here, the server can throttle or block access to these requesters, and the success rate is very low.
You make a possibly faulty assumption here.
Just like spam runs can be spread across hundreds of thousands of machines, so can dictionary attacks. Which makes it a lot harder to block or throttle access to random IPs.
Do they still do that? I know from a distant past they tried it with smaller providers too, but haven't seen them for a long time. As far as I can tell, spammers do still use malware which harvests/sniffs email-address directly from peoples computers.
This is a definite tactic. I see it all the time on a mail server that I administer. From the results, there are definitely spammers that monitor user's e-mail, address book, or other sources of e-mail addresses on their computer. (Basically, on a brand new e-mail address, the user started getting spam within a few hours of contacting someone else.)
But we still see dictionary attacks on our mail server, so that's a popular tactic too.
1) Move the external SSH port to some other port number. Easy change. A minor road-block, but one that works well at cutting out about 99% of the attack attempts. Plus, if they have to port scan the server to find the SSH port, it gives you an opportunity to detect the scans and shut them down.
2) Don't allow root to login over SSH. You'll still see a lot of people who haven't disabled that...
3) Force users to use SSH keys to authenticate (no password-based authentication). Yes, it doesn't address this particular issue of stolen SSH private keys, but it does put a pretty large roadblock in the way of brute-force attacks.
4) Limit the # of accounts on publicly available servers.
We haven't tried deny-hosts yet. I keep meaning to, but haven't had time to go digging into it yet.
Size isn't the only consideration. Smaller disks have faster seek times. It actually can effect gaming performance.
Except that for the 250GB vs the 500GB... both disks are the SAME physical size. They (probably) both use the exact same internal platters, except that the 500GB drive has twice as many platters inside.
Two things affect seek time:
1) Physical platter size (radius or diameter measurement). A 2.5" platter vs 3.5" platter means that the drive head has a smaller total travel distance to seek from the outer track to the inner track. Which is why higher end disks use smaller platters (like the 2.5" SAS platters). Also why the Raptor drives used smaller physical platters (in addition to the issue that you can't spin a 3.5" disk platter at 10k RPM).
2) Rotational speed. Which puts a sector in a particular track under the data heads sooner. That's why higher end disks spin at faster RPMs (10k or 15k RPM).
Drive capacity has no bearing on seek time. Except for one caveat. If you take a 500GB drive, and only use the first 50% of the drive, it's possible that you'll simulate a smaller platter size. Which will make your seek time lower.
(The assumption there is that the drive's internal layout is such that using only the first 50% of the disk results in the drive using the inner tracks on the physical platters. Which is not a guarantee with Logical Block Addressing or LBA.)
I ran a pair of GeForce 7950s for a while (2yrs or so), and recent upgraded to a pair of GeForce 8800 GT 512MBs. The GeForce 8800 GT 512MB cards are only about $180 now, which is a darn good price.
SLI works just fine. In fact, it works so well that you'll find some games become CPU constrained. Whoops, time to upgrade that old CPU. It's cost-efficient if you can buy 2 cards that are at the knee of the price curve. Cards that are fast but inexpensive (like the 8800 GT 512MB units) are good choices.
The only downside to SLI is that you can't run multiple monitors when you switch to SLI mode. (At least with the 7950 and 8800 series video cards.) So, if you really want dual-monitor support, while gaming, SLI won't do you any good.
(I generally just reboot into SLI mode when I want to do a bit of gaming.)
But yes... it's a bit of a luxury. OTOH, if one of my cards fries, I'm only out $200 instead of seeing a $500 card go up in smoke.
For a good, power-gamer box, you can build a fairly screaming machine for a reasonable price these days.
22" 1680x1050 LCD - yes, they're more expensive then the 19", but not that much more. Figure about $275-$300 for this.
CPU - pick something in the $200-$250 range. Probably a quad-core now.
Motherboard - figure $150 for a good SLI or Crossfire board.
RAM - 4GB of fast memory is only $150 or so right now.
(2) hard drives - $150 or so - use the 2nd one to store images of the 1st one
(2) video cards - $180-$200 each. Much more then that, and you're probably overpaying for the video card. The GeForce 8800 512MB cards are a good bang for the buck. They're also pretty good at bang for the watt too.
