Well I've been using a new Tungsten C for a couple of weeks. The unit is fantastic, the one downer: Graffiti2 (aka Jot).
It's inaccurate, slow and constantly gets confused. I used to be able to keep minutes on the thing, but no more, because the strokes just don't pick up fast enough. Forget about PocketPCs, boycott Xerox!
Sure, but there are only 3 non-overlapping channels on a "b" set (therefore also a "g" set). Since multiple AP's have to be on differing channels, you quickly run out of available space.
I'm no signaling expert, but couldn't the loss in range be compensated for by a better antenna?
No, I disagree. Given the fact that 802.11a gives a higher rate (than a mixed network), and does not need to drop back to the slower speeds for reservation, and finally, the fact that a/b cards are dropping in price quickly, I would say that an a|b network works better in the long run.
This allows segmenting the slower clients out of the higher speed network, and increases your overall bandwidth. Older clients use the (somewhat crowded) 2.5 GHz spectrum, high demand clients use the 5GHz bandwidth. Neither interfere with each other, and in fact both can transmit at the same time. You also gain more available channels on the 5GHz spectrum, which makes deployment easier.
I do Solaris admin. as well, and I run Linux as both my primary WS at work, and at home. I do use a Mac OS laptop, but I find it's underlining Unix a real bear to deal with. Apple sliced and diced just enough to make it a pain in the butt to use.
What I wish Sun would do is start shipping all units out w/ dual processors min. The SPARC is designed around multiprocessor, you pay a price for that in a single chip config. The should set their price points to push the fact that you should have more than one chip.
It's unfortunate that the real opportunity has been missed during the anti-trust trial in the U.S.
As part of the remedy, it should have been put down that MS must document all files used by any program they produce. This documentation should be public, and should require at least six months notice for any change.
This would have broken the true monopoly on the desktop, MS's ability to hide how your own data is stored. It's your document, why shouldn't you know this? MS's monopoly depends on shaking off the competition, using copyright laws and secrecy as a bludgeon. It takes a disinfecting light to combat them.
Well, main focus is on running out of IPs, though there are other improvements in speed and utilization.
I'd say that another major reason that the US can put this off is that the major use is from outside the US to inside the US, not the other way around. Culturally the US has less interest in accessing the rest of the world. IPv6 only saves you IPs if you don't need to have IPv4 node addressability, ie. the machine is only an IPv6 node.
Now if the US wanted to have compatability to access the remote nodes, they would need to upgrade. The logical reasoning is that a law be passed requiring all adult sites to implement only IPv6. Bing! Instant compatability.
This sort of vehicle need to hit below $15,000 USD, perhaps as low as $10.000 USD in order for people to make the obvious trade-offs in flexibility, cargo handling and passenger space, even if fuel prices doubled. (Which given the current US administration, is not going to happen.)
It's always a tough battle when you against the entrenched market.
Just another neat gadget that loses so much potential, 'cause it only works on Windoze. When will the HW guys get it, that there are other poeple who use this stuff?
It's not just this drive. Even things with the most basic interfaces like labelers and signs, even if they wrote their little gizmo interfaces in Java. Sure here you probably need a driver or an ioctl(), but it's not rocket science.
The major problem with using VPNs are that they are implemented differently by different vendors. The Netgear box doesn't connect to the Sun box. Does the Windows machine connect to the OpenBSD gateway? Who knows? Even if it does you need to explain to everyone how it works (ie. support), and you need to distribute keys for people to use.
Let's face it, the solutions that most people have suggested here are the techies doing the typing. You need to have a basic way of authenticating, distributing keys, and establishing connections, and most importantly, it needs to be easy or automatic and effecient to use.
Actually, I'd like to know what other projects are using. I think that some of the fustration with C/C++/Java also springs from the fact that this is a Windows development effort. I used C++ on Windows for a few projects and it -- is -- a -- bear. Doing any of this type of development for Unix/Linux or cross platform support I think changes the rules.
Not mentioned is the final langauge they used. Is the code in C/C++?
