1. Buy high end task furniture (Haworth, Herman-Miller, etc.) but buy it "used". It's 1/2 or less the price, and often you can get the used high-end stuff for less than commodity new. 2. Get a telephone system that doesn't suck. This is harder than you might think. Today, I'd build something with Asterix/VOIP integrated with a customer database to do some real-time CTI. In the past, I've used Aspect successfully as well. Cisco's VOIP gear is nice, but overpriced. 3. Everyone gets their own . Whether it be a headset, keyboard, etc. Trust me, it makes sense. 4. Lockers outside the NOC for staff. Make them nice, tall and big, and nobody shares. 5. Plan for actual breaks from operations. Nobody can stare at a computer screen that many hours and stay alert.
1. How much money do you really have? Lots of people think they have money, but run out when it comes to all the details. 2. Do you want flash, or functionality? The two are sometimes complementary, but often one trumps the other. 3. How many people will staff? What's the schedule? This helps you figure out workstation configurations. 4. Are you putting multiple tiers in the same room? This is "best practice" if you do it right. 5. Are you handling customer calls directly? Do you deal with customers?
Basically, you need to figure out a lot of goals. A true "global" NOC can cost $50M easily for a telecom or comparable organization.
I've been a big fan of Barco for large projectors, and their IP-based solution is quite powerful. Recently, I rolled out a "public safety" SOC (security operations center) with 8 SXGA+ rear-projection displays. The largest I've worked on was 40+ of that style of display. Your garden-variety projector isn't cut out to handle this kind of duty-cycle. They're not cheap, but they're designed to operate 24x7x365, and many models have multiple lamps, etc. so that you can service them while they're online. So here's a few more things to think about:
* What goes on the "big screen" has to be useful. It must be grokable in a very short period of time. If you can't look at it for 2 seconds and get a good idea of what's going on, it's too complicated. * Multiple displays per operations person * Operational "graphs" that show overall statistics that matter to the people working, not to management. * Good task lighting. Good lighting period is everything. Pay a real designer to do this. * Good seating. We have let operations people pick chairs that fit their needs. Expect to spend $800-1k/person on seating. * Sound deadening/management. NOCs get loud, and managing the acoustics is important to make sure that people can "think" and they can interact with one another. * Ticketing is everything. Look at systems that are available commercially and for free. Consider writing your own if needed. If the system is streamlined to your own business, it will always be an impediment to getting the job done, which means people won't use it. If they don't use it, lots of knowledge is lost and post-mortems are more difficult.
Also, a few things that seem superfluous, but ended up being critical in some places I've worked (not all these were at the same place):
* Virtualized desktops (think RDP, X11, etc.) so that people can move and maintain their setup * Color-shifting lighting to compensate for normal rhythms of people on weird shifts. Turns out green is effective after lunch at helping people maintain focus. This isn't cheap, but it sure does have a big impact. * Keep your customers OUT OF THE NOC. A glass wall into the NOC is fine, but actually letting them in is distracting, and depending, can come with legal issues around privacy, HIPAA, etc. Best to keep them at a distance. * Before you let customers see the NOC, you warn people. We had a blinking lighting strip under the displays that was linked into the Crestron system so that you couldn't flip the LCD-glass for 10 seconds to give NOC operators a warning. You don't want customers seeing people picking their nose.:)
Finally, as nice as good facilities are, if you don't have the process and people, it's useless. People people people people. Good people create good processes. Promote from within, and develop a strategy to give people a career path. Otherwise, you'll burn people out, and get huge turnover. That sucks for everyone.
A man-trap, in the physical security world, is a "room" (loosely defined here) which has control points on both sides. Often you have to use two different forms of authorization, one for entry (i.e. a badge) and another for exit (biometrics, let's say). This allows it to *trap* anyone who tries to sneak through the system. What the article is really talking about is not a man-trap, but the anti-"bum" measures that banks use in many cities around ATMs inside a building. You have to put your ATM card into a slot, but it really doesn't read the card, it just verifies that you stuck a magstrip card into the slot. You then use your ATM card to access the ATM where it is presumably verified.
Setting anything in this method is absurd, and the physical security people should be fired on the spot for this kind of kindergarten mistake. While what likely happened is that it was turned this way when installed so that you could teach people to use it without having to deal with the slowdown of people actually being blocked, it's a bad way to behave, and shouldn't have been even turned on the first time this way. It may also be that, in fact, it was turned this way because of a problem with reliability of magstripe cards (they fail pretty regularly), and instead the system should have been converted to another form of identification -- Wiegand, RF proxy, etc.
I grew up in Austin, and I just got back. The construction on the interchange of US 71 (Ben White Blvd) and I-35 has now been under construction for almost a decade, it's massively behind schedule, last I heard, massively over budget, and is actually not that big of a construction project compared to the mixing bowl of I-495, I-395, I-95 in DC where I live now.
