Software Libre on merits alone, when you can poison the community's message boards and frighten newbies looking for help. We [the majority, anyway] blindly assume that egg troll is actually a consultant, that he/she is posting in honesty, and even assume that he/she is actually a human.
Interesting attack on the Open Source/Free Software community. You just have to keep us from recognizing it.:)
Yes, there must be an investment in terms of staff to investigate and try out any new software (building is not a factor, because you can always download precompiled packages) This is a common factor in all systems administration. You have to decide what software to use, and how to use it. The difference between Open Source / Free Software is that the software is free, and you have the source. That is all (of the first-order effects).
If you're talking the cost of adding desired functionality to an existing project, then you're still wrong. It might be prohibitively expensive for one small business to add a specific bit of functionality, but they still have options. They can petition the community to build it for them, possibly making a contest out of it with some (relative to the business) cheap prize. They can group together with other like-minded businesses to work on the problem, or, which is also likely, staff experiencing the difficulties could work on it in their spare time, at no cost to the company. Finally, they can contribute a small amount to the community, and wait for others to contribute their small bits, until the feature is added (you seem to be foregetting that, for any one idea, there are lots of others in 6 billion people who will also find it a good idea).
Contrast this with the proprietary model where you can:
plead with the company to get the feature implemented or
plead with the company to get the feature implemented
The beauty of OSS/FS is that the cost of the development is shared amongst the community, with nobody taking a share as profit!.
Furthermore, your third-from-final statement is incomplete. It should read, "In contrast, the cost of accessibility is spread among all of a proprietary developer's customers, as long as someone convinces that developer that the feature needs to be implemented." So "niche" features are not more easily integrated, because they're niche features, and the very much finite developers must choose which features they're going to work for, given the very limited number of hours each of them can work. Say you work for a software company. You know your developers' time is limited, so which are you going to work on developing: features that 95% of the population will want, or features that 5% of the population will want?
Just because you want a feature and it's an OSS/FS project doesn't mean that you have to do all the work yourself. There are lots of developers out there who will aide you.
The scratch-an-itch argument you present is somewhat accurate, but only when shallowly examined.
The point is that anyonecan develop to scratch any itch. If Windows didn't develop to your tastes, you have to convince them that it's something you need and that they should develop it. If you can't convince them, you have no recourse, since you have no source code.
You seem to be forgetting that Open Source/Free Software is, at its core, simply laying your source code open to the public for review and enhancement. There is absolutely zip preventing a business with handicapped persons (or a collection of them, or a collection of Concerned Citizens, or whatever) from taking the source code, adding the desired functionality, and submitting patches for it. Whereas in the proprietary model, you have to convince a business that it's in their interest to implement feature X. In Open Source / Free Software, you always have the option of maintaining just patches or a completely separate tree.
Bottom line--Open Source / Software Libre does not fail here. Anyone can fix problems in the software and make the fixes available to the public. The problem thus far has been a lack of awareness in the active community, and the lack of wider acceptance of many Software Libre / Open Source projects. The more widely distributed OSS/FS projects, the larger the pool of developers.
If you wish, grab a copy of KDE source and go to town. Or write documentation. Or help the developers see what direction to develop if you don't want to write any code or documentation.
The kernel has to know where the root partition is. The information is stored in the kernel (and in lilo.conf). However, when you swap disks around, this can change. To get back in, you need to tell the kernel where the root parition is. To do this, either wait for the LILO prompt to come up, or pop in your rescue floppy, reboot, and type the following in at the LILO prompt:
boot linux root=newroot
[newroot is the new device holding your root partition, named via the standard Linux naming scheme.]
You should be able to get in then. See LILO documentation on all the nifty things you can do with it.
Basically, on a Debian system, programs are installed in packages that contain default configuration files, information files (man pages and info pages, for example), libraries needed, and programs. Now, you can trust the package or not, but we won't deal with that yet. Let's assume you trust the packages.
