Nonsense. All programs MAY have bugs. The likelyhood of bugs increases (very rapidly) with program complexity. Trivial programs have a very low probability of bugs, and when bug free can be proven so.
Irrelevant; *all* programs in daily use by normal users have reached so far, far up your complexity scale that my generalization is fair. You're talking about things like "helo world", and even then 1) you're running it through a compiler or vm that is very complex and may add bugs; 2) you're using a complex io-library that may contain bugs; 3) your output is somehow shown on screen through graphic drivers (and we know how buggy they use to be...)
ANY TIME you use any program, you risk a bug being triggered (either unintentional or, if you are unlucky, due some some sinister intent). After that, all bets are off -- a malware payload may do anything that you can do with your computer [...]
Ridiculous. Not all bugs can lead to execution of arbitrary code. Only those that do really get much publicity.
Except for "ridiculous" how does what you say go against anything I said? Reread: "risk being triggered" and "a malware payload may do anything". Also note; its not just "execution of arbitary code" that may be dangerous, even other bugs may still be expolitable in a potentially harmful way.
The *point* I'm trying to get through here is that you always run A RISK. That is, NO PROGRAM IS SAFE TO RUN operating on any data. You never know if the code you run contains expolitable bugs or not.
I've been telling people for years "no, you can't get a virus from things like a JPEG picture. You're fine."
Where do all these misconceptions come from? "It's safe to view images", "It's safe to read pdf:s", "It's safe to open text documents in emacs", etc. etc.
ALL programs have bugs. Each bug is potentially exploitable. ANY TIME you use any program, you risk a bug being triggered (either unintentional or, if you are unlucky, due some some sinister intent). After that, all bets are off -- a malware payload may do anything that you can do with your computer -- erase all you files, copy your private emails, replace your internet browser to snoop on your login to your internet bank, etc. etc. Even a bug triggered by mistake may run a recursive remove on all your files just by pure bad luck.
And then this... (from the top post): So all those times you told your parents/friends that looking at images was safe - well, not anymore."
"not anymore"? To be precise; it was NEVER safe. It still (after an update) isn't "safe". The bug has been lurking there basically forever. Who knows if some blackhat hacker have known this long before it became publically known today? Who are *sure* there are no other parsing bugs lurking in the code that will be reported next year? A better summing up would have been: "So all those times you told your parents/friends that looking at images was safe -- you were obviously driveling and had no idea what you were talking about".
On a related note, I have previously (http://slashdot.org/comments.pl?sid=95331&cid=817 0288) pushed for extending resource management to strengthen the protection against all kinds of virus, malware and catastrophic bugs.
I'm wondering how useful to us Sun's move really is if the code will not be put out under a GPL-like or BSD-like license
Well, your definition of "useful" seems a bit narrow to me. Everyone does not define useful as "being able to use it for advancing our position in the strategic game towards making GPL software taking over the world."
HOWEVER, open sourced "non free" code can be fantastically useful to people for other reasons. As in being able to fix problems yourself (or pay someone else to do it) if the vendor abandons the product; As being able to extend the functionality for a specific task that it was not intended for; etc, etc.
Windows XP is a stable operating system when it's running good quality, name brand software/hardware. At least the desktop is, no comment on server stuff.
The core operating system is "stable", but I believe the desktop is still surprisingly buggy. On my xp system I have folders linked from the desktop which I can rename files in, without the changes being shown until I use 'refresh'; It happens regulary during smb network browsing that shell windows lock up and I have to kill them; I regulary see xp pause for a moment and for no apparent reason rebuild the desktop icons; etc, etc. A few of these things might be driver issues, but I have too many friends reporting similar issues to belive its only driver problems.
You never 'start' an application. You always open (read:look at) a document by clicking on it.
I am not the original poster, but I have been thinking along similar lines.
What about when I sign on to AIM or IRC?
Then you open the document "my AIM session" which opens your AIM window with your preferences and configs. You could also have a 'my alter-ego AIM session' document to open AIM with a completely different configuration.
What about when I start emacs to read its built-in documentation?
You open an emacs document and pick 'help' from the context menu. If you do not yet have any emacs document you create one from the emacs-document-template.