If it's any consolation, it'll likely cost 4-10x what they claim, and require 200 years to finish.
Personally, I carry a Namiki Vanishing Point retractable fountain pen. They run about $100 or so for the entry level. I prefer the black with rhodium, and it's a few dollars more.
While they don't write quite as well as the $500+ fountain pens I have used, they're infinately more convenient, and because they're retractable, you can shove it in your pocket easily. The one thing I've found is that if I buy nice pens, I don't lose them, where-as the cheap Uniball pens, I forget all over the place.
There are those who will say fountain pens are "old fashioned," and they're probably right, but they're easy enough to write with, and honestly, the quality and "feel" can't be beat. When I'm writing, I want decent paper, and a good pen, not some recycled toilet paper and a 5 cent mass produced piece of plastic.
That's odd, as my company had several thousand of the PD-series drives in the field---mostly the half height 1.5Gb drives (double sided, but you had to manually flip them). Mostly, they were used by banks, and media companies to store huge files that needed either WORM functionality, or re-writable functionality but with audit (something our interface controller did). I remember having to recover 20-30 disks that had been damaged because of a big proofing ink-jet printer at a NY media house. Basically the printer had atomized ink into the air and it got sucked into the case of the drive through the fan blowing the wrong way (but the way Panasonic designed it:/). THis meant that it deposited on the disc itself.
Fortunately, a little alcohol cleaned the disks and fixed them. We saw failure rate of drives be comparable to other drives (i.e. HP/Maxtor/IBM re-writable MO drives) and orders of magnitude better than the popular Pinnacle drives of the time. Disc failure rate, if handled correctly, was virtually zero across the customer base. We kept stats internally for product development feedback.
Honestly, there are of course always bad drives out there, and bad disc batches. Also, failure rates of the optical drives were much higher than CD-R. But, they were also pushed much harder than CD-R drives are today. I suspect if you ran in CD-RW all the time, you'd see similar failure rates because of heating of the laser diode. That was the most common failure.
The other poster is correct. While we did not examine CD-RW, as they didn't really exist at the time, and also violated a rule for archiving for the group we were doing work for, which is that you can change them undetectably, I would imagine they suffer from the same problem as other mediums. Panasonic's phase-change designs were excellent, and substantially more stable than MO-RW, but in the end, they were VERY sensitive to UV, which is why they were in cases that had shutters that had interlocks. CD-RWs don't have this.
I used to work for a company in Austin, TX whose speciality was optical drives (not CDs, but WORM mostly), and one of our customers was the National Archives. This was when CD-Rs were just coming out, and the NA was interested in a cost/benefit analysis of whether or not they could replace their expensive 14" WORM systems with cheap CD-Rs.
The first thing to understand is that WORM systems, true WORM systems, not the Magnetic-Optical pseudo-WORM systems, are built on ablation of material in the disc itself. In other words, you burn holes in the disc revealing a lower layer that is reflective. In the case of most discs, and Kodak especially, they were gold on the reflective layer for long-term stability. Various tests of accelerated degradation were performed in both climate stabilized and non-stabilized situations, and at worst, the discs were stable for 100 years before any error correction was necessary.
We decided to perform the same kind of evaluation of CD-Rs, and found that brand varied greatly. The best were stable for 3-4 years, the worst only 6-8 months if the climate changed dramatically. In addition, UV exposure had a radical impact on the life-span of the disc. Further research found out that the problem was the natural instability of the organic dyes that were used in the disc layers.
Basically, if the disc wasn't perfectly sealed (look at the work done in the referenced article, and how it starts at the edges), oxygen would get in and react with the dye, which would change it's characteristics relatively quickly. It doesn't take much before the dye structure collapses, and data becomes unreadable after a short period. While I suspect the dyes have gotten better over time, they're still organic last I knew, and still subject to degradation by contact with air. Quality control is the only thing that will get you anything here, and I suspect even the best dye-based discs can't make it past 20 years unless exposure to UV is totally eliminated.
What Kodak had developed was what they called "Century Discs", which were basically scaled down WORM discs, but in CD-ROM format. They were gold inside, non-reactive, and well made. They did, however, require a very expensive writer because they needed more power than a CD-R drive could ever hope to provide to force the burn away the spots. They were, however, readable in a normal drive.
That's just my experience, but everytime I've seen an organization talking about "archiving" on CD-R, I have issues with it. It's fine for "backup," where the data cycle is shorter, but true archival purposes (for example, financial data), it won't cut it. You either need to use WORM, or tape. Tape is, however, subject to problems over the cycles as well, witness the failing properties of 9-track tapes written by NASA in the 1970s (heard first hand, not sure where to find it written up). Linear-write systems are better than helical.
