Of course a 64 bit pointer is 2x the size of a 32 bit pointer... 32 bit pointers only need 4 byte alignment, and thus pack nicely. So 64 bit pointers will take twice the cache space.
And... the pointers have to be loaded. It will take more address bits in the instructions to build constants. More cache used.
It is NOT highly likely that 64-bit alignment is done when optimizing. In fact, that's just wrong.
All new gear that present recordable signals in HD will come with "Constrainded Image" (Open Cable Licensing Agreement). All non-HDCP encrypted video shall be reduced to 520,000 pixels.
The good stuff will come out of the DV output plug, encrypted, of course. If you use component video, you get the constrained signal (unless the HD decode is built into your set).
If you go component video, you end up with 960x540 effective, if you use an external HD decoder.
If someone knows of an HD decoder that doesn't have this "feature", let me know...
If Microsoft wanted to change SMB (and lock out SAMBA), they would need some other network file system.
So, maybe, SFU is being released to allow claim that Microsoft based servers can share with NFS, and that SMB can now be modified to add additional (Longhorn) features, while locking the *nix world out (say, by encrypting the traffic and not telling how).
SAMBA becomes useless; and if executed properly, Windows shares become completely MS proprietary. Limited access to "competitive" OSs provided by NFS.
New features available to Windows ONLY.
If I were doing a strategy for MS, this is the path I would take. Also, ensure that NFS is available for a LONG lead time, and that it runs well.
Nope, that wouldn't give me a RIGHT to use the car...
But, if you left your car on the street, keys in, with a sign that said: "Please use this car. I don't care WHO you are, just use it...". Then someone takes the car.... Is that theft?
Patently NO. Now, after the car has been used, you discover that it has been painted -- and you say "Hey, that's not fair -- I only meant for it to be DRIVEN", should you have a legal case?
We start with one wood chipper, one wood chipper operator and a pile of wood. We can chip (whatever) per unit time.
We make the chipper faster, and can do more (increase clock speed of processor), but at some point the operator can't bring us the wood. So, we use a wheelbarrow to transport more wood in a go, and we keep the stack next to the chipper (a cache).
Now, there's plenty of wood, so we get a SECOND chipper. The operator can stick wood into whatever chipper is free (multiple ALU units, out of order execution).
Add a third chipper, and a separate wheel-barrow operator.
This is what we have (pre-"hyperthreading").
Add a second wood chipper operator. If one of them gets tired, the second can take over stuffing the chippers (hyperthreading).
*You* put resources on the Internet. Obviously, for *some* reason.
Normally, the reason you would do that is to provide some service to users. Usually anonymous, given that this is the Internet, and not your private Intranet. If you want it private, don't put it on the Internet.
And, in putting in on the Internet, the resource is available for use.
What you *haven't* done is contracted with *me* as to how to use the service or resource.
Let's put this in simpler terms -- if you have a 20 dollar bill in your pocket, it's yours. If someone takes it that's probably theft.
If you put the same bill out in a public place (say, on a public sidewalk) and then go away, and someone takes, it's probably NOT theft.
When does a resource stop being the "property" of someone? The simplest answer is when they have no control on that resource. Another/may/ be when the police do not need a warrant.
Currently, legislation is trying to make a distrinction between "authorized" and "unauthorized" use of such a service or resource. "unathorized" if the provider of the resource doesn't like the way its used. [Of course, that's very slippery slope.]
Yes, the BIOS needs a TCP/IP stack. At the least, it should support TFTP to allow for network boot. A lot of BIOSs have this. You should be able to boot from the net, or any attached device, such as tape drives. As well, a (possibly primitive) bitmap display should be presented, allowing a graphics window manager to handle new hardware without the need to install a driver. Of course, installing a driver would give you more performance. As well, simple patches to hardware drivers should be installable. There is an implementation of this -- check SUNs firmware. And, to boot, its independent of the processor. Its byte coded FORTH.
I run an email server. For the family. Total of four email addresses. My server directly delivers to other servers. How is this to be taxed?
Obviously, the computer SENDING the email pays the tax. But this means that some form of compliance checking will have to be put into place. Which means a change to the email protocols. But, other countries may not comply. Of course, running an email service for sending may simply be declared illegal, forcing all emails through a centralized point. This solution also has its problems. I guess the tax revenue collected could be used to run the central email servers.
