Exchange is good software. It Just Works. And it performs exceptionally well.
If by "Just Works" you mean "Barely Works", sure. Our tuned [1] Exchange server was absolutely crawling until I put a much-lower-specced Postfix/ClamAV/Greylist server in front of it to take the brunt of what the Internet was throwing at us.
Of course, it actually works reasonably well now (since it only has to handle pre-filtered traffic), so maybe that's what you meant: a firewalled Exchange is adequate in many cases. If so, then I apologize and withdraw my clarification.
[1] Our Windows IT guy actually knows what he's doing - if he says it's tuned, then it is.
They're talking about 4 or 8 cores on a single CPU, which might be nice for blades but not so useful for a laptop or a gamer.
...as long as they in the habit of only running one program at a time. I personally can't wait to transcode DVDs in the background while still enjoying a nimble, responsive desktop. Multi-core/CPU is pretty nice in general usage, not just as server iron.
(5*10^8)/(5*10^5) = 1000 watts per home. As long as everyone uses no more than 10 bright lightbulbs simultaneously, your estimate is reasonable. Throw in a few stoves, air conditioners, and home computers and that number drops very quickly. Anybody know what the average usage per household is in America?
Theyboth list quite a few ports, but the trick is in deciding which list is actually longer. If you count complete hardware platforms, then it looks like NetBSD might take the lead since, in my opinion, just booting the Linux kernel doesn't really qualify as "running Linux". That's a point that has to be decided, though. On the other hand, if you count CPU architectures instead, then Linux might be ahead. That depends on how loosely you aggregate similar chips - for example, is "Intel IA32-compatibles (Cyrix MediaGX, STMicroelectronics STPC, ZF Micro ZFx86)", listed under "Diverse PDA / embedded / microcontroller / router devices", really different than "Intel IA32 family"?
My point is that it's not entirely clear which OS supports more platforms, since "supports" and "platforms" are both variables that would need to be nailed down before the conversation even begins. My own first impression is that NetBSD is still the winner, since you can actually boot into each of its listed platforms, install software from pkgsrc, and generally treat them as equals except for the obvious performance differences. Others could effectively argue the opposite, I'm sure.
This is the average memory consumption [...] 1,746,752 bytes
Let me be the first to say that your benchmarking methodology sucks. Does that 1.7MB of RAM consist of 1.6MB of constant overhead +.1MB of script usage? Who knows - you certainly don't seem to have bothered to test it. Does eAccelerator use the extra memory for results caching that result in higher throughput and overall lower resource usage? Again, you don't seem to have gotten past "it's teh bloat" so we have no way to venture a guess based on your numbers.
Frankly, if you measure performance in terms of "module
A uses 770KB of RAM, but module B only uses 320KB", then PHP is clearly the wrong language for you. I don't particularly like PHP, either, but your "reasons" certainly would not have convinced me if I felt otherwise.
Re:To buy or not to buy, the reviewer doesn't know
on
Spring Into PHP 5
·
· Score: 2, Interesting
This book is clearly not intended to be one of those PHP + MySQL combo books that have proven so popular during the past few years.
[...]
Another potential point of criticism could be that the book does not adequately explain how to use PHP with the various available database systems, only covering MySQL (the industry's favorite for use with PHP).
But yet it does explain MySQL? Which is it? I'm not going to buy the book if the reviewer can't figure it out.
In other words, the book isn't centered around building DB-driven apps using PHP+MySQL, even if it does discuss the topic. The reviewer wishes it also had chapters on PHP+PostgreSQL or PHP+Oracle, but it doesn't. Seems pretty reasonable and consistent to me.
Where do you live that you have to chase deer? We can't keep 'em away around here. Or maybe your idea of "chase" is "scream like a little girl after the animal jumps directly into your velocity vector and immediately freezes so as to avoid doing less than the maximum possible amount of damage to your vehicle and its occupants"? If that's the case, then I chase deer all the damn time.
If law and order ever breaks down where I live, I'm mounting an infrared-aimed chain gun and snowplow to the wife's minivan.
HD-DVDs keep an internal list of watermarks that have been "revoked", and every new HD-DVD printed will contain an up-to-date copy of that "revoked" list which the HD-DVD player must update every time you put in an HD-DVD.
Sweet! Assuming that the watermark is at least 128 bits (surely it'd take at least that much?), it only takes 68 million cracked DVDs to fill a gig of RAM with nothing but the revoked key list. If that sounds like a lot, consider that US DVD sales were about $18 billion dollars last year. Assuming $15/DVD gives roughly 1.2 billion DVDs. If 1% of those disks are ripped, released, and revoked, then that 1 gig gets used in six years.