What about when I start a performance tracker like ps3 or gkrellm?
Same thing as with AIM above.
What about when I start aterm to type a command?
You open a 'terminal' document. The terminal opens in precisely the same state as when you last 'closed' it. If you want a terminal with a fresh state, you create one from the aterem-templates.
Why can't I just exit an application when I know i'm done with it, freeing up system resources then instead of later?
In the system described, there is no difference between "hiding away" and "closing". As for system resources, they will be freed up if they are needed. If they are not needed, you have the benefit of cached access when you re-activate the document.
I'm pretty sure all the SB Live! and Audigy cards do real-time encoding.
You are wrong. There are NO consumer-level soundcards except for the nvidia nforce 2 chipset that do dolby digital *encoding* (= you can listen to suround sound games on your dolby digital hardware, such a receiver).
"Dolby digital" soundcards such as Audigy typically do dolby digital *decoding*, meaning that you can listen on dolby digital DVD:s without an external receiver. But you won't be able to use that to hook your computer up to your home cinema equipment to play games.
Third rule of passwords is:
DON'T use your pet's name as password.
or more elegantly expressed: "left to their own ways, some people will still use cute doggie names as passwords"// Grammp, F. T., and R. H. Morris. 'UNIX Operating System Security.'
Come on, the sort of people who run programs they download from strangers can't even find them after they've saved them to their disks because they're buried 6 directory levels deep in the "My Documents" directory. God, I hate that dumb-down "My Documents" crap...
Have you watched any such Windows users in action? My impression is that they, much like everyone else, have this autonomous reflex to click 'my desktop' + 'save' upon seeing any 'where to save this file' dialog...
There are three problems with your proposal: (1) it would be difficult/impossible to implement in a satisfactory manner
As you say, it is already somewhat available for Java VM applications. And, as I mentioned in my original comment, the resource barries I talk about are already available between different users in multi user systems.
(3) this would pose extra difficulties in production environments where code must run unattended or users must be handed quick-fixes in the form of batch files, etc.
Let say that code that must run unatended asks for the needed priviledges during installation. I would say 'yes' to grant my disk defragmenter access to any file on my disk at any time; I would not do that to the 'fancy-screen-saver.exe' I just got in the mail.
Quick-fixes deliverd as batch-files etc: These are based on your customer completely trusting you. You will anyway have explain how to run your "batch script" etc. I don't see the problem with including something like: "trust me, select 'yes' when the big scary dialog pops up giving you a huge warning about how much this looks like a virus."
(2) users would be very prone just to OK their way through the warning messages [much like they do when they launch the attachment in the first place];
Yes, this is the main obstacle. Still, I have the impression that the difference is to bring up *unexpected* dialogs that are *rare*, big and scary enough.
The regular email clients pops up warnings for each attached "word file" a user opens. When you have seen that dialog 10 times for legetime emails, its natural to have the 'ok'-click learnt as a reflex. But if a user opens his 'contents.zip' with something that looks like a txt-file inside; he expects notepad to start with the file, but if this time an unexpected red warning dialog states something along the lines of "This application tries to search through all files in your home directory and may do as it please with any information therein, is this okay with you?". Hopefully the fact that this warning is unexpected makes enough users stop and think. That would at least somewhat dampen the propagation of our usual emails worms...
What do you want to do - specify exactly what regions of memory and resources an executable can access when you run it?
There is already a memory protection barrier between running binaries on most OS:es, and yet you don't need to specify "regions of memory" for those to run in. I think most other resource barriers can be handled as transparently, but one has to discuss them on a case-to-case basis. The primary resource under discussion here (which should be clear from my previous comment) is the actual disk storage space for a user; so lets examine that in the following:
Let a binary access files in an own storage (like a home directory for the application, much like the common '~/.name/*' in unix). Now, an UI application that want to access ANY OTHER file should HAVE TO go through an OS API that brings up a STANDARDLIZED 'open file' dialog or similar.