The other day I received a spam. The only thing noteworthy about this was that the sender used my domain as the return address (sales (at) weigel-mohamed (dot) org). This is upsetting, so I wanted to track her down.
I went to the web site -- it offered life insurance brokering. I put in a fake quotation request, assuming that I would here from the life insurance company. Which happened a few hours later.
I then had the life insurance company try to track the spammer from their end -- but the "company" had vanished. If the company isn't in business LONG ENOUGH TO COLLECT, how can they make money?
Out of curousity, I have tried to track some of these companies over the past few years. Most are disconnected before there is a chance to make money. My estimate is that any payoff must come within HOURS of posting the SPAM (3 to 6 hours).
Display driver -- that was interesting. Blue screen of death until I installed OS followed by display. May not have been, but all I got was the blue screen.
After that, I didn't trust any of the drivers to be restored; installed Windows again, and all drivers, rather than rely on restore.
Some application defaults were not restored properly (extension bindings) -- no biggy, and this *may* have been my fault (in that they may have been wrong at backup time, I wouldn't really know).
So, it was installation of OS, drivers, and updates (3 or 4 hours), followed by user data restore (1 hour). This is problematic because this should have take 1 hour.
The idea here is a FULL backup. This restores the machine to the same point as when the backup was made. OS, applications, user data.
No hoops. No muss. No fuss.
Also, your instructions don't actually work. Some software must be re-installed. I've struggled with this (and I *won't* buy "ghost" or whatever -- this should be part of the OS).
The main problem with biometric schemes is that it is (almost) impossible to contest.
The card holds your biometric data (say finger-print and iris scan). If card production is cracked, the cards can then be forged -- making the biometric data useless. This is comparable to pasting in another picture on a drivers license.
To prevent this, the biometric data can be stored elsewhere. One copy (the one that can't be repudiated) is on the person. One copy may be on the card (if the card is secure). Typically, one *more* copy is on a computer.
The "client" is scanned, and biometric data is compared against the stored copy. Hack attempts portrayed on movies have the "bad guys" using cut off fingers, etc. to beat the system. But this isn't the attack point of choice.
If the biometric data is modified in the stored computer file, we have a problem. Someone makes a change saying "this person is a terrorist". Or another identity change. You CAN'T change your biometric data, and governments aren't likely to reveal what is in the "secret" files.
So, a hacker seeds data in a computer somewhere, and the next time you travel, BAM, you are arrested with no way of proving that it ISN'T you. Of course its you, the biometric data matches.
Any compromise in the system is very bad. This is a very bad thing. The privacy thing is a canard -- not being able to repudiate the biometric data makes it almost impossible to correct records, and reclaim identity through government layers.
It is NOT illegal to make a copy of a sound recording in the US and Canada (no, I don't know about other jurisdictions).
It is illegal to mass duplicate music - and I agree with that.
The Copying rights for Linux are CLEARLY labeled, and every user is CLEARLY informed of them. It is illegal to not supply this information.
The Copying rights for CDs aren't particularly clearly stated. It takes a bit of digging to actually figure them out.
Sound recordings occupy a special place in Copyright law. Computer source programs are NOT the same thing at all.
Now, all of this has been hashed to death, but the thing that makes you a Bozo is:
The RIAA does NOT hold sound recording copyrights. The RIAA is simply a cartel representing its members.
Linux is not even a cartel. It is simply a trademark. There is NO cartel; copyright is held and defended by the individual authors, or, in some cases, has been assigned to the FSF.
- First, register a domain name. godaddy.com would do nicely. This will cost you ten bucks a year (prepay for a couple of years).
- Next, you need DNS services. Here there is a problem: (1) if your IP isn't really static and (2) you need 2 (two) different machines serving DNS. So go to someone who can solve both problems. dyndns.org would do nicely. Give money to dyndns.org to do your DNS (you want CUSTOM service) -- they will give you the IP addresses of the DNS servers, and you supply those to godaddy.com. dnydns.org used to have a one-time contribution to do the dns (30 bucks?).
(3) Any time your IP address changes, simply go to the dyndns.org website and update your record. If you have a DHCP assigned IP, there are scripts that will automatically update dyndns.org whenever it changes.