Now, I doubt that any new DVD player is going to come with 1GB of nonvolatile memory in the near future (I don't see them carrying hard drives any time soon, and RAM would drive the product price up). If they had 256MB of revocation list memory, then they only get 1.5 years of use before it overflows.
In other words, the industry is actually providing an incentive for us to violate copyrights. The sooner we meet our quota, the sooner we overload the DRM in the hardware we've bought. Ironically, this is one of the few things they could do to make me want to make illicit copies. I have a pretty big collection of movies that I bought and paid for - I have no problem forking over a few bucks for entertainment - but this is enough to switch me to the Dark Side (from their point of view). Good going, fellas.
Basically, her union didn't do jack. As far as I can tell, they just collected dues and used them to finance political candidates. I don't think I ever heard her telling a story of the union sticking up for a member, but I heard plenty about them going along gladly with whatever new asshat-ery the railroad had come up with that month.
My mom worked at a smaller railroad that got swallowed up by BN. She rose through the ranks to become Wire Chief (think electrical engineering supervisor), the first and probably only female in that job at that railroad. As she grew closer to retirement, though, her new job assignments become increasingly horrific. For example, her last job included cleaning toilets in the crew shanties in the railyard.
Was it because she was a bad employee? Nope - her work record was spotless and her evaluations were impeccable. No, it was because BN-SF went out of their way to try to make people quit before they reached retirement. They did this to everyone in hopes of avoiding paying those hard-earned pensions.
Therefore, it doesn't surprise me at all to hear that they're trying to screw over yet another set of employees. That's been their SOP for years, so I can't imagine they'd turn tail now.
By the way, if you want an example of a completely incompetent union, there you have it. I'm not pro-union to begin with, but I'd expect one to at least try to help its members.
However, in exploring it not that long ago, I found one glaring omission was journaling file system support.
Are you sure you need journalling, though? FreeBSD's softupdates cover most of the advantages of a journal, and the background fsck (which mainly makes sure that unallocated blocks are actually marked as such) lets you boot quickly.
I certainly don't think journals are a bad thing, but I've honestly not missed them.
Regarding vinum: what do you like better about Linux's logical volume management? gvinum (GEOM-based vinum) has been pleasant enough for what I've asked of it.
I've started using pgrep and pkill which are fairly new to the base system. The versions in 6.0 have a "-j" flag to specify the jail ID to restrict the action to.
You mean the one posted a whopping 9 minutes before mine was posted (and probably 2 minutes before I started writing mine)? When the editors start apologizing for releasing dupe stories hours after the originals, I'll apologize for my 9 minute blind spot.:-)
Using VPC? Our IT guy would've loved to have seen that. Ours seemed to run OK for a few hours, but would fall over after receiving a bunch of network connections - the whole session would freeze and require an "end task". The VMWare sessions, though, have been cranking away for several weeks without a hitch.
BTW, this was running Win2K inside the sessions and WS2003 as the host OS, if that makes any difference.
Their main competitor is about to give away a free copy of Virtual PC with every upper level version of Visual Studio 2005.
We just tried (and failed) to virtualize four machines on a single dual-Xeon system with 3GB of RAM using VPC. After dicking around with it for a week, we switched to VMWare and had it running well in about an hour. Save for the shoddy documentation, VMWare trounced it in every way.
I don't think I'd consider VPC a competitor to VMWare today. Next year, maybe. Right now? Not from what I can tell.
According to an older explanation (photographers don't seem to jump on new technology very fast, so I imagine that's still reasonably accurate), 4"x5" film works out to about 24 megapixels. You can't magically add extra detail by running the prints through a hi-res scanner. By your logic, my little Sony digital is really 110MP, since I could make an 8"x10" print and scan it on my 1200DPI scanner.
Well, right. The quoted text sounded silly to me; was the alternative to save 2% some days and nothing on others? Of course you'd save an average of $(average savings) EACH DAY (emphasis theirs) by definition.
I completely agree that 1% isn't worth the hassle. Giving everyone a free CF bulb and asking them to use it would probably do a lot more in the long run, even though I'd find that equally preposterous.
I wish the president would have had the gumption to just extend Daylight Savings Time to all year long and ditch the date changes entirely.
Easy for you Southerners to say, but many of us in the north (particularly toward the western edge of their zones) would prefer it the other way. I'm glad the President doesn't have the gumption to unnecessarily inconvenience much of the population in this way.
Here is a reason, you can decide on your own whether or not this is a 'good' reason.
It's not, and the person who wrote it is a jackass:
Studies done by the U.S. Department of Transportation show that we trim the entire country's electricity usage by about one percent EACH DAY with Daylight Saving Time.
Hey, if we extend it to go year-round, we can save an extra 160%! Also:
Daylight Saving Time saves lives and prevents traffic injuries. The earlier Daylight Saving Time allowed more people to travel home from work and school in daylight, which is much safer than darkness. And except for the months of November through February, Daylight Saving Time does not increase the morning hazard for those going to school and work.
