"Adding a second disc drive to my Mac Pro took less than five minutes, including shutdown, opening the case, drive installation, rebooting, disk formatting and creating a file system, and adding the new disk volumes. That was simply amazing... they're setting a new standard for serviceability." -- -- Robert J. "Bob" Burke, from the article
That is impressive. Indeed, I don't believe it (more correctly, I could believe it, but I am shocked).
Yes, I am a "square" user. I put files on a RAID5 server, based on Redhat 9. To add a new drive, I need to shutdown, and I give several days notice. I open the case (seconds) install the drive (seconds) and then bring up the box (a couple of minutes). I then surface scan (a couple of hours). I then reboot the box again, (if this was a replacement drive) and inform users that the file system will running at reduced capacity for a day.
If its a new drive, I rebuild the RAID array after surface scanning is complete -- and I have already informed users that the array will not be available for a day.
Si, it takes at least 2 hours, with reduced functionality for a day, or just over a day.
(sarcasm on) This Mac stuff is impressive! (sarcasm off). Seriously, I get people spouting "Bob" style facts (it works for me!), and trying to convince me that I should give up ___ (fill in the blank) and move to that new-fangled good technology. Happens a lot, actually. Fortunately, they also like to have "the latest and greatest" which tends to avoid longer term issues.
Vista vs. OS X? Neither, right now. See me again in a couple of years, when I am upgrading some infrastructure, and we'll talk.
NFS client and server seem to be supported -- I need client.
NIS authentication does NOT seem to be supported, but setup to allow yypasswd and NIS map push to update Windows password seems to be supported. Also, change password on Windows can be set to push to the master NIS server. So that's ok -- I hope I don't have to install an "Active Directory" server, though.
Using directory maps from NIS; I just don't know (or couldn't find it). I'll dig into Unix Services to see. Since NFS mounting can be done by command line, I can generate the appropriate Windows mount args. Worst comes to worst, I can cook up a script using Unix Services to ypcat the maps, find differences, and generate the appropriate NFSADMIN commands to keep the mounts up to date.
The problem is that I have many mounts that are dynamic (not established on the server until actually used). Specifically, all my data CDs are stored this way. I would like to be able to defer the mount request until use -- My "CD Server" is currently set to only allow 8 concurrent dynamic mounts.
But the good news is that, with Unix Services, it looks like Windows may be able to reasonably play.
I develop for Windows (Currently working on driver to add features to USB keyboard). And Solaris. And Linux. And other platforms.
Most of my home office infrastructure is based on standard technologies. For signon, I use NIS. Automounter for publishing directories (along with NIS maps). NFS for file sharing. NTP for time services.
This works fine with all platforms I work with, with the exception of Windows. I can understand why this stuff isn't in Windows 98, but it is also not supported in XP SP2. XP does support NTP (thank heavens).SMB support provides Windows with access to file systems.
I heard a rumour that VISTA will support NFS "out of the box". Is this true? And my big question: is NIS supported by VISTA? If so, can VISTA use a password map from NIS with standard password hashes (or MD5 hashes)? Can VISTAs user home directory be mounted over NFS?
If I get a "yes" on the NFS and NIS, I would jump on VISTA immediately. It would fit into my development environment perfectly (when working on Microsoft projects, I use WINE to run CL.EXE and the rest of the MSVC7 toolchain integrated with standard "make" and GNU toolchain).
But then I am an extreme edge case. So I expect the answers to be (generally) "no".
Intel made some design errors in the original 386 architecture that made virtualization very difficult -- stuff like popf behaviour, and the design of its paging.
This caused emulator (virtulization) writers a lot of grief over the years (including me). But we muddle through with a combination of scanning/interpretation and native execution, eventually resulting in VMware. XEN didn't really count.
The chip architects apparently responded with changes that permitted virtualization. On the other hand, I didn't read up on the "latest and greatest". I just thought that they would learn from IBM virtualization (a lesson only 40ish years old), and implement an efficient self virtualization architecture that is able to again be self-virtualized. After all, we can contemplate an OS design that is purely a hypervisor; why not run that under another hypervisor? Isn't the point here not to worry about it?
And this simplifies the hypervisor and makes it more robust in a real sense. If things are implemented this way, there is never reason to support more than two guests under a single host. If you need more guests, simply run another copy of the hypervisor. This eliminates the need for most, if not all, dynamic memory allocation and most resource control, as well as the use of advanced data structures from the hypervisor. (And, I am not seriously suggesting 2, but a fixed limit can be imposed).
It would be useful (in the architecture) to specify that the hypervisor extension is not available, but this can be a function of the hypervisor itself (generate a trap instead, and disable the feature selection). In a sense, that was what I thought the OP problem was!
Unfortunately, I seem to be very much out of date here. I guess I have to study the Intel and AMD designs before buying/recommending processors for hypervisor use. And here I thought that this was a solved problem!
[and here I thought the fancy stuff would come from hypervisor->hypervisor virtualization punting.]
To summarize (since you don't like reading the long post): Don't insult people. Windows is good; mostly because lots of people use it. Linux is good because it has excellent driver support built-in and lots of useful applications out of the box.
Long stuff:
I get a call from a Windows user. The following must be done: read a "pst" file and extract emails. User: lawyer, *not* LUSER. Certainly has a reasonable attention span. Local IT help: useless (and this isn't forensic, thank the deity). Of course I bill for the phone conversation (she did). Indeed, I have a timer on my desk for these types of issues.