No information on your system is sent to Debian. How? Investigate the process, something you are unable to do in the WindowsUpdate world. The upgrade process as I understand it consists of three major steps: downloading the current package list, downloading the packages that are installed and have changed since the last update, and then installing the downloaded packages. You can find the files it downloads yourself if you go to, for instance http.us.debian.org/debian/dists/stable/main/binary -i386/Packages.gz
This file is grabbed for every source you list in/etc/apt/sources (I don't have my Debian system handy right now, so I can't verify this), and the results are combined into/etc/lib/dpkg/Available. The old Available file is renamed into Available.old. A quick diff is probably run (or something similar), and the packages that have changed will appear. Don't like my analysis? Think I'm full of it? Then you can get the apt and dpkg source and analyze it yourself. Once you've convinced yourself that it's kosher, then you can compile it and run it until the specs change or a bug becomes too annoying for you. Then you can download the source, analyze, and compile it again. The point is that you *have* the source and can analyze it. Oh, and it's legal for you to reverse-engineer it, too.:)
Apt and dselect understand a variety of protocols, including http and ftp for getting packages and package information. If you want, you can download the packages manually one by one and install them with a version of dpkg you've inspected yourself. Maybe even altered to make it better for your particular system.
Naturally, dpkg can ask you questions the package maintainer needs to know in order to get the new configurations in place. This can be automated via switches to apt (or dpkg too, I think). After initial installation, pretty much the questions are "install this new configuration, yes or no?" and you can safely say "no" if your config works fine.
The point is, you have the source and can check it out if you need. There are other security concerns (how can you trust your mirror?, for instance) but I can't address them now. I have to run.:)
Mmmm. Yes. And, besides that, the user has the ability to:
Select which SSL version to use
Select which ciphers in which strengths to use
Selectively block cookies based upon site address
BLOCK CERTAIN SITES [or just by default; much more configurable, though, read the link] FROM BEING ABLE TO LAUNCH POPUPS [or various other bits of JavaScript]
Encrypt saved passwords and have a master password for access to them
Save form data
Work with certificates and their validation
stay memory resident like IE and just launch new windows [this is why IE "loads" so quickly]
the ability to not be memory-resident [unlike IE]
and onward. There's more to it than that. Download it and check it out.
To MODERATORS: Please moderate up the person with whom I'm conversing. His or her commentary is quite relevant to the discussion at hand, and presents new information that could be quite educational to the slashdot reader. Also, since he or she is an AC, his or her comments are, by default, suppresed, and they are probably even more relevant than my personal revelations at a Score of +2:).
Back on track, then
Yeah. That "gravitationally bound" vs "spatially bound" was what I objected to.:)
Hrm. I definitely need to read up on general relativity.
I still wonder why we limit oourselves to the physical measuring system we have. Just because we can't measure it with a ruler doesn't make it any less relevant. In actuality, we seem to have two coordinate systems: one physical one (rulers), and one spatial one. The physical one stays constant to us; the spatial one expands with space. My basic question is how these two are related and interact. I still see a system of two particles held at a fixed distance via physical measuring devices as defining a physical measuring radius, and with it, a spatial "flux"; we just can't observe the flux. *sigh* I guess I just need to read up more. Dang. And I enjoy reading, too. What a loss.;)
On the side, I just now found out why "space" as an adjective is "spatial", (it is derived from the Latin spatium, according to Miriam-Webster), with which I until now disagreed.
If you want to continue the conversation, please email jap3003+redshift@ksu.edu, since I don't think the average slashdot reader will care about the topic, nor is communicating via slashdot very efficient.:) --
Slashdot Mirror
V4L/V4L2 (Video 4 Linux, Video 4 Linux 2). Ask google about it.
See also, ALSA and OSS.
Regarding your Win32 bit, I dunno what the deal is on your machine. I only see win32 in conjunction with Mozilla source, Java, and various wines on my machine. That, and my distro doesn't know anything about 'mplayer'. Maybe this is a wine-wrapped Windows Media Player (which would definitely explain the reliance on DLLs). I use Xine, mpg123, and others.
heh. I bought a Cisco Aironet 352 PCMCIA card, too. :)
:)
Seriously, though, it's better than a lot of manufacturers (although, Cisco, you still have a ways to go!). The annoying part, though, is that I keep getting several "weird status" notices, such when rebooting or unloading the drivers. (particularly annoying is the one I get on our wireless network here, which floods the virtual consoles with "got weird status 1000" (0x1000) messages. Luckily, I could edit the driver and make it stop annoying me (actually, I named 0x1000 EV_ANNOY
Unfortunately, nobody knew what status 0x1000 was, and Cisco wasn't about to tell me. Crappy "proprietary" information. Sheesh. Any cisco employees (or execs) want to "accidentally" leak the info, so that I can at least enjoy about as much use of the hardware I purchased as the Windows people get? Better yet, make it a company policy to completely open the specs! Sheesh.