Just a few thoughts, but this is not an easy issue. You have to understand what you're storing, and how long it has to be readable before you consider an actual medium for storage.
Today, normal enterprise/campus networks don't work with any amount of power, quite honestly. Horizontal building cable is almost exclusively copper or multi-mode fiber, and riser/inter-building cable is single-mode, but relatively cheap stuff. Where this kind of thing comes into play is in the long-haul networks of companies like AT&T, Level 3, Sprint, etc., where you have 100Km+ between OA (Optical Amplifier) sites.
Many people are working to extend the OA interval to 600Km through doped and Raman amplifiers, which are giving you launch powers in the 30db+ range, and are starting to approach the powers that can do this. However, as someone pointed out, none of this happens with normal correct fiber installation. I know my company, which runs a large (tens of thousands of miles) network has reams of paper describing exact splice tray designs, stress on cables, bend angles. It goes down to how you support things going in and out of a OA, etc., and addresses the radius, which I believe we try and keep around 15-20cm minimum.
If you follow smart rules, these don't matter. If you don't, well, it probably won't affect anyone who is working outside the large telco space. The cost of an EDFA (Erbium doped fiber amplifier) is tens of thousands of dollars.
I've dropped my 15.2" PowerBook G4 about oh, 15 times, and other than a ding on the case, it's never had any damage. The cover was always closed. This includes knocking it off the table onto a carpeted floor. I've never had a notebook this well made, and tolerant of abuse that I'm stupid enough to subject it to.
Close, but not quite. AppleLink was only partially developed by Apple, and it was run by GE for a *VERY* hefty fee, even for Apple employees to use it. So, in order to get rid of GE, and get away from some of the idiocyncracies of AppleLink, Apple decided to write their own software. Unfortunately, about 90% of the way there, it got killed for being "outside our business," and Quantum bought the rights to the whole thing.
That's why in the early days, AOL was *VERY* Mac-centric and friendly. Often new releases for the Mac came out months before the PC releases, and had more features. This slowly changed, and eventually it got so bad that Apple felt they needed to offer a competitor. What to do? Well, they ended up having to license the executable code back from AOL for them to use in creating eWorld, but they never got the rights to the source again, and so anytime eWorld wanted a change, they had to pay AOL for it.
Pretty messed up, in both directions. This is the history as told to me by someone who worked for eWorld during their entire "lifespan".
BBN actually has 2 natural Class A addresses (4/8 and 8/8), which were transfered to GTE Internetworking, then Genuity, then to Level 3 during the acquisition. Very long story, but you kinda get to assign whatever you need when you get to be AS1 as well. Anyway, 4/8 is heavily divided up and assigned out to customers as well as being used for the internal network. During the integration by Level3, my understanding is that a lot of these will be renumbered into 4/8 from the Level3 blocks, just as Level3 will likely renumber to AS1. It's simply easier, and has a bit of cachet.
8/8, on the other hand, has never been used as far as I know, but is held in reserve, because simply getting that kind of address space flexibility is impossible in this day and age. Yeah, probably not the "right thing," to do, but there it is.
The funny thing is, even with only 1 or 5 star ratings, if you get enough of them, you begin to approximate a fair score. Love / hate eventually becomes a gray that you can judge more finely. In fact, one might even argue that only giving users 3 options (love/like/hate), would encourage more people to score albums, and thereby eventually give a more clear real-world indication of the quality.
No offense, but Bob Taylor is not the most disinterested source to quote. If you read Dealers in Lightning, you'll get a better view of what was going on at PARC at the time. I've met a bunch of people at PARC at various points, and most understand that the biggest flaw was the disconnect between PARC's goals and Xerox' goals. PARC was very long-term, and focused on innovation, where-as Xerox was very focused on what would help them next quarter.
For many reasons, Xerox was never going to capitalize on the Star. I've owned various D* machines (my last a Dandelion), and they were great, but they were $16,000 new, and made the Lisa look zippy and cheap. Xerox lost this game pretty fair and square. Bob Taylor was brilliant, but never ever to articulate to management what he was doing, and more importantly, how Xerox as a Fortune 10 company could use it to build a better marketplace.
It's a lot of sour apples, no pun intended, if you ask me.
This article is pretty much bunk, as everyone else has pointed out. The top ISPs in the world don't operate in a manner consistent with idiot security people's "guesses". Let's talk about some real risks to the routing fabric of the Internet:
Homogenous software. This doesn't sound like a huge problem, but it is. I'd say most Tier-1 ISPs run Juniper core routers, Cisco on the edge---long reasoning, but has to do with who does what best. This means that on the whole, most ISPs will be vulerable to the same attack. This is the same destructon issues that Microsoft presents. To give you an example, many moons ago, UUnet (before Worldcom) injected a bogus route entry accidentally, and it caused every Cisco router that heard it to reset the BGP session with that router. This continued forever until it it was identified. The only backbone not impacted was ANSnet, because they were using customer routers based on IBM hardware and gated.