Typical kind of operational Window issue. Why does the law firm use Windows? Because their clients do. That is the network effect. The problems that ensue? Hey, I don't mind. Microsoft likes lock-in; doesn't matter to me -- I make money from it. No need to call the lawyer a LUSER -- what the fuck is a ".pst" file anyway?
I guess that with so many users, Windows is "easy to support", but there are motherboards which REQUIRE OEM XP installations. Go figure, I happen to have one. Load XP from CD, the CD disappars on first boot. XP cannot then load the "motherboard driver" which is needed to enable networking or CD. Figure that. Is your contention that an average user can slipstream the needed drivers into XP? Without access to a running XP? Just wondering. Exercise: compare and contrast to setting up a Linux MythTV box.
Windows is preset by manufactures to "just work". By itself? Its more of a crapshoot than Linux. Linux comes with a boatload more driver support "out-of-the-box". It much more probable that a new user would be able to set up a useful Linux (Ubuntu / Fedora Core) system EASIER than setting up a Windows XP system. (XP doesn't come with DVD playback either -- you actually have to BUY this feature)
In no particular order - online systems currently use text as the primary communications medium. Not mouse gestures and clicks. Windows used to come with a "macro recorder" -- I imagine that something similar could be cooked up, and the GUI events captured, canonicalized into a text format, added to a post, and fed back. Roughly, I imagine the sequence would look like:
MOUSE MOVE ABSOLUTE 15 10 LEFT CLICK MOUSE MOVE ABSUTE 50 100 LEFT CLICK..
and so on for a bit. Not that useful, when compared to cut and paste of mostly readable commands. More recent Windows versions have dropped the Recorder application, in favour of scripting, by the way, which is closer to the Linux (Unix) way.
After all, you seem to understand most of it, except "-ivh" (which stands for "install,verbose, hash", if it matters to you).
Cut, paste: which is what a GUI is good for. Symbolic, dense commands: which is what a CLI is good for. This example is a happy medium, not a negative example. The correct comparision is to obscure RegEdit keys; against that, -ivh wins. How to determine "-ivh"? By lore (as you have just done), reading the RPM guide, typing "rpm", looking on the web, etc.
As to livna. Generally, found by Google for the query: "how do i play dvd in fedora 6" and following the first link.
However, the end user (who is NOT a J. Random Luser, and generally has greater than 25 second attention span) should use a distribution with local support (someone around who can help) if they are not computer savy. This is one of the "network effects" and is CERTAINLY the case with Windows adoption. If this is the case, a simple personal question would do.
Generally, most Linux distributions are simple enough to use that a network query will work as well (as it did for this case).
My "Grandma" (Grandma to my kids) doesn't know that. All Grandma knows is that a functional computer that does email, light web browsing (looking for discount air fares), and webcam is $30. She also knows that a Windows XP license for that computer is $230. Office (for some light WP and bill payments) is hundreds more.
As a result, Grandma uses Windows 98 (which license CAME with the computer), or Linux. Pretty much the only application of interest is OpenOffice, and "MSN Messenger".
Grandma was very happy with Linux, until she wanted to run MSN Messenger with Webcam. She then put the old Windows 98 license back into use. And reloads the computer every 6 or so months. She is getting to be quite the tech. Number 1 computer complaint from Grandma? When will MSN Messenger with Webcam be out for Linux... Oh, and Grandma even uses Gmail, for ease of rebuilding computer.
[tech note - yes I know amsn, etc. but required is support for logitech webcam with only v4l2, not 1 support. So it doesn't quite work yet... maybe next update to amsn (and I don't have time to fix it personally, sorry Grandma).
Broadband may be new; indeed that isn't even relevant. The only effective means of getting "broadband" to my hourse, or neighbourhood is via buried cables (copper, or fiber -- pick one). Right now, "wireless" isn't a choice. (Although it could be, and we will revisit that point).
To provide a reliable service, easements are needed. Neither the phone company OR the cable company came to me to negotiate a right of way, or usage on my land. Instead, they went to the local government, which negotiated for the tax-payers. Part of the negotiation was for -- hold it -- network neutrality. AT&T argued it, and lost. Now it comes up again.
As to "wireless": we want RELIABLE access. This means that use of "open" 802.11a/b frequencies are out. At best "patchwork" best effort service can be offered. Plus, if a licensed operator complains, the 802.11a/b user will have to shut off. For limited use, yes, this may work. But for more general use? I like cables here, because I am running my office phone on VOIP.
I'll tell you what I think -- Sony supports "Blu-Ray" -- and they hosed a computer of mine (ok, a relatives) with a root-kit. Disney seems to support "Blu-Ray" -- just advertised all their movie-wares on Blu-Ray. I don't like their attitude toward Copyright Extension. (But I do like Squeak, go figure).
So, I lean to HD-DVD, just to pimp-alap Sony and Disney a bit. You know?
Do I care about the GB? As long as its lots, and (reasonably) reliable, no. Now, if they came up with a disc/changer that was affordable (say $500 US) that would burn a terrabyte without attention overnight. I would kiss them, no, I would get a sex-change and have their babies. Oh, that was a bit extreme... But I would definitely kiss them.
The current attack vectors on cryptographic based "DRM" schemes are (1) accidental key leakage, (2) the key exchange system or (3) the fact that the data must be eventually decoded.
Note that (3) is what makes DRM systems very dumb. It also follows that the Operating System must get involved in order to so hide the data.