So, let us now re-examine your grandma installing the legetime screensaver you sent her in an email. She double-clicks the screensaver and a dialog appears:
The following application:
coward-screen-saver-setup.exe Wants to edit one of your files:
~/.gnome/user-config That file have the following description:
"Configuration file for the behaviour and apperence of your workspace." The application gave the following explanation:
"This nifty screensaver wants to replace your old screensaver in your workspace configuration file." Is it okay for the application to edit the above named file? [Yes] [No] [Help]
I guess this interface can be somewhat more streamlized; but, I don't think dialogs like this would get in the way. Under normal opening of say "word files", applications will have to use the standalized 'open file' dialog; which also grants access to the file. Exceptional needs (like access to the whole disk for, say, a disk defragmenter) can be asked for and granted by the user during installation.
As many other has commented, the ability to click-and-run executables from the email clients is not the only reason for virus/worms spreading. Even with only 'save to disk' functionality, people will still run these binaries.
It is often said that what users fail to understand is that they should not run "untrusted binaries". But in my opinion this is the greatest shortcoming of all modern operating systems. I want my operating system to shield resources beloninging to one binary from another. Much in the same way it shields the actions of one user on a muliti user unix system from affecting another user. Why can't the same basic ideas be used when I run 'nice-screen-saver.exe' to NOT allow it access to 'email-addresses.txt'?!
Instead of having this functionality, I am told that the solution is to only run "trusted binaries"? But come on, it's not like I can personally audit all the code I run; and even if that would be possible it is easy to miss small bugs that eventually will run 'rm -rf' in my home directory. The point of this discussion is that NO binary should be ALLOWED to do 'rm -rf' in my home directory without me externally authorizing that operation. Exactly as I cannot read or delete user 'joe's files without his authorization.
I agree; people writing worm filters that bounce to forged addresses are as bad as their worm writing counterparts!
I mean, what happens when user 'joe' gets a couple of "WARNING: You sent me a virus" in their email? They come running to me "just to make sure", and I will have to explain for them how the email protocol works... AGAIN... sigh... for, what is it, the 10:th time that day.
Here is a hint to people writing these crappy anti-virus/worm filter: make sure you **ONLY** send a bounce IF the detected virus is on A **WHITELIST** for viruses that always send themselves WITHOUT A FORGED SENDER ADDRESS. If you send *any* other bounces, you are a part of the problem -- not the solution...
I could install Interix on my Windows NT box, and then create an account on the NT box with the same name as any solaris account. Then I could connect to any resources on network drives using the Interix tools and have full read/write/delete/etc. privledges on files with that person's permissions.
This sounds like the usual problem with the unexisting authentication in old NFS. Basically: you must configure your network to only allow trusted machines to mount NFS, because if they can mount NFS they also have access to all users files and it is completly up to the software on that machine to prevent a user from messing with other users files.
Now, if someone pulls the network cable of one of your computers and put their own laptop there instead it can be configured to mimic the removed computer -- ouch -- Or if someone hacks one of your computers, you can basically regard all you network files as toast... You know, this is why things like AFS and kerberos was invented...
I don't think the RAID algorithm is the right way to syncronize all your data, when applied on the larger scale. I imagine that what a person really want to do is to unify all his accounts, on slow and fast links all over the world, to look like a huge syncronized partition which stores the data throughout the accounts with sufficient redundancy (meaning something like 'keep copies of all data on at least three different locations). I think using RAID for this would give horrible performance and not be nearly flexible enough in how data is distributed through the different locations.
A new networked file system is needed. I am working on such a solution on my spare time (but it is still in the design phase).
The main idea is to unify cache and storage. This means that the least used files are deleted when an account is running out of storage, but under the constraint that a mimum number of copies of the files are kept online. (Hence, data will propagate to the nodes that actually use it). Upon a data request the filesystem goes out and fetch the data. Preferably in some P2P-like way where it is fetched simultaniously from all locations that has copies of that data.
If someone knows a solution that already works something like this, please tell me.
but the license forbids people from distributing patched releases...
Actually, I think the license is the major reason qmail is so secure. It avoids making qmail a patchwork of "features" coded by all these people who fail to perform even the most basic validation of user-data. Sure, in the mail servers with more open licenses, such bugs are eventually removed by the community; but that also means you really have to watch out for upcoming patches...