Case - good cases are about $75. Get a good case, and it will last you through 3+ upgrades.
PSU - good PSUs are $100 or so for all that power
Misc parts - figure $75 for DVD, spare fans, other bits
That's $1650, plus the cost of the O/S and other software. The base unit is only $1350 if you already have a monitor. Tack on another $100 for an audio card if you want something better then the (usually) decent onboard sound.
And if you want to do it on the cheap... save yourself $200 by not buying the 2nd video card, which may keep you from also needing the better PSU (another $100). You could probably shave another $100 by buying a slightly less expensive CPU and only 2GB of RAM. Which would drop the costs down to about $950 for the base system.
Which would still be a very decent mid-range machine. With the option to toss that 2nd video card, 2GB RAM and a better PSU in it down the road as an inexpensive upgrade.
If you want to be able to downgrade to WinXP from Vista, you'll have to buy either Vista Ultimate or Vista Business, which will run you about $150 or so for OEM versions.
Such things would require a prohibitively high number of actual persons playing NPCs, and the amount of coordination between them would make this extremely buggy.
Which basically means that you WILL have players who figure out how to ruin events and storylines for other people. That's why a lot of quests are persistent, and ever un-changing.
They usually don't get used directly because they're much more expensive computationally than AES and the like, and potentially vulnerable to chosen-plaintext attacks.
I was with you up until the last bit. Do you have any information regarding public-key ciphers being vulnerable to chosen plaintext attacks? And which of the various public-key encryption algorithms are vulnerable?
EQ (probably looks dated... but EQ2 was a lot more graphic-intensive then WoW or EQ1. In fact, it was so graphic intensive, that most video cards of the day had difficulty running EQ2 on anything higher then medium settings.
I loved the look of the world. I just didn't like:
- The smallness of the zones at launch
- The limited number of zones at launch
- The world felt small, like a bunch of shoeboxes strung together
- Developers without a central vision
- Poor QA process
I have a Samsung ML2150, and have noticed the same thing. Lights flicker, etc. whenever I submit a print job and the printer transitions from standby to active. The various UPSes in my office sense that, and respond with clicks and beeps.
If the lights are flickering when a laser printer spins up, then you (very likely) have overloaded circuits (and probably/possibly overloaded cabling). Spend the money and get the electricians to run a separate 20A (or 30A) run for that laser printer.
(Otherwise, you're likely risking more catastrophic issues down the road. Such as fire.)
Re:Can Oscar's be given posthumously?
on
Batman Discussion
·
· Score: 1
Why was it so easy for the Joker to turn cops evil? Ramirez loves her mom so much that's she's willing to help the Joker kidnap Gordon's family?
Ramirez was owned by the mob. If you read between the lines at the end, she accepted a bribe from the mob back when hospital bills were mounting up. Which gave them leverage on her, after which point, she was a corrupt cop.
And it wasn't the Joker who kidnapped Gordon's family. It was Harvey, holding a gun on Ramirez, to force the family out of the house into the open while the cop cars were pulled off by Ramirez.
Re:any chance of an unrated dvd release ?
on
Batman Discussion
·
· Score: 2, Interesting
I think the movie did fantastic by hiding just about every single drop of blood.
It made the scenes scarier when your mind filled in the details for you.
That there, is I think a big reason why the movie is going to be worth watching a few more times when it comes out on DVD. There were so many points when you first meet the Joker that you're thinking "oh hell, here comes the blood spurts".
It didn't seem like there was a lot of swearing either. Or at least, it wasn't obnoxious like a lot of action movies are (with F-this and F-that).
Basically, they did a pretty good job of not overplaying their hand and grossing the audience out.
Coding guidelines should focus instead on the techniques that help reduce the number of bugs in code. How is that done? It takes someone (typically a senior person) looking at the the bugs that have been found in the code, categorizing their cause, devising a way to prevent those bugs from occurring, then putting that into the guidelines.
Ding! Most definitely!
