I've been suggesting this for a long time, and I usually only get blank stares. I have yet to find ONE good reason to maintain the "traditional" unix filesystem layout on a desktop machine (well, even server, but let's not go there).
The Unix tradition of splitting up applications by *type of content* instead of *application* is crazy. Thre are two bad reasons: 1) "Hey, I can throw every little binary in 'bin', go me!" 2) "Hey, I can throw every little library in 'lib', go me!". Parts of an application are hardly ever dealt with seperately. Does anybody install only the binaries of an app, and not, say, it's libraries?? or it's docs? No, these all belong to a cohesive unit that should be installed, uninstalled, moved, and run together.
As for #1, when your primary interface to the OS is a GUI desktop, having every piddly executable on your system in one directory doesn't really confer any benefit. As for #2, not all applicatinos need to use all other applications to begin with, and for those who do, why should those libraries not then be considered reusable common libraries, and then and only then, linked or put in a common place?
The system i'd propose would look something like this:
all applications have a structure like:
[appname]/[bin,lib,doc,conf]
All user applications live in:/apps/[appname]
You may choose to symlink the nested app dirs into/apps/[bin,lib,doc,conf] if you wish, like Stow does.
All "system" apps (e.g. stuff that is typically in/sbin) live in a mirror structure at:/sys/[appname]
Again, any utility binaries or common libs *may* be symlinked into the base/sys dir.
Application configuration would live with each app, no more throwing every fscking config file into the mud pit of/etc.
Things like 'man' would index *into* the seperate app dirs, not the other way around.
Oddly enough, I don't think the police would have much sympathy for anyone who's house got burgled like this...
On the contrary, if there were a strange figure that was rifling through my house and refused to identify himself, I would sure as hell hope that the police would concern themselves, despite the fact that my doormat says "Welcome"...
You are equating technology with law, and that is a very dangerous thing to do. That I have a technological means to commit a crime does not invalidate the fact that it is a crime. That this person puts copyrighted material on the web does not annul the copyright.
It seems like attaching a simple copyright license to the web site that prohibits this behavior would be a legitimate counter. (Cyveillance, you DID read the license on my site didn't you? Oh, you didn't? Here's your subpeona.)
Like I said, the performance wasn't all that bad anyway. Object-Relational Mapping does not (typically) solve a performance problem, it solves the impedence mismatch between Object Oriented Programming (the way most anything important is written these days) and relational databases and their declarative[ish] languages like SQL.
I also had the good fortune that the application I was writing had VERY little data, so the iteration was very fast, even when hitting the db each time, and was 99.999% read-only (and my final write solution was "blow away the cache entirely and don't even bother with integrity").
There are several Object-Relational mapping packages for Java out there. Hibernate is another one.
However, I have to say, that I have not found any complete O/R mapping package that implements everything transparently. In specific, I'm talking about recursive map structures, and "long transactions". Most of these packages are aimed towards short lived transactions on one object, or on a unidirectional tree of relationships.
I sunk absolutely OBSCENE amounts of time and effort into trying to make various packages work with a many-to-many self-recursive database structure only to realize it cannot currently be done with the packages out there. I would be glad to be proven wrong.
I would love to be able to keep a cache of a recursive map structure in memory (indefinately!) and have modifications automatically cascade the required updates to all nodes and revoke/expire any checked out nodes.
Finally I gave up trying to cache the actual structure in memory and now I just cache data, descend over the structure in the database for each request (it's not so bad, it turns out to be rather fast anyway).
Then again, I do NOT want arbitrary games to have complete superuser access to all my hardware. What an absolute nightmare. Some sort of "special" kernel mode that disables many things and gives utmost priority to one process might suffice. But then again, you can sort of already get this with a modular kernel, process nicing, and an intelligent schedular.
Developments are already underway for the sequel, "Return to the Gates of LFS". "LFS 3D" is planned in Q4 2004. It is rumored "LFS: Vice City" will follow.
Ok, I read the article but I still don't understand what Space Channel 5 is, or Chu Chu Rocket, or any of these games or what they have to do with each other. WTF is the "rhythm music" genre, or "rhythm action title"? Is this a GAME genre or what?
Re:What this is about
on
Linus on DRM
·
· Score: 1
If the "signature" is part of the kernel binary, then the user has rights under the GPL to obtain that signature along with source. If the signature is NOT part of the kernel binary, is it not an academic exercise to simply create your own kernel with accompanying, seperate signature?
Am I missing something here? I imagine the signature would include some sort of hash of the binary being protected, but if you have access to the kernel, can't you circumvent that anyway? (e.g. load a copy of the "real" kernel wherever the hardware wants to do the hash...)?
Re:Right tool for the job
on
Linus on DRM
·
· Score: 1
"When ideals get in the way of actually achieving your goals they are doing more harm than good for the cause."
Slashdot Mirror
Point.
But when they did offer such an OS, they did it right.
Here is a mirror:
Google to Create Blog Search Engine?
Hype or reality? Total hype. I mean why would people want to use software to say, "post their ideas" on a "shared forum". Totally ludicrous.
I've been suggesting this for a long time, and I usually only get blank stares. I have yet to find ONE good reason to maintain the "traditional" unix filesystem layout on a desktop machine (well, even server, but let's not go there).
/apps/[appname]
/apps/[bin,lib,doc,conf] if you wish, like Stow does.
