At one point I worked at a company that had one particular project team trying out a monetary penalty for uttering the words "it works on my machine". (The fine was something like $5 and went into a project team party fund.) Another clever penalty was that if you broke the build, not only was there a small fine, but your mug shot photo was taped to the window of the main project meeting room, as a sort of "build breaker's wall of shame". When I was walking by one day I saw one person's picture up there at least 4 times.:)
Well, I think the expected Unix way to do stuff like this is to just lean really hard on the kernel's disk cache. If you don't *have* to have it in RAM 100% of the time, use a decent algorithm / data structure that lets you know where exactly it is on disk, and ask the OS for it every time you need it. If you need it a lot (which could vary based on user activity), it'll be in cache anyway, so no biggie. (Whether you have to have it in RAM is a judgement call that could be made based on performance testing.)
But, you get into trouble when using a 1600x1200x32-bit screen that needs 8MB just for a single frame buffer. Add double-buffering and backing stores for every window and all of a sudden you're talking about a lot of memory required just to draw a bunch of windows on the screen. (Don't forget that gorgeous desktop wallpaper!) Open 20 fullscreen windows, and what do you expect is going to happen? Oh my god, why is it taking 160MB? (The/. chorus chants "bloat! bloat! bloat!") Well, because it has to, unless you want to wait as every window is redrawn from scratch every time you move something.
A side note: I've noticed this on Windows particularly, but it probably applies to OS X as well: minimize windows as much as you can; it makes your system more responsive. It sounds dumb, but using [Windows]-D often means that moving or minimizing a window doesn't result in 50 overlapping windows (and/or the apps they belong to) needing to be un-swapped so they can be redrawn now that you've minimized the topmost window. I'm not sure which apps choose to repaint "from scratch" vs. which ones just keep a private bitmap of what's supposed to be in the window, but subjectively, it seems like a lot of apps have gone in the "keep a private bitmap" direction instead of repainting every time. If so, that means the developers have chosen to use memory instead of CPU, which backfires if the window has to be repainted from a bitmap that's been swapped out. Or maybe it's just that the application code is swapped out and needs to be swapped back in before it can repaint the window... anybody know more about this?
The Safari leak you describe is something I experienced too: Use it heavily for a few hours (lots of open windows, lots of downloads, lots of tabs), then close all windows. Run "top" in Terminal. It's taking 1GB of memory (including swap), whoa! Quit, and it can take 30 seconds or more. Ouch.
OTOH, force-quitting it takes about 1 second. So, I wouldn't blame the OS - blame Safari for having a crazy memory leak that makes it take a long time to gracefully shut down (and at 100% CPU, thrashing the heck out of the swap volume, so basically the Mac is ultra slow for the time being). Force-quit or kill -9 gets rid of it much more quickly.
Now that you mention it, I haven't encountered this problem in the last couple of weeks. One of those "delete these preferences files" articles from www.info.apple.com seems to have made it go away permanently.
Parent is the most generic, uninformed anti-Mac troll I've seen in quite a while.
Let's see if we can count the cliches, shall we? 1) Nobody sells Macs in stores 2) Macs are incredibly expensive 2a) Steve Jobs could just change his mind and make Macs as cheap as PCs, but he chooses not to. Damn him!! 3) Everybody is abandoning the Mac platform for PCs because they're cheaper 4) Apple should just give up on PPC and port to x86. It would be easy and it would solve the price and compatibility problems, because the PPC is the reason that Macs are sooo expensive 5) Apple needs to make game developers port more games to the Mac 6) Comparison to much, much less successful OSs that failed for totally different reasons from the ones listed as problems with Apple 7) It's all going to be over soon... but there's a chance that Apple will survive, if it just abandons the PPC and takes on Dell and HP and Compaq and Microsoft and all the Linux distributors simultaneously and tries to beat them on their own turf. None of those companies are a threat as long as Apple switches to x86.
Did I miss anything?
By the way, most of these arguments have been made ever since the Macintosh was introduced. Apple has been "doomed to go out of business real soon now" for almost 20 years because of this same set of fatal and urgent issues. (Exception: it used to be 68K instead of PPC.)
Seriously, it sucks that your G5 is crashy. I take it this is even with the 10.2.8 G5 update? I suppose if it is a firmware thing that they'll sort it out soon enough, but that's not much consolation when you've shelled out big bucks to get some serious work done and the thing crashes all the time.
(The best support site I've seen is www.macintouch.com, by the way. When there are things that are dodgy that vendors aren't really admitting, that's where I see them transformed from rumors and odd grumblings into actual reproducible patterns of failure and conversations about how to address them.)
>I like to piece mine together using exactly the parts I want.
I know this is heresy on/., but...
Don't you ever worry that this might be a colossal waste of your time? Just keeping on top of the nitty gritty details of each CPU that you might buy is a pain, but then you have to match that with a mobo, a fan, RAM, etc. etc.