If the Operating System allows a debugger to run AT THE SAME TIME as the "DRM", its attackable. If the OS allows "unsigned" drivers to run, its attackable.
The OS (for example, Vista) will (eventually) not allow unsigned drivers. It must also "kick out" or "suspend" all non-DRM (unsigned) software when DRM content is played.
This behaviour falls into Mr. Coxs patent.
Now, if (Vista) doesn't implement the scheme, it remains vulnerable. So, the problem must be solved another way.
My suggestion then is to ALSO patent (or disallow) by widely publishing the idea that a hypervisor (VM supervisor) can be used for DRM control as well, and can also be used to suspend, terminate or otherwise control applications that could be used to attack DRM software.
May I quote "Both Apple and Microsoft are selling closed and proprietary systems in their consumer electronics products. However, Apple typically sells unlocked hardware but manages its hardware and software as an "integrated experience." Microsoft seeks to lock down its hardware, but leaves third party applications largely unmanaged"
This bombshell dropped by the author really floored me. (and this was about Apple, OS X and the new phone)
How about his excuse as to why Apple doesn't support EVDO? He asks "which one" and also wonders why Apple would compete with itself. Say WHAT?!?
More quality opinions coming, I am sure. (that was sarcastic).
Sorry, I can't make out what you are trying to say. "Cooperated"?
Sure SUN did. What processor was the Standard (published)? Yes, that would have been SUNs. Why didn't the "Asia hardware makers" build something? Right, because Intel was more cooperative and gave proper designs -- not.
Because Microsoft was willing to give source code to those devs -- not.
Because Microsoft is smart -- ok, I'll grant that one. And, because of WinHEC... Why would you think such a thing is good? Given a choice between reasonable documentation, and souce availability, vs. WinHEC, most devs would take the former.
Now, SUN is giving you access to binaries where they are mailed to you without charge, and downloadable source code, and access to the standard dev kit, and you are saying... what; the people who take SUN up on this offer have less brains, and will have less fun (quoted from your message),
Are you serious? Can't you give up on the fanboi attitude for just a wee bit?
Find a lawyer in the right jurisdiction. Hopefully the US. Basically, you are after treble damages on songwriter royalties. For something THIS close, other royalties as well (it was sampled).
It will run on a few years -- stick with it. This is a good hunk of your retirement we are talking about here (remeber, treble?, and hopefully with someone selling millions of albums). You also need an accountant along with the lawyer.
Note that the songwriter typically takes 50% of the publishers share in co-publishing arrangements. But, you would not have signed up to that, so take it all. At three times, you can now argue for 7- to 100% of the publishers take on the album.
In my jurisdiction, we pay a media tariff. As a benefit we have a personal copying priviledge.
Now, it would have been "illegal" to copy music for friends before the personal copying provision was put into place. Currently, this provision does not extend to videos or audio books (for two examples).
Let's say that a personal copying provision is put into place for videos. How does DRM get retracted now that it is legal to copy? At that point, it is still either not possible or very difficult to copy the material.
An example closer to home -- I just tried to copy a CD that was "copy protected". Wasn't possible (with the software in place). I guess the "DRM" worked, right? But it was perfectly my right to do what I tried to do. Interesting that the CD had purchased after the personal copy provision was made law.
An explanation for those who don't want to read the patent:
An anlog joystick is pretty much a variable resistor. In order to convert this into a digital domain, the resistance must be measured. We know that the resistor itself can be used to change the discharge time of a capacitor -- this is common. Meauring the discharge time means determining the time the capacitor goes from one voltage to a lower voltage. Which, by its very nature is a "pulse".
And this patent seems to cover all such interfaces in the "joystick" domain.
Now, I can come up with alternate methods for reading the resistance. First one (off the top of my head), is to use a series of resistors controlled by a latched value to produce a reference voltage which is then sent through the joystick resistance. The final output is run through a gate which triggers at a reference voltage. Via search, we can determine the target joystick resistance. (I would probably use a binary search). But this is not any where NEAR as simple or obvious as the R-C approach.
Now, the R-C approach has been used for other variable resistors (prior to 1998) -- the "joystick" application is the only new thing. I used it myself in the 80's. Just never for a joystick (not being into gaming). Its main benefits are that it needs only a single input pin and the circuit is simple. But, given those constraints it is obvious.
Oh well -- go patent trolls. Sure glad I am not in the US.
I had a "simple" optimization project. It came down to one critical function (ISO JBIG compression). I coded the thing by hand in assembler, carefully manually scheduling instructions. It took me days. Managed to beat GNU gcc 2 and 3 by a reasonable margin. The latest Microsoft C compiler? Blew me away. I looked at the assembler it produced -- and I don't get where the gain is coming from. The compiler understands the machine better than I do.
Go figure -- I hung up my assembler badge. Still a useful skill for looking at core dumps, though. And for dealing with micro-controllers.
So, have you had at it and benchmarked your assembler vs. a compilers?
10,000 seems like a lot of money. Its not. I expect a driver dev to get $70+ per hour, this pays for 143 hours.
You are not going to get a driver in that amount of time.
But, I will give you clues. The nVidia chip is pretty high on the OpenGL stack. The chip itself handles most OpenGL primitive operations. It just won't do contexts (nor will the ATI). I don't know the underlying protocol to communicate with the chip, but I would guess it is packet based. Registers would prove far too slow. I would imagine that for OpenGL, VGA, video, and mode support you are looking at almost a thousand "registers" or eqivalents.