Limited skill set distribution. The number of people who truly understand routing at the massive scale that the Tier-1s have to deal with is tiny. In any backbone, it's maybe 5-6 people, maybe less, that truly understand it in detail. At my old company we always argued that the fastest attack against the backbone was to puncture the tired of a car carrying the engineers to lunch.
Those really are the two biggest issues from my perspective. BGPv4 is attackable, always has been. SecureBGP deals with some aspects of the problem, but certainly not DOS issues. Huge routing updates consume CPU resources, and even with route dampening and other tricks, you can artificially (though rarely naturally) kill a router through BGP.
It's unfortunate that an article purporting to cover risks doesn't bother with the real ones and instead sticks to sensationalistic strategies.
If you want to understand BGP at a detailed level, and all the tricks that are used (or at least most of them), John Stewart's book "BGP-4 Inter-Domain Routing in the Internet" is the book I give people, and it's excellent. John definately is on the side of the world that gets it.
180 petabits per day? What kind of measurement is that? Where was it measured? How was it measured? Who was included? Were bits counted twice?
Just to give you an idea, I work for a large IP carrier, and we peak around oh, 200Gbps aggregate traffic entering the network. Gigabits/second is a good measurement of traffic, as is total gigabytes/terabytes... but to use the term petabit, implies they're using bandwidth, not data, and that asks where that was measured and how? There's not a lot of 200Gbps networks in the world.
Would you please define "complete and finished" for me? And BTW, many things in Jaguar are not OSS, in fact 90% of it isn't OSS. People in the Open Source community seem to revel in pointing out that Apple is based on some Open Source (mostly the command line tools), but that is not 90% of the OS, it is in fact probably 10% of the code base.
I don't buy OSX for Apache, or Perl, or anything like that. I buy it for a rich UI, and many other applications that I use that are not OSS. Maybe that's just me.
I think you need to leave the religion at the church, and quit letting it blind your understanding of commerce and reality.
To quote "they have suddenly introduced charges for stuff that people assumed would be free".
Note, that assumptions make an ass out of you and me. I'm a dedicated Apple user, and I even used to work for them, but never once did I expect that all upgrades would be free. Without getting into other companies who charge for "updates" that are barely more than bug fixes, Apple has traditionally only charged for the major releases. 10.1 cost some money, but not much, 10.2 cost money, but 10.2.1, 10.2.2, 10.2.3 didn't, some of which introduced new features like logging on file systems.
I do think the hardware prices are not comparable to the cheapest in the PC world, but they are comparable (esp the notebooks) to the comparable hardware integration. Would I pay for iMovie? Nope, I don't do movies. Would I pay for iPhoto? Probably, since it currently has 1500+ photos of mine in it. Would I pay for iSync+iCal+iSomethingNotReleasedYet, probably.
As many people have pointed out, they're not saying it won't come free, you'll just have to pay for upgrades. Many people think Windows comes for free, and you have to pay for upgrades. It may not be "rigdyÖs-OSS-deems-right" but it is "normal."
Funny, there also was never a 15" CRT Studio Display, only 17" (I own 2) and 19". As other people have pointed out, I suspect this person never owned an actual Cube, but instead just regurgitated everything he'd heard from various trade rags.
BTW, if you're planning to flip a computer upside down, shove cables in, etc., personally, I'd power it down, rather than spin the gyroscope (hard disc) around a lot, and risk damaging it. But hey, that's just me.
Ok, how about : I had to keep the clutch half open if i wanted to go slower then 80..
A nice lie, but also BS.
But i do invite you, get a fireblade or a duc 996 or TL or so... tune it to perfection... and then show me that you can keep it at legal speeds...
I've owned a 916 and had no problems racing it on the track, and riding legally during the day. Funny, it does actually go 20mph just fine. Not particularly comfortable, but hey, that's the design. I've also ridden TL1000S, no problem, Aprilia Mille, not a problem, funny that. Sounds like operator error.
But i'm talking about a bike that goes from 0 - 100 (kph) in 2.5 secs, that in first gear goes up to 140 kph... it's not the same... try using a formula 1 car in every day traffic...
Honestly, this is just more distraction from personal responsibility. Plenty of people manage to ride safely and sanely. That you choose not to is not a reflection of the bike, but of the rider.
On the other had, it's just plain impossible to respect the speed limits with any stock supersport bike these days.
Bullsh*t. It's called self-control. I ride quickly on the open roads, as do all my friends, but that doesn't mean we ride at 100MPH+ on the streets. That's what track days are for.
I'm sorry, but you and your friends were nothing more than squids who took stupid risks, and paid the price. Personal responsibility, buster. It's just like people who choose to ride without a helmet, in shorts and a set of flip flops on a bike that can do 200MPH. Yeah, there's using your noggin.