Don't you actually have something that you're trying to accomplish with your PC, other than trying to hot rod it and show off to your friends?
Don't overlook the amount of time you then have to waste getting all the drivers to work, updating firmware, and on and on and on, tweaking it and getting it just perfect. You could spend your whole life just making a PC 100% up to date and patched... to what end?
I guess if you make $7 an hour and want to squeeze every last bit of performance out of a $500 PC, and you have lots and lots of free time on your hands to figure it all out, it kinda makes sense. But in my opinion if you need a PC you should just order a damn Dell and get on with the part where you actually USE it for something. You might spend a few extra bucks or get a setup that's not quite 100% bleeding edge, costing you 5% in system performance, but it's worth it if your time is worth anything.
Jesus H tap-dancing Christ. Please tell me you were trying to earn a (Score: 5, Funny).
Are you not aware of what a self-fulfilling prophecy that is? You won't donate to an organization that explicitly fights to protect your privacy, because you're afraid that they won't protect your privacy?
Are you also afraid to vote for anybody other than Bush because Ashcroft might find out and come gitcha?
You might as well be afraid to go to a restaurant during your lunch hour because you might not have enough time left over to eat at your desk.
Ask Be how successful the switch-to-x86 strategy is. Just because lots of folks wish they could play around with Jaguar on their trusty el cheapo white-box x86 machine doesn't make that a winning product strategy for Apple, as a software or hardware manufacturer. After all, Microsoft and Dell wouldn't exactly roll over and die the day that Apple decided to go x86. There are plenty of carcasses of x86 vendors that got stomped by Dell because they couldn't operate as efficiently as Dell does, and plenty of x86 based OSs that Microsoft has sent to the grave. If they were to switch, they wouldn't *become* Dell, they'd have to compete directly *against* Dell. As it is, they can force you to buy their hardware in order to get their OS, and thus they can make their OS work well on their hardware. And maybe, if IBM delivers the goods, they can keep up on the CPU front enough that the people who think human productivity is measured in GHz and spec_int won't jeer too loudly.
Apple made lots of other mistakes (no clones, high margins, very limited product availability among them) that have led them to their current market share position. OTOH, market share isn't profitability; Apple is profitable.
(bringing this back on-topic...) Sun, on the other hand, seems to lack a strategy altogether. Their core business is big mofo servers, but Linux is chipping away at that. IMO they should abandon the low-end (1 or 2 CPU boxes) entirely, which means dumping Cobalt gear; commit to both Linux on x86 and Solaris on SPARC in the midrange (overlapping products with different CPUs, maybe even offering CPU family as a configurable option in an otherwise identical box); and stick with SPARC on the high end. I personally find it hard to believe that now that grids exist, there are no problems that still require a big server. I'm not an expert on that but that's my hunch. If that's correct then Sun would do well to keep pushing SPARC into that niche, keep making premium-priced ultra high end servers, and keep investing in Solaris as the only OS that will actually work in that sort of environment. Linux is going to continue to cannibalize the midrange, so there's really nowhere else for Sun to go. Hanging around pretending that Linux isn't a threat because Solaris is better on the same hardware is just denial and will lead them to ruin.
Of course Sun can also make a bundle of dough by selling storage and backup and other related hardware. I think they're already doing pretty well there.
Re:What's wrong with national IDs?
on
Beyond Fear
·
· Score: 1
Ever heard of a Social Security Number? Yeah, I too can find all sorts of documents that prove that companies and governmental organizations don't have a right to demand my SSN, but the recourse for that is that they simply deny you service / business. So you can live in a little shack in Montana and not give anybody your SSN, but in the real world, everybody requires it and will just tell you to sod off if you refuse. See, your rights haven't been abridged because no one is forcing you to leave your shack in Montana...
A privacy amendment would pretty much be the only way to counteract this IMO.
Re:What's wrong with national IDs?
on
Beyond Fear
·
· Score: 1
You've just done a great job of supporting the anti-privacy argument - only criminals have anything to hide, so privacy is bad.
Next time, pay the extra $1 (well, $1.25 now) to use another turnstile that actually gives you a transfer. Or, if you like taking the risk, at least go to court when you get a ticket. Failing to pay a ticket or show up for a court date just because you couldn't be bothered to pay $1 is pretty pathetic, and is far from a valid reason for not wanting a national ID card.
>Didn't you know? Face to face is better for business. Ah, I didn't realize you meant *video*phones. Don't use IM, use videophones. Practical advice; kudos!
>How slow is your network that it's not realtime on your intranet? "Didn't you know?" Some companies are so large that they have _multiple offices_. Some companies even let people work from outside of the office!
>I've never worked anywhere where IM was used internally Well, that's your problem. It's a really nice thing to have, and you can't really have multiple different conversations at the same time on the phone like you can with IM. Also, you can't send links over the phone effectively.