One of the things that I learned from working over the years is that the most useful processes and guidelines have built-in safeties against simple errors. They're not guidelines for the sake of making things pretty, the underlying goal is to prevent easily avoided problems. Or making it easy to spot problems.
A simple example of this is that any two identifiers should differ by at least two characters. Which prevents a single typo from resulting in using the wrong identifier. A rule that was more useful in the days before IDEs and auto-complete, but is still useful when it comes to naming project files or data files.
At the company I used to work for there were about 5 developers and we used Microsoft Visual SourceSafe, which didn't actually seem to be that safe.
(Laughs at the story... unfortunately, source code control systems seem to be rare. I think, mostly because they were expensive or difficult to use until things like Mercurial / SVN / GIT hit the scene. It feels, to me, like SCC system usage is a lot more common then it used to be.)
We used VSS at our location for 5 developers for quite a few years. In our case, we made it a lot safer by purchasing SourceOffSite and using that to interact with the VSS database. (Basically, nobody used VSS.) It seemed to keep corruption to an absolute minimum because it was the only program mucking around with the VSS storage files.
(We've since moved to Subversion a few years back and have been very happy.)
Interesting, I never played DAoC. As usual, Blizzard did a good job of picking the best feature concepts from the other MMOs, and then bundling them up into a tightly polished game.
You missed one big feature of WoW - the Talent system and being able to re-spec in some cases to a completely different play style.
For example: Tired of your life as a healing priest? Re-spec as Shadow Priest and turn yourself into a DPS caster. Or the druids, which can be either healers, tanks, melee DPS or caster DPS. Some classes can be spec'd out different more then others (Warlocks and Mages are always caster DPS, they just re-spec for different styles).
All without having to re-roll your character from level 1.
Microsoft can't be excised from the IT world. If they, for the sake of argument, collapsed next week, there would be a worldwide IT company crash of epic proportions. We would all suffer.
Not us. We'd simply move the rest of our Windows servers over to Linux and use projects like Mono/WINE to ease the pain of migration. Same thing for the desktops, switch everyone over to Linux or OS X and move on.
A few years ago - that would not have been a possible choice. But WINE & Mono and the maturation of distros like Ubuntu are really coming close to being "good enough". Good enough in that we wouldn't have to replace ALL of our applications, just the ones that don't work properly.
Are we ready to make the move yet? Hell no. We just finished upgrading everyone to WinXP and plan on staying on WinXP for another 5 years or so. Our next O/S upgrade will not happen before 2011-2012. Vista will not be used at our workplace until at least 2010.
But if push came to shove, Linux + WINE would be the direction we'd move in.
Some years ago I remember reading an article that argued that Microsoft should dump Windows and shift to Linux. Specifically it argued that MS should code the Windows desktop as a window interface running on top of a Linux core. At the time I dismissed it as the ravings of a Linux fan, but I wonder more and more if there isn't some value in the argument.
Frankly, I think there's a lot of value in that approach.
Take Linux, which is a very good operating system, already developed with wide acceptance. No need to re-invent the wheel there. Then start contributing to the WINE project. Spend all that money that you'd waste on Windows 7 or whatever and spend it making WINE more compatible with the top 1000 applications.
Then either take Gnome or KDE or write your own GUI on top of it and you'd be off to the races. Sell the whole thing as a support package, make it better then KDE or Gnome, and go back to giving away the software for free (just like they ignored the pirates in the Win9x days).
Unfortunately, Microsoft management is too much the "NIH" type. They simply don't grasp the concept that things simply work better if you follow the standards rather then "embrace, extend, extinguish". They're also locked into the mindset that the Windows monopoly has to be there in order to also lock people into the MS Office monopoly.
(In fact, their investors might object if Microsoft didn't attempt to lock people in.)
The best thing that could have happened to Microsoft would have been to be a forced company split back around 2000. Where the DoJ should have forced them to split their business into an OS company, an applications company, and a device company. The OS folks would then not have their hands tied making sure that only Windows could run MS Office properly (and vice-versa).
Overall, I give the HP series about a B, maybe a B+. The first book was great for wish fulfillment, the second two were only so-so follow-ups. Things get better in book 4 (Goblet of Fire) which is one of the best of the series. There was a bit too much of "Harry and crew learn something that conveniently helps them later". Still things, plod along nice in 5 & 6.