So, if you prioritize features and "all free open source"-styled license over not having to patch your mail server, there are other mail servers available providing this. Just don't complain over the license of qmail: it is one of the things making it what it is! (that is: a very secure mail server)
how do I get all mail going through my qmail system [...] to go through SpamAssassin, but with per-user settings - i.e. after the decision has been made on who to deliver the mail to - without losing the ability to use.qmail files?
Why not provide users with default.qmail files for a fully per-user opt-in/opt-out:able filtering soultion? We use a setup like this on our systems, and it should work on a completly unpatched qmail setup:
Every Treo communicator comes with Blazer(tm)--the award-winning wireless web browser
Good thing: If this trend with 'web-browser in your hand computer thingy' catches on, perhaps sites starts providing pages which I can surf without resorting to Netscape 4.5:s "future proof" mode (view source, cut, paste).
Why Netscape 4.5? In digital unix you kind of take what's offered to you...:-)
Yes; this is good - avoid bad releases
on
Simsville Canceled
·
· Score: 2, Interesting
I think it is really good news that someone has the nerve to pull the emergency break in this way.
I really don't understad the idea of pushing out unfinised games. I would guess it only hurts the company itself, and the poor people that buys the product. I mean: if you bought a full-price game that really felt "unfinished" I guess you'll *never* buy a game from that company again.
As an example: I picked up 'Global domination' from the low-price bin one day; thinking that it looked like a simplistic strategy game somewhat like 'risk'. Gee -- guess if I was fooled? The gameplay is basically to click as fast as you can (that sends missiles...) on other countries. It's just *that more fun* than to senslessly move around icons on the desktop! I will *never* buy a game from thoose people again...
Well; back to the point; this is a good move from maxis, and I hope other companies will follow their example.
I totally agree that innovation makes the greatest games.
I think there has only been a few truly original games, spawning whole eras of gaming. During an era all games looks the same and works the same. And almost anyone thinks that this is how games always have looked and worked. Needless to say; today we are within the FPS era. Based on the above I would say that, of the FPS games you mention, only Wolfenstein 3D was truly innovative.
Here are some other games I belive were truly innovative (but I might be entirely wrong, please give any examples of earlier games with the same ideas):
Super-Mario (the beginning of the plateau era, or was that lode-runner?)
Pitstop II (beginning of racing games)
Decathlon (beginning of sport games)
Interceptor (beginning of flight simulators as we know them today)
I would like to add the first shoot-em-up game here; but I have no idea what that was. I would like to say Uridium (wonderful game, that); but I don't truly belive that to be the first. Does anyone know?
I think another 'lightweight' aspect of twm (and it's close desendants), makes it extremely useful. You can either run twm or download the source and compile on almost all possible computer system with a valid X-environment! It will compile nicely and run without much hassle setting upp the tons of libraries required by Gnome/KDE and enlightenment.
Have you ever tried to compile Gnome/KDE or enlightenment on unprepared alphastations with digital unix, improperly installed sun stations or in hp-unix? Sure, eventually you get it to work, but with a properly configured.twmrc or.vtwmrc you have a working desktop in minutes (the twm defaults are in my opinion unusable)
If you work on many different systems (and they keep changing) and wish to have a common environment on them all; I think twm or vtwm is the best choise available.
Maybe the next step (after trashing the
uniformity of DNS) will be to free the IP-
numbers? Think of it; who wouldn't like 100.100.100.100 ?
Or what about 6.6.6.6 for a sex-related web
service? So why not start an alternative IP-number allocating service and start handing out
thoose goodies for some $$ ?
I had the same problem until I found that the FAQ at tivocommunity
actually points to another FAQ at www.tivo.com:
(From the faq:)
"TiVo is the most advanced and easy-to-use personal TV service available. But don't take our word for it. Our
approach has won thumbs up approval by reviewers from major business, technology, consumer electronics, and
home entertainment publications. TiVo digitally records television shows, without videotape, so you can watch
what you want, when you want. The TiVo service allows you to control your TV viewing in ways never before
possible-now you can pause, slow motion, or instant replay live TV. Or, with the simple click of a button, you can
tell TiVo to automatically find and record your favorite programs every time they air-so you won't ever miss a
moment of great TV again. You can even tell TiVo what you like and don't like, so it can suggest other shows you
may want to record and watch, based on your interests."