The head is well protected, the neck/spine is not as much so. Dianese has some stuff for racing for this. For the back, though, nothing beats armor, such as the stuff I wear from Vanson. This helps protect the spine, and in many cases is fully articulated. Not the most comfortable thing, but it can be very helpful in a crash.
Slashdot Mirror
Oh a few more thoughts:
1. Buy high end task furniture (Haworth, Herman-Miller, etc.) but buy it "used". It's 1/2 or less the price, and often you can get the used high-end stuff for less than commodity new.
2. Get a telephone system that doesn't suck. This is harder than you might think. Today, I'd build something with Asterix/VOIP integrated with a customer database to do some real-time CTI. In the past, I've used Aspect successfully as well. Cisco's VOIP gear is nice, but overpriced.
3. Everyone gets their own . Whether it be a headset, keyboard, etc. Trust me, it makes sense.
4. Lockers outside the NOC for staff. Make them nice, tall and big, and nobody shares.
5. Plan for actual breaks from operations. Nobody can stare at a computer screen that many hours and stay alert.
There's a million more details.
So, a couple things to think about:
1. How much money do you really have? Lots of people think they have money, but run out when it comes to all the details.
2. Do you want flash, or functionality? The two are sometimes complementary, but often one trumps the other.
3. How many people will staff? What's the schedule? This helps you figure out workstation configurations.
4. Are you putting multiple tiers in the same room? This is "best practice" if you do it right.
5. Are you handling customer calls directly? Do you deal with customers?
Basically, you need to figure out a lot of goals. A true "global" NOC can cost $50M easily for a telecom or comparable organization.
I've been a big fan of Barco for large projectors, and their IP-based solution is quite powerful. Recently, I rolled out a "public safety" SOC (security operations center) with 8 SXGA+ rear-projection displays. The largest I've worked on was 40+ of that style of display. Your garden-variety projector isn't cut out to handle this kind of duty-cycle. They're not cheap, but they're designed to operate 24x7x365, and many models have multiple lamps, etc. so that you can service them while they're online. So here's a few more things to think about:
* What goes on the "big screen" has to be useful. It must be grokable in a very short period of time. If you can't look at it for 2 seconds and get a good idea of what's going on, it's too complicated.
* Multiple displays per operations person
* Operational "graphs" that show overall statistics that matter to the people working, not to management.
* Good task lighting. Good lighting period is everything. Pay a real designer to do this.
* Good seating. We have let operations people pick chairs that fit their needs. Expect to spend $800-1k/person on seating.
* Sound deadening/management. NOCs get loud, and managing the acoustics is important to make sure that people can "think" and they can interact with one another.
* Ticketing is everything. Look at systems that are available commercially and for free. Consider writing your own if needed. If the system is streamlined to your own business, it will always be an impediment to getting the job done, which means people won't use it. If they don't use it, lots of knowledge is lost and post-mortems are more difficult.
Also, a few things that seem superfluous, but ended up being critical in some places I've worked (not all these were at the same place):
* Virtualized desktops (think RDP, X11, etc.) so that people can move and maintain their setup :)
* Color-shifting lighting to compensate for normal rhythms of people on weird shifts. Turns out green is effective after lunch at helping people maintain focus. This isn't cheap, but it sure does have a big impact.
* Keep your customers OUT OF THE NOC. A glass wall into the NOC is fine, but actually letting them in is distracting, and depending, can come with legal issues around privacy, HIPAA, etc. Best to keep them at a distance.
* Before you let customers see the NOC, you warn people. We had a blinking lighting strip under the displays that was linked into the Crestron system so that you couldn't flip the LCD-glass for 10 seconds to give NOC operators a warning. You don't want customers seeing people picking their nose.
Finally, as nice as good facilities are, if you don't have the process and people, it's useless. People people people people. Good people create good processes. Promote from within, and develop a strategy to give people a career path. Otherwise, you'll burn people out, and get huge turnover. That sucks for everyone.
A man-trap, in the physical security world, is a "room" (loosely defined here) which has control points on both sides. Often you have to use two different forms of authorization, one for entry (i.e. a badge) and another for exit (biometrics, let's say). This allows it to *trap* anyone who tries to sneak through the system. What the article is really talking about is not a man-trap, but the anti-"bum" measures that banks use in many cities around ATMs inside a building. You have to put your ATM card into a slot, but it really doesn't read the card, it just verifies that you stuck a magstrip card into the slot. You then use your ATM card to access the ATM where it is presumably verified.
Setting anything in this method is absurd, and the physical security people should be fired on the spot for this kind of kindergarten mistake. While what likely happened is that it was turned this way when installed so that you could teach people to use it without having to deal with the slowdown of people actually being blocked, it's a bad way to behave, and shouldn't have been even turned on the first time this way. It may also be that, in fact, it was turned this way because of a problem with reliability of magstripe cards (they fail pretty regularly), and instead the system should have been converted to another form of identification -- Wiegand, RF proxy, etc.