But the 7th and last book suffers from being crammed into a timeline of exactly one year. Some of the climaxes were either overly obvious or a bit iffy.
So, it's a good series, but not a great series. And sometimes that's enough to hit the sweet spot of market acceptance.
Portage is a great piece of software, and I stuck with Gentoo for many years because because of its strengths. But portage is not what lets Gentoo down. The complete lack of QA on the official tree that leads to dependency blocks, updated libraries in the stable tree that break ABI compatability with previous software and the general cavalier attitude to pushing any old crap into the stable release are what kills Gentoo.
The lack of consistent QA is what killed off Gentoo for us on servers. It was a much better decision (though less of a "cool" factor) to switch our servers over to RedHat/CentOS. Not that we ever had issues with the Gentoo servers that we used.
As much as people may dislike Red Hat, their business model centers around servers and making sure that stuff doesn't break. Which is worth more in the long run. And if it does break, it's a lot easier to get paid support for Red Hat / CentOS from a local company.
Gentoo can't seem to decide if it's a hobby distro, a desktop distro, a LFS style distro, a server distro, or a flavor of the month distro. I'll remember it fondly for the lessons that it taught me, but at the end of the day I need a distro that works without fiddling.
Re:I understand why you`d want to go pre-built
on
What NAS To Buy?
·
· Score: 1
.. but unfortunately all the pre-built NAS cubes I`ve seen are way over priced. They usually end up costing about as much as a home built file server _without_ the drives.
Yes and no...
There are quite a few ~$600 NAS cubes. Which is not that expensive once you consider the labor cost of assembling a unit out of parts (plus the cost of the parts).
Overall, it's a wash. I've used the little NAS cubes and they're fine for small (under 10 people) offices. As much as I'd prefer a redundant server box with dual-PSUs and lots of blinky lights, sometimes the NAS cubes are a better choice.
Re:RAID5 is stupid, RAID 10 or no RAID
on
What NAS To Buy?
·
· Score: 1
The big advantage of RAID1+0 is that rebuild time is based on the size of an individual drive element. In RAID5, rebuild time increases as the overall size of the array increases.
How does RAID6 compare for rebuild time?
(For the TRULY paranoid and wealthy, you can do RAID0 over top of 3-element RAID1 arrays. Meaning that you build a bunch of 3-disk RAID1 mirrors, then lay RAID0 over top. Which gives you only 1/3 net space - hugely wasteful - but you're protected for sure against double-failures.)
Re:RAID5 is stupid, RAID 10 or no RAID
on
What NAS To Buy?
·
· Score: 1
Using RAID 1+0, you get almost 4 times the performance for reads, and 2 times for writes.
Using RAID5, you get maybe 3 times the performance for reads (if you're lucky), and writes can be slower than a single drive due to parity calculations.
That may be true for high-level raid controllers, but Linux Software RAID, you'll only see a 2x read bandwidth on RAID1+0. And that's because the underlying RAID1 code doesn't seem to provide additional read power based on the number of drives in the RAID1 mirror.
Basically, what we've seen, is that if a single drive gives 60MB/s, you'll only see 60MB/s out of a Software RAID1 array. But it does scale up for RAID0 and RAID1+0 based on the number of sets of spindles.
(This was using SATA II on modern 64bit 2GHz CPUs.)
During that time, I was also given access to a Windows box via RDP, from a Windows client. It was on the same pipe, but what struck me was how awfully, awfully slow it was. I could watch the mouse trails crawl across the desktop. Obviously, they may have improved their compression over the past few years.
I suspect that you may have had 15/16/32 bit color enabled then.
Comparing RDP vs VNC for the same color-depth, you'll find that RDP is simply a lot more responsive over the same connection.
There was a time when Microsoft were the good guys, where people suddenly found that they could write a product for DOS and it would run on almost any computer. That meant it was possible to become a software house with a lot less effort and money than before.
And then they got greedy and started competing in a lot of other markets (office software, financial software, bookkeeping software).