It may be possible to catch the kernel level packet interfaces -- mode setting and VGA extension should be reversable via emulation. But this won't tell you what any of the commands do. You could try iterating OpenGL and comparing generated packets... but...
Modern chips typically DON'T implement a fixed-function pipeline. So you will have to figure out how OpenGL shader compiler for the chip works (because you have to know the "machine code").
Good luck for a 4 week driver project. The shader compiler itself is almost a C++ compiler which has to be reversed, the communications format and the packet streams. I would give 10 man-years as a first estimate.
Or, you could try to get the vendors to "be nice".
"When did people start equating rudimentary copy protection with Digital Rights Management?
The term has lost all meaning."
DRM is not Copy Protection. It is Copy Prevention under defined circumstances (most of them)*. Copy Protection is Copy Prevention under all circumstances. Rudimentary Copy Protection is -- my guess -- Copy Protection that doesn't work.
Anyway, that all means that "Copy Protection == DRM for original media".
* DRM defines a series of "rights" you have to the content. Each application of a "digital right" begins with the allowing of a copy, or degraded copy. Or not allowing the copy (no access). So, the "rights" part of DRM can by replaced by "copy". Which is fundamentally what DRM controls -- Copyright. Further, "Copy Protection" (prevention) controls the same right. Except that instead of using fancy crypto, it is supposed to verify the original media only. Same job. The reason that "DRM" is separated from "Copy Protection" is that "Copy Protection" as defined here cannot be applied unless original media is available. Consider application to downloaded content.
"I want to do more than just "play" the DVD, I want an option to "rip" the DVD and store it. But you can't bundle that right now (legal issues). I want to play music, rip music, download music and podcasts and connect to subscription services all in one. But this stuff is still independent from the services that play movies.
And for the second generation, I want to hook up a second PC in the basement and have it talk to the first PC upstairs. And then I want these guys to share a media library. I want multiple output option so that I can stream music to different rooms via the same remote. But this is still in the dreamer and prototype stage."
Yes, call it "prototype" stage. It is unlikely that you will get what you want from AMD, Intel or Apple, though. They have to maintain relationships with the "content" companies.
As a prototype, MythTV does what you consider to be the first and second generation level stuff (play, rip and share media library). It doesn't do the "connect to subscription service" thing, nor the "same remote" thing. The same remote bit could be added (but would take quite a bit of fiddling).
Now, MythTV is not yet all there -- it definitely qualifies as a "prototype". But people who have seen my rig are interested (most thought that what it does is "impossible" -- especially playing back mpeg2 video via wireless on a PII 400 client without problem).
All told, I view this as a business opportunity -- I can undercut the VVendor! (sic) by $1000 and provide systems with more functionality. Its a "mom and pop" business, but there is certainly money to be made.
I am counting.... 5 wall-warts, 11 plugs, and one "in-line" wart-on-the-floor. And a UPS on the floor (only two plugs, no warts). Three power bars, and two separate wall outlets.
And that's just in my office space.
I am sick of this (unsightly) rigging. Yes, I will tolerate a higher power bill to eliminate it.
You are correct. The value in Sealand is being able to create embassies of Sealand. It should be easy to find nations willing to host Sealand embassies: if only to tweak the noses of the "first world".
Sealand in itself is too dependent on the UK government. But really, the "data centre" or other services can easily be provided at the embassy level.
The remaining problem is that this approach is expensive -- the embassy hosting nations require investment in infrastructure. Or Sealand would have to own its own satellite...
Archival meaning -- read-only. Multiple OS support meaning -- standard.
This cuts the field down. ISO 9660 would be a good bet, but is a bit "overkill". TAR format (which can be viewed as a "primitive" filesystem) would be my choice. Simple, can be read on all your target systems. If a tar client is not (for whatever reason) easily available, the data can still be simply extracted.
Bad point: the "directory" can only be obtained by scanning the entire byte stream. If that is tolerable (and, by indexing the files stored, is mostly just fine), its the one.
If you need fast directory searching as well, consider ISO 9660. Again, clients to read the format are available (although they may be limited to 700MB at a time).
"tar files" (a bit of a misnomer -- should be "tar format byte stream") can be recorded on any device - floppy, disc, tape, CD, DVD, USB stick. It may even provide amusement -- if someone trys to read a recordable CD with a tar image on it, rather than an ISO image:)
Insert is then no longer a "mode". It is simply a command, terminated by ESC. In TECO, all commands are terminated by ESC (except for a limited number of "immediate" keys). In VI, i is terminated by ESC.
And, with this definition, VI is no longer modal. Also, some commands now make sense. To generate the banner box: 80i*ESC makes sense: it is, after all, a command.
Slashdot Mirror
"Adding a second disc drive to my Mac Pro took less than five minutes, including shutdown, opening the case, drive installation, rebooting, disk formatting and creating a file system, and adding the new disk volumes. That was simply amazing... they're setting a new standard for serviceability." -- -- Robert J. "Bob" Burke, from the article
That is impressive. Indeed, I don't believe it (more correctly, I could believe it, but I am shocked).
Yes, I am a "square" user. I put files on a RAID5 server, based on Redhat 9. To add a new drive, I need to shutdown, and I give several days notice. I open the case (seconds) install the drive (seconds) and then bring up the box (a couple of minutes). I then surface scan (a couple of hours). I then reboot the box again, (if this was a replacement drive) and inform users that the file system will running at reduced capacity for a day.
If its a new drive, I rebuild the RAID array after surface scanning is complete -- and I have already informed users that the array will not be available for a day.