While y'all are arguing about the GPL versus BSD versus other licenses, or the evillity of Its Billness, here's another, more important point to ponder:
This speech was given at a Microsoft convention for Governments. There, Your Representatives get hob-nobbed, pampered, as well as get their ears bent by Microsoft. That is, Microsoft gets a wonderful chance to come off as Great People® and gets to put some of their thoughts and opinions in the heads of Your Representatives.
So, when are we GPL, BSD, and Other Software Libre/Open Source Software people going to create our own conference for hobnobbing Your Representatives?
Indeed. Someone was hassling me on a newsgroup I read, and I just pointed out that Debian has over 6-7 thousand packages (note: not programs and libraries; there'd be vastly more) to support, so the failure rate is less than .3% so far this year. Compared with however they count MS vulnerabilities, which apparently didn't include the 8 new IIS ones. Sheesh. Silly people.
You have more programs, you increase your chance of any one program in the system being gronky. Including the 8-12 web servers, 15 or so FTP servers, etc. [numbers firmly pulled from my rear; anyone want to actually count?]
Oh, yeah. The guy hasn't responded to my post yet. Maybe shamed him into actually thinking about it?
No, I have a simpler and more probable explanation for the millions of lines of code, given said code's performance: Millions and millions of monkeys, each with a keyboard. That and some smattering of probability theory.
Just think of it: they only need a cage and a banana every once in a while! The actual human Microsoft employees merely try to compile the output of the monkeys' typing, and save what compiles and works halfway. Easy!
That's why you either do IMAP with pre-authentication over an SSH connection, simply require users to switch from telnet to ssh, or requring a secured IMAP connection.
Does Exchange do IMAP pre-authentication and secured IMAP connections? I know the University of Washington's IMAP server does.
Running analysis...
* Site is discussing punishment of Microsoft, due to monopolistic practices.
* URLs end in ".htm".
* The following text appears at the beginning of the page http://www.usdoj.gov/atr/cases/ms-major.htm:
<meta content="text/html; charset=windows-1252" http-equiv=Content-Type><!-- META instructs browser to insure latest version of this page is loaded: -->
<meta content="Mon, 01 Jan 1996 23:59:59 GMT" http-equiv=Expires>
<meta content="MSHTML 5.00.2014.210" name=GENERATOR></head>
Preliminary Analysis returns POSITIVE check for ironic content.
re-run ironyanalysis -d <level> to run a deeper irony analysis.
Provided it's used properly, I don't see the confusion.
operator overloading can be critical to some applications. It's annoying to have to use my_complex_number.multiply(another_complex_number) instead of simply doing my_complex_number*another_complex_number. Same with vectors (although you need to make sure you know what * does, is it vector or scalar product), usw. Operator overloading, when done correctly, can vastly simplify programming, especially for scientific application.
Thus, I present:
Trelane's Installation Scheme for the Fairly Paranoid
Naturally, there are still holes in this scheme, but I'll leave that to Trelane's Installation Scheme for the More Paranoid
Heh. If you see the string "Linux engineers are weenies" or "seineew era sreenigne xuniL" in it, then let slip the dogs of conspiracy theories. ;)
How is the Linux kernel dead last at usability? After all Linux only refers to the kernel. The "usability" you probably are referring to is the GUI that the user sees.
My friend, that is not Linux. Linux is the kernel. The fluff that you use to interact with the kernel is the operating environment, to borrow from Sun's terminology. It's the distribution that the user sees. While you can whine about the "usability" of various distros, it is really a matter of taste. I like the command line. Many do not. For me, Debian is the distribution. For a Windows user looking around, I'd recommend Red Hat or especially Mandrake, since they are very much GUI-oriented. Seeing the RH7 installer made me very happy, since I could finally show the Windows users something they could relate to. For some reason, block graphics turn people off faster than, well, some very fast things.
Linux is not just for servers. I've run it successfully on several desktops and my laptop. It's just what I could ask for. It can be point-and-drool if you want. It can be all command-line if you want. I'd say that it has better usability than some *ahem* other Operating Environments out there, since it gives you a choice and makes both of them just as good as the other. You're not herded into anything.