Back in the mid-late 90s, ISVs lived in fear that Microsoft would decide to take an interest in their market segment. They'd come in with an inferior product, and keep throwing money at it until you went out of business. Or they'd buy you out and retire your product from the market, leaving only their solution.
Slashdot Mirror
here, the server can throttle or block access to these requesters, and the success rate is very low.
You make a possibly faulty assumption here.
Just like spam runs can be spread across hundreds of thousands of machines, so can dictionary attacks. Which makes it a lot harder to block or throttle access to random IPs.
Do they still do that? I know from a distant past they tried it with smaller providers too, but haven't seen them for a long time. As far as I can tell, spammers do still use malware which harvests/sniffs email-address directly from peoples computers.
This is a definite tactic. I see it all the time on a mail server that I administer. From the results, there are definitely spammers that monitor user's e-mail, address book, or other sources of e-mail addresses on their computer. (Basically, on a brand new e-mail address, the user started getting spam within a few hours of contacting someone else.)
But we still see dictionary attacks on our mail server, so that's a popular tactic too.
Easier tricks for us...
1) Move the external SSH port to some other port number. Easy change. A minor road-block, but one that works well at cutting out about 99% of the attack attempts. Plus, if they have to port scan the server to find the SSH port, it gives you an opportunity to detect the scans and shut them down.
2) Don't allow root to login over SSH. You'll still see a lot of people who haven't disabled that...
3) Force users to use SSH keys to authenticate (no password-based authentication). Yes, it doesn't address this particular issue of stolen SSH private keys, but it does put a pretty large roadblock in the way of brute-force attacks.
4) Limit the # of accounts on publicly available servers.
We haven't tried deny-hosts yet. I keep meaning to, but haven't had time to go digging into it yet.
Size isn't the only consideration. Smaller disks have faster seek times. It actually can effect gaming performance.
Except that for the 250GB vs the 500GB... both disks are the SAME physical size. They (probably) both use the exact same internal platters, except that the 500GB drive has twice as many platters inside.
Two things affect seek time:
1) Physical platter size (radius or diameter measurement). A 2.5" platter vs 3.5" platter means that the drive head has a smaller total travel distance to seek from the outer track to the inner track. Which is why higher end disks use smaller platters (like the 2.5" SAS platters). Also why the Raptor drives used smaller physical platters (in addition to the issue that you can't spin a 3.5" disk platter at 10k RPM).
2) Rotational speed. Which puts a sector in a particular track under the data heads sooner. That's why higher end disks spin at faster RPMs (10k or 15k RPM).
Drive capacity has no bearing on seek time. Except for one caveat. If you take a 500GB drive, and only use the first 50% of the drive, it's possible that you'll simulate a smaller platter size. Which will make your seek time lower.
(The assumption there is that the drive's internal layout is such that using only the first 50% of the disk results in the drive using the inner tracks on the physical platters. Which is not a guarantee with Logical Block Addressing or LBA.)
I ran a pair of GeForce 7950s for a while (2yrs or so), and recent upgraded to a pair of GeForce 8800 GT 512MBs. The GeForce 8800 GT 512MB cards are only about $180 now, which is a darn good price.
SLI works just fine. In fact, it works so well that you'll find some games become CPU constrained. Whoops, time to upgrade that old CPU. It's cost-efficient if you can buy 2 cards that are at the knee of the price curve. Cards that are fast but inexpensive (like the 8800 GT 512MB units) are good choices.
The only downside to SLI is that you can't run multiple monitors when you switch to SLI mode. (At least with the 7950 and 8800 series video cards.) So, if you really want dual-monitor support, while gaming, SLI won't do you any good.
(I generally just reboot into SLI mode when I want to do a bit of gaming.)
But yes... it's a bit of a luxury. OTOH, if one of my cards fries, I'm only out $200 instead of seeing a $500 card go up in smoke.
For a good, power-gamer box, you can build a fairly screaming machine for a reasonable price these days.
22" 1680x1050 LCD - yes, they're more expensive then the 19", but not that much more. Figure about $275-$300 for this.
CPU - pick something in the $200-$250 range. Probably a quad-core now.
Motherboard - figure $150 for a good SLI or Crossfire board.
