I read the article itself. I do not need you to interpret it for me. I encourage everyone to click on your link and judge for themselves. You're full of it.
You didn't read the article you cited, did you? It made the exact opposite point of what you argue - that iTunes is immensely profitable for the firm. Read the link you posted.
Hi Apple! Been a big fan since I dumped Linux in '04 or so. I've really enjoyed MacOS X and my laptops. You sold me a good UNIX with support for Microsoft Office - which I need. But you know what? Ever since you started making money hands over fist with iTunes, you've started REALLY SUCKING as a company. I don't want to buy from you any longer. My next phone will NOT be an iPhone. My next laptop will NOT be another Macbook. I think I'll be perfectly happy with an unlocked Nexus and a laptop running FreeBSD. So... FUCK OFF, Apple. For me, your time has come and gone. -M
To remove my credit card from their database. Also, to remove my customer record. And finally, to let them know just how displeased I am with their business practices.
I was a customer for over ten years. Spent well over ten thousand dollars there in books and other items. But for the last several years their customer support has declined, their partner businesses engage in numerous disreputable practices that mirror the abuses at ebay, their manipulation of book rankings on so-called adult material (gay), and they seem intent on monopolizing the epublishing trade. I closed my account and won't look back.
Yes, the Kindle-DX looks like a nice machine. But what one gives up in basic rights as a reader is more than enough to keep me buying used books printed on dead trees for some time. And I can always scan the books I buy to load on an ereader with less virulent DRM limitations and corporate controls. I own an iRex iLiad, that while not the best manufacturer, at least they offer a free Linux development environment to download and install. Users are hacking new software on that platform. Does anyone here expect Amazon to allow that? Not me.
BTW: closing my account with Amazon took several phone calls and numerous transfers from one department to the next. They don't like it when customers attempt to leave them and make the process as difficult as possible. Yet another reason to never give them my money again.
There's a huge difference between publishing a legitimate paper by an unknown and an author knowingly writing falsehoods for a journal he or she knows is a sham. When academics knowingly violate standards of truth, those academics risk losing their credentials. It really does happen. Ph.Ds, M.Ds and licenses to practice are really revoked by accredited institutions and state licensing boards in certain circumstances like these.
Make an example of this academic sophistry. Yank everyone's credentials who knowingly published false information in a false academic journal. Do the world a favor.
Any scientist or doctor found to have knowingly written a paper making false scientific claims in a propaganda journal should lose their academic standing. Period. Ph.D? Revoked by the granting institution. M.D? Gone. Along with his or her medical license.
That's how the scientific and medical communities ought to fix this. Because when bogus science is published in medical journals, some innocent people needlessly die.
If the Kindle viewer could display LaTeX, you'd have a point. But entering in plain text Tex as annotation, without any rendering back in real math symbols, is not what I call user friendly.
It's not. The iRex DR1000SW is the only device that would meet your needs right now. And it's significantly more expensive, manufactured by an almost fly-by-night outfit to rather poor tolerances, and the software sucks. But, it does real pen input for annotations, allowing you to save and merge math and non alphanumeric characters in your PDFs.
I'd wait for a device from Sony or Apple and pray for good pen input.
Annotating math with that keyboard strikes me as... less fun than banging my head against a brick wall. But that's just me.
Battlestar Galactica is a television show. It's a fine enough show I guess, but it is not worthy of wasting the time of a body that meets ostensibly to diplomatically resolve real world conflicts, forge various international agreements, and - at times - deploy troops for peacekeeping. That television show is fantasy. What's going on at Darfur is the real thing.
I'm happy with 3-ware cards and we have lots of experience with them. We're not going RAID 6, we're doing mirrored pairs with a manual data spread of AFS volumes. RAID 6 might have been a good solution too, but we don't need the additional capacity RAID 6 offers nor do we want the write performance penalties associated with calculating checksums.
RE: HA, that reply was in error. We'll have an HA failover system off-site with the heartbeat and data networks connected via VLAN. I actually noted that there is a failover system in the first paragraph of my top comment, though I didn't go into detail. But, yeah... you'd have to be nuts to set up a single point of failure like that without some kind of automated HA solution in place.