/sbin) live in a mirror structure at: /sys/[appname]
/sys dir.
/etc.
The Unix tradition of splitting up applications by *type of content* instead of *application* is crazy. Thre are two bad reasons: 1) "Hey, I can throw every little binary in 'bin', go me!" 2) "Hey, I can throw every little library in 'lib', go me!". Parts of an application are hardly ever dealt with seperately. Does anybody install only the binaries of an app, and not, say, it's libraries?? or it's docs? No, these all belong to a cohesive unit that should be installed, uninstalled, moved, and run together.
As for #1, when your primary interface to the OS is a GUI desktop, having every piddly executable on your system in one directory doesn't really confer any benefit. As for #2, not all applicatinos need to use all other applications to begin with, and for those who do, why should those libraries not then be considered reusable common libraries, and then and only then, linked or put in a common place?
The system i'd propose would look something like this:
all applications have a structure like:
[appname]/[bin,lib,doc,conf]
All user applications live in:
You may choose to symlink the nested app dirs into
All "system" apps (e.g. stuff that is typically in
Again, any utility binaries or common libs *may* be symlinked into the base
Application configuration would live with each app, no more throwing every fscking config file into the mud pit of
Things like 'man' would index *into* the seperate app dirs, not the other way around.
"will the directory structure still be in english?"
...
'bin', 'var', 'usr', 'local', 'opt'
Is the *current* de facto standard even in any recognizable language?
...how about installing security cameras in the Congress and making citizens watch for "suspicious activity"...
oh wait, it's called CSPAN, and nobody watches...
You are equating technology with law, and that is a very dangerous thing to do. That I have a technological means to commit a crime does not invalidate the fact that it is a crime. That this person puts copyrighted material on the web does not annul the copyright.
It seems like attaching a simple copyright license to the web site that prohibits this behavior would be a legitimate counter. (Cyveillance, you DID read the license on my site didn't you? Oh, you didn't? Here's your subpeona.)
I dunno, maybe because software does so much more today? Are you really suggesting writing database-backed applications in assembler?
Like I said, the performance wasn't all that bad anyway. Object-Relational Mapping does not (typically) solve a performance problem, it solves the impedence mismatch between Object Oriented Programming (the way most anything important is written these days) and relational databases and their declarative[ish] languages like SQL.
I also had the good fortune that the application I was writing had VERY little data, so the iteration was very fast, even when hitting the db each time, and was 99.999% read-only (and my final write solution was "blow away the cache entirely and don't even bother with integrity").
There are several Object-Relational mapping packages for Java out there. Hibernate is another one.
However, I have to say, that I have not found any complete O/R mapping package that implements everything transparently. In specific, I'm talking about recursive map structures, and "long transactions". Most of these packages are aimed towards short lived transactions on one object, or on a unidirectional tree of relationships.
I sunk absolutely OBSCENE amounts of time and effort into trying to make various packages work with a many-to-many self-recursive database structure only to realize it cannot currently be done with the packages out there. I would be glad to be proven wrong.
I would love to be able to keep a cache of a recursive map structure in memory (indefinately!) and have modifications automatically cascade the required updates to all nodes and revoke/expire any checked out nodes.
Finally I gave up trying to cache the actual structure in memory and now I just cache data, descend over the structure in the database for each request (it's not so bad, it turns out to be rather fast anyway).
1) take over world
2) enslave humanity and use them as "energy cells"
3) wire each unit with IPv6
Also, presumably if someone steals your finger, you will be aware of the security breach rather quickly!
...get a bridge and position a guard to ask:
What is your favorite color?
Wow...welcome to 1993.
I beat all of you, my Whitespace implementation is below:
MS created a neat little game called Terrarium to showcase .NET. So they probably do envision a future for games on .NET.
Terrarium
Then again, I do NOT want arbitrary games to have complete superuser access to all my hardware. What an absolute nightmare. Some sort of "special" kernel mode that disables many things and gives utmost priority to one process might suffice. But then again, you can sort of already get this with a modular kernel, process nicing, and an intelligent schedular.
Developments are already underway for the sequel, "Return to the Gates of LFS". "LFS 3D" is planned in Q4 2004. It is rumored "LFS: Vice City" will follow.
Ok, I read the article but I still don't understand what Space Channel 5 is, or Chu Chu Rocket, or any of these games or what they have to do with each other. WTF is the "rhythm music" genre, or "rhythm action title"? Is this a GAME genre or what?
"Americans who travel to fight in illegal combat with other terrorists get no sympathy from me"
"5. slashdot.slashdot.org (figure this one out!)"
This is where sites referenced on slashdot can redirect users, correct?
...one large step for INVINCIBLE MICE
In Soviet Russia, Helicoptor drops YOU!
If the "signature" is part of the kernel binary, then the user has rights under the GPL to obtain that signature along with source. If the signature is NOT part of the kernel binary, is it not an academic exercise to simply create your own kernel with accompanying, seperate signature?
Am I missing something here? I imagine the signature would include some sort of hash of the binary being protected, but if you have access to the kernel, can't you circumvent that anyway? (e.g. load a copy of the "real" kernel wherever the hardware wants to do the hash...)?
"When ideals get in the way of actually achieving your goals they are doing more harm than good for the cause."
Are you sure you meant that?