[...] some utterly brainless idiot at Microsoft Corp. decided that ascii is the default for FTP-servers.
Sorry, but this just isn't Microsoft's fault. It goes back MUCH farther than that, to the early days of FTP. In fact, on non-free UNIX-alikes, "TYPE A" remains the default to this day. I've had "alias ftp=ftp -inv" for years, long before Microsoft had an IP stack, let alone an FTP server. (Well, maybe in Minix....)
I'm not even sure the server gets to decide the default type. If your client doesn't request TYPE I, it shouldn't count on anything. I'd look up the RFCs, but it's lunch time.
Keep in mind when the ARPANET was young, most machines weren't binary-compatible. (Of the few that were wired together.) Binary transfers were rare. With compressed and archive formats, that's no longer the case, even for plain text... but the defaults remain.
But for fun, do price-matching between Future Shop and Best Buy.
"You know Best Buy is selling this $20 cheaper... uh, don't they own you guys now?"
Watching the guy completely fail to be able to access Best Buy's website from the in-store computer was worth it. The computer worked fine, he just couldn't type "www.bestbuycanada.ca". (It wouldn't have made a difference, Best Buy Canada doesn't have a catalog on the 'web.)
Of course, Logitech's new trackball tracks at 60 degrees to where it is comfortable to hold, so after all the price-matching fun I got my money back. Hehehehe.
"display" is a CSS thing, which would mean 'style="display:none"' in this case. "hidden" is a type of input widget, along with "text", "password", "checkbox" and has been in HTML for ages, at least since 2.0. (I didn't do any work with forms in older versions of HTML.)
If you rely on CSS to hide your hidden field (I've used them for session IDs on generated pages, instead of using cookies), then it will "unhide" on non-CSS aware clients. Also, when unhidden, it will probably also be input-able; readonly is pretty new. "type=hidden" gets all the semantics right all at once (style="display:none" type="string" readonly), and it is valid HTML 2.0.
Graceful degredation and all that. Of course, one must also fail to trust the client, so be careful what goes in anything the client sends back to you, hidden or not.
(And you can omit the quotes on attribute values in certain cases; check http://www.w3.org/TR/html401/intro/sgmltut.html#at tributes.)
There is no "PACK(1)" prgma for c structures on a mac.
[...]there is no penalty for use on later CPUs[...]
But on the G3, misaligned loads are corrected by software after taking a SIGBUS. So you really, really, really, really, really, really, really, really don't want to pack structures; if you want to save space, rearrange them instead.
(I don't how the G4 feels about misaligned loads--the only other PowerPCs I can get to are IBM POWER4.)
NVidia has a kernal patch so that they isolate the kernel-version-dependent code and can distribute that as source. It simply glues in the binary-only part, and allows you to recompile it if you patch your kernel. That's why there's so many kernel glue layer packages from NVidia, they're all built for a very specific kernel version.
But you can save some typing, leave off the ",1", tape drive is the default. Load much faster. Of course, shift-STOP will load and run the next thing on tape.
And if you do go see a movie, you'll find out when you got back the program was _earlier_ on the tape, and now you have to rewind and start all over.
There's got to be something better I could be using these brain cells for.
You can't get the same density at higher rotational speeds; it requires more "design" in the heads and sense amps to read the signals reliably enough. So you have to trade capacity for high RPM... and you have to hope the density is _enough_ better that you still get a transfer-rate advantage with the higher RPM.
Since they don't even make SCSI drives with the same rotational speed as IDE these days, you can't compare, there are too many differences. So you get the "new research" model with SCSI, and once they have it figured out, the IDE versions come out.
So you've got more than just www.scsita.org putting up the prices, you've also got the bleeding-edge factor too.
I used to like SCSI. I still like the command set, and since FireWire supports serial SCSI (SBP-2), I get the best of both: cheap IDE drives in a FireWire box and I can use SCSI drivers to control them.
23-pin D-sub? Amiga floppy and video; very unstandard size connector. Hacksaw and a D-25 was my usual solution....
Don't forget 15-pin D-subs are (OK, were) also used for Ethernet AUI connectors; also hardly ever seen these days. I would expect some cases built to support on-board networking to have these.