Email is not a real-time medium, and everybody else seems to know this since they send all sorts of messages that don't need to be dealt with right this second. It's not practical to mix realtime messages with asynchronous messages, because employees will always have to stop what they're doing and read the message... in order to find out if they actually needed to stop and read the message right now. People can be much more productive if they can *ignore* email for large periods of time, concentrate on work, and only answer phone calls and IMs which are real-time media that carry with them an implied urgency. And no, IM is not interchangeable with the phone; phones carry even higher urgency and are more disruptive. It's best to have all three.
No, it's the way the Apple ][ series worked (and PCs, for that matter), except with older, fixed hardware. We all migrated to hard disks for a number of reasons including SPEED and storage. Nowadays we have more hardware, which means more drivers, bigger OSs, which mean more stuff to load when booting, and much much bigger games. Call it "bloat" because it's/. and we love to forget that we assume that all our hardware will be supported and all our favorite apps will be included in our OS, instead of there being just 1 video card (ooo, 80 column card!) and a stack of floppies containing one (or part of one) app each. But no matter what you call it, the code we run is bigger, to go with our bigger media.
I can't wait to boot one CD just to load the OS into a giant RAM disk, and then another CD for the game. Or maybe I can notch the CD with a hole puncher and flip it over so I can store twice as much info on one disk?
In case you're curious to see what the educational discount price actually is, you can do so through the Apple Store very easily:
- go to http://store.apple.com/ (get redirected to the store home page) - on the left side under "Interests...' click Education - under "Shop for your School" pick "Find Your College or University" - Pick "Virginia" and enter "Blacksburg", then click "Find" - Select "Virginia Polytechnic Inst. & State Univ" (the only option) and click Continue Voila, you get redirected to the Apple Store home page, but this time you are seeing the educational discount prices that VT departments would get.
Pick the G5 dual-CPU, no customization, and put Qty 1100 in your shopping cart. Click Update Subtotal. $2,968,900.00, will that be Visa or Mastercard?
I'm still trying to mirror a mirror so I can't read this for myself... but, do these systems even have enough memory that the "not really 64-bit" gripe is even relevant? IIRC, if they have 2GB or less, it wouldn't.
Do you mean if I poison your DNS server with a fake FQDN of "intranet.internal" that I can start harvesting passwords from your VPN users, as they log into my faked-up copy of your intranet server? Can you say "man in the middle attack"? Only when your help desk gets a flood of calls asking why the intranet server doesn't seem to have the right info on it will you know that something is amiss.
Actually, I don't think that if a user searches for (intentional typo:) intarnet.intrenal.foo.org, and there is an authoritative source for foo.org, that Verisign can even hijack that. It'll go to your server for foo.org where it will fail. Only if people are looking servers up by "intranet.internal' and relying on domain search paths to be set properly will Verisign get to intercept it. And for that matter, any sec^H^H^Hobscurity you got from not publishing that DNS entry is diminished since DNS queries for "intranet.internal" will be sent across the internet every time somebody does that. And it's slower to wait for it to fail than to just do the FQDN search. And if you use SSL the browser will bitch anyway because the cert was signed for "intranet.internal.foo.org", not "intranet.internal".
Am I not understanding something about your setup, or Windows, or something sneaky Verisign is doing? It sounds like the problem you described is more a problem with your DNS setup than with anything Verisign is doing.
1- get the RPM, which won't rebuild successfully because you don't have GTK 2 installed. ncftpget ftp://ftp.easynet.be/openssh/portable/rpm/SRPMS/op enssh-3.7p1-1.src.rpm (remove the stupid space that/. added to "openssh" in the above URL where it says "op enssh")
2- install the sources so you can edit the RPM spec file rpm -ihv openssh-3.7p1-1.src.rpm
3- cd to where the spec file was installed cd/usr/src/redhat/SPECS/
4- patch the spec file so that it doesn't depend on GTK 2 anymore. (Or just manually change the 1 on line 24 to a 0.) patch EOF --- openssh.spec-orig Wed Sep 17 15:08:44 2003 +++ openssh.spec Wed Sep 17 15:39:04 2003 @@ -20,8 +20,8 @@
# Do we want smartcard support (1=yes 0=no)
%define scard 0
-# Use GTK2 instead of GNOME in gnome-ssh-askpass -%define gtk2 1 +# Use GNOME instead of GTK2 in gnome-ssh-askpass +%define gtk2 0
# Is this build for RHL 6.x?
%define build6x 0 EOF
5- build the SRPM with the new spec file. (It will appear in/usr/src/redhat/SRPMS so you might want to make sure the original isn't also in there.) rpm -bs openssh.spec
OK, now you can rebuild and install the SRPM... which is the part that failed before.