Slashdot Mirror
ALL programs have bugs.
Nonsense. All programs MAY have bugs. The likelyhood of bugs increases (very rapidly) with program complexity. Trivial programs have a very low probability of bugs, and when bug free can be proven so.
Irrelevant; *all* programs in daily use by normal users have reached so far, far up your complexity scale that my generalization is fair. You're talking about things like "helo world", and even then 1) you're running it through a compiler or vm that is very complex and may add bugs; 2) you're using a complex io-library that may contain bugs; 3) your output is somehow shown on screen through graphic drivers (and we know how buggy they use to be...)
ANY TIME you use any program, you risk a bug being triggered (either unintentional or, if you are unlucky, due some some sinister intent). After that, all bets are off -- a malware payload may do anything that you can do with your computer [...]
Ridiculous. Not all bugs can lead to execution of arbitrary code. Only those that do really get much publicity.
Except for "ridiculous" how does what you say go against anything I said? Reread: "risk being triggered" and "a malware payload may do anything". Also note; its not just "execution of arbitary code" that may be dangerous, even other bugs may still be expolitable in a potentially harmful way.
The *point* I'm trying to get through here is that you always run A RISK. That is, NO PROGRAM IS SAFE TO RUN operating on any data. You never know if the code you run contains expolitable bugs or not.
I've been telling people for years "no, you can't get a virus from things like a JPEG picture. You're fine."
7 0288)
Where do all these misconceptions come from? "It's safe to view images", "It's safe to read pdf:s", "It's safe to open text documents in emacs", etc. etc.
ALL programs have bugs. Each bug is potentially exploitable. ANY TIME you use any program, you risk a bug being triggered (either unintentional or, if you are unlucky, due some some sinister intent). After that, all bets are off -- a malware payload may do anything that you can do with your computer -- erase all you files, copy your private emails, replace your internet browser to snoop on your login to your internet bank, etc. etc. Even a bug triggered by mistake may run a recursive remove on all your files just by pure bad luck.
And then this... (from the top post):
So all those times you told your parents/friends that looking at images was safe - well, not anymore."
"not anymore"? To be precise; it was NEVER safe. It still (after an update) isn't "safe". The bug has been lurking there basically forever. Who knows if some blackhat hacker have known this long before it became publically known today? Who are *sure* there are no other parsing bugs lurking in the code that will be reported next year? A better summing up would have been:
"So all those times you told your parents/friends that looking at images was safe -- you were obviously driveling and had no idea what you were talking about".
On a related note, I have previously (http://slashdot.org/comments.pl?sid=95331&cid=81
pushed for extending resource management to strengthen the protection against all kinds of virus, malware and catastrophic bugs.
I'm wondering how useful to us Sun's move really is if the code will not be put out under a GPL-like or BSD-like license
Well, your definition of "useful" seems a bit narrow to me. Everyone does not define useful as "being able to use it for advancing our position in the strategic game towards making GPL software taking over the world."
HOWEVER, open sourced "non free" code can be fantastically useful to people for other reasons. As in being able to fix problems yourself (or pay someone else to do it) if the vendor abandons the product; As being able to extend the functionality for a specific task that it was not intended for; etc, etc.
Windows XP is a stable operating system when it's running good quality, name brand software/hardware. At least the desktop is, no comment on server stuff.
The core operating system is "stable", but I believe the desktop is still surprisingly buggy. On my xp system I have folders linked from the desktop which I can rename files in, without the changes being shown until I use 'refresh'; It happens regulary during smb network browsing that shell windows lock up and I have to kill them; I regulary see xp pause for a moment and for no apparent reason rebuild the desktop icons; etc, etc. A few of these things might be driver issues, but I have too many friends reporting similar issues to belive its only driver problems.
You never 'start' an application. You always open (read:look at) a document by clicking on it.
I am not the original poster, but I have been thinking along similar lines.
What about when I sign on to AIM or IRC?
Then you open the document "my AIM session" which opens your AIM window with your preferences and configs. You could also have a 'my alter-ego AIM session' document to open AIM with a completely different configuration.