Those RT connectors carried over into the early RS/6000s, which meant every machine came with a "book" containing 1' long cables with the weird connector on one side and the standard connector on the other. (Plus the wrap plugs for self-diagnosis.)
I finally decided to drop iCab and switch to Mozilla. I really like a lot of things in iCab, but I'm using CSS2 for all my documents now, and iCab just doesn't do CSS2.
I would like to see a Gecko renderer with the iCab user interface and filtering features.
I remember pie menus from way back in '86, there were some articles in the Amiga mags about their benefits. The only problem was, the machines of the day didn't have enough CPU grunt to do the rendering, and you couldn't fit much information around the pie because everyone's pixels were so big. On a 640x200 display, there just isn't room.
*senil yvaw*
One thing pie menus do is give an "infinite depth" target area, like placing icons right at the edge of the screen.
This allows the novice user to make a large motion, to where the pie-slice is "wider", and hit the item. But as you get used to the menu, you can make a smaller motion and hit the narrower part of your choice.
Because this implementation pops up submenus for some directions, it doesn't _quite_ reach that ideal. But now I don't know if I want to learn gestures or pies. Maybe both together; I've got lots of mouse buttons....
But you still retain the advantage all context-menus have in that the menu is "right here", instead of off at the top of the screen (and infinately tall) or at the top of the window.
Very spiffy. I've waited 16 years for this, it seems.
Those weren't seven independent processors, those were seven chips making up a single POWER CPU. The original POWER CPUs were NOT a microprocessor, they were multi-chip; 7 to 12 or so depending on the model. The functional units (decode, float, fixed, and so on) each had their own chip to live on. Even the POWER2 CPU was multi-chip, but they were all mounted in a single package (with the biggest heatsink I'd ever seen, long before Intel and AMD started playing MHz battles.)
That is why the PowerPC was a big deal in 1993; the POWER instruction architecture on a single-chip microprocessor.
There was a single-chip POWER CPU made for the 7011-220 through 7011-240, but nobody wants to think about those. The 7011-250 was a PowerPC 601, and worlds different from the other types in that model line.
Objective C uses a very different way of getting messages passed between objects. It doesn't have any of the binary interface issues C++ does, particularly name mangling and very touchy binary object layout issues.
It takes a bit of a twist to get your mind around Objective C, as it is very much a dynamic, run-time-determined kind of language.
I guess your ISP does not block port 25. And the recipient's server has to be set up to accept mail from dial-up pools; many do not. At least, when I tried that a couple of years ago, the server I tried to use wouldn't let me in from the dial-up-like block my DSL address is in.
And ISPs blocking port 25 is pretty common; mine's done it, so I can't access an authenticated SMTP server directly, I have to proxy to it on port 26 because some mail clients hard-code the port number.
None of this would be needed if spam wasn't a problem.
But what's cute is with my USB Compact Flash reader it shows up as a SCSI drive, mount cp & umount. Even easier on Mac and Windows--great way to transfer big files. And if you don't need to use the card in your camera, you could even partition it however you want and have different kinds of filesystems.
Single-track spiral media is just dumb for computer data. But we're stuck with it, it seems.
Older US discs would also jump to the feature on auto-play. You had to press "menu" if you really wanted it. MGM started offering pan & scan and widescreen on the same side, so they needed a menu before the feature--though it would timeout and default to widescreen if you left it.
When you buy the Sol 8/9 (Example) media kits, you can install on either 32 bit or 64 bit sparc boxen.
It's even better than that. When you install on a 64-bit SPARC (or AIX on a 64-bit POWER or PowerPC), you can still run 32-bit code fully natively. No emulation.
You have to craft your OS carefully to allow 32- and 64-bit applications to co-exist on the system. But you usually have a kernel that runs in one mode and the appropriate userland code to call into it from either.
The easiest way to get started, of course, is to have a 32-bit kernel support 64-bit user tasks. Depending on the architecture, there may not be any advantage to re-building the kernel in 64-bit native. ("modprobe 64bit" anyone? AIX has the equivelent to turn on 64-bit user processes in a 32-bit kernel.)
Of course, the SPARC and PowerPCs have been doing this for... uh... how many years now? 6 at least, right? Intel going to a "pure" 64-bit architecture seemed to me to be just plain arrogant. "Forget your old code, buy our new 100-watt CPUs!"