6- rebuild the binary RPM (This takes 3 1/2 minutes on my PIII-500 box) rpm --rebuild/usr/src/redhat/SRPMS/openssh-3.7p1-1.src.rpm
7- Check the last few lines of output. Hopefully about 15 lines from the end, it says: Wrote:/usr/src/redhat/RPMS/i386/openssh-3.7p1-1.i386.rpm Wrote:/usr/src/redhat/RPMS/i386/openssh-clients-3.7p1-1. i386.rpm Wrote:/usr/src/redhat/RPMS/i386/openssh-server-3.7p1-1.i 386.rpm Wrote:/usr/src/redhat/RPMS/i386/openssh-askpass-3.7p1-1. i386.rpm Wrote:/usr/src/redhat/RPMS/i386/openssh-askpass-gnome-3. 7p1-1.i386.rpm
Now you have binary RPMs that you can install. If this failed, you have some other problem to fix.
8- "Freshen" OpenSSH on your system using the binary RPM you just made cd/usr/src/redhat/RPMS/i386/ rpm -Fvh openssh-3.7p1-1.i386.rpm openssh-askpass-3.7p1-1.i386.rpm \
openssh-askpass-gnome-3.7p1-1.i386.rpm openssh-clients-3.7p1-1.i386.rpm \
openssh-server-3.7p1-1.i386.rpm
9- [Re]start sshd/etc/rc.d/init.d/sshd restart
10 - Raise arms above head and shout "Woo hoo!"
11- Undo any temporary measures (firewall rule changes) that you put in place to keep from getting 0wn3d:)
The innovation is that there's no wireless base station sitting on your desk, because it uses Bluetooth instead of infrared or some proprietary RF thingy. It's not earth shattering but I guess if I hadn't already committed to having a USB hub and a whole bunch of cables (that whole "Digital Hub" thing really does describe my laptop setup when it's plugged in at home) then I might care. But I don't because 1 more cable neatly routed under the desk isn't tragic. Switching batteries all the time is a PITA, though.
Rechargeable batteries drain on their own at an alarming rate - a few percent per day. Using them in long-term applications like remote controls or wireless keyboards is not a good application because they won't last as long. Rechargeable batteries are good for certain applications (like my beloved Rio 500 MP3 player) that kill batteries in a day or two but not for stuff like this... or the emergency flashlight that sits in a drawer for 6 months. It'll be mostly drained before you get around to turning it on.
So, try not to get too upset that Apple isn't recommending rechargeables for these gizmos. Disposables will last longer. And I agree, some kind of recharger dock would be better... perhaps USB power? Oh wait.
Good thinking! I agree. We should concentrate all our access control into a single layer that does everything. All traffic will have to pass through that layer. That will not only be faster but more flexible, and more robust in case there's a bug.
We shall call it Aleiob, which is short for All Eggs In One Basket. 'Cuz that makes _so_ much sense.
If you can't keep track of your network configuration, draw a picture, and/or write some scripts that generate product-specific rulesets from a centralized, abstracted ruleset.
When your company is based on hiring smart people right out of college, you end up with a lot of smart people, but they also [a] don't know anything about how the rest of the world has solved the same problems before and [b] don't think that the world could possibly have solved any problems before, or at least not as well as they could have. So you get a bunch of half-assed solutions to already-solved problems, but this time with key requirements overlooked, lots of bugs and design mistakes, and new terminology. We're all very fortunate (seriously) that billg and his underlings have finally started to see the light of interoperability, and started to phase out the flawed MS/IBM technologies in favor of stuff that the not-working-for-Bill smart people worked out.
There are still problems that have yet to be solved convincingly by anybody, of course, but I don't see MS being as magnanimous as Apple has suddenly started to be about trying to solve those problems in an open and standardized way. (In this case I'm referring to Zeroconf, Rendezvous, etc., which are so new and weird that I'm not sure it they're good solutions yet, but they seem to be, and they're definitely not the same sort of well-intentioned but ultimately flawed and inflexible technologies as stuff like the AppleTalk stack.)
Just because you don't want to see yourself / your brother / your father naked doesn't mean that all men are ugly. Ever hear women complain about how the women that're always shown in bikinis / porn / movies aren't "real"? Same problem... Aunt Gertie may not be a hottie but that doesn't mean all women are icky. Maybe making 1000 pictures of the most "ideal".01% of women makes it seem like women are better looking on average...
Slashdot Mirror
This is a variation of the well-known mantra of developers everywhere:
:)
"IT WORKS ON MY MACHINE"
See also: http://c2.com/cgi/wiki?BugsArentVoodoo
At one point I worked at a company that had one particular project team trying out a monetary penalty for uttering the words "it works on my machine". (The fine was something like $5 and went into a project team party fund.) Another clever penalty was that if you broke the build, not only was there a small fine, but your mug shot photo was taped to the window of the main project meeting room, as a sort of "build breaker's wall of shame". When I was walking by one day I saw one person's picture up there at least 4 times.