RAM - 4GB of fast memory is only $150 or so right now.
(2) hard drives - $150 or so - use the 2nd one to store images of the 1st one
(2) video cards - $180-$200 each. Much more then that, and you're probably overpaying for the video card. The GeForce 8800 512MB cards are a good bang for the buck. They're also pretty good at bang for the watt too.
Case - good cases are about $75. Get a good case, and it will last you through 3+ upgrades.
PSU - good PSUs are $100 or so for all that power
Misc parts - figure $75 for DVD, spare fans, other bits
That's $1650, plus the cost of the O/S and other software. The base unit is only $1350 if you already have a monitor. Tack on another $100 for an audio card if you want something better then the (usually) decent onboard sound.
And if you want to do it on the cheap... save yourself $200 by not buying the 2nd video card, which may keep you from also needing the better PSU (another $100). You could probably shave another $100 by buying a slightly less expensive CPU and only 2GB of RAM. Which would drop the costs down to about $950 for the base system.
Which would still be a very decent mid-range machine. With the option to toss that 2nd video card, 2GB RAM and a better PSU in it down the road as an inexpensive upgrade.
If you want to be able to downgrade to WinXP from Vista, you'll have to buy either Vista Ultimate or Vista Business, which will run you about $150 or so for OEM versions.
Such things would require a prohibitively high number of actual persons playing NPCs, and the amount of coordination between them would make this extremely buggy.
The bigger issue is the "Internet Fuckwad Theory".
Which basically means that you WILL have players who figure out how to ruin events and storylines for other people. That's why a lot of quests are persistent, and ever un-changing.
They usually don't get used directly because they're much more expensive computationally than AES and the like, and potentially vulnerable to chosen-plaintext attacks.
I was with you up until the last bit. Do you have any information regarding public-key ciphers being vulnerable to chosen plaintext attacks? And which of the various public-key encryption algorithms are vulnerable?
EQ (probably looks dated... but EQ2 was a lot more graphic-intensive then WoW or EQ1. In fact, it was so graphic intensive, that most video cards of the day had difficulty running EQ2 on anything higher then medium settings.
I loved the look of the world. I just didn't like:
- The smallness of the zones at launch
- The limited number of zones at launch
- The world felt small, like a bunch of shoeboxes strung together
- Developers without a central vision
- Poor QA process
WoW wide-open world is much more appealing.
I have a Samsung ML2150, and have noticed the same thing. Lights flicker, etc. whenever I submit a print job and the printer transitions from standby to active. The various UPSes in my office sense that, and respond with clicks and beeps.
If the lights are flickering when a laser printer spins up, then you (very likely) have overloaded circuits (and probably/possibly overloaded cabling). Spend the money and get the electricians to run a separate 20A (or 30A) run for that laser printer.
(Otherwise, you're likely risking more catastrophic issues down the road. Such as fire.)
Why was it so easy for the Joker to turn cops evil? Ramirez loves her mom so much that's she's willing to help the Joker kidnap Gordon's family?
Ramirez was owned by the mob. If you read between the lines at the end, she accepted a bribe from the mob back when hospital bills were mounting up. Which gave them leverage on her, after which point, she was a corrupt cop.
And it wasn't the Joker who kidnapped Gordon's family. It was Harvey, holding a gun on Ramirez, to force the family out of the house into the open while the cop cars were pulled off by Ramirez.
I think the movie did fantastic by hiding just about every single drop of blood.
It made the scenes scarier when your mind filled in the details for you.
That there, is I think a big reason why the movie is going to be worth watching a few more times when it comes out on DVD. There were so many points when you first meet the Joker that you're thinking "oh hell, here comes the blood spurts".
It didn't seem like there was a lot of swearing either. Or at least, it wasn't obnoxious like a lot of action movies are (with F-this and F-that).
Basically, they did a pretty good job of not overplaying their hand and grossing the audience out.
Coding guidelines should focus instead on the techniques that help reduce the number of bugs in code. How is that done? It takes someone (typically a senior person) looking at the the bugs that have been found in the code, categorizing their cause, devising a way to prevent those bugs from occurring, then putting that into the guidelines.