I can bet you that if I went up to someone in the street and said, if you lost your copy of a CD, or PS2 game, or if your child snapped the disc in half, should you be entitled to a free new one?
For me to accept this, they must stop selling a "license to use" the software, and just charge for the media. As it is, the media charge is a tiny component of the total price, most of which is the software license.
A failure of the media does not invalidate that license to use the software.
And your car analogy. I get my car fixed when it is broken... I even get the tires fixed. You can fix, to some extent, analogue audio and video tape if the damage is minor enough. (You'll lose a few frames around the splice, but hey, you can still watch the movie or listen to the album.)
But software is different from hardware. It's about the ideas, not the device containing them--that's why it is called software.
I've been using cdrecord and mkisofs on Mac OS X for a while now. The latest alphas of cdrtools have full Mac OS X support--if Apple Disc Burning works, you can use it. So that gets you MMC SCSI, FireWire, IDE and USB recorders all in one. The device addresses are... different, but if you've only got one drive that's no problem either.
With very little work, I was able to take Schily's newest libraries from cdrtools and drop it in to cdrdao to get that going too. All because my FireWire recorder rips better than my SCSI recorder, and I got tired of booting into Linux just to copy.
The one oddity is the drive sharing stuff. Because you don't start these programs until AFTER you insert the disc, you get Finder's offer to prepare the disc. You have to tell Finder to ignore it, then hit continue in the really annoying "second bad disc dialog".
I'm going to update the programs so they reserve the drive and wait for the recorder to do tray close, if I can figure out how. Maybe someone else already has; it works well enough that I haven't looked for updates recently.
The irony is, that I'm running IE6 and it's identified as follows:
You are using: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)
I truly despise browser makers that do that. I set up a trap on my webserver to catch Netscape 4 users, because some of my style sheets cause pages to be completely illegible in NS4. Of course, it caught IE 6 too, which handled the page more-or-less OK. (IE has a LOT of work to do with CSS2, Mozilla is OK, Opera has different bugs from Mozilla (and I like Mozilla's failings better). But Netscape 4 is just WRONG with style-sheets.)
It's ironic, but due to NS4's very serious CSS flaws, sites can look better in Netscape 3 than they do in 4--NS 3 ignores the things it doesn't understand, NS 4 doesn't understand it either, but it tries and gets it just... so... very... wrong.
Slashdot Mirror
Sorry, but this just isn't Microsoft's fault. It goes back MUCH farther than that, to the early days of FTP. In fact, on non-free UNIX-alikes, "TYPE A" remains the default to this day. I've had "alias ftp=ftp -inv" for years, long before Microsoft had an IP stack, let alone an FTP server. (Well, maybe in Minix....)
I'm not even sure the server gets to decide the default type. If your client doesn't request TYPE I, it shouldn't count on anything. I'd look up the RFCs, but it's lunch time.
Keep in mind when the ARPANET was young, most machines weren't binary-compatible. (Of the few that were wired together.) Binary transfers were rare. With compressed and archive formats, that's no longer the case, even for plain text... but the defaults remain.
So did I, so I sent it to abuse@rational.com and abuse@ibm.com.
But for fun, do price-matching between Future Shop and Best Buy.
"You know Best Buy is selling this $20 cheaper... uh, don't they own you guys now?"
Watching the guy completely fail to be able to access Best Buy's website from the in-store computer was worth it. The computer worked fine, he just couldn't type "www.bestbuycanada.ca". (It wouldn't have made a difference, Best Buy Canada doesn't have a catalog on the 'web.)
Of course, Logitech's new trackball tracks at 60 degrees to where it is comfortable to hold, so after all the price-matching fun I got my money back. Hehehehe.
They don't make 'em like they used to.
"display" is a CSS thing, which would mean 'style="display:none"' in this case. "hidden" is a type of input widget, along with "text", "password", "checkbox" and has been in HTML for ages, at least since 2.0. (I didn't do any work with forms in older versions of HTML.)
t tributes.)
If you rely on CSS to hide your hidden field (I've used them for session IDs on generated pages, instead of using cookies), then it will "unhide" on non-CSS aware clients. Also, when unhidden, it will probably also be input-able; readonly is pretty new. "type=hidden" gets all the semantics right all at once (style="display:none" type="string" readonly), and it is valid HTML 2.0.