Well, I think the expected Unix way to do stuff like this is to just lean really hard on the kernel's disk cache. If you don't *have* to have it in RAM 100% of the time, use a decent algorithm / data structure that lets you know where exactly it is on disk, and ask the OS for it every time you need it. If you need it a lot (which could vary based on user activity), it'll be in cache anyway, so no biggie. (Whether you have to have it in RAM is a judgement call that could be made based on performance testing.)
/. chorus chants "bloat! bloat! bloat!") Well, because it has to, unless you want to wait as every window is redrawn from scratch every time you move something.
But, you get into trouble when using a 1600x1200x32-bit screen that needs 8MB just for a single frame buffer. Add double-buffering and backing stores for every window and all of a sudden you're talking about a lot of memory required just to draw a bunch of windows on the screen. (Don't forget that gorgeous desktop wallpaper!) Open 20 fullscreen windows, and what do you expect is going to happen? Oh my god, why is it taking 160MB? (The
A side note: I've noticed this on Windows particularly, but it probably applies to OS X as well: minimize windows as much as you can; it makes your system more responsive. It sounds dumb, but using [Windows]-D often means that moving or minimizing a window doesn't result in 50 overlapping windows (and/or the apps they belong to) needing to be un-swapped so they can be redrawn now that you've minimized the topmost window. I'm not sure which apps choose to repaint "from scratch" vs. which ones just keep a private bitmap of what's supposed to be in the window, but subjectively, it seems like a lot of apps have gone in the "keep a private bitmap" direction instead of repainting every time. If so, that means the developers have chosen to use memory instead of CPU, which backfires if the window has to be repainted from a bitmap that's been swapped out. Or maybe it's just that the application code is swapped out and needs to be swapped back in before it can repaint the window... anybody know more about this?
The Safari leak you describe is something I experienced too: Use it heavily for a few hours (lots of open windows, lots of downloads, lots of tabs), then close all windows. Run "top" in Terminal. It's taking 1GB of memory (including swap), whoa! Quit, and it can take 30 seconds or more. Ouch.
OTOH, force-quitting it takes about 1 second. So, I wouldn't blame the OS - blame Safari for having a crazy memory leak that makes it take a long time to gracefully shut down (and at 100% CPU, thrashing the heck out of the swap volume, so basically the Mac is ultra slow for the time being). Force-quit or kill -9 gets rid of it much more quickly.
Now that you mention it, I haven't encountered this problem in the last couple of weeks. One of those "delete these preferences files" articles from www.info.apple.com seems to have made it go away permanently.
Why isn't RMS lobbying for: ...?
- GNU FreeBSD
- GNU OpenBSD
- GNU NetBSD
- GNU Mac OS X
There has been a rumor for some time that Apple maintains internal builds of OS X on x86 hardware "just in case".
But, imagine how wonderful it would be if you could have the same level of driver availability you have for Solaris x86, but with a Mac UI.
You'd uninstall it as soon as you realized that none of your hardware was recognized, and that the vendors weren't ever going to do anything about it.
Why would you spend the extra money to get a Mac and then eliminate the whole reason for buying a Mac in the first place by running Linux on it?
Parent is the most generic, uninformed anti-Mac troll I've seen in quite a while.
Let's see if we can count the cliches, shall we?
1) Nobody sells Macs in stores
2) Macs are incredibly expensive
2a) Steve Jobs could just change his mind and make Macs as cheap as PCs, but he chooses not to. Damn him!!
3) Everybody is abandoning the Mac platform for PCs because they're cheaper
4) Apple should just give up on PPC and port to x86. It would be easy and it would solve the price and compatibility problems, because the PPC is the reason that Macs are sooo expensive
5) Apple needs to make game developers port more games to the Mac
6) Comparison to much, much less successful OSs that failed for totally different reasons from the ones listed as problems with Apple
7) It's all going to be over soon... but there's a chance that Apple will survive, if it just abandons the PPC and takes on Dell and HP and Compaq and Microsoft and all the Linux distributors simultaneously and tries to beat them on their own turf. None of those companies are a threat as long as Apple switches to x86.
Did I miss anything?
By the way, most of these arguments have been made ever since the Macintosh was introduced. Apple has been "doomed to go out of business real soon now" for almost 20 years because of this same set of fatal and urgent issues. (Exception: it used to be 68K instead of PPC.)
Just because nobody else has said it yet...
"OK, then give it to me, man, I'll take it!"
Seriously, it sucks that your G5 is crashy. I take it this is even with the 10.2.8 G5 update? I suppose if it is a firmware thing that they'll sort it out soon enough, but that's not much consolation when you've shelled out big bucks to get some serious work done and the thing crashes all the time.