Anywho, the point is wake up. Linux isn't just for geeks anymore. Anyone who would tell you otherwise is probably trying to sell you something (e.g. XP)
CSS Validator
A few more Points for Moz:
[on the subject of making things work right]
Now go use them and stop whining about stuff you don't know is IE specific (*unlike* the Moz standard of adding "-moz-" in front of things to alert you to the fact they're not standard)
If your page doesn't validate, you're wrong. Fix it.
I read up on the other comments and knew that it was a troll. I just had an interesting idea [the posting in order to poison the data] and wanted to see what other people thought.
Actually, this does bring up an interesting question. I'll betcha Microsoft is out there, posting FUD under Anonymous Cowards and fake identities. Anyone have concrete evidence?
:)
Why waste time working on conquering Software Libre on merits alone, when you can poison the community's message boards and frighten newbies looking for help. We [the majority, anyway] blindly assume that egg troll is actually a consultant, that he/she is posting in honesty, and even assume that he/she is actually a human.
Interesting attack on the Open Source/Free Software community. You just have to keep us from recognizing it.
If you're talking the cost of adding desired functionality to an existing project, then you're still wrong. It might be prohibitively expensive for one small business to add a specific bit of functionality, but they still have options. They can petition the community to build it for them, possibly making a contest out of it with some (relative to the business) cheap prize. They can group together with other like-minded businesses to work on the problem, or, which is also likely, staff experiencing the difficulties could work on it in their spare time, at no cost to the company. Finally, they can contribute a small amount to the community, and wait for others to contribute their small bits, until the feature is added (you seem to be foregetting that, for any one idea, there are lots of others in 6 billion people who will also find it a good idea).
Contrast this with the proprietary model where you can:
or
The beauty of OSS/FS is that the cost of the development is shared amongst the community, with nobody taking a share as profit!.
Furthermore, your third-from-final statement is incomplete. It should read, "In contrast, the cost of accessibility is spread among all of a proprietary developer's customers, as long as someone convinces that developer that the feature needs to be implemented." So "niche" features are not more easily integrated, because they're niche features, and the very much finite developers must choose which features they're going to work for, given the very limited number of hours each of them can work. Say you work for a software company. You know your developers' time is limited, so which are you going to work on developing: features that 95% of the population will want, or features that 5% of the population will want?
Just because you want a feature and it's an OSS/FS project doesn't mean that you have to do all the work yourself. There are lots of developers out there who will aide you.
The scratch-an-itch argument you present is somewhat accurate, but only when shallowly examined.
The point is that anyone can develop to scratch any itch. If Windows didn't develop to your tastes, you have to convince them that it's something you need and that they should develop it. If you can't convince them, you have no recourse, since you have no source code.
You seem to be forgetting that Open Source/Free Software is, at its core, simply laying your source code open to the public for review and enhancement. There is absolutely zip preventing a business with handicapped persons (or a collection of them, or a collection of Concerned Citizens, or whatever) from taking the source code, adding the desired functionality, and submitting patches for it. Whereas in the proprietary model, you have to convince a business that it's in their interest to implement feature X. In Open Source / Free Software, you always have the option of maintaining just patches or a completely separate tree.
Bottom line--Open Source / Software Libre does not fail here. Anyone can fix problems in the software and make the fixes available to the public. The problem thus far has been a lack of awareness in the active community, and the lack of wider acceptance of many Software Libre / Open Source projects. The more widely distributed OSS/FS projects, the larger the pool of developers.
If you wish, grab a copy of KDE source and go to town. Or write documentation. Or help the developers see what direction to develop if you don't want to write any code or documentation.
The kernel has to know where the root partition is. The information is stored in the kernel (and in lilo.conf). However, when you swap disks around, this can change. To get back in, you need to tell the kernel where the root parition is. To do this, either wait for the LILO prompt to come up, or pop in your rescue floppy, reboot, and type the following in at the LILO prompt:
boot linux root=newroot
[newroot is the new device holding your root partition, named via the standard Linux naming scheme.]
You should be able to get in then. See LILO documentation on all the nifty things you can do with it.
Really? Interesting.
I guess, if it worked for IE, why not Office?
Make your stuff *appear* to load faster, even though the slow part is at the beginning and consumes RAM even when inactive. Whee!