I grew up in Austin, and I just got back. The construction on the interchange of US 71 (Ben White Blvd) and I-35 has now been under construction for almost a decade, it's massively behind schedule, last I heard, massively over budget, and is actually not that big of a construction project compared to the mixing bowl of I-495, I-395, I-95 in DC where I live now.
If it's any consolation, it'll likely cost 4-10x what they claim, and require 200 years to finish.
Personally, I carry a Namiki Vanishing Point retractable fountain pen. They run about $100 or so for the entry level. I prefer the black with rhodium, and it's a few dollars more.
While they don't write quite as well as the $500+ fountain pens I have used, they're infinately more convenient, and because they're retractable, you can shove it in your pocket easily. The one thing I've found is that if I buy nice pens, I don't lose them, where-as the cheap Uniball pens, I forget all over the place.
There are those who will say fountain pens are "old fashioned," and they're probably right, but they're easy enough to write with, and honestly, the quality and "feel" can't be beat. When I'm writing, I want decent paper, and a good pen, not some recycled toilet paper and a 5 cent mass produced piece of plastic.
That's odd, as my company had several thousand of the PD-series drives in the field---mostly the half height 1.5Gb drives (double sided, but you had to manually flip them). Mostly, they were used by banks, and media companies to store huge files that needed either WORM functionality, or re-writable functionality but with audit (something our interface controller did). I remember having to recover 20-30 disks that had been damaged because of a big proofing ink-jet printer at a NY media house. Basically the printer had atomized ink into the air and it got sucked into the case of the drive through the fan blowing the wrong way (but the way Panasonic designed it :/). THis meant that it deposited on the disc itself.
Fortunately, a little alcohol cleaned the disks and fixed them. We saw failure rate of drives be comparable to other drives (i.e. HP/Maxtor/IBM re-writable MO drives) and orders of magnitude better than the popular Pinnacle drives of the time. Disc failure rate, if handled correctly, was virtually zero across the customer base. We kept stats internally for product development feedback.
Honestly, there are of course always bad drives out there, and bad disc batches. Also, failure rates of the optical drives were much higher than CD-R. But, they were also pushed much harder than CD-R drives are today. I suspect if you ran in CD-RW all the time, you'd see similar failure rates because of heating of the laser diode. That was the most common failure.
The other poster is correct. While we did not examine CD-RW, as they didn't really exist at the time, and also violated a rule for archiving for the group we were doing work for, which is that you can change them undetectably, I would imagine they suffer from the same problem as other mediums. Panasonic's phase-change designs were excellent, and substantially more stable than MO-RW, but in the end, they were VERY sensitive to UV, which is why they were in cases that had shutters that had interlocks. CD-RWs don't have this.
I used to work for a company in Austin, TX whose speciality was optical drives (not CDs, but WORM mostly), and one of our customers was the National Archives. This was when CD-Rs were just coming out, and the NA was interested in a cost/benefit analysis of whether or not they could replace their expensive 14" WORM systems with cheap CD-Rs.
The first thing to understand is that WORM systems, true WORM systems, not the Magnetic-Optical pseudo-WORM systems, are built on ablation of material in the disc itself. In other words, you burn holes in the disc revealing a lower layer that is reflective. In the case of most discs, and Kodak especially, they were gold on the reflective layer for long-term stability. Various tests of accelerated degradation were performed in both climate stabilized and non-stabilized situations, and at worst, the discs were stable for 100 years before any error correction was necessary.
We decided to perform the same kind of evaluation of CD-Rs, and found that brand varied greatly. The best were stable for 3-4 years, the worst only 6-8 months if the climate changed dramatically. In addition, UV exposure had a radical impact on the life-span of the disc. Further research found out that the problem was the natural instability of the organic dyes that were used in the disc layers.
Basically, if the disc wasn't perfectly sealed (look at the work done in the referenced article, and how it starts at the edges), oxygen would get in and react with the dye, which would change it's characteristics relatively quickly. It doesn't take much before the dye structure collapses, and data becomes unreadable after a short period. While I suspect the dyes have gotten better over time, they're still organic last I knew, and still subject to degradation by contact with air. Quality control is the only thing that will get you anything here, and I suspect even the best dye-based discs can't make it past 20 years unless exposure to UV is totally eliminated.
What Kodak had developed was what they called "Century Discs", which were basically scaled down WORM discs, but in CD-ROM format. They were gold inside, non-reactive, and well made. They did, however, require a very expensive writer because they needed more power than a CD-R drive could ever hope to provide to force the burn away the spots. They were, however, readable in a normal drive.