Si, it takes at least 2 hours, with reduced functionality for a day, or just over a day.
(sarcasm on) This Mac stuff is impressive! (sarcasm off). Seriously, I get people spouting "Bob" style facts (it works for me!), and trying to convince me that I should give up ___ (fill in the blank) and move to that new-fangled good technology. Happens a lot, actually. Fortunately, they also like to have "the latest and greatest" which tends to avoid longer term issues.
Vista vs. OS X? Neither, right now. See me again in a couple of years, when I am upgrading some infrastructure, and we'll talk.
Bob
NFS client and server seem to be supported -- I need client.
NIS authentication does NOT seem to be supported, but setup to allow yypasswd and NIS map push to update Windows password seems to be supported. Also, change password on Windows can be set to push to the master NIS server. So that's ok -- I hope I don't have to install an "Active Directory" server, though.
Using directory maps from NIS; I just don't know (or couldn't find it). I'll dig into Unix Services to see. Since NFS mounting can be done by command line, I can generate the appropriate Windows mount args. Worst comes to worst, I can cook up a script using Unix Services to ypcat the maps, find differences, and generate the appropriate NFSADMIN commands to keep the mounts up to date.
The problem is that I have many mounts that are dynamic (not established on the server until actually used). Specifically, all my data CDs are stored this way. I would like to be able to defer the mount request until use -- My "CD Server" is currently set to only allow 8 concurrent dynamic mounts.
But the good news is that, with Unix Services, it looks like Windows may be able to reasonably play.
Thanks for your help
So, thanks
I develop for Windows (Currently working on driver to add features to USB keyboard). And Solaris. And Linux. And other platforms.
Most of my home office infrastructure is based on standard technologies. For signon, I use NIS. Automounter for publishing directories (along with NIS maps). NFS for file sharing. NTP for time services.
This works fine with all platforms I work with, with the exception of Windows. I can understand why this stuff isn't in Windows 98, but it is also not supported in XP SP2. XP does support NTP (thank heavens).SMB support provides Windows with access to file systems.
I heard a rumour that VISTA will support NFS "out of the box". Is this true? And my big question: is NIS supported by VISTA? If so, can VISTA use a password map from NIS with standard password hashes (or MD5 hashes)? Can VISTAs user home directory be mounted over NFS?
If I get a "yes" on the NFS and NIS, I would jump on VISTA immediately. It would fit into my development environment perfectly (when working on Microsoft projects, I use WINE to run CL.EXE and the rest of the MSVC7 toolchain integrated with standard "make" and GNU toolchain).
But then I am an extreme edge case. So I expect the answers to be (generally) "no".
Intel made some design errors in the original 386 architecture that made virtualization very difficult -- stuff like popf behaviour, and the design of its paging.
This caused emulator (virtulization) writers a lot of grief over the years (including me). But we muddle through with a combination of scanning/interpretation and native execution, eventually resulting in VMware. XEN didn't really count.
The chip architects apparently responded with changes that permitted virtualization. On the other hand, I didn't read up on the "latest and greatest". I just thought that they would learn from IBM virtualization (a lesson only 40ish years old), and implement an efficient self virtualization architecture that is able to again be self-virtualized. After all, we can contemplate an OS design that is purely a hypervisor; why not run that under another hypervisor? Isn't the point here not to worry about it?
And this simplifies the hypervisor and makes it more robust in a real sense. If things are implemented this way, there is never reason to support more than two guests under a single host. If you need more guests, simply run another copy of the hypervisor. This eliminates the need for most, if not all, dynamic memory allocation and most resource control, as well as the use of advanced data structures from the hypervisor. (And, I am not seriously suggesting 2, but a fixed limit can be imposed).
It would be useful (in the architecture) to specify that the hypervisor extension is not available, but this can be a function of the hypervisor itself (generate a trap instead, and disable the feature selection). In a sense, that was what I thought the OP problem was!
Unfortunately, I seem to be very much out of date here. I guess I have to study the Intel and AMD designs before buying/recommending processors for hypervisor use. And here I thought that this was a solved problem!
[and here I thought the fancy stuff would come from hypervisor->hypervisor virtualization punting.]
Your thoughts?
To summarize (since you don't like reading the long post): Don't insult people. Windows is good; mostly because lots of people use it. Linux is good because it has excellent driver support built-in and lots of useful applications out of the box.
Long stuff:
I get a call from a Windows user. The following must be done: read a "pst" file and extract emails. User: lawyer, *not* LUSER. Certainly has a reasonable attention span. Local IT help: useless (and this isn't forensic, thank the deity). Of course I bill for the phone conversation (she did). Indeed, I have a timer on my desk for these types of issues.
Typical kind of operational Window issue. Why does the law firm use Windows? Because their clients do. That is the network effect. The problems that ensue? Hey, I don't mind. Microsoft likes lock-in; doesn't matter to me -- I make money from it. No need to call the lawyer a LUSER -- what the fuck is a ".pst" file anyway?
I guess that with so many users, Windows is "easy to support", but there are motherboards which REQUIRE OEM XP installations. Go figure, I happen to have one. Load XP from CD, the CD disappars on first boot. XP cannot then load the "motherboard driver" which is needed to enable networking or CD. Figure that. Is your contention that an average user can slipstream the needed drivers into XP? Without access to a running XP? Just wondering. Exercise: compare and contrast to setting up a Linux MythTV box.