What about when I start emacs to read its built-in documentation?
You open an emacs document and pick 'help' from the context menu. If you do not yet have any emacs document you create one from the emacs-document-template.
What about when I start a performance tracker like ps3 or gkrellm?
Same thing as with AIM above.
What about when I start aterm to type a command?
You open a 'terminal' document. The terminal opens in precisely the same state as when you last 'closed' it. If you want a terminal with a fresh state, you create one from the aterem-templates.
Why can't I just exit an application when I know i'm done with it, freeing up system resources then instead of later?
In the system described, there is no difference between "hiding away" and "closing". As for system resources, they will be freed up if they are needed. If they are not needed, you have the benefit of cached access when you re-activate the document.
I'm pretty sure all the SB Live! and Audigy cards do real-time encoding.
You are wrong. There are NO consumer-level soundcards except for the nvidia nforce 2 chipset that do dolby digital *encoding* (= you can listen to suround sound games on your dolby digital hardware, such a receiver).
"Dolby digital" soundcards such as Audigy typically do dolby digital *decoding*, meaning that you can listen on dolby digital DVD:s without an external receiver. But you won't be able to use that to hook your computer up to your home cinema equipment to play games.
>>> The really best way is RAID 1 + a third drive for backups, on another system.
>>At a different site.
> In a galaxy far, far away..
On a computer that is unplugged when not doing backups.
Third rule of passwords is:
// Grammp, F. T., and R. H. Morris. 'UNIX Operating System Security.'
DON'T use your pet's name as password.
or more elegantly expressed:
"left to their own ways, some people will still use cute doggie names as passwords"
Come on, the sort of people who run programs they download from strangers can't even find them after they've saved them to their disks because they're buried 6 directory levels deep in the "My Documents" directory. God, I hate that dumb-down "My Documents" crap...
Have you watched any such Windows users in action? My impression is that they, much like everyone else, have this autonomous reflex to click 'my desktop' + 'save' upon seeing any 'where to save this file' dialog...
There are three problems with your proposal: (1) it would be difficult/impossible to implement in a satisfactory manner
As you say, it is already somewhat available for Java VM applications. And, as I mentioned in my original comment, the resource barries I talk about are already available between different users in multi user systems.
(3) this would pose extra difficulties in production environments where code must run unattended or users must be handed quick-fixes in the form of batch files, etc.
Let say that code that must run unatended asks for the needed priviledges during installation. I would say 'yes' to grant my disk defragmenter access to any file on my disk at any time; I would not do that to the 'fancy-screen-saver.exe' I just got in the mail.
Quick-fixes deliverd as batch-files etc: These are based on your customer completely trusting you. You will anyway have explain how to run your "batch script" etc. I don't see the problem with including something like: "trust me, select 'yes' when the big scary dialog pops up giving you a huge warning about how much this looks like a virus."
(2) users would be very prone just to OK their way through the warning messages [much like they do when they launch the attachment in the first place];
Yes, this is the main obstacle. Still, I have the impression that the difference is to bring up *unexpected* dialogs that are *rare*, big and scary enough.
The regular email clients pops up warnings for each attached "word file" a user opens. When you have seen that dialog 10 times for legetime emails, its natural to have the 'ok'-click learnt as a reflex. But if a user opens his 'contents.zip' with something that looks like a txt-file inside; he expects notepad to start with the file, but if this time an unexpected red warning dialog states something along the lines of "This application tries to search through all files in your home directory and may do as it please with any information therein, is this okay with you?". Hopefully the fact that this warning is unexpected makes enough users stop and think. That would at least somewhat dampen the propagation of our usual emails worms...
There is already a memory protection barrier between running binaries on most OS:es, and yet you don't need to specify "regions of memory" for those to run in. I think most other resource barriers can be handled as transparently, but one has to discuss them on a case-to-case basis. The primary resource under discussion here (which should be clear from my previous comment) is the actual disk storage space for a user; so lets examine that in the following:
Let a binary access files in an own storage (like a home directory for the application, much like the common '~/.name/*' in unix). Now, an UI application that want to access ANY OTHER file should HAVE TO go through an OS API that brings up a STANDARDLIZED 'open file' dialog or similar.