That person who wrote that is a damn liar. According to the US Navy (who take this whole time thing quite seriously), sunrise on November 3, 2007 at my house will be at 7:05AM CST, or 8:05AM CDT. In other words, I'm at work and the kids are in school before the sun ever rises. It won't be as bad when it goes into effect that spring (6:49/7:49 instead of 6:11/7:11 under the old plan), but that's still obnoxious. The poor kids in Fargo, ND have it even worse.
What they really mean is "the majority of the people who would be ready to lynch us for pulling this boneheaded maneuver will be only slightly inconvenienced". Miamians gets light 35 minutes earlier (within their timezone) than us northerners, so I'm sure this seems a lot more reasonable to the AARP crowd who're most likely to vote.
I think you picked badly. I bought an HP LaserJet 1200 for $300.00 nearly four years ago. Minus $50 "every year or so", it'll soon be cheaper for me to buy a brand new one outright than for you to have continued your hyper-expensive warranty coverage.
Maybe it's time you expected more from your hardware. If you're willing to bet nearly 20% of the price that it'll break in the next year or so, then perhaps you should consider a different model.
Slashdot Mirror
If by "Just Works" you mean "Barely Works", sure. Our tuned [1] Exchange server was absolutely crawling until I put a much-lower-specced Postfix/ClamAV/Greylist server in front of it to take the brunt of what the Internet was throwing at us.
Of course, it actually works reasonably well now (since it only has to handle pre-filtered traffic), so maybe that's what you meant: a firewalled Exchange is adequate in many cases. If so, then I apologize and withdraw my clarification.
[1] Our Windows IT guy actually knows what he's doing - if he says it's tuned, then it is.
...as long as they in the habit of only running one program at a time. I personally can't wait to transcode DVDs in the background while still enjoying a nimble, responsive desktop. Multi-core/CPU is pretty nice in general usage, not just as server iron.
(5*10^8)/(5*10^5) = 1000 watts per home. As long as everyone uses no more than 10 bright lightbulbs simultaneously, your estimate is reasonable. Throw in a few stoves, air conditioners, and home computers and that number drops very quickly. Anybody know what the average usage per household is in America?
..and it's sitting on top of an Amiga 2000. I love you guys, seriously - viva la NetBSD!
My point is that it's not entirely clear which OS supports more platforms, since "supports" and "platforms" are both variables that would need to be nailed down before the conversation even begins. My own first impression is that NetBSD is still the winner, since you can actually boot into each of its listed platforms, install software from pkgsrc, and generally treat them as equals except for the obvious performance differences. Others could effectively argue the opposite, I'm sure.
Let me be the first to say that your benchmarking methodology sucks. Does that 1.7MB of RAM consist of 1.6MB of constant overhead + .1MB of script usage? Who knows - you certainly don't seem to have bothered to test it. Does eAccelerator use the extra memory for results caching that result in higher throughput and overall lower resource usage? Again, you don't seem to have gotten past "it's teh bloat" so we have no way to venture a guess based on your numbers.
Frankly, if you measure performance in terms of "module A uses 770KB of RAM, but module B only uses 320KB", then PHP is clearly the wrong language for you. I don't particularly like PHP, either, but your "reasons" certainly would not have convinced me if I felt otherwise.
But yet it does explain MySQL? Which is it? I'm not going to buy the book if the reviewer can't figure it out.
In other words, the book isn't centered around building DB-driven apps using PHP+MySQL, even if it does discuss the topic. The reviewer wishes it also had chapters on PHP+PostgreSQL or PHP+Oracle, but it doesn't. Seems pretty reasonable and consistent to me.
Where do you live that you have to chase deer? We can't keep 'em away around here. Or maybe your idea of "chase" is "scream like a little girl after the animal jumps directly into your velocity vector and immediately freezes so as to avoid doing less than the maximum possible amount of damage to your vehicle and its occupants"? If that's the case, then I chase deer all the damn time.
If law and order ever breaks down where I live, I'm mounting an infrared-aimed chain gun and snowplow to the wife's minivan.
Sweet! Assuming that the watermark is at least 128 bits (surely it'd take at least that much?), it only takes 68 million cracked DVDs to fill a gig of RAM with nothing but the revoked key list. If that sounds like a lot, consider that US DVD sales were about $18 billion dollars last year. Assuming $15/DVD gives roughly 1.2 billion DVDs. If 1% of those disks are ripped, released, and revoked, then that 1 gig gets used in six years.
Now, I doubt that any new DVD player is going to come with 1GB of nonvolatile memory in the near future (I don't see them carrying hard drives any time soon, and RAM would drive the product price up). If they had 256MB of revocation list memory, then they only get 1.5 years of use before it overflows.