Windows is preset by manufactures to "just work". By itself? Its more of a crapshoot than Linux. Linux comes with a boatload more driver support "out-of-the-box". It much more probable that a new user would be able to set up a useful Linux (Ubuntu / Fedora Core) system EASIER than setting up a Windows XP system. (XP doesn't come with DVD playback either -- you actually have to BUY this feature)
Wow!
..
In no particular order - online systems currently use text as the primary communications medium. Not mouse gestures and clicks. Windows used to come with a "macro recorder" -- I imagine that something similar could be cooked up, and the GUI events captured, canonicalized into a text format, added to a post, and fed back. Roughly, I imagine the sequence would look like:
MOUSE MOVE ABSOLUTE 15 10
LEFT CLICK
MOUSE MOVE ABSUTE 50 100
LEFT CLICK
and so on for a bit. Not that useful, when compared to cut and paste of mostly readable commands. More recent Windows versions have dropped the Recorder application, in favour of scripting, by the way, which is closer to the Linux (Unix) way.
After all, you seem to understand most of it, except "-ivh" (which stands for "install,verbose, hash", if it matters to you).
Cut, paste: which is what a GUI is good for. Symbolic, dense commands: which is what a CLI is good for. This example is a happy medium, not a negative example. The correct comparision is to obscure RegEdit keys; against that, -ivh wins. How to determine "-ivh"? By lore (as you have just done), reading the RPM guide, typing "rpm", looking on the web, etc.
As to livna. Generally, found by Google for the query: "how do i play dvd in fedora 6" and following the first link.
However, the end user (who is NOT a J. Random Luser, and generally has greater than 25 second attention span) should use a distribution with local support (someone around who can help) if they are not computer savy. This is one of the "network effects" and is CERTAINLY the case with Windows adoption. If this is the case, a simple personal question would do.
Generally, most Linux distributions are simple enough to use that a network query will work as well (as it did for this case).
My "Grandma" (Grandma to my kids) doesn't know that. All Grandma knows is that a functional computer that does email, light web browsing (looking for discount air fares), and webcam is $30. She also knows that a Windows XP license for that computer is $230. Office (for some light WP and bill payments) is hundreds more.
As a result, Grandma uses Windows 98 (which license CAME with the computer), or Linux. Pretty much the only application of interest is OpenOffice, and "MSN Messenger".
Grandma was very happy with Linux, until she wanted to run MSN Messenger with Webcam. She then put the old Windows 98 license back into use. And reloads the computer every 6 or so months. She is getting to be quite the tech. Number 1 computer complaint from Grandma? When will MSN Messenger with Webcam be out for Linux... Oh, and Grandma even uses Gmail, for ease of rebuilding computer.
[tech note - yes I know amsn, etc. but required is support for logitech webcam with only v4l2, not 1 support. So it doesn't quite work yet... maybe next update to amsn (and I don't have time to fix it personally, sorry Grandma).
I feel like a broken record. But...
Broadband may be new; indeed that isn't even relevant. The only effective means of getting "broadband" to my hourse, or neighbourhood is via buried cables (copper, or fiber -- pick one). Right now, "wireless" isn't a choice. (Although it could be, and we will revisit that point).
To provide a reliable service, easements are needed. Neither the phone company OR the cable company came to me to negotiate a right of way, or usage on my land. Instead, they went to the local government, which negotiated for the tax-payers. Part of the negotiation was for -- hold it -- network neutrality. AT&T argued it, and lost. Now it comes up again.
As to "wireless": we want RELIABLE access. This means that use of "open" 802.11a/b frequencies are out. At best "patchwork" best effort service can be offered. Plus, if a licensed operator complains, the 802.11a/b user will have to shut off. For limited use, yes, this may work. But for more general use? I like cables here, because I am running my office phone on VOIP.
I'll tell you what I think -- Sony supports "Blu-Ray" -- and they hosed a computer of mine (ok, a relatives) with a root-kit. Disney seems to support "Blu-Ray" -- just advertised all their movie-wares on Blu-Ray. I don't like their attitude toward Copyright Extension. (But I do like Squeak, go figure).
So, I lean to HD-DVD, just to pimp-alap Sony and Disney a bit. You know?
Do I care about the GB? As long as its lots, and (reasonably) reliable, no. Now, if they came up with a disc/changer that was affordable (say $500 US) that would burn a terrabyte without attention overnight. I would kiss them, no, I would get a sex-change and have their babies. Oh, that was a bit extreme... But I would definitely kiss them.
The current attack vectors on cryptographic based "DRM" schemes are (1) accidental key leakage, (2) the key exchange system or (3) the fact that the data must be eventually decoded.
Note that (3) is what makes DRM systems very dumb. It also follows that the Operating System must get involved in order to so hide the data.
If the Operating System allows a debugger to run AT THE SAME TIME as the "DRM", its attackable. If the OS allows "unsigned" drivers to run, its attackable.
The OS (for example, Vista) will (eventually) not allow unsigned drivers. It must also "kick out" or "suspend" all non-DRM (unsigned) software when DRM content is played.
This behaviour falls into Mr. Coxs patent.
Now, if (Vista) doesn't implement the scheme, it remains vulnerable. So, the problem must be solved another way.
My suggestion then is to ALSO patent (or disallow) by widely publishing the idea that a hypervisor (VM supervisor) can be used for DRM control as well, and can also be used to suspend, terminate or otherwise control applications that could be used to attack DRM software.
Got that? It's now published.
After wasting time reading TFM.