And you are off to the races... your own email, web hosting, whatever. Total cost: ~$50US for two years.
The dongle is NOT to prevent crackers from duplicating the software.
I used to write and sell low volume high margin software. Price was 6k to 30k US for the software (typically, through in the computer for free). 800 to 1k per year for support.
I burned dongles for each client, and keyed option modules to each indivdual dongle/client. If a client wanted it, I would give them a "dongle-free" version. Same price. After a simple explantion, NO ONE TOOK ME UP ON THE OFFER.
And here is the explanation... You (the client) have just spent 20k or 30k on this software. It is VITAL to your business. If there is no dongle, an employee could simply walk with a copy of the goodies, and set up shop. If you WANT it, we will give you a non-dongle version, and we escrow the source anyway.
If someone can use the software without paying for it, they will gain a serious advantage. So don't let them.
Same thing applies if I (for example) modify GCC to support a BRAND X microcontroller. I would charge a reasonable (or not) amount for the port. The client would then be in the position to (1) release the source that they received and lose competitive advantage, or (2) keep it a secret. Most would go with option (2) here. If it (the software) works well, why share the benefit?
1 - PC emulating serial printer from an IBM DisplayWrite (current-loop). Recover documents and convert to standard format.
1a - direct media i/o from PC to legacy system (8" floppy controller and driver software for PC, including target filesystem - CP/M, Wang OIS, Philips P2000, IBM DisplayWrite, AM Comp/Edit and Comp/Set). Custom PC controller card.
2 - Parallel interface to mini. Emulate Lino 202N Photocomposer. Recover documents and print onto laserprinter.
3 - Same as (2), but CG8600, and APS u5.
Send me the model of the printer you want emulated. fred_weigel (at) hotmail.com I'll work up a quotation for you (hardware, custom hardware, software).
No it doesn't... Go get a Aviator 900Mhz wireless device, or a D-Link DMP-90 and plug that into your shiny Windows XP. What about NVidia drivers? How about mother-board specific sound drivers?
I have HAD it with people telling me that "driver support is better with Windows". It isn't. Linux supports more devices out of the box. Linux supports more architectures out of the box.
If I here that piece of revisionist crap one more time, I'll have to strangle the perp. Windows may be good for something, but it sure isn't device support. Because drivers under Windows are generally closed source, the drivers live just as long as the maintainers have interest in keeping the device alive. If the company goes out of business, or the product is obsoleted - tough luck.
If the driver is open-sourced (as most Linux drivers are), the product can be supported until the last piece of hardware dies.
1 - Microsoft Windows is NOT the only Operating Environment out there.
2 - Yes, you DO want to run 32 bit apps on your 64 bit system. Not every application improves when built as 64 bit. Stuff takes more space in cache (for instance).
3 - Why are we not ready for 64 bit desktops? I've had one for years. YOU may not be ready, but there are applications which are difficult without the address space.
4 - As to "splitting" the market... This is probably a problem with binary-only programs. Some of us don't care (we use MIPS, SPARC, ULTRA and other architectures). A simple recompile resolves the issue.
5 - Ok, you don't have faith in AMD anymore. About the only valid point in your post.
I was ready to moderate, but decided to reply instead --
Virtual 86 mode is missing from the '286 but it isn't the most important thing...
The '286 is a 16 bit machine. This makes porting Unix software (which generally assumes a 32 bit 'int' type) a bit difficult.
Also, the '286 does not support paging. Only segmentation. It is very difficult to get a "steady state" swapping working with variable length segments. Also, there were only 8K user and 8K system segments. Which rendered the idea of a separate segment for each data item moot.
Also, the '286 was limited to 16MB of real memory. Which is lightweight for a GUI (fine for console based applications, though).
Microsoft DID do a "lot" with the 286, but the user base of the time didn't see the need to play along. OS/2 1.x was actually quite choice, and exploited the '286 well.
However, it was expensive (compared to DOS), used too much memory (not many computers had more than 640K, 2MB was considered a lot) and there weren't many compelling applications for the new protected mode. Also, OS/2 was late. VERY late.
Doom and gloom for the 286.
No, it wasn't "Intel's fault". But Intel just made the chips. The '286 was sampling in 1982. Personally, I like it that way. Give us the chips, and we will decide what can fly on the market and what must die...