(The best support site I've seen is www.macintouch.com, by the way. When there are things that are dodgy that vendors aren't really admitting, that's where I see them transformed from rumors and odd grumblings into actual reproducible patterns of failure and conversations about how to address them.)
>I like to piece mine together using exactly the parts I want.
/., but...
I know this is heresy on
Don't you ever worry that this might be a colossal waste of your time? Just keeping on top of the nitty gritty details of each CPU that you might buy is a pain, but then you have to match that with a mobo, a fan, RAM, etc. etc.
Don't you actually have something that you're trying to accomplish with your PC, other than trying to hot rod it and show off to your friends?
Don't overlook the amount of time you then have to waste getting all the drivers to work, updating firmware, and on and on and on, tweaking it and getting it just perfect. You could spend your whole life just making a PC 100% up to date and patched... to what end?
I guess if you make $7 an hour and want to squeeze every last bit of performance out of a $500 PC, and you have lots and lots of free time on your hands to figure it all out, it kinda makes sense. But in my opinion if you need a PC you should just order a damn Dell and get on with the part where you actually USE it for something. You might spend a few extra bucks or get a setup that's not quite 100% bleeding edge, costing you 5% in system performance, but it's worth it if your time is worth anything.
Jesus H tap-dancing Christ. Please tell me you were trying to earn a (Score: 5, Funny).
Are you not aware of what a self-fulfilling prophecy that is? You won't donate to an organization that explicitly fights to protect your privacy, because you're afraid that they won't protect your privacy?
Are you also afraid to vote for anybody other than Bush because Ashcroft might find out and come gitcha?
You might as well be afraid to go to a restaurant during your lunch hour because you might not have enough time left over to eat at your desk.
Ask Be how successful the switch-to-x86 strategy is. Just because lots of folks wish they could play around with Jaguar on their trusty el cheapo white-box x86 machine doesn't make that a winning product strategy for Apple, as a software or hardware manufacturer. After all, Microsoft and Dell wouldn't exactly roll over and die the day that Apple decided to go x86. There are plenty of carcasses of x86 vendors that got stomped by Dell because they couldn't operate as efficiently as Dell does, and plenty of x86 based OSs that Microsoft has sent to the grave. If they were to switch, they wouldn't *become* Dell, they'd have to compete directly *against* Dell. As it is, they can force you to buy their hardware in order to get their OS, and thus they can make their OS work well on their hardware. And maybe, if IBM delivers the goods, they can keep up on the CPU front enough that the people who think human productivity is measured in GHz and spec_int won't jeer too loudly.
Apple made lots of other mistakes (no clones, high margins, very limited product availability among them) that have led them to their current market share position. OTOH, market share isn't profitability; Apple is profitable.
(bringing this back on-topic...)
Sun, on the other hand, seems to lack a strategy altogether. Their core business is big mofo servers, but Linux is chipping away at that. IMO they should abandon the low-end (1 or 2 CPU boxes) entirely, which means dumping Cobalt gear; commit to both Linux on x86 and Solaris on SPARC in the midrange (overlapping products with different CPUs, maybe even offering CPU family as a configurable option in an otherwise identical box); and stick with SPARC on the high end. I personally find it hard to believe that now that grids exist, there are no problems that still require a big server. I'm not an expert on that but that's my hunch. If that's correct then Sun would do well to keep pushing SPARC into that niche, keep making premium-priced ultra high end servers, and keep investing in Solaris as the only OS that will actually work in that sort of environment. Linux is going to continue to cannibalize the midrange, so there's really nowhere else for Sun to go. Hanging around pretending that Linux isn't a threat because Solaris is better on the same hardware is just denial and will lead them to ruin.
Of course Sun can also make a bundle of dough by selling storage and backup and other related hardware. I think they're already doing pretty well there.
Ever heard of a Social Security Number? Yeah, I too can find all sorts of documents that prove that companies and governmental organizations don't have a right to demand my SSN, but the recourse for that is that they simply deny you service / business. So you can live in a little shack in Montana and not give anybody your SSN, but in the real world, everybody requires it and will just tell you to sod off if you refuse. See, your rights haven't been abridged because no one is forcing you to leave your shack in Montana...
A privacy amendment would pretty much be the only way to counteract this IMO.
You've just done a great job of supporting the anti-privacy argument - only criminals have anything to hide, so privacy is bad.
Next time, pay the extra $1 (well, $1.25 now) to use another turnstile that actually gives you a transfer. Or, if you like taking the risk, at least go to court when you get a ticket. Failing to pay a ticket or show up for a court date just because you couldn't be bothered to pay $1 is pretty pathetic, and is far from a valid reason for not wanting a national ID card.
>Didn't you know? Face to face is better for business.
Ah, I didn't realize you meant *video*phones. Don't use IM, use videophones. Practical advice; kudos!
>How slow is your network that it's not realtime on your intranet?
"Didn't you know?" Some companies are so large that they have _multiple offices_. Some companies even let people work from outside of the office!