On the other hand, I might argue that (if you could only find a way to manage it) spreading the load over a larger number of identical servers would enable you to minimize the cost of the loss of a single server, and at the same time at least provide a single reference platform with identical support needs across all nodes.
At the cost of either:
A) expensive RAM purchase for each cluster node, in order to run critical filesystems and dbs out of RAMdisk.
B) a couple to three orders of magnitude additional latency and at least an order of magnitude data bandwidth loss for RAID disk access compared to system RAM.
Assuming the same budget I doubt I'd be able to buy more than one additional machine with RAM savings (assuming I fit the host with 4GB or 8GB RAM and ran off disk). The servers have tripple redundant power supplies (though I'd have been happy with dual). In all honesty, the RAM was only about 1/3 the cost of the total for the Xen instance server in question.
I think to go your approach would have required considerably more money for the same effective performance.
Otherwise I think breaking both storage and servers up a bit more would have made more sense given your constraints.
Breaking data volumes up across multiple fileservers is exactly what AFS does very well. I would have preferred that approach (a cluster of AFS fileservers) over a cluster of Xen instance servers had enough funds been available. In all honestly, what I'm doing is upgrading the totality of our server infrastructure with the least funds possible.
This whole approach was not chosen to try out some crazy home-brew IT solution for the fun of it, we're really facing major funding constraints. I wouldn't recommend this approach for anyone with enough money to do it the traditional way. And yeah, if I screw up badly, it's my ass. Such is life in IT.
Perhaps you could add a few tens of thousands of dollars to my budget and then I'll "do it better". Until then, I'll decide what is possible given my available funds.
No. I have two primary servers and two failovers. I did not have enough funds to purchase two sets of identical machines, so I will use older hardware at a remote site connected via a VLAN as the failover system.
Total disk capacity is not the issue. I need I/O capacity on the main server, which is why I went with two 8 port SATA cards on the file server. However, one or both goes down and the failover takes control, users will experience noticeable performance degradation until the primary comes back online. That is due to budgetary constraints of the project.
Rereading my initial comment I note that I did specify a failover server. Perhaps you did NOT read the first paragraph of my top level comment and thus your post is total bullshit. For the last sentence of the first paragraph says thusly:
Those deltas will then be copied to a remote fail-over server periodically as well.
No. It is accurate that I did not list the failover plan. However, It is also true that there is a plan for remote failover hardware, I just didn't think it was worth pointing out in the initial message. But, of course, this is slashdot. Someone has to find something wrong with whatever is said in order to drive bullshit "discussion" (that goes nowhere).
We'll have a remote failover system in place. But you're absolutely right, what's driving this approach is entirely due to budgetary concerns. We're broke and don't have enough funds to replace all the old hardware we currently have in production. This was the best solution I could find that fit my available resources this FY. And given the economy, I doubt we'll see significant budgetary increases any time soon.
Nope. I have a fail-over system ready to put in place as well. We'll copy the deltas over periodically to a server at a remote site connected by VLAN to the same network.
BTW: my comment was not a proposal for you to fund, just a partial description for the community.
No, not for personal use or gaming. It will run Linux with a Xen kernel and is intended to replace nearly all of our old individual servers. Everything from the piddly servers like DNS, LDAP, Kerberos, and our minimal web services to the AFS db servers. No file services on that beast though, I'm not crazy - no disk I/O-RAM access contention please. My plan is to copy an entire OS image of/usr into a RAMFS filesystem in the top level Dom 0 domain and then cross mount that as RO in each Xen instance. We'll also stick small SQL server and other dbs copies in local tempfs RAMdisks too. Everything in RAM will be snapshotted and saved to physical disk periodically. Those deltas will then be copied to a remote fail-over server periodically as well.
It should be both reasonably stable and blindingly fast.
Another machine will handle AFS and some NFS file services, which has up to sixteen SATA disks attached to two 8 port 3-Ware RAID cards, thus spreading I/O load across two PCI buses. No, we don't need all that disk space - we need the I/O performance. It too should be reasonably fast. We're gearing up to connect that either by several channel bonded 1Gb to a CISCO 6509, or - if we're lucky - we'll just go 10Gb optical. We'll see how the finances work out there.