Mmm, have you used apt, dselect, or dpkg?
y -i386/Packages.gz
/etc/apt/sources (I don't have my Debian system handy right now, so I can't verify this), and the results are combined into /etc/lib/dpkg/Available. The old Available file is renamed into Available.old. A quick diff is probably run (or something similar), and the packages that have changed will appear. Don't like my analysis? Think I'm full of it? Then you can get the apt and dpkg source and analyze it yourself. Once you've convinced yourself that it's kosher, then you can compile it and run it until the specs change or a bug becomes too annoying for you. Then you can download the source, analyze, and compile it again. The point is that you *have* the source and can analyze it. Oh, and it's legal for you to reverse-engineer it, too. :)
:)
Basically, on a Debian system, programs are installed in packages that contain default configuration files, information files (man pages and info pages, for example), libraries needed, and programs. Now, you can trust the package or not, but we won't deal with that yet. Let's assume you trust the packages.
No information on your system is sent to Debian. How? Investigate the process, something you are unable to do in the WindowsUpdate world. The upgrade process as I understand it consists of three major steps: downloading the current package list, downloading the packages that are installed and have changed since the last update, and then installing the downloaded packages. You can find the files it downloads yourself if you go to, for instance http.us.debian.org/debian/dists/stable/main/binar
This file is grabbed for every source you list in
Apt and dselect understand a variety of protocols, including http and ftp for getting packages and package information. If you want, you can download the packages manually one by one and install them with a version of dpkg you've inspected yourself. Maybe even altered to make it better for your particular system.
Naturally, dpkg can ask you questions the package maintainer needs to know in order to get the new configurations in place. This can be automated via switches to apt (or dpkg too, I think). After initial installation, pretty much the questions are "install this new configuration, yes or no?" and you can safely say "no" if your config works fine.
The point is, you have the source and can check it out if you need. There are other security concerns (how can you trust your mirror?, for instance) but I can't address them now. I have to run.
BLOCK CERTAIN SITES [or just by default; much more configurable, though, read the link] FROM BEING ABLE TO LAUNCH POPUPS [or various other bits of JavaScript]
Save form data
and onward. There's more to it than that. Download it and check it out.
- Permanent
- The C Programming Language, Kernighan & Ritchie
- Advanced Programming in the Unix Environment, Stevens
- All 3 volumes of TCP/IP Illustrated, Stevens
- Both volumes of Unix Network Programming by Stevens
- Stroustrup's C++ book (I forget what its name is exactly)
- All of the The Art of Programming series by Knuth (from hearsay; I haven't read any of them)
- Design and Implementation of the 4.4 BSD Operating System by McKusick, Bostic, Karels, and Quarterman
- Computer Networks by Tannenbaum
- Unix System Administration Handbook by Nemeth, Snyder, Seebass, and Hein
- Applied Cryptography, Schneier
- More Short-Term Lifetime:
- Programming Perl, Wall, Christiansen, and Schwartz
- Linux-Kernel Programmierung (is Addison-Wesley)
- Perl/Tk, Walsh
Hrm.. I seem to be running out of ideas.--
To MODERATORS: Please moderate up the person with whom I'm conversing. His or her commentary is quite relevant to the discussion at hand, and presents new information that could be quite educational to the slashdot reader. Also, since he or she is an AC, his or her comments are, by default, suppresed, and they are probably even more relevant than my personal revelations at a Score of +2 :).
Back on track, then
Yeah. That "gravitationally bound" vs "spatially bound" was what I objected to. :)
Hrm. I definitely need to read up on general relativity.
I still wonder why we limit oourselves to the physical measuring system we have. Just because we can't measure it with a ruler doesn't make it any less relevant. In actuality, we seem to have two coordinate systems: one physical one (rulers), and one spatial one. The physical one stays constant to us; the spatial one expands with space. My basic question is how these two are related and interact. I still see a system of two particles held at a fixed distance via physical measuring devices as defining a physical measuring radius, and with it, a spatial "flux"; we just can't observe the flux. *sigh* I guess I just need to read up more. Dang. And I enjoy reading, too. What a loss. ;)
On the side, I just now found out why "space" as an adjective is "spatial", (it is derived from the Latin spatium, according to Miriam-Webster), with which I until now disagreed.
If you want to continue the conversation, please email jap3003+redshift@ksu.edu, since I don't think the average slashdot reader will care about the topic, nor is communicating via slashdot very efficient. :)
--