That's just my experience, but everytime I've seen an organization talking about "archiving" on CD-R, I have issues with it. It's fine for "backup," where the data cycle is shorter, but true archival purposes (for example, financial data), it won't cut it. You either need to use WORM, or tape. Tape is, however, subject to problems over the cycles as well, witness the failing properties of 9-track tapes written by NASA in the 1970s (heard first hand, not sure where to find it written up). Linear-write systems are better than helical.
Just a few thoughts, but this is not an easy issue. You have to understand what you're storing, and how long it has to be readable before you consider an actual medium for storage.
Today, normal enterprise/campus networks don't work with any amount of power, quite honestly. Horizontal building cable is almost exclusively copper or multi-mode fiber, and riser/inter-building cable is single-mode, but relatively cheap stuff. Where this kind of thing comes into play is in the long-haul networks of companies like AT&T, Level 3, Sprint, etc., where you have 100Km+ between OA (Optical Amplifier) sites.
:-)
Many people are working to extend the OA interval to 600Km through doped and Raman amplifiers, which are giving you launch powers in the 30db+ range, and are starting to approach the powers that can do this. However, as someone pointed out, none of this happens with normal correct fiber installation. I know my company, which runs a large (tens of thousands of miles) network has reams of paper describing exact splice tray designs, stress on cables, bend angles. It goes down to how you support things going in and out of a OA, etc., and addresses the radius, which I believe we try and keep around 15-20cm minimum.
If you follow smart rules, these don't matter. If you don't, well, it probably won't affect anyone who is working outside the large telco space. The cost of an EDFA (Erbium doped fiber amplifier) is tens of thousands of dollars.
No story, move along.
I've dropped my 15.2" PowerBook G4 about oh, 15 times, and other than a ding on the case, it's never had any damage. The cover was always closed. This includes knocking it off the table onto a carpeted floor. I've never had a notebook this well made, and tolerant of abuse that I'm stupid enough to subject it to.
Close, but not quite. AppleLink was only partially developed by Apple, and it was run by GE for a *VERY* hefty fee, even for Apple employees to use it. So, in order to get rid of GE, and get away from some of the idiocyncracies of AppleLink, Apple decided to write their own software. Unfortunately, about 90% of the way there, it got killed for being "outside our business," and Quantum bought the rights to the whole thing.
That's why in the early days, AOL was *VERY* Mac-centric and friendly. Often new releases for the Mac came out months before the PC releases, and had more features. This slowly changed, and eventually it got so bad that Apple felt they needed to offer a competitor. What to do? Well, they ended up having to license the executable code back from AOL for them to use in creating eWorld, but they never got the rights to the source again, and so anytime eWorld wanted a change, they had to pay AOL for it.
Pretty messed up, in both directions. This is the history as told to me by someone who worked for eWorld during their entire "lifespan".
BBN actually has 2 natural Class A addresses (4/8 and 8/8), which were transfered to GTE Internetworking, then Genuity, then to Level 3 during the acquisition. Very long story, but you kinda get to assign whatever you need when you get to be AS1 as well. Anyway, 4/8 is heavily divided up and assigned out to customers as well as being used for the internal network. During the integration by Level3, my understanding is that a lot of these will be renumbered into 4/8 from the Level3 blocks, just as Level3 will likely renumber to AS1. It's simply easier, and has a bit of cachet.
8/8, on the other hand, has never been used as far as I know, but is held in reserve, because simply getting that kind of address space flexibility is impossible in this day and age. Yeah, probably not the "right thing," to do, but there it is.
The funny thing is, even with only 1 or 5 star ratings, if you get enough of them, you begin to approximate a fair score. Love / hate eventually becomes a gray that you can judge more finely. In fact, one might even argue that only giving users 3 options (love/like/hate), would encourage more people to score albums, and thereby eventually give a more clear real-world indication of the quality.
No offense, but Bob Taylor is not the most disinterested source to quote. If you read Dealers in Lightning, you'll get a better view of what was going on at PARC at the time. I've met a bunch of people at PARC at various points, and most understand that the biggest flaw was the disconnect between PARC's goals and Xerox' goals. PARC was very long-term, and focused on innovation, where-as Xerox was very focused on what would help them next quarter.
For many reasons, Xerox was never going to capitalize on the Star. I've owned various D* machines (my last a Dandelion), and they were great, but they were $16,000 new, and made the Lisa look zippy and cheap. Xerox lost this game pretty fair and square. Bob Taylor was brilliant, but never ever to articulate to management what he was doing, and more importantly, how Xerox as a Fortune 10 company could use it to build a better marketplace.
It's a lot of sour apples, no pun intended, if you ask me.
What, not The Iron Works? For beef ribs anyway. :-) Who needs A/C in Texas? That's for wimps.