This is how departmental IT is done. Or, at least, it's how it *should* be done. I spent less than $25K on these two computers and they will replace well a couple hundred thousand dollars worth of accumulated hardware purchased over the last ten years and now fully depreciated.
You're saying the Kindle sucks because it doesn't support the features you need - which are available when you're willing to pay twice the cost; and that you're not in the target market for Kindle but you are for the DR1000.
Yes. That's what I'm saying. The Kindle sucks because as a product it doesn't even come close to meeting my needs. It doesn't matter if it sells for $1000, $350, or $.25. If it doesn't perform the functions I need, to buy it is a waste of money. Period.
I need a device that would replace a physical 8.5" x 11" notebook and allow me to manipulate electronic documents as if I were holding a printed copy. I want to write in eink because using a keyboard does not allow for good math annotation. That feature is critical to me. I would gladly spend $1000 for a light device that replaced the many pounds of dead trees I carry around in my backpack / shoulder bag every day.
That's great! so uh, remind me why you bothered posting?
It's a 1280x1024 10.2" diagonal screen, capable of displaying a full 8 1/2" x 11" page with good quality large fonts. You're right about PDF support on the Sony device, however.
It also supports eink annotation and native PDF rendering. It's the only ereader device appropriate for academics and technical professionals on the market today. The price tag is irrelevant given its feature set. I could care less that it costs a grand if it saves me significant project time.
When I read a PDF I need quality image support for interpreting graphs and other types of visual data. The Kindle doesn't come close. Yes, Amazon offers a PDF "conversion" service. In the process, formatting and image support is either lost or horribly mangled.
Never mind the total lack of touch support for eink annotations makes the thing worthless for serious use. Fine if you want to spend $350 for a device to read novels on the train. But if you want to read technical papers and annotate in math, the Kindle doesn't come close to being a useful device.
The only thing out there that does meet that need is:
Slashdot Mirror
I read the article itself. I do not need you to interpret it for me. I encourage everyone to click on your link and judge for themselves. You're full of it.
You didn't read the article you cited, did you? It made the exact opposite point of what you argue - that iTunes is immensely profitable for the firm. Read the link you posted.
Excellent point. Fortunately, iTunes doesn't install on FreeBSD. -M
Cite that please. You're arguing that they make little to no money from content distribution through iTunes? I think you're full of it.
Hi Apple! Been a big fan since I dumped Linux in '04 or so. I've really enjoyed MacOS X and my laptops. You sold me a good UNIX with support for Microsoft Office - which I need. But you know what? Ever since you started making money hands over fist with iTunes, you've started REALLY SUCKING as a company. I don't want to buy from you any longer. My next phone will NOT be an iPhone. My next laptop will NOT be another Macbook. I think I'll be perfectly happy with an unlocked Nexus and a laptop running FreeBSD. So... FUCK OFF, Apple. For me, your time has come and gone. -M
To remove my credit card from their database. Also, to remove my customer record. And finally, to let them know just how displeased I am with their business practices.
I was a customer for over ten years. Spent well over ten thousand dollars there in books and other items. But for the last several years their customer support has declined, their partner businesses engage in numerous disreputable practices that mirror the abuses at ebay, their manipulation of book rankings on so-called adult material (gay), and they seem intent on monopolizing the epublishing trade. I closed my account and won't look back.
Yes, the Kindle-DX looks like a nice machine. But what one gives up in basic rights as a reader is more than enough to keep me buying used books printed on dead trees for some time. And I can always scan the books I buy to load on an ereader with less virulent DRM limitations and corporate controls. I own an iRex iLiad, that while not the best manufacturer, at least they offer a free Linux development environment to download and install. Users are hacking new software on that platform. Does anyone here expect Amazon to allow that? Not me.
BTW: closing my account with Amazon took several phone calls and numerous transfers from one department to the next. They don't like it when customers attempt to leave them and make the process as difficult as possible. Yet another reason to never give them my money again.