May I quote "Both Apple and Microsoft are selling closed and proprietary systems in their consumer electronics products. However, Apple typically sells unlocked hardware but manages its hardware and software as an "integrated experience." Microsoft seeks to lock down its hardware, but leaves third party applications largely unmanaged"
This bombshell dropped by the author really floored me. (and this was about Apple, OS X and the new phone)
How about his excuse as to why Apple doesn't support EVDO? He asks "which one" and also wonders why Apple would compete with itself. Say WHAT?!?
More quality opinions coming, I am sure. (that was sarcastic).
Sorry, I can't make out what you are trying to say. "Cooperated"?
Sure SUN did. What processor was the Standard (published)? Yes, that would have been SUNs. Why didn't the "Asia hardware makers" build something? Right, because Intel was more cooperative and gave proper designs -- not.
Because Microsoft was willing to give source code to those devs -- not.
Because Microsoft is smart -- ok, I'll grant that one. And, because of WinHEC... Why would you think such a thing is good? Given a choice between reasonable documentation, and souce availability, vs. WinHEC, most devs would take the former.
Now, SUN is giving you access to binaries where they are mailed to you without charge, and downloadable source code, and access to the standard dev kit, and you are saying... what; the people who take SUN up on this offer have less brains, and will have less fun (quoted from your message),
Are you serious? Can't you give up on the fanboi attitude for just a wee bit?
Sue.
Find a lawyer in the right jurisdiction. Hopefully the US. Basically, you are after treble damages on songwriter royalties. For something THIS close, other royalties as well (it was sampled).
It will run on a few years -- stick with it. This is a good hunk of your retirement we are talking about here (remeber, treble?, and hopefully with someone selling millions of albums). You also need an accountant along with the lawyer.
Note that the songwriter typically takes 50% of the publishers share in co-publishing arrangements. But, you would not have signed up to that, so take it all. At three times, you can now argue for 7- to 100% of the publishers take on the album.
We are talking about millions here.
In quotes, for a reason.
In my jurisdiction, we pay a media tariff. As a benefit we have a personal copying priviledge.
Now, it would have been "illegal" to copy music for friends before the personal copying provision was put into place. Currently, this provision does not extend to videos or audio books (for two examples).
Let's say that a personal copying provision is put into place for videos. How does DRM get retracted now that it is legal to copy? At that point, it is still either not possible or very difficult to copy the material.
An example closer to home -- I just tried to copy a CD that was "copy protected". Wasn't possible (with the software in place). I guess the "DRM" worked, right? But it was perfectly my right to do what I tried to do. Interesting that the CD had purchased after the personal copy provision was made law.
Just Saying
(In answer to the question -- DRM is always bad).
but not in a good sense.
An explanation for those who don't want to read the patent:
An anlog joystick is pretty much a variable resistor. In order to convert this into a digital domain, the resistance must be measured. We know that the resistor itself can be used to change the discharge time of a capacitor -- this is common. Meauring the discharge time means determining the time the capacitor goes from one voltage to a lower voltage. Which, by its very nature is a "pulse".
And this patent seems to cover all such interfaces in the "joystick" domain.
Now, I can come up with alternate methods for reading the resistance. First one (off the top of my head), is to use a series of resistors controlled by a latched value to produce a reference voltage which is then sent through the joystick resistance. The final output is run through a gate which triggers at a reference voltage. Via search, we can determine the target joystick resistance. (I would probably use a binary search). But this is not any where NEAR as simple or obvious as the R-C approach.
Now, the R-C approach has been used for other variable resistors (prior to 1998) -- the "joystick" application is the only new thing. I used it myself in the 80's. Just never for a joystick (not being into gaming). Its main benefits are that it needs only a single input pin and the circuit is simple. But, given those constraints it is obvious.
Oh well -- go patent trolls. Sure glad I am not in the US.
I had a "simple" optimization project. It came down to one critical function (ISO JBIG compression). I coded the thing by hand in assembler, carefully manually scheduling instructions. It took me days. Managed to beat GNU gcc 2 and 3 by a reasonable margin. The latest Microsoft C compiler? Blew me away. I looked at the assembler it produced -- and I don't get where the gain is coming from. The compiler understands the machine better than I do.
Go figure -- I hung up my assembler badge. Still a useful skill for looking at core dumps, though. And for dealing with micro-controllers.
So, have you had at it and benchmarked your assembler vs. a compilers?
10,000 seems like a lot of money. Its not. I expect a driver dev to get $70+ per hour, this pays for 143 hours.
You are not going to get a driver in that amount of time.
But, I will give you clues. The nVidia chip is pretty high on the OpenGL stack. The chip itself handles most OpenGL primitive operations. It just won't do contexts (nor will the ATI). I don't know the underlying protocol to communicate with the chip, but I would guess it is packet based. Registers would prove far too slow. I would imagine that for OpenGL, VGA, video, and mode support you are looking at almost a thousand "registers" or eqivalents.
It may be possible to catch the kernel level packet interfaces -- mode setting and VGA extension should be reversable via emulation. But this won't tell you what any of the commands do. You could try iterating OpenGL and comparing generated packets... but...
Modern chips typically DON'T implement a fixed-function pipeline. So you will have to figure out how OpenGL shader compiler for the chip works (because you have to know the "machine code").
Good luck for a 4 week driver project. The shader compiler itself is almost a C++ compiler which has to be reversed, the communications format and the packet streams. I would give 10 man-years as a first estimate.