Ding! Most definitely!
One of the things that I learned from working over the years is that the most useful processes and guidelines have built-in safeties against simple errors. They're not guidelines for the sake of making things pretty, the underlying goal is to prevent easily avoided problems. Or making it easy to spot problems.
A simple example of this is that any two identifiers should differ by at least two characters. Which prevents a single typo from resulting in using the wrong identifier. A rule that was more useful in the days before IDEs and auto-complete, but is still useful when it comes to naming project files or data files.
At the company I used to work for there were about 5 developers and we used Microsoft Visual SourceSafe, which didn't actually seem to be that safe.
(Laughs at the story... unfortunately, source code control systems seem to be rare. I think, mostly because they were expensive or difficult to use until things like Mercurial / SVN / GIT hit the scene. It feels, to me, like SCC system usage is a lot more common then it used to be.)
We used VSS at our location for 5 developers for quite a few years. In our case, we made it a lot safer by purchasing SourceOffSite and using that to interact with the VSS database. (Basically, nobody used VSS.) It seemed to keep corruption to an absolute minimum because it was the only program mucking around with the VSS storage files.
(We've since moved to Subversion a few years back and have been very happy.)
Interesting, I never played DAoC. As usual, Blizzard did a good job of picking the best feature concepts from the other MMOs, and then bundling them up into a tightly polished game.
You missed one big feature of WoW - the Talent system and being able to re-spec in some cases to a completely different play style.
For example: Tired of your life as a healing priest? Re-spec as Shadow Priest and turn yourself into a DPS caster. Or the druids, which can be either healers, tanks, melee DPS or caster DPS. Some classes can be spec'd out different more then others (Warlocks and Mages are always caster DPS, they just re-spec for different styles).
All without having to re-roll your character from level 1.
Microsoft can't be excised from the IT world. If they, for the sake of argument, collapsed next week, there would be a worldwide IT company crash of epic proportions. We would all suffer.
Not us. We'd simply move the rest of our Windows servers over to Linux and use projects like Mono/WINE to ease the pain of migration. Same thing for the desktops, switch everyone over to Linux or OS X and move on.
A few years ago - that would not have been a possible choice. But WINE & Mono and the maturation of distros like Ubuntu are really coming close to being "good enough". Good enough in that we wouldn't have to replace ALL of our applications, just the ones that don't work properly.
Are we ready to make the move yet? Hell no. We just finished upgrading everyone to WinXP and plan on staying on WinXP for another 5 years or so. Our next O/S upgrade will not happen before 2011-2012. Vista will not be used at our workplace until at least 2010.
But if push came to shove, Linux + WINE would be the direction we'd move in.
Some years ago I remember reading an article that argued that Microsoft should dump Windows and shift to Linux. Specifically it argued that MS should code the Windows desktop as a window interface running on top of a Linux core. At the time I dismissed it as the ravings of a Linux fan, but I wonder more and more if there isn't some value in the argument.
Frankly, I think there's a lot of value in that approach.
Take Linux, which is a very good operating system, already developed with wide acceptance. No need to re-invent the wheel there. Then start contributing to the WINE project. Spend all that money that you'd waste on Windows 7 or whatever and spend it making WINE more compatible with the top 1000 applications.
Then either take Gnome or KDE or write your own GUI on top of it and you'd be off to the races. Sell the whole thing as a support package, make it better then KDE or Gnome, and go back to giving away the software for free (just like they ignored the pirates in the Win9x days).
Unfortunately, Microsoft management is too much the "NIH" type. They simply don't grasp the concept that things simply work better if you follow the standards rather then "embrace, extend, extinguish". They're also locked into the mindset that the Windows monopoly has to be there in order to also lock people into the MS Office monopoly.
(In fact, their investors might object if Microsoft didn't attempt to lock people in.)
The best thing that could have happened to Microsoft would have been to be a forced company split back around 2000. Where the DoJ should have forced them to split their business into an OS company, an applications company, and a device company. The OS folks would then not have their hands tied making sure that only Windows could run MS Office properly (and vice-versa).