There's a huge difference between publishing a legitimate paper by an unknown and an author knowingly writing falsehoods for a journal he or she knows is a sham. When academics knowingly violate standards of truth, those academics risk losing their credentials. It really does happen. Ph.Ds, M.Ds and licenses to practice are really revoked by accredited institutions and state licensing boards in certain circumstances like these.
Make an example of this academic sophistry. Yank everyone's credentials who knowingly published false information in a false academic journal. Do the world a favor.
Any scientist or doctor found to have knowingly written a paper making false scientific claims in a propaganda journal should lose their academic standing. Period. Ph.D? Revoked by the granting institution. M.D? Gone. Along with his or her medical license.
That's how the scientific and medical communities ought to fix this. Because when bogus science is published in medical journals, some innocent people needlessly die.
If the Kindle viewer could display LaTeX, you'd have a point. But entering in plain text Tex as annotation, without any rendering back in real math symbols, is not what I call user friendly.
It's not. The iRex DR1000SW is the only device that would meet your needs right now. And it's significantly more expensive, manufactured by an almost fly-by-night outfit to rather poor tolerances, and the software sucks. But, it does real pen input for annotations, allowing you to save and merge math and non alphanumeric characters in your PDFs.
I'd wait for a device from Sony or Apple and pray for good pen input.
Annotating math with that keyboard strikes me as ... less fun than banging my head against a brick wall. But that's just me.
Battlestar Galactica is a television show. It's a fine enough show I guess, but it is not worthy of wasting the time of a body that meets ostensibly to diplomatically resolve real world conflicts, forge various international agreements, and - at times - deploy troops for peacekeeping. That television show is fantasy. What's going on at Darfur is the real thing.
WTF?
I'm happy with 3-ware cards and we have lots of experience with them. We're not going RAID 6, we're doing mirrored pairs with a manual data spread of AFS volumes. RAID 6 might have been a good solution too, but we don't need the additional capacity RAID 6 offers nor do we want the write performance penalties associated with calculating checksums.
RE: HA, that reply was in error. We'll have an HA failover system off-site with the heartbeat and data networks connected via VLAN. I actually noted that there is a failover system in the first paragraph of my top comment, though I didn't go into detail. But, yeah ... you'd have to be nuts to set up a single point of failure like that without some kind of automated HA solution in place.
At the cost of either:
A) expensive RAM purchase for each cluster node, in order to run critical filesystems and dbs out of RAMdisk.
B) a couple to three orders of magnitude additional latency and at least an order of magnitude data bandwidth loss for RAID disk access compared to system RAM.
Assuming the same budget I doubt I'd be able to buy more than one additional machine with RAM savings (assuming I fit the host with 4GB or 8GB RAM and ran off disk). The servers have tripple redundant power supplies (though I'd have been happy with dual). In all honesty, the RAM was only about 1/3 the cost of the total for the Xen instance server in question.
I think to go your approach would have required considerably more money for the same effective performance.
Breaking data volumes up across multiple fileservers is exactly what AFS does very well. I would have preferred that approach (a cluster of AFS fileservers) over a cluster of Xen instance servers had enough funds been available. In all honestly, what I'm doing is upgrading the totality of our server infrastructure with the least funds possible.
This whole approach was not chosen to try out some crazy home-brew IT solution for the fun of it, we're really facing major funding constraints. I wouldn't recommend this approach for anyone with enough money to do it the traditional way. And yeah, if I screw up badly, it's my ass. Such is life in IT.
Perhaps you could add a few tens of thousands of dollars to my budget and then I'll "do it better". Until then, I'll decide what is possible given my available funds.
No. I have two primary servers and two failovers. I did not have enough funds to purchase two sets of identical machines, so I will use older hardware at a remote site connected via a VLAN as the failover system.
Total disk capacity is not the issue. I need I/O capacity on the main server, which is why I went with two 8 port SATA cards on the file server. However, one or both goes down and the failover takes control, users will experience noticeable performance degradation until the primary comes back online. That is due to budgetary constraints of the project.