>I've never worked anywhere where IM was used internally
Well, that's your problem. It's a really nice thing to have, and you can't really have multiple different conversations at the same time on the phone like you can with IM. Also, you can't send links over the phone effectively.
Email is not a real-time medium, and everybody else seems to know this since they send all sorts of messages that don't need to be dealt with right this second. It's not practical to mix realtime messages with asynchronous messages, because employees will always have to stop what they're doing and read the message... in order to find out if they actually needed to stop and read the message right now. People can be much more productive if they can *ignore* email for large periods of time, concentrate on work, and only answer phone calls and IMs which are real-time media that carry with them an implied urgency. And no, IM is not interchangeable with the phone; phones carry even higher urgency and are more disruptive. It's best to have all three.
>It is a revolutionary way of using PCs.
/. and we love to forget that we assume that all our hardware will be supported and all our favorite apps will be included in our OS, instead of there being just 1 video card (ooo, 80 column card!) and a stack of floppies containing one (or part of one) app each. But no matter what you call it, the code we run is bigger, to go with our bigger media.
No, it's the way the Apple ][ series worked (and PCs, for that matter), except with older, fixed hardware. We all migrated to hard disks for a number of reasons including SPEED and storage. Nowadays we have more hardware, which means more drivers, bigger OSs, which mean more stuff to load when booting, and much much bigger games. Call it "bloat" because it's
I can't wait to boot one CD just to load the OS into a giant RAM disk, and then another CD for the game. Or maybe I can notch the CD with a hole puncher and flip it over so I can store twice as much info on one disk?
>Having dual 2ghz processors is not as fast as having one 4ghz
Right. It's much faster than having one 4GHz processor.
In case you're curious to see what the educational discount price actually is, you can do so through the Apple Store very easily:
- go to http://store.apple.com/
(get redirected to the store home page)
- on the left side under "Interests...' click Education
- under "Shop for your School" pick "Find Your College or University"
- Pick "Virginia" and enter "Blacksburg", then click "Find"
- Select "Virginia Polytechnic Inst. & State Univ" (the only option) and click Continue
Voila, you get redirected to the Apple Store home page, but this time you are seeing the educational discount prices that VT departments would get.
Pick the G5 dual-CPU, no customization, and put Qty 1100 in your shopping cart. Click Update Subtotal. $2,968,900.00, will that be Visa or Mastercard?
I'm still trying to mirror a mirror so I can't read this for myself... but, do these systems even have enough memory that the "not really 64-bit" gripe is even relevant? IIRC, if they have 2GB or less, it wouldn't.
Do you mean if I poison your DNS server with a fake FQDN of "intranet.internal" that I can start harvesting passwords from your VPN users, as they log into my faked-up copy of your intranet server? Can you say "man in the middle attack"? Only when your help desk gets a flood of calls asking why the intranet server doesn't seem to have the right info on it will you know that something is amiss.
Actually, I don't think that if a user searches for (intentional typo:) intarnet.intrenal.foo.org, and there is an authoritative source for foo.org, that Verisign can even hijack that. It'll go to your server for foo.org where it will fail. Only if people are looking servers up by "intranet.internal' and relying on domain search paths to be set properly will Verisign get to intercept it. And for that matter, any sec^H^H^Hobscurity you got from not publishing that DNS entry is diminished since DNS queries for "intranet.internal" will be sent across the internet every time somebody does that. And it's slower to wait for it to fail than to just do the FQDN search. And if you use SSL the browser will bitch anyway because the cert was signed for "intranet.internal.foo.org", not "intranet.internal".
Am I not understanding something about your setup, or Windows, or something sneaky Verisign is doing? It sounds like the problem you described is more a problem with your DNS setup than with anything Verisign is doing.
1- get the RPM, which won't rebuild successfully because you don't have GTK 2 installed.p enssh-3.7p1-1.src.rpm /. added to "openssh" in the above URL where it says "op enssh")
/usr/src/redhat/SPECS/
/usr/src/redhat/SRPMS so you might want to make sure the original isn't also in there.)
/usr/src/redhat/SRPMS/openssh-3.7p1-1.src.rpm
/usr/src/redhat/RPMS/i386/openssh-3.7p1-1.i386.rpm /usr/src/redhat/RPMS/i386/openssh-clients-3.7p1-1. i386.rpm /usr/src/redhat/RPMS/i386/openssh-server-3.7p1-1.i 386.rpm /usr/src/redhat/RPMS/i386/openssh-askpass-3.7p1-1. i386.rpm /usr/src/redhat/RPMS/i386/openssh-askpass-gnome-3. 7p1-1.i386.rpm
/usr/src/redhat/RPMS/i386/
/etc/rc.d/init.d/sshd restart
:)
ncftpget ftp://ftp.easynet.be/openssh/portable/rpm/SRPMS/o
(remove the stupid space that
2- install the sources so you can edit the RPM spec file
rpm -ihv openssh-3.7p1-1.src.rpm
3- cd to where the spec file was installed
cd
4- patch the spec file so that it doesn't depend on GTK 2 anymore. (Or just manually change the 1 on line 24 to a 0.)