The article is sooooo right. You are allowed to make a copy of a musical work, in whole or in part for your perosnal use. The law does not indicate that you must have copyright, or any other interest in the work. This is NOT "fair use". This is "personal copying".
It must be made to a media that is not proscribed, and is commonly used for this purpose (hard disk, flash, cdr would all qualify).
You may NOT make the copy if the copy is to be publically telecommunicated, sold, rented.
The question that remains is whether the downloader or provider actually makes the copy. If its the downloader, the RIAA shouldn't be going after the kazaa sites. These kinds of definitions tend to leak across the border... If it IS the downloader, then Kazaa is completely legal in Canada. If not (the provider is making the copy), then dowloading and MINIMAL uploading is ok in Canada (watch out for that "telecommunications" clause). The simple way to get around that is to put the files on (say) an FTP server, where the downloader is explicitly informed that THEY are making the copy. Completely anonymous or multi-site downloading would probably be forbidden.
Anyway, it hasn't come to a legal test yet.
And the Canadian Recording Association doesn't seem to want to put it to the test yet, either.
Of course media codecs aren't installed in RH9. Not even MP3.
If *you* are the "IT person" (geek) who is responsible for maintaining the systems, then you need to install the standard pieces that make it work. Otherwise, use Microsoft Windows.
WIN2K doesn't come with needed codecs either. It can't play DVDs "out of the box".
As to a "user friendly" desktop. That's your job. Really.
Can't play MP3 files from a Samba share? Works for me... as does playing re-coded MP3s via Apache at work from my home server.
Now, on to the tools that are "missing" from RH9, that you probably want to download and install:
ddclient.tar.gz - Update your dynamic IP with dyndns.org. Installs as a standard RH9 service
gotmail_0.7.10.tar.gz - Fetch all hotmail and transfer to your local mail service.
install_flash_player_6_linux.tar.gz - Add flash to mozilla.
j2re-1_4_1_04-linux-i586-rpm.bin - Add Java to mozilla.
lame-3.93.1.tar.gz - MP3 encoding or recoding.
mpg123-pre0.59s.tar.gz - MP3 command line playback
xmms-mpg123-1.2.7-13.i386.rpm - Add MP3 format to xmms player.
wine-0.20030709-1.i386.rpm - Allow some windows executables (I use this to run MS MSVC6 for cross-builds).
XINE video and DVD player (you may want to skip the mozilla plugin).
And that should do it! Not all of these install cleanly, but they all do work.
Now for the bonus section. I have a cable modem, and want to listen to my tunes at work. I decided on a simple web-based "click on the song to listen" approach (at least initially). The problem is that my cable modem only does 128Kbps uplink, and most of my collection is stored at a higher bit rate. I don't want to completely download the music first. So, I decided to recode the MP3 on the server first (to 64Kbps -- and yes, I use cheap headphones at the office):
Now, restart Apache (service httpd restart), and any "mp3" files will be recoded to 64Kbps.
How to do this with Win2K? Damned if I know. Maybe you can give me a bit of guidance. Would it simply be 7 lines of text-based scripting, and a commonly available command line program? Hopefully, something even simpler (although, I must admit, I have a hard time envisioning that -- but I could be wrong!).
Slashdot Mirror
Of course a 64 bit pointer is 2x the size of a 32 bit pointer... 32 bit pointers only need 4 byte alignment, and thus pack nicely. So 64 bit pointers will take twice the cache space.
And... the pointers have to be loaded. It will take more address bits in the instructions to build constants. More cache used.
It is NOT highly likely that 64-bit alignment is done when optimizing. In fact, that's just wrong.
Yes, cache performance suffers.
No, x86 does not require data alignment for 4 byte memory load.
Back in the day... a company commerialized some software (in the '70s). The company was called the "Itty Bitty Machine" company.
No big deal...
Ratboy
All new gear that present recordable signals in HD will come with "Constrainded Image" (Open Cable Licensing Agreement). All non-HDCP encrypted video shall be reduced to 520,000 pixels.
The good stuff will come out of the DV output plug, encrypted, of course. If you use component video, you get the constrained signal (unless the HD decode is built into your set).
If you go component video, you end up with 960x540 effective, if you use an external HD decoder.
If someone knows of an HD decoder that doesn't have this "feature", let me know...