Rereading my initial comment I note that I did specify a failover server. Perhaps you did NOT read the first paragraph of my top level comment and thus your post is total bullshit. For the last sentence of the first paragraph says thusly:
It's a minor detail anyway.
No. It is accurate that I did not list the failover plan. However, It is also true that there is a plan for remote failover hardware, I just didn't think it was worth pointing out in the initial message. But, of course, this is slashdot. Someone has to find something wrong with whatever is said in order to drive bullshit "discussion" (that goes nowhere).
We'll have a remote failover system in place. But you're absolutely right, what's driving this approach is entirely due to budgetary concerns. We're broke and don't have enough funds to replace all the old hardware we currently have in production. This was the best solution I could find that fit my available resources this FY. And given the economy, I doubt we'll see significant budgetary increases any time soon.
Nope. I have a fail-over system ready to put in place as well. We'll copy the deltas over periodically to a server at a remote site connected by VLAN to the same network.
BTW: my comment was not a proposal for you to fund, just a partial description for the community.
No, not for personal use or gaming. It will run Linux with a Xen kernel and is intended to replace nearly all of our old individual servers. Everything from the piddly servers like DNS, LDAP, Kerberos, and our minimal web services to the AFS db servers. No file services on that beast though, I'm not crazy - no disk I/O-RAM access contention please. My plan is to copy an entire OS image of /usr into a RAMFS filesystem in the top level Dom 0 domain and then cross mount that as RO in each Xen instance. We'll also stick small SQL server and other dbs copies in local tempfs RAMdisks too. Everything in RAM will be snapshotted and saved to physical disk periodically. Those deltas will then be copied to a remote fail-over server periodically as well.
It should be both reasonably stable and blindingly fast.
Another machine will handle AFS and some NFS file services, which has up to sixteen SATA disks attached to two 8 port 3-Ware RAID cards, thus spreading I/O load across two PCI buses. No, we don't need all that disk space - we need the I/O performance. It too should be reasonably fast. We're gearing up to connect that either by several channel bonded 1Gb to a CISCO 6509, or - if we're lucky - we'll just go 10Gb optical. We'll see how the finances work out there.
This is how departmental IT is done. Or, at least, it's how it *should* be done. I spent less than $25K on these two computers and they will replace well a couple hundred thousand dollars worth of accumulated hardware purchased over the last ten years and now fully depreciated.
Yes. That's what I'm saying. The Kindle sucks because as a product it doesn't even come close to meeting my needs. It doesn't matter if it sells for $1000, $350, or $.25. If it doesn't perform the functions I need, to buy it is a waste of money. Period.
I need a device that would replace a physical 8.5" x 11" notebook and allow me to manipulate electronic documents as if I were holding a printed copy. I want to write in eink because using a keyboard does not allow for good math annotation. That feature is critical to me. I would gladly spend $1000 for a light device that replaced the many pounds of dead trees I carry around in my backpack / shoulder bag every day.
You're an ass.
It's a 1280x1024 10.2" diagonal screen, capable of displaying a full 8 1/2" x 11" page with good quality large fonts. You're right about PDF support on the Sony device, however.
It also supports eink annotation and native PDF rendering. It's the only ereader device appropriate for academics and technical professionals on the market today. The price tag is irrelevant given its feature set. I could care less that it costs a grand if it saves me significant project time.
When I read a PDF I need quality image support for interpreting graphs and other types of visual data. The Kindle doesn't come close. Yes, Amazon offers a PDF "conversion" service. In the process, formatting and image support is either lost or horribly mangled.
Never mind the total lack of touch support for eink annotations makes the thing worthless for serious use. Fine if you want to spend $350 for a device to read novels on the train. But if you want to read technical papers and annotate in math, the Kindle doesn't come close to being a useful device.
The only thing out there that does meet that need is:
The IREX Digital Reader 1000:
https://www.irexshop.com/product_info.php?cPath=22_35&products_id=69
That is the first device to come on the market which exceeds the eReader feature set available on the Apple's old Newton MP2x00 from 1998.
Pathetic.