So, let us now re-examine your grandma installing the legetime screensaver you sent her in an email.
She double-clicks the screensaver and a dialog appears:
I guess this interface can be somewhat more streamlized; but, I don't think dialogs like this would get in the way. Under normal opening of say "word files", applications will have to use the standalized 'open file' dialog; which also grants access to the file. Exceptional needs (like access to the whole disk for, say, a disk defragmenter) can be asked for and granted by the user during installation.
As many other has commented, the ability to click-and-run executables from the email clients is not the only reason for virus/worms spreading. Even with only 'save to disk' functionality, people will still run these binaries.
It is often said that what users fail to understand is that they should not run "untrusted binaries". But in my opinion this is the greatest shortcoming of all modern operating systems. I want my operating system to shield resources beloninging to one binary from another. Much in the same way it shields the actions of one user on a muliti user unix system from affecting another user. Why can't the same basic ideas be used when I run 'nice-screen-saver.exe' to NOT allow it access to 'email-addresses.txt'?!
Instead of having this functionality, I am told that the solution is to only run "trusted binaries"? But come on, it's not like I can personally audit all the code I run; and even if that would be possible it is easy to miss small bugs that eventually will run 'rm -rf' in my home directory. The point of this discussion is that NO binary should be ALLOWED to do 'rm -rf' in my home directory without me externally authorizing that operation. Exactly as I cannot read or delete user 'joe's files without his authorization.
I agree; people writing worm filters that bounce to forged addresses are as bad as their worm writing counterparts!
I mean, what happens when user 'joe' gets a couple of "WARNING: You sent me a virus" in their email? They come running to me "just to make sure", and I will have to explain for them how the email protocol works... AGAIN... sigh... for, what is it, the 10:th time that day.
Here is a hint to people writing these crappy anti-virus/worm filter: make sure you **ONLY** send a bounce IF the detected virus is on A **WHITELIST** for viruses that always send themselves WITHOUT A FORGED SENDER ADDRESS. If you send *any* other bounces, you are a part of the problem -- not the solution...
I could install Interix on my Windows NT box, and then create an account on the NT box with the same name as any solaris account. Then I could connect to any resources on network drives using the Interix tools and have full read/write/delete/etc. privledges on files with that person's permissions.
This sounds like the usual problem with the unexisting authentication in old NFS. Basically: you must configure your network to only allow trusted machines to mount NFS, because if they can mount NFS they also have access to all users files and it is completly up to the software on that machine to prevent a user from messing with other users files.
Now, if someone pulls the network cable of one of your computers and put their own laptop there instead it can be configured to mimic the removed computer -- ouch -- Or if someone hacks one of your computers, you can basically regard all you network files as toast... You know, this is why things like AFS and kerberos was invented...
I don't think the RAID algorithm is the right way to syncronize all your data, when applied on the larger scale. I imagine that what a person really want to do is to unify all his accounts, on slow and fast links all over the world, to look like a huge syncronized partition which stores the data throughout the accounts with sufficient redundancy (meaning something like 'keep copies of all data on at least three different locations). I think using RAID for this would give horrible performance and not be nearly flexible enough in how data is distributed through the different locations.
A new networked file system is needed. I am working on such a solution on my spare time (but it is still in the design phase).
The main idea is to unify cache and storage. This means that the least used files are deleted when an account is running out of storage, but under the constraint that a mimum number of copies of the files are kept online. (Hence, data will propagate to the nodes that actually use it). Upon a data request the filesystem goes out and fetch the data. Preferably in some P2P-like way where it is fetched simultaniously from all locations that has copies of that data.
If someone knows a solution that already works something like this, please tell me.
but the license forbids people from distributing patched releases...
.qmail files?
.qmail files for a fully per-user opt-in/opt-out:able filtering soultion? We use a setup like this on our systems, and it should work on a completly unpatched qmail setup:
.qmail:
./Maildir/
.qmail-spam: /dev/null;ISSPAM="$?"; if [ "$ISSPAM" = "1" ]; then exit 99; fi; exit 0;
./Maildir/.spam/
Actually, I think the license is the major reason qmail is so secure. It avoids making qmail a patchwork of "features" coded by all these people who fail to perform even the most basic validation of user-data. Sure, in the mail servers with more open licenses, such bugs are eventually removed by the community; but that also means you really have to watch out for upcoming patches...