Ratboy
If Microsoft wanted to change SMB (and lock out SAMBA), they would need some other network file system.
So, maybe, SFU is being released to allow claim that Microsoft based servers can share with NFS, and that SMB can now be modified to add additional (Longhorn) features, while locking the *nix world out (say, by encrypting the traffic and not telling how).
SAMBA becomes useless; and if executed properly, Windows shares become completely MS proprietary. Limited access to "competitive" OSs provided by NFS.
New features available to Windows ONLY.
If I were doing a strategy for MS, this is the path I would take. Also, ensure that NFS is available for a LONG lead time, and that it runs well.
But maybe I'm just a paranoid.
Ratboy.
Car running, keys in...
Nope, that wouldn't give me a RIGHT to use the car...
But, if you left your car on the street, keys in, with a sign that said: "Please use this car. I don't care WHO you are, just use it...". Then someone takes the car.... Is that theft?
Patently NO. Now, after the car has been used, you discover that it has been painted -- and you say "Hey, that's not fair -- I only meant for it to be DRIVEN", should you have a legal case?
That is what we should be talking about.
Ratboy.
More correct:
We start with one wood chipper, one wood chipper operator and a pile of wood. We can chip (whatever) per unit time.
We make the chipper faster, and can do more (increase clock speed of processor), but at some point the operator can't bring us the wood. So, we use a wheelbarrow to transport more wood in a go, and we keep the stack next to the chipper (a cache).
Now, there's plenty of wood, so we get a SECOND chipper. The operator can stick wood into whatever chipper is free (multiple ALU units, out of order execution).
Add a third chipper, and a separate wheel-barrow operator.
This is what we have (pre-"hyperthreading").
Add a second wood chipper operator. If one of them gets tired, the second can take over stuffing the chippers (hyperthreading).
Is that a bit clearer?
Ratboy
WTF?!?
/may/ be when the police do not need a warrant.
-- You need to think about what "property" is --
*You* put resources on the Internet. Obviously, for *some* reason.
Normally, the reason you would do that is to provide some service to users. Usually anonymous, given that this is the Internet, and not your private Intranet. If you want it private, don't put it on the Internet.
And, in putting in on the Internet, the resource is available for use.
What you *haven't* done is contracted with *me* as to how to use the service or resource.
Let's put this in simpler terms -- if you have a 20 dollar bill in your pocket, it's yours. If someone takes it that's probably theft.
If you put the same bill out in a public place (say, on a public sidewalk) and then go away, and someone takes, it's probably NOT theft.
When does a resource stop being the "property" of someone? The simplest answer is when they have no control on that resource. Another
Currently, legislation is trying to make a distrinction between "authorized" and "unauthorized" use of such a service or resource. "unathorized" if the provider of the resource doesn't like the way its used. [Of course, that's very slippery slope.]
Ratboy.
Yes, the BIOS needs a TCP/IP stack. At the least, it should support TFTP to allow for network boot. A lot of BIOSs have this. You should be able to boot from the net, or any attached device, such as tape drives. As well, a (possibly primitive) bitmap display should be presented, allowing a graphics window manager to handle new hardware without the need to install a driver. Of course, installing a driver would give you more performance. As well, simple patches to hardware drivers should be installable. There is an implementation of this -- check SUNs firmware. And, to boot, its independent of the processor. Its byte coded FORTH.
Implement that on a PC, and I'd be happy.
Ratboy
I run an email server. For the family. Total of four email addresses. My server directly delivers to other servers. How is this to be taxed?
Obviously, the computer SENDING the email pays the tax. But this means that some form of compliance checking will have to be put into place. Which means a change to the email protocols. But, other countries may not comply. Of course, running an email service for sending may simply be declared illegal, forcing all emails through a centralized point. This solution also has its problems. I guess the tax revenue collected could be used to run the central email servers.
Ratboy
The other day I received a spam. The only thing noteworthy about this was that the sender used my domain as the return address (sales (at) weigel-mohamed (dot) org). This is upsetting, so I wanted to track her down.
I went to the web site -- it offered life insurance brokering. I put in a fake quotation request, assuming that I would here from the life insurance company. Which happened a few hours later.
I then had the life insurance company try to track the spammer from their end -- but the "company" had vanished. If the company isn't in business LONG ENOUGH TO COLLECT, how can they make money?