patch EOF
--- openssh.spec-orig Wed Sep 17 15:08:44 2003
+++ openssh.spec Wed Sep 17 15:39:04 2003
@@ -20,8 +20,8 @@
# Do we want smartcard support (1=yes 0=no)
%define scard 0
-# Use GTK2 instead of GNOME in gnome-ssh-askpass
-%define gtk2 1
+# Use GNOME instead of GTK2 in gnome-ssh-askpass
+%define gtk2 0
# Is this build for RHL 6.x?
%define build6x 0
EOF
5- build the SRPM with the new spec file. (It will appear in
rpm -bs openssh.spec
OK, now you can rebuild and install the SRPM... which is the part that failed before.
6- rebuild the binary RPM (This takes 3 1/2 minutes on my PIII-500 box)
rpm --rebuild
7- Check the last few lines of output. Hopefully about 15 lines from the end, it says:
Wrote:
Wrote:
Wrote:
Wrote:
Wrote:
Now you have binary RPMs that you can install. If this failed, you have some other problem to fix.
8- "Freshen" OpenSSH on your system using the binary RPM you just made
cd
rpm -Fvh openssh-3.7p1-1.i386.rpm openssh-askpass-3.7p1-1.i386.rpm \
openssh-askpass-gnome-3.7p1-1.i386.rpm openssh-clients-3.7p1-1.i386.rpm \
openssh-server-3.7p1-1.i386.rpm
9- [Re]start sshd
10 - Raise arms above head and shout "Woo hoo!"
11- Undo any temporary measures (firewall rule changes) that you put in place to keep from getting 0wn3d
The innovation is that there's no wireless base station sitting on your desk, because it uses Bluetooth instead of infrared or some proprietary RF thingy. It's not earth shattering but I guess if I hadn't already committed to having a USB hub and a whole bunch of cables (that whole "Digital Hub" thing really does describe my laptop setup when it's plugged in at home) then I might care. But I don't because 1 more cable neatly routed under the desk isn't tragic. Switching batteries all the time is a PITA, though.
Rechargeable batteries drain on their own at an alarming rate - a few percent per day. Using them in long-term applications like remote controls or wireless keyboards is not a good application because they won't last as long. Rechargeable batteries are good for certain applications (like my beloved Rio 500 MP3 player) that kill batteries in a day or two but not for stuff like this... or the emergency flashlight that sits in a drawer for 6 months. It'll be mostly drained before you get around to turning it on.
So, try not to get too upset that Apple isn't recommending rechargeables for these gizmos. Disposables will last longer. And I agree, some kind of recharger dock would be better... perhaps USB power? Oh wait.
Good thinking! I agree. We should concentrate all our access control into a single layer that does everything. All traffic will have to pass through that layer. That will not only be faster but more flexible, and more robust in case there's a bug.
We shall call it Aleiob, which is short for All Eggs In One Basket. 'Cuz that makes _so_ much sense.
If you can't keep track of your network configuration, draw a picture, and/or write some scripts that generate product-specific rulesets from a centralized, abstracted ruleset.
No, it just sounds like Microsoft.
When your company is based on hiring smart people right out of college, you end up with a lot of smart people, but they also [a] don't know anything about how the rest of the world has solved the same problems before and [b] don't think that the world could possibly have solved any problems before, or at least not as well as they could have. So you get a bunch of half-assed solutions to already-solved problems, but this time with key requirements overlooked, lots of bugs and design mistakes, and new terminology. We're all very fortunate (seriously) that billg and his underlings have finally started to see the light of interoperability, and started to phase out the flawed MS/IBM technologies in favor of stuff that the not-working-for-Bill smart people worked out.
There are still problems that have yet to be solved convincingly by anybody, of course, but I don't see MS being as magnanimous as Apple has suddenly started to be about trying to solve those problems in an open and standardized way. (In this case I'm referring to Zeroconf, Rendezvous, etc., which are so new and weird that I'm not sure it they're good solutions yet, but they seem to be, and they're definitely not the same sort of well-intentioned but ultimately flawed and inflexible technologies as stuff like the AppleTalk stack.)
>guys don't really look good naked.
.01% of women makes it seem like women are better looking on average...
Tell that to Michelangelo.
Just because you don't want to see yourself / your brother / your father naked doesn't mean that all men are ugly. Ever hear women complain about how the women that're always shown in bikinis / porn / movies aren't "real"? Same problem... Aunt Gertie may not be a hottie but that doesn't mean all women are icky. Maybe making 1000 pictures of the most "ideal"