So, if you prioritize features and "all free open source"-styled license over not having to patch your mail server, there are other mail servers available providing this. Just don't complain over the license of qmail: it is one of the things making it what it is! (that is: a very secure mail server)
how do I get all mail going through my qmail system [...] to go through SpamAssassin, but with per-user settings - i.e. after the decision has been made on who to deliver the mail to - without losing the ability to use
Why not provide users with default
| bash -c "/usr/bin/spamassassin -e | forward $LOCAL-spam;if [ \"\${PIPESTATUS[0]}\" != \"0\" ];then exit 99;else exit 0;fi"
| grep "X-Spam-Flag: YES" >
Now, if there is a recipe doing de-css in the program language 'Chef'...
... would cooking and serving a meal based on that recipe be illegal?
I think it would be hilarious to refer to a meal as "a collection of food carrying an illegal decryption algoritm"...
Every Treo communicator comes with Blazer(tm)--the award-winning wireless web browser
:-)
Good thing: If this trend with 'web-browser in your hand computer thingy' catches on, perhaps sites starts providing pages which I can surf without resorting to Netscape 4.5:s "future proof" mode (view source, cut, paste).
Why Netscape 4.5? In digital unix you kind of take what's offered to you...
I think it is really good news that someone has the nerve to pull the emergency break in this way.
I really don't understad the idea of pushing out unfinised games. I would guess it only hurts the company itself, and the poor people that buys the product. I mean: if you bought a full-price game that really felt "unfinished" I guess you'll *never* buy a game from that company again.
As an example: I picked up 'Global domination' from the low-price bin one day; thinking that it looked like a simplistic strategy game somewhat like 'risk'. Gee -- guess if I was fooled? The gameplay is basically to click as fast as you can (that sends missiles...) on other countries. It's just *that more fun* than to senslessly move around icons on the desktop! I will *never* buy a game from thoose people again...
Well; back to the point; this is a good move from maxis, and I hope other companies will follow their example.
I think there has only been a few truly original games, spawning whole eras of gaming. During an era all games looks the same and works the same. And almost anyone thinks that this is how games always have looked and worked. Needless to say; today we are within the FPS era. Based on the above I would say that, of the FPS games you mention, only Wolfenstein 3D was truly innovative. Here are some other games I belive were truly innovative (but I might be entirely wrong, please give any examples of earlier games with the same ideas):
Super-Mario (the beginning of the plateau era, or was that lode-runner?)
Pitstop II (beginning of racing games)
Decathlon (beginning of sport games)
Interceptor (beginning of flight simulators as we know them today)
I would like to add the first shoot-em-up game here; but I have no idea what that was. I would like to say Uridium (wonderful game, that); but I don't truly belive that to be the first. Does anyone know?
Have you ever tried to compile Gnome/KDE or enlightenment on unprepared alphastations with digital unix, improperly installed sun stations or in hp-unix? Sure, eventually you get it to work, but with a properly configured .twmrc or .vtwmrc you have a working desktop in minutes (the twm defaults are in my opinion unusable)
If you work on many different systems (and they keep changing) and wish to have a common environment on them all; I think twm or vtwm is the best choise available.
Oh - Collisions? - heh - oops! :-)
I had the same problem until I found that the FAQ at tivocommunity actually points to another FAQ at www.tivo.com:
(From the faq:)
"TiVo is the most advanced and easy-to-use personal TV service available. But don't take our word for it. Our approach has won thumbs up approval by reviewers from major business, technology, consumer electronics, and home entertainment publications. TiVo digitally records television shows, without videotape, so you can watch what you want, when you want. The TiVo service allows you to control your TV viewing in ways never before possible-now you can pause, slow motion, or instant replay live TV. Or, with the simple click of a button, you can tell TiVo to automatically find and record your favorite programs every time they air-so you won't ever miss a moment of great TV again. You can even tell TiVo what you like and don't like, so it can suggest other shows you may want to record and watch, based on your interests."