Out of curousity, I have tried to track some of these companies over the past few years. Most are disconnected before there is a chance to make money. My estimate is that any payoff must come within HOURS of posting the SPAM (3 to 6 hours).
How the hell do they make any money at this?
Ratboy
Display driver -- that was interesting. Blue screen of death until I installed OS followed by display. May not have been, but all I got was the blue screen.
After that, I didn't trust any of the drivers to be restored; installed Windows again, and all drivers, rather than rely on restore.
Some application defaults were not restored properly (extension bindings) -- no biggy, and this *may* have been my fault (in that they may have been wrong at backup time, I wouldn't really know).
So, it was installation of OS, drivers, and updates (3 or 4 hours), followed by user data restore (1 hour). This is problematic because this should have take 1 hour.
Annoyed me.
Ratboy
Slim:
Um... you hit the problem on the head...
To restore you
1 - Install Windows
2 - Insert tape and...
Its Step 1 that shouldn't be needed.
The idea here is a FULL backup. This restores the machine to the same point as when the backup was made. OS, applications, user data.
No hoops. No muss. No fuss.
Also, your instructions don't actually work. Some software must be re-installed. I've struggled with this (and I *won't* buy "ghost" or whatever -- this should be part of the OS).
Ratboy
The main problem with biometric schemes is that it is (almost) impossible to contest.
The card holds your biometric data (say finger-print and iris scan). If card production is cracked, the cards can then be forged -- making the biometric data useless. This is comparable to pasting in another picture on a drivers license.
To prevent this, the biometric data can be stored elsewhere. One copy (the one that can't be repudiated) is on the person. One copy may be on the card (if the card is secure). Typically, one *more* copy is on a computer.
The "client" is scanned, and biometric data is compared against the stored copy. Hack attempts portrayed on movies have the "bad guys" using cut off fingers, etc. to beat the system. But this isn't the attack point of choice.
If the biometric data is modified in the stored computer file, we have a problem. Someone makes a change saying "this person is a terrorist". Or another identity change. You CAN'T change your biometric data, and governments aren't likely to reveal what is in the "secret" files.
So, a hacker seeds data in a computer somewhere, and the next time you travel, BAM, you are arrested with no way of proving that it ISN'T you. Of course its you, the biometric data matches.
Any compromise in the system is very bad. This is a very bad thing. The privacy thing is a canard -- not being able to repudiate the biometric data makes it almost impossible to correct records, and reclaim identity through government layers.
Ratboy
It is NOT illegal to make a copy of a sound recording in the US and Canada (no, I don't know about other jurisdictions).
It is illegal to mass duplicate music - and I agree with that.
The Copying rights for Linux are CLEARLY labeled, and every user is CLEARLY informed of them. It is illegal to not supply this information.
The Copying rights for CDs aren't particularly clearly stated. It takes a bit of digging to actually figure them out.
Sound recordings occupy a special place in Copyright law. Computer source programs are NOT the same thing at all.
Now, all of this has been hashed to death, but the thing that makes you a Bozo is:
The RIAA does NOT hold sound recording copyrights. The RIAA is simply a cartel representing its members.
Linux is not even a cartel. It is simply a trademark. There is NO cartel; copyright is held and defended by the individual authors, or, in some cases, has been assigned to the FSF.
Ratboy
Ok -- here's the steps:
- First, register a domain name. godaddy.com would do nicely. This will cost you ten bucks a year (prepay for a couple of years).
- Next, you need DNS services. Here there is a problem: (1) if your IP isn't really static and (2) you need 2 (two) different machines serving DNS. So go to someone who can solve both problems. dyndns.org would do nicely. Give money to dyndns.org to do your DNS (you want CUSTOM service) -- they will give you the IP addresses of the DNS servers, and you supply those to godaddy.com. dnydns.org used to have a one-time contribution to do the dns (30 bucks?).
(3) Any time your IP address changes, simply go to the dyndns.org website and update your record. If you have a DHCP assigned IP, there are scripts that will automatically update dyndns.org whenever it changes.
And you are off to the races... your own email, web hosting, whatever. Total cost: ~$50US for two years.
Ratboy.
The dongle is NOT to prevent crackers from duplicating the software.