I do agree with the binary protocols vs. text protocols, but I'm not sure that is a recent revelation. Most of the Internet protocols use text based protocols.
This goes back to a conversation I was having the other day over the design philosophy of secure TCP. I would prefer the design of IPsec, where the security is layered into the network itself. The problem arises from the fact that there is no way programmatically to ask for a secure connection. All of the work needs to be done by an admin., and all of the details need to be worked out ahead of time (what keys, distribution, etc.) IPsec also doesn't have user key support last time I checked.
The ideal secure connection would allow secure authentication, and no data confidentiality, like Kerberos; a known host connection w/ no user auth; or a fully secured connection (like ssh), or open security (ie. no security). The problem is that the socket abstraction (which has a few basic flaws in it) no longer stretches to compensate for the network underpinnings. ssh/ssl suffer from this in that the application has to take on the security problem.
All of the network security methods lack a good, fundamental authentication model. I think that this is where Kerberos works as a good add-on. Until a better trust model can be designed network security seems doomed to be a complicated mess.
Re:Privacy -- money example
on
RFID Explained
·
· Score: 1
I agree. In the story, the example of placing the RFID inside of money is seen as a "threat to privacy." While I'm not sure I would want someone to be able to tell how much money I have in my pocket at any time, the "anonymity" of money is already a subjective thing. All US currency have serial numbers on them, and it would be rather trivial for a bank machine to record what range it had been loaded with, and which bills had been given out. "What do you have in your wallet -- Arr?"
The law does not at this point take into account that Windows is a platform. When developing for a platform, certain rules should apply, as there is an implicit contract between the developer and the platform supplier. It's seem ridiculous that MS has the ability to hedge it's control of the platform to destroy any market segment it feel threatened by. (Java is a platform, and would be governed in the same fashion.)
It is quite clear that a law establishing what a platform is, and how it should be treated legally is long overdue.
This is naive. IP under its current implementaion assumes a secure local net. This is obviously flawed in a public environment, even if you are not using Wi-Fi. Adding IPSec to the mix would be useful, if there were a good universal authentication structure.
But even then, most people use Wi-Fi w/ DHCP. Ah, but how do you know that the correct DHCP server has responed? Are you connected to another client creating an evil network?
So, even out of the box we end up with:
Problems with what network we are talking to (wirelessly).
Problems with what network we are talking through (wired).
Malicious clients messing with our arp table.
To be fair, the problem is that no one wants to fix the fundamental security issues. They're too busy trying to hedge against the competition to bother fixing the customer's problem.
True. But now that Firewire 800Mb/s is out, the answer to all of this is: if you care about USB-480 (or USB Hi Speed, etc.) you should be using Firewire instead.
I got the 5000D, and the lower memory hurts. I'm not sure that that was the main point I was making though. The question is: does the abstraction of a general OS work when the basic assumptions about the environment have changed so radically?
Great! I think that IPs should be available, and cheaply too. If the transition takes a decade, then who cares? We've had a lot of time to work out the migration path, and IPv6 includes a number of additional benefits [better routing, IPsec, etc.].
Better that we start to think about it now, so we don't fall into the traps when it counts.
True. I would hope that AOL would take the long term view and continue developing a stratigic "wedge" agains MS.
I would have liked to see something along the line of "no packaging IE" w/ a default config. Even to remove the icons to the program would be a great help. Of course this was a company's private suit, not a govermental body.
Well, mabey Bush'll choke on a pretzel or something, and we'll get a real trust buster in.
Slashdot Mirror
It's inaccurate, slow and constantly gets confused. I used to be able to keep minutes on the thing, but no more, because the strokes just don't pick up fast enough. Forget about PocketPCs, boycott Xerox!
I'm no signaling expert, but couldn't the loss in range be compensated for by a better antenna?
This allows segmenting the slower clients out of the higher speed network, and increases your overall bandwidth. Older clients use the (somewhat crowded) 2.5 GHz spectrum, high demand clients use the 5GHz bandwidth. Neither interfere with each other, and in fact both can transmit at the same time. You also gain more available channels on the 5GHz spectrum, which makes deployment easier.