Or, you could try to get the vendors to "be nice".
But I won't do it for 10 grand. Sorry.
"When did people start equating rudimentary copy protection with Digital Rights Management?
The term has lost all meaning."
DRM is not Copy Protection. It is Copy Prevention under defined circumstances (most of them)*. Copy Protection is Copy Prevention under all circumstances. Rudimentary Copy Protection is -- my guess -- Copy Protection that doesn't work.
Anyway, that all means that "Copy Protection == DRM for original media".
* DRM defines a series of "rights" you have to the content. Each application of a "digital right" begins with the allowing of a copy, or degraded copy. Or not allowing the copy (no access). So, the "rights" part of DRM can by replaced by "copy". Which is fundamentally what DRM controls -- Copyright. Further, "Copy Protection" (prevention) controls the same right. Except that instead of using fancy crypto, it is supposed to verify the original media only. Same job. The reason that "DRM" is separated from "Copy Protection" is that "Copy Protection" as defined here cannot be applied unless original media is available. Consider application to downloaded content.
"I want to do more than just "play" the DVD, I want an option to "rip" the DVD and store it. But you can't bundle that right now (legal issues). I want to play music, rip music, download music and podcasts and connect to subscription services all in one. But this stuff is still independent from the services that play movies.
And for the second generation, I want to hook up a second PC in the basement and have it talk to the first PC upstairs. And then I want these guys to share a media library. I want multiple output option so that I can stream music to different rooms via the same remote. But this is still in the dreamer and prototype stage."
Yes, call it "prototype" stage. It is unlikely that you will get what you want from AMD, Intel or Apple, though. They have to maintain relationships with the "content" companies.
As a prototype, MythTV does what you consider to be the first and second generation level stuff (play, rip and share media library). It doesn't do the "connect to subscription service" thing, nor the "same remote" thing. The same remote bit could be added (but would take quite a bit of fiddling).
Now, MythTV is not yet all there -- it definitely qualifies as a "prototype". But people who have seen my rig are interested (most thought that what it does is "impossible" -- especially playing back mpeg2 video via wireless on a PII 400 client without problem).
All told, I view this as a business opportunity -- I can undercut the VVendor! (sic) by $1000 and provide systems with more functionality. Its a "mom and pop" business, but there is certainly money to be made.
Ratboy
I am counting.... 5 wall-warts, 11 plugs, and one "in-line" wart-on-the-floor. And a UPS on the floor (only two plugs, no warts). Three power bars, and two separate wall outlets.
And that's just in my office space.
I am sick of this (unsightly) rigging. Yes, I will tolerate a higher power bill to eliminate it.
You are correct. The value in Sealand is being able to create embassies of Sealand. It should be easy to find nations willing to host Sealand embassies: if only to tweak the noses of the "first world".
Sealand in itself is too dependent on the UK government. But really, the "data centre" or other services can easily be provided at the embassy level.
The remaining problem is that this approach is expensive -- the embassy hosting nations require investment in infrastructure. Or Sealand would have to own its own satellite...
Archival meaning -- read-only. Multiple OS support meaning -- standard.
:)
This cuts the field down. ISO 9660 would be a good bet, but is a bit "overkill". TAR format (which can be viewed as a "primitive" filesystem) would be my choice. Simple, can be read on all your target systems. If a tar client is not (for whatever reason) easily available, the data can still be simply extracted.
Bad point: the "directory" can only be obtained by scanning the entire byte stream. If that is tolerable (and, by indexing the files stored, is mostly just fine), its the one.
If you need fast directory searching as well, consider ISO 9660. Again, clients to read the format are available (although they may be limited to 700MB at a time).
"tar files" (a bit of a misnomer -- should be "tar format byte stream") can be recorded on any device - floppy, disc, tape, CD, DVD, USB stick. It may even provide amusement -- if someone trys to read a recordable CD with a tar image on it, rather than an ISO image
"The hardware is severely under-powered; it cannot be reasonably used for most modern experimenting with computers"
I am a professional developer. Lets look at what I use, and compare to the OLPC:
OLPC 500Mhz geode, 128MB RAM, 512..1GB storage
My main server (external email, httpd, user accounts, time keeping, external portal, ftp server):
IBM PC365, 200Mhz x 2, 128MB RAM
My "workstation" (dev, web browsing, email client, media playback (MythTV client): PII 400, 256MB
Now, my storage is centralized: RAID 5, PII 266 (IBM GL300), 128MB, 4 x 250GB disk.
I have one faster machine, used as a PVR -- the OLPC however could be a replacement for any of my "main squeeze" boxes.
The biggest problem the OLPC has is lack of external storage.
Study TECO first. Look at VI again.
Insert is then no longer a "mode". It is simply a command, terminated by ESC. In TECO, all commands are terminated by ESC (except for a limited number of "immediate" keys). In VI, i is terminated by ESC.
And, with this definition, VI is no longer modal. Also, some commands now make sense. To generate the
banner box: 80i*ESC makes sense: it is, after all, a command.
Wrong on the self-modifying code.
... elide ...
... elide ...
... elide ...
Consider ALTER: it rewrites the targets of GO TO statements.
1-PART.
GO TO 2-PART.
2-PART.
ALTER 1-PART TO 2-PART.
And we scream in pain!
But the ISP wants to encourage the development of such cliques. It can be directed to keep traffic inside the ISPs bounds.
Interestingly, if bittorrent clients start "cheating", ISPs will be happier, and you will see less throttling.
Ratboy