Graceful degredation and all that. Of course, one must also fail to trust the client, so be careful what goes in anything the client sends back to you, hidden or not.
(And you can omit the quotes on attribute values in certain cases; check http://www.w3.org/TR/html401/intro/sgmltut.html#a
(I don't how the G4 feels about misaligned loads--the only other PowerPCs I can get to are IBM POWER4.)
NVidia has a kernal patch so that they isolate the kernel-version-dependent code and can distribute that as source. It simply glues in the binary-only part, and allows you to recompile it if you patch your kernel. That's why there's so many kernel glue layer packages from NVidia, they're all built for a very specific kernel version.
But you can save some typing, leave off the ",1", tape drive is the default. Load much faster. Of course, shift-STOP will load and run the next thing on tape.
And if you do go see a movie, you'll find out when you got back the program was _earlier_ on the tape, and now you have to rewind and start all over.
There's got to be something better I could be using these brain cells for.
You can't get the same density at higher rotational speeds; it requires more "design" in the heads and sense amps to read the signals reliably enough. So you have to trade capacity for high RPM... and you have to hope the density is _enough_ better that you still get a transfer-rate advantage with the higher RPM.
Since they don't even make SCSI drives with the same rotational speed as IDE these days, you can't compare, there are too many differences. So you get the "new research" model with SCSI, and once they have it figured out, the IDE versions come out.
So you've got more than just www.scsita.org putting up the prices, you've also got the bleeding-edge factor too.
I used to like SCSI. I still like the command set, and since FireWire supports serial SCSI (SBP-2), I get the best of both: cheap IDE drives in a FireWire box and I can use SCSI drivers to control them.
23-pin D-sub? Amiga floppy and video; very unstandard size connector. Hacksaw and a D-25 was my usual solution....
Don't forget 15-pin D-subs are (OK, were) also used for Ethernet AUI connectors; also hardly ever seen these days. I would expect some cases built to support on-board networking to have these.
Those RT connectors carried over into the early RS/6000s, which meant every machine came with a "book" containing 1' long cables with the weird connector on one side and the standard connector on the other. (Plus the wrap plugs for self-diagnosis.)
I finally decided to drop iCab and switch to Mozilla. I really like a lot of things in iCab, but I'm using CSS2 for all my documents now, and iCab just doesn't do CSS2.
I would like to see a Gecko renderer with the iCab user interface and filtering features.
*wavy lines*
I remember pie menus from way back in '86, there were some articles in the Amiga mags about their benefits. The only problem was, the machines of the day didn't have enough CPU grunt to do the rendering, and you couldn't fit much information around the pie because everyone's pixels were so big. On a 640x200 display, there just isn't room.
*senil yvaw*
One thing pie menus do is give an "infinite depth" target area, like placing icons right at the edge of the screen.
This allows the novice user to make a large motion, to where the pie-slice is "wider", and hit the item. But as you get used to the menu, you can make a smaller motion and hit the narrower part of your choice.
Because this implementation pops up submenus for some directions, it doesn't _quite_ reach that ideal. But now I don't know if I want to learn gestures or pies. Maybe both together; I've got lots of mouse buttons....
But you still retain the advantage all context-menus have in that the menu is "right here", instead of off at the top of the screen (and infinately tall) or at the top of the window.
Very spiffy. I've waited 16 years for this, it seems.
Those weren't seven independent processors, those were seven chips making up a single POWER CPU. The original POWER CPUs were NOT a microprocessor, they were multi-chip; 7 to 12 or so depending on the model. The functional units (decode, float, fixed, and so on) each had their own chip to live on. Even the POWER2 CPU was multi-chip, but they were all mounted in a single package (with the biggest heatsink I'd ever seen, long before Intel and AMD started playing MHz battles.)
That is why the PowerPC was a big deal in 1993; the POWER instruction architecture on a single-chip microprocessor.
There was a single-chip POWER CPU made for the 7011-220 through 7011-240, but nobody wants to think about those. The 7011-250 was a PowerPC 601, and worlds different from the other types in that model line.