Those really are the two biggest issues from my perspective. BGPv4 is attackable, always has been. SecureBGP deals with some aspects of the problem, but certainly not DOS issues. Huge routing updates consume CPU resources, and even with route dampening and other tricks, you can artificially (though rarely naturally) kill a router through BGP.
It's unfortunate that an article purporting to cover risks doesn't bother with the real ones and instead sticks to sensationalistic strategies.
If you want to understand BGP at a detailed level, and all the tricks that are used (or at least most of them), John Stewart's book "BGP-4 Inter-Domain Routing in the Internet" is the book I give people, and it's excellent. John definately is on the side of the world that gets it.
Well, how can I confirm a figure that's not got any math behind it?
I can say pretty authoritatively there's not 10 networks that can do 200+Gbps of traffic at a time. So how did this number get created?
180 petabits per day? What kind of measurement is that? Where was it measured? How was it measured? Who was included? Were bits counted twice?
Just to give you an idea, I work for a large IP carrier, and we peak around oh, 200Gbps aggregate traffic entering the network. Gigabits/second is a good measurement of traffic, as is total gigabytes/terabytes... but to use the term petabit, implies they're using bandwidth, not data, and that asks where that was measured and how? There's not a lot of 200Gbps networks in the world.
Would you please define "complete and finished" for me? And BTW, many things in Jaguar are not OSS, in fact 90% of it isn't OSS. People in the Open Source community seem to revel in pointing out that Apple is based on some Open Source (mostly the command line tools), but that is not 90% of the OS, it is in fact probably 10% of the code base.
I don't buy OSX for Apache, or Perl, or anything like that. I buy it for a rich UI, and many other applications that I use that are not OSS. Maybe that's just me.
I think you need to leave the religion at the church, and quit letting it blind your understanding of commerce and reality.
To quote "they have suddenly introduced charges for stuff that people assumed would be free".
Note, that assumptions make an ass out of you and me. I'm a dedicated Apple user, and I even used to work for them, but never once did I expect that all upgrades would be free. Without getting into other companies who charge for "updates" that are barely more than bug fixes, Apple has traditionally only charged for the major releases. 10.1 cost some money, but not much, 10.2 cost money, but 10.2.1, 10.2.2, 10.2.3 didn't, some of which introduced new features like logging on file systems.
I do think the hardware prices are not comparable to the cheapest in the PC world, but they are comparable (esp the notebooks) to the comparable hardware integration. Would I pay for iMovie? Nope, I don't do movies. Would I pay for iPhoto? Probably, since it currently has 1500+ photos of mine in it. Would I pay for iSync+iCal+iSomethingNotReleasedYet, probably.
As many people have pointed out, they're not saying it won't come free, you'll just have to pay for upgrades. Many people think Windows comes for free, and you have to pay for upgrades. It may not be "rigdyÖs-OSS-deems-right" but it is "normal."
Funny, there also was never a 15" CRT Studio Display, only 17" (I own 2) and 19". As other people have pointed out, I suspect this person never owned an actual Cube, but instead just regurgitated everything he'd heard from various trade rags.
BTW, if you're planning to flip a computer upside down, shove cables in, etc., personally, I'd power it down, rather than spin the gyroscope (hard disc) around a lot, and risk damaging it. But hey, that's just me.
Ok, how about : I had to keep the clutch half open if i wanted to go slower then 80 ..
... tune it to perfection ... and then show me that you can keep it at legal speeds ...
... it's not the same ... try using a formula 1 car in every day traffic ...
A nice lie, but also BS.
But i do invite you, get a fireblade or a duc 996 or TL or so
I've owned a 916 and had no problems racing it on the track, and riding legally during the day. Funny, it does actually go 20mph just fine. Not particularly comfortable, but hey, that's the design. I've also ridden TL1000S, no problem, Aprilia Mille, not a problem, funny that. Sounds like operator error.
But i'm talking about a bike that goes from 0 - 100 (kph) in 2.5 secs, that in first gear goes up to 140 kph
Honestly, this is just more distraction from personal responsibility. Plenty of people manage to ride safely and sanely. That you choose not to is not a reflection of the bike, but of the rider.
Bullsh*t. It's called self-control. I ride quickly on the open roads, as do all my friends, but that doesn't mean we ride at 100MPH+ on the streets. That's what track days are for.
I'm sorry, but you and your friends were nothing more than squids who took stupid risks, and paid the price. Personal responsibility, buster. It's just like people who choose to ride without a helmet, in shorts and a set of flip flops on a bike that can do 200MPH. Yeah, there's using your noggin.
Darwinian evolution. Get out of the gene pool.
The head is well protected, the neck/spine is not as much so. Dianese has some stuff for racing for this. For the back, though, nothing beats armor, such as the stuff I wear from Vanson. This helps protect the spine, and in many cases is fully articulated. Not the most comfortable thing, but it can be very helpful in a crash.