In other words, the industry is actually providing an incentive for us to violate copyrights. The sooner we meet our quota, the sooner we overload the DRM in the hardware we've bought. Ironically, this is one of the few things they could do to make me want to make illicit copies. I have a pretty big collection of movies that I bought and paid for - I have no problem forking over a few bucks for entertainment - but this is enough to switch me to the Dark Side (from their point of view). Good going, fellas.
Basically, her union didn't do jack. As far as I can tell, they just collected dues and used them to finance political candidates. I don't think I ever heard her telling a story of the union sticking up for a member, but I heard plenty about them going along gladly with whatever new asshat-ery the railroad had come up with that month.
Was it because she was a bad employee? Nope - her work record was spotless and her evaluations were impeccable. No, it was because BN-SF went out of their way to try to make people quit before they reached retirement. They did this to everyone in hopes of avoiding paying those hard-earned pensions.
Therefore, it doesn't surprise me at all to hear that they're trying to screw over yet another set of employees. That's been their SOP for years, so I can't imagine they'd turn tail now.
By the way, if you want an example of a completely incompetent union, there you have it. I'm not pro-union to begin with, but I'd expect one to at least try to help its members.
Are you sure you need journalling, though? FreeBSD's softupdates cover most of the advantages of a journal, and the background fsck (which mainly makes sure that unallocated blocks are actually marked as such) lets you boot quickly.
I certainly don't think journals are a bad thing, but I've honestly not missed them.
Regarding vinum: what do you like better about Linux's logical volume management? gvinum (GEOM-based vinum) has been pleasant enough for what I've asked of it.
I've started using pgrep and pkill which are fairly new to the base system. The versions in 6.0 have a "-j" flag to specify the jail ID to restrict the action to.
Not very. Irony implies surprise, and I think most people actually expected this one.
You mean the one posted a whopping 9 minutes before mine was posted (and probably 2 minutes before I started writing mine)? When the editors start apologizing for releasing dupe stories hours after the originals, I'll apologize for my 9 minute blind spot. :-)
The dumbest part is that they probably could've used BSD code (eg FreeBSD's Linux emulation layer) and done it legally.
BTW, this was running Win2K inside the sessions and WS2003 as the host OS, if that makes any difference.
We just tried (and failed) to virtualize four machines on a single dual-Xeon system with 3GB of RAM using VPC. After dicking around with it for a week, we switched to VMWare and had it running well in about an hour. Save for the shoddy documentation, VMWare trounced it in every way.
I don't think I'd consider VPC a competitor to VMWare today. Next year, maybe. Right now? Not from what I can tell.
According to an older explanation (photographers don't seem to jump on new technology very fast, so I imagine that's still reasonably accurate), 4"x5" film works out to about 24 megapixels. You can't magically add extra detail by running the prints through a hi-res scanner. By your logic, my little Sony digital is really 110MP, since I could make an 8"x10" print and scan it on my 1200DPI scanner.
Well, right. The quoted text sounded silly to me; was the alternative to save 2% some days and nothing on others? Of course you'd save an average of $(average savings) EACH DAY (emphasis theirs) by definition.
I completely agree that 1% isn't worth the hassle. Giving everyone a free CF bulb and asking them to use it would probably do a lot more in the long run, even though I'd find that equally preposterous.
[1] I'm a Master Mason.
[2] My wife's and firstborn's birthday.
Easy for you Southerners to say, but many of us in the north (particularly toward the western edge of their zones) would prefer it the other way. I'm glad the President doesn't have the gumption to unnecessarily inconvenience much of the population in this way.
It's not, and the person who wrote it is a jackass:
Hey, if we extend it to go year-round, we can save an extra 160%! Also:
That person who wrote that is a damn liar. According to the US Navy (who take this whole time thing quite seriously), sunrise on November 3, 2007 at my house will be at 7:05AM CST, or 8:05AM CDT. In other words, I'm at work and the kids are in school before the sun ever rises. It won't be as bad when it goes into effect that spring (6:49/7:49 instead of 6:11/7:11 under the old plan), but that's still obnoxious. The poor kids in Fargo, ND have it even worse.
What they really mean is "the majority of the people who would be ready to lynch us for pulling this boneheaded maneuver will be only slightly inconvenienced". Miamians gets light 35 minutes earlier (within their timezone) than us northerners, so I'm sure this seems a lot more reasonable to the AARP crowd who're most likely to vote.
Maybe it's time you expected more from your hardware. If you're willing to bet nearly 20% of the price that it'll break in the next year or so, then perhaps you should consider a different model.
Another sad story of a life wasted to piracy. Will the tragedy never end?