Objective C uses a very different way of getting messages passed between objects. It doesn't have any of the binary interface issues C++ does, particularly name mangling and very touchy binary object layout issues.
It takes a bit of a twist to get your mind around Objective C, as it is very much a dynamic, run-time-determined kind of language.
I guess your ISP does not block port 25. And the recipient's server has to be set up to accept mail from dial-up pools; many do not. At least, when I tried that a couple of years ago, the server I tried to use wouldn't let me in from the dial-up-like block my DSL address is in.
And ISPs blocking port 25 is pretty common; mine's done it, so I can't access an authenticated SMTP server directly, I have to proxy to it on port 26 because some mail clients hard-code the port number.
None of this would be needed if spam wasn't a problem.
But what's cute is with my USB Compact Flash reader it shows up as a SCSI drive, mount cp & umount. Even easier on Mac and Windows--great way to transfer big files. And if you don't need to use the card in your camera, you could even partition it however you want and have different kinds of filesystems.
Single-track spiral media is just dumb for computer data. But we're stuck with it, it seems.
You pick the words to go with the cool name.
Enfoce In-Order Execution of Input/Output anyone?
You know someone wanted to be able to have an instruction called "eieio".
0ddc 0ffe ebad f00d
Older US discs would also jump to the feature on auto-play. You had to press "menu" if you really wanted it. MGM started offering pan & scan and widescreen on the same side, so they needed a menu before the feature--though it would timeout and default to widescreen if you left it.
Now they're getting silly.
What size is an int? In my life I've been on machines where it's 8, 16, 32 and 64 bits. What will it be in the future?
It's even better than that. When you install on a 64-bit SPARC (or AIX on a 64-bit POWER or PowerPC), you can still run 32-bit code fully natively. No emulation.
You have to craft your OS carefully to allow 32- and 64-bit applications to co-exist on the system. But you usually have a kernel that runs in one mode and the appropriate userland code to call into it from either.
The easiest way to get started, of course, is to have a 32-bit kernel support 64-bit user tasks. Depending on the architecture, there may not be any advantage to re-building the kernel in 64-bit native. ("modprobe 64bit" anyone? AIX has the equivelent to turn on 64-bit user processes in a 32-bit kernel.)
Of course, the SPARC and PowerPCs have been doing this for... uh... how many years now? 6 at least, right? Intel going to a "pure" 64-bit architecture seemed to me to be just plain arrogant. "Forget your old code, buy our new 100-watt CPUs!"
For me to accept this, they must stop selling a "license to use" the software, and just charge for the media. As it is, the media charge is a tiny component of the total price, most of which is the software license.
A failure of the media does not invalidate that license to use the software.
And your car analogy. I get my car fixed when it is broken... I even get the tires fixed. You can fix, to some extent, analogue audio and video tape if the damage is minor enough. (You'll lose a few frames around the splice, but hey, you can still watch the movie or listen to the album.)
But software is different from hardware. It's about the ideas, not the device containing them--that's why it is called software.
Ooooh, got to get one of those for Tempest!
I've been using cdrecord and mkisofs on Mac OS X for a while now. The latest alphas of cdrtools have full Mac OS X support--if Apple Disc Burning works, you can use it. So that gets you MMC SCSI, FireWire, IDE and USB recorders all in one. The device addresses are... different, but if you've only got one drive that's no problem either.
With very little work, I was able to take Schily's newest libraries from cdrtools and drop it in to cdrdao to get that going too. All because my FireWire recorder rips better than my SCSI recorder, and I got tired of booting into Linux just to copy.
The one oddity is the drive sharing stuff. Because you don't start these programs until AFTER you insert the disc, you get Finder's offer to prepare the disc. You have to tell Finder to ignore it, then hit continue in the really annoying "second bad disc dialog".
I'm going to update the programs so they reserve the drive and wait for the recorder to do tray close, if I can figure out how. Maybe someone else already has; it works well enough that I haven't looked for updates recently.
But just try and rent Photoshop for $5/day.
It's ironic, but due to NS4's very serious CSS flaws, sites can look better in Netscape 3 than they do in 4--NS 3 ignores the things it doesn't understand, NS 4 doesn't understand it either, but it tries and gets it just... so... very... wrong.