Overall, I give the HP series about a B, maybe a B+. The first book was great for wish fulfillment, the second two were only so-so follow-ups. Things get better in book 4 (Goblet of Fire) which is one of the best of the series. There was a bit too much of "Harry and crew learn something that conveniently helps them later". Still things, plod along nice in 5 & 6.
But the 7th and last book suffers from being crammed into a timeline of exactly one year. Some of the climaxes were either overly obvious or a bit iffy.
So, it's a good series, but not a great series. And sometimes that's enough to hit the sweet spot of market acceptance.
Portage is a great piece of software, and I stuck with Gentoo for many years because because of its strengths. But portage is not what lets Gentoo down. The complete lack of QA on the official tree that leads to dependency blocks, updated libraries in the stable tree that break ABI compatability with previous software and the general cavalier attitude to pushing any old crap into the stable release are what kills Gentoo.
The lack of consistent QA is what killed off Gentoo for us on servers. It was a much better decision (though less of a "cool" factor) to switch our servers over to RedHat/CentOS. Not that we ever had issues with the Gentoo servers that we used.
As much as people may dislike Red Hat, their business model centers around servers and making sure that stuff doesn't break. Which is worth more in the long run. And if it does break, it's a lot easier to get paid support for Red Hat / CentOS from a local company.
Gentoo can't seem to decide if it's a hobby distro, a desktop distro, a LFS style distro, a server distro, or a flavor of the month distro. I'll remember it fondly for the lessons that it taught me, but at the end of the day I need a distro that works without fiddling.
.. but unfortunately all the pre-built NAS cubes I`ve seen are way over priced. They usually end up costing about as much as a home built file server _without_ the drives.
Yes and no...
There are quite a few ~$600 NAS cubes. Which is not that expensive once you consider the labor cost of assembling a unit out of parts (plus the cost of the parts).
Overall, it's a wash. I've used the little NAS cubes and they're fine for small (under 10 people) offices. As much as I'd prefer a redundant server box with dual-PSUs and lots of blinky lights, sometimes the NAS cubes are a better choice.
The big advantage of RAID1+0 is that rebuild time is based on the size of an individual drive element. In RAID5, rebuild time increases as the overall size of the array increases.
How does RAID6 compare for rebuild time?
(For the TRULY paranoid and wealthy, you can do RAID0 over top of 3-element RAID1 arrays. Meaning that you build a bunch of 3-disk RAID1 mirrors, then lay RAID0 over top. Which gives you only 1/3 net space - hugely wasteful - but you're protected for sure against double-failures.)
Using RAID 1+0, you get almost 4 times the performance for reads, and 2 times for writes.
Using RAID5, you get maybe 3 times the performance for reads (if you're lucky), and writes can be slower than a single drive due to parity calculations.
That may be true for high-level raid controllers, but Linux Software RAID, you'll only see a 2x read bandwidth on RAID1+0. And that's because the underlying RAID1 code doesn't seem to provide additional read power based on the number of drives in the RAID1 mirror.
Basically, what we've seen, is that if a single drive gives 60MB/s, you'll only see 60MB/s out of a Software RAID1 array. But it does scale up for RAID0 and RAID1+0 based on the number of sets of spindles.
(This was using SATA II on modern 64bit 2GHz CPUs.)
During that time, I was also given access to a Windows box via RDP, from a Windows client. It was on the same pipe, but what struck me was how awfully, awfully slow it was. I could watch the mouse trails crawl across the desktop. Obviously, they may have improved their compression over the past few years.
I suspect that you may have had 15/16/32 bit color enabled then.
Comparing RDP vs VNC for the same color-depth, you'll find that RDP is simply a lot more responsive over the same connection.
(I've been using both for quite a few years now.)
There was a time when Microsoft were the good guys, where people suddenly found that they could write a product for DOS and it would run on almost any computer. That meant it was possible to become a software house with a lot less effort and money than before.
And then they got greedy and started competing in a lot of other markets (office software, financial software, bookkeeping software).
Back in the mid-late 90s, ISVs lived in fear that Microsoft would decide to take an interest in their market segment. They'd come in with an inferior product, and keep throwing money at it until you went out of business. Or they'd buy you out and retire your product from the market, leaving only their solution.