I used to write and sell low volume high margin software. Price was 6k to 30k US for the software (typically, through in the computer for free). 800 to 1k per year for support.
I burned dongles for each client, and keyed option modules to each indivdual dongle/client. If a client wanted it, I would give them a "dongle-free" version. Same price. After a simple explantion, NO ONE TOOK ME UP ON THE OFFER.
And here is the explanation... You (the client) have just spent 20k or 30k on this software. It is VITAL to your business. If there is no dongle, an employee could simply walk with a copy of the goodies, and set up shop. If you WANT it, we will give you a non-dongle version, and we escrow the source anyway.
If someone can use the software without paying for it, they will gain a serious advantage. So don't let them.
Same thing applies if I (for example) modify GCC to support a BRAND X microcontroller. I would charge a reasonable (or not) amount for the port. The client would then be in the position to (1) release the source that they received and lose competitive advantage, or (2) keep it a secret. Most would go with option (2) here. If it (the software) works well, why share the benefit?
Ratboy.
yes...
1 - PC emulating serial printer from an IBM DisplayWrite (current-loop). Recover documents and convert to standard format.
1a - direct media i/o from PC to legacy system
(8" floppy controller and driver software for PC, including target filesystem - CP/M, Wang OIS, Philips P2000, IBM DisplayWrite, AM Comp/Edit and Comp/Set). Custom PC controller card.
2 - Parallel interface to mini. Emulate Lino 202N Photocomposer. Recover documents and print onto laserprinter.
3 - Same as (2), but CG8600, and APS u5.
Send me the model of the printer you want emulated. fred_weigel (at) hotmail.com
I'll work up a quotation for you (hardware, custom hardware, software).
Ratboy
No it doesn't... Go get a Aviator 900Mhz wireless device, or a D-Link DMP-90 and plug that into your shiny Windows XP. What about NVidia drivers? How about mother-board specific sound drivers?
I have HAD it with people telling me that "driver support is better with Windows". It isn't. Linux supports more devices out of the box. Linux supports more architectures out of the box.
If I here that piece of revisionist crap one more time, I'll have to strangle the perp. Windows may be good for something, but it sure isn't device support. Because drivers under Windows are generally closed source, the drivers live just as long as the maintainers have interest in keeping the device alive. If the company goes out of business, or the product is obsoleted - tough luck.
If the driver is open-sourced (as most Linux drivers are), the product can be supported until the last piece of hardware dies.
1 - Microsoft Windows is NOT the only Operating Environment out there.
2 - Yes, you DO want to run 32 bit apps on your 64 bit system. Not every application improves when built as 64 bit. Stuff takes more space in cache (for instance).
3 - Why are we not ready for 64 bit desktops? I've had one for years. YOU may not be ready, but there are applications which are difficult without the address space.
4 - As to "splitting" the market... This is probably a problem with binary-only programs. Some of us don't care (we use MIPS, SPARC, ULTRA and other architectures). A simple recompile resolves the issue.
5 - Ok, you don't have faith in AMD anymore. About the only valid point in your post.
Ratboy
So go there...
Remove your blinders, and find out what SUN or IBM have to offer you.
For SUN gear, a server handling 40 users, and SUN Ray terminals wouldn't be bad...
Ratboy666
I was ready to moderate, but decided to reply instead --
Virtual 86 mode is missing from the '286 but it isn't the most important thing...
The '286 is a 16 bit machine. This makes porting Unix software (which generally assumes a 32 bit 'int' type) a bit difficult.
Also, the '286 does not support paging. Only segmentation. It is very difficult to get a "steady state" swapping working with variable length segments. Also, there were only 8K user and 8K system segments. Which rendered the idea of a separate segment for each data item moot.
Also, the '286 was limited to 16MB of real memory. Which is lightweight for a GUI (fine for console based applications, though).
Microsoft DID do a "lot" with the 286, but the user base of the time didn't see the need to play along. OS/2 1.x was actually quite choice, and exploited the '286 well.
However, it was expensive (compared to DOS), used too much memory (not many computers had more than 640K, 2MB was considered a lot) and there weren't many compelling applications for the new protected mode. Also, OS/2 was late. VERY late.
Doom and gloom for the 286.