What I wish Sun would do is start shipping all units out w/ dual processors min. The SPARC is designed around multiprocessor, you pay a price for that in a single chip config. The should set their price points to push the fact that you should have more than one chip.
As part of the remedy, it should have been put down that MS must document all files used by any program they produce. This documentation should be public, and should require at least six months notice for any change.
This would have broken the true monopoly on the desktop, MS's ability to hide how your own data is stored. It's your document, why shouldn't you know this? MS's monopoly depends on shaking off the competition, using copyright laws and secrecy as a bludgeon. It takes a disinfecting light to combat them.
You don't need to have a phone.
That's it.
Not if you're using an iBook.
I'd say that another major reason that the US can put this off is that the major use is from outside the US to inside the US, not the other way around. Culturally the US has less interest in accessing the rest of the world. IPv6 only saves you IPs if you don't need to have IPv4 node addressability, ie. the machine is only an IPv6 node.
Now if the US wanted to have compatability to access the remote nodes, they would need to upgrade. The logical reasoning is that a law be passed requiring all adult sites to implement only IPv6. Bing! Instant compatability.
(As opposed to the pretty one? :/ )
It's always a tough battle when you against the entrenched market.
It's not just this drive. Even things with the most basic interfaces like labelers and signs, even if they wrote their little gizmo interfaces in Java. Sure here you probably need a driver or an ioctl(), but it's not rocket science.
I'll buy the one w/ Linux/Unix/Mac support.
Let's face it, the solutions that most people have suggested here are the techies doing the typing. You need to have a basic way of authenticating, distributing keys, and establishing connections, and most importantly, it needs to be easy or automatic and effecient to use.
I'll check back next week . . . :)
Not mentioned is the final langauge they used. Is the code in C/C++?
I do agree with the binary protocols vs. text protocols, but I'm not sure that is a recent revelation. Most of the Internet protocols use text based protocols.
The ideal secure connection would allow secure authentication, and no data confidentiality, like Kerberos; a known host connection w/ no user auth; or a fully secured connection (like ssh), or open security (ie. no security). The problem is that the socket abstraction (which has a few basic flaws in it) no longer stretches to compensate for the network underpinnings. ssh/ssl suffer from this in that the application has to take on the security problem.
All of the network security methods lack a good, fundamental authentication model. I think that this is where Kerberos works as a good add-on. Until a better trust model can be designed network security seems doomed to be a complicated mess.
Look ma' -- no mod!
I agree. In the story, the example of placing the RFID inside of money is seen as a "threat to privacy." While I'm not sure I would want someone to be able to tell how much money I have in my pocket at any time, the "anonymity" of money is already a subjective thing. All US currency have serial numbers on them, and it would be rather trivial for a bank machine to record what range it had been loaded with, and which bills had been given out. "What do you have in your wallet -- Arr?"
It is quite clear that a law establishing what a platform is, and how it should be treated legally is long overdue.
But even then, most people use Wi-Fi w/ DHCP. Ah, but how do you know that the correct DHCP server has responed? Are you connected to another client creating an evil network?
So, even out of the box we end up with:
To be fair, the problem is that no one wants to fix the fundamental security issues. They're too busy trying to hedge against the competition to bother fixing the customer's problem.
True. But now that Firewire 800Mb/s is out, the answer to all of this is: if you care about USB-480 (or USB Hi Speed, etc.) you should be using Firewire instead.
Sounds a lot like Chrysler Motors. Mabey we can sell Micron to the Germans! :)
I got the 5000D, and the lower memory hurts. I'm not sure that that was the main point I was making though. The question is: does the abstraction of a general OS work when the basic assumptions about the environment have changed so radically?
Better that we start to think about it now, so we don't fall into the traps when it counts.
I would have liked to see something along the line of "no packaging IE" w/ a default config. Even to remove the icons to the program would be a great help. Of course this was a company's private suit, not a govermental body.
Well, mabey Bush'll choke on a pretzel or something, and we'll get a real trust buster in.