No, it wasn't "Intel's fault". But Intel just made the chips. The '286 was sampling in 1982. Personally, I like it that way. Give us the chips, and we will decide what can fly on the market and what must die...
Ratboy.
The article is sooooo right. You are allowed to make a copy of a musical work, in whole or in part for your perosnal use. The law does not indicate that you must have copyright, or any other interest in the work. This is NOT "fair use". This is "personal copying".
It must be made to a media that is not proscribed, and is commonly used for this purpose (hard disk, flash, cdr would all qualify).
You may NOT make the copy if the copy is to be publically telecommunicated, sold, rented.
The question that remains is whether the downloader or provider actually makes the copy. If its the downloader, the RIAA shouldn't be going after the kazaa sites. These kinds of definitions tend to leak across the border... If it IS the downloader, then Kazaa is completely legal in Canada. If not (the provider is making the copy), then dowloading and MINIMAL uploading is ok in Canada (watch out for that "telecommunications" clause). The simple way to get around that is to put the files on (say) an FTP server, where the downloader is explicitly informed that THEY are making the copy. Completely anonymous or multi-site downloading would probably be forbidden.
Anyway, it hasn't come to a legal test yet.
And the Canadian Recording Association doesn't seem to want to put it to the test yet, either.
Ratboy
Of course media codecs aren't installed in RH9. Not even MP3.
/var/www/cgi-bin I added an executable
/usr/local/bin/lame -S --mp3input -q 7 -b 64 "$PATH_TRANSLATED" -
/etc/httpd/conf/httpd.conf file added:
/cgi-bin/nph-mp3recode
If *you* are the "IT person" (geek) who is responsible for maintaining the systems, then you need to install the standard pieces that make it work. Otherwise, use Microsoft Windows.
WIN2K doesn't come with needed codecs either. It can't play DVDs "out of the box".
As to a "user friendly" desktop. That's your job. Really.
Can't play MP3 files from a Samba share? Works for me... as does playing re-coded MP3s via Apache at work from my home server.
Now, on to the tools that are "missing" from RH9, that you probably want to download and install:
ddclient.tar.gz - Update your dynamic IP with dyndns.org. Installs as a standard RH9 service
gotmail_0.7.10.tar.gz - Fetch all hotmail and transfer to your local mail service.
install_flash_player_6_linux.tar.gz - Add flash to mozilla.
j2re-1_4_1_04-linux-i586-rpm.bin - Add Java to mozilla.
lame-3.93.1.tar.gz - MP3 encoding or recoding.
mpg123-pre0.59s.tar.gz - MP3 command line playback
xmms-mpg123-1.2.7-13.i386.rpm - Add MP3 format to xmms player.
wine-0.20030709-1.i386.rpm - Allow some windows executables (I use this to run MS MSVC6 for cross-builds).
XINE: (RealPlayer9-9.0.7.151-4.i386.rpm w32codec-0.52-1.i386.rpm xine-mozilla-plugin-0.2-030528.i586.rpm libdvdcss-1.2.6-2.network.i386.rpm xine-ui-0.9.21cvs-030528.i586.rpm libxine1-1_cvs-030528.i586.rpm)
XINE video and DVD player (you may want to skip the mozilla plugin).
And that should do it! Not all of these install cleanly, but they all do work.
Now for the bonus section. I have a cable modem, and want to listen to my tunes at work. I decided on a simple web-based "click on the song to listen" approach (at least initially). The problem is that my cable modem only does 128Kbps uplink, and most of my collection is stored at a higher bit rate. I don't want to completely download the music first. So, I decided to recode the MP3 on the server first (to 64Kbps -- and yes, I use cheap headphones at the office):
In
nph-mp3recode:
#!/usr/bin/sh
echo "HTTP/1.0 200 Document follows"
echo "Content-type: audio/mpeg"
echo ""
In my
Action audio/mpeg
to one of my virtual hosts.
Now, restart Apache (service httpd restart),
and any "mp3" files will be recoded to 64Kbps.
How to do this with Win2K? Damned if I know. Maybe you can give me a bit of guidance. Would it simply be 7 lines of text-based scripting, and a commonly available command line program? Hopefully, something even simpler (although, I must admit, I have a hard time envisioning that -- but I could be wrong!).
Ratboy.
Naw...
I went and typed this:
% man 0 Allah
Ratboy