So businesses and govornments are going to use software that anyone can see the source code for. Does anyone else see this as a security risk?
Of course knowing exactly what software a government agency uses poses a potential security risk. At the same time this gives people who are monitoring security risks a list of the contacts that it makes a lot of sense to notify when a vulnerability in that code comes up.
As a comparison point, when code red, nimda, and slammer came out, was there any kind of list of agencies or businesses that should be notified of that fact? I suspect that well over 90% of the agencies and businesses were made aware of the existence of the problem by their own systems responding unusually.
The clasic example of an "incompatible monitor" would be one for which the refresh rate was not supported by the video card. Usually this would only be a problem on 21" or larger fixed frequency monitors, but I don't know what monitor the history of this thread is refering to.
Another way that a monitor could be incompatible is if it responds to the pnp enquiry with information indicating that it is monitor XYZ1504, for which the detection software has not compariable reference for. Either it would be too new, or just have too small of a market share to have an entry in the monitor database.
Another recomendation. If you are going to have one or more printers, that you would like two or more computers in your house to use, get a network print server for your printer.
I have used multi-port print servers, and printer network cards, and have found both to be acceptable for most tasks. I have even seen them built into wi-fi APs so you may be able to use a single device to provide both services. (If you do I recomend that you do not use that same device as a broadband router, unless the documentation gives you a nice easy way to prevent or restrict remote users from printing to you printer.)
I am currently using a print server with one parallel and two usb ports, with a usb-parallel adapter to support two parallel port printers. (already had the usb-parallel adapter.) I went this route because it was less expensive (even if I were to buy two usb-parallel adapters to go with it) than the three parallel port print servers. It's small, sits in the office with the printers, is low power (meaning that I am not leaving a computer powered up to make these printers available) and is a workable target for every operating system I am using as an lpd and iss print server. It even supports AppleTalk if I have the printer drivers for my Macs.
Why do I bring this up? Once you install a printer on the network, (other than possibly some win-printers) you pretty much never have to worry about what printer port mode (std, ecp, epp) the printer supports or allows.
Caveat, while you are welcome to test this on the Lexmark Zx1 and Zx2 printer lines, I would recomend making sure that you can return the print server, or the printer at your option. These are WinPrinters and while they may work with Linux on a parallel port, I have not had good luck with them on print servers of any type. Fourtunately the low end of these are generally disposable as the replacement ink cartridge for them is more expensive than the printer itself is. (Z11, Z12, Z32 as examples, the Z5x set start to get a bit more expensive.)
Then again, all this is just my opinion. Make your own choices based uopn your own opinions and observations.
More accuratly the small population of idiots that are part of the entire population of the military is neither controlling, or particularly idiotic.
They also tend to be self limiting in that they are the ones that end up becoming targets first, which helps the remainder of the troops figure out where the enemy happens to be. As a result the population also tends to become very controled.
Sure we can. But what fun would that be? I suspect that some poor soul would think that hell had frozen over.
I use several different OSs. I have yet to find anything that Windows does better over the long haul. Over the next couple of years it will even loose it's current edge in gaming.
Then again that's puerly my opinion. Maintain and support your own, even if I happen to think they are wrong.
On topic, OpenZaurus, on top of either th 5500 or the 5600 supports all of Debian. This means that you can install anything that has been compiled for the processor included. Or install and compile from source
Do you mean that you will believe it when Microsoft releases "OpenWindows (TM)" and reports that all the doors have been secured?
-Rusty
prices will become "affordable"...
on
LCD Price Fixing?
·
· Score: 1
... when the people building and selling these devices realize that they can make more money overall by selling more at a lower margin than they can make by selling fewer at a higher margin.
If the monitor costs $300 to manufacture, and the maker sells 1000 at $500 each, he earns $200,000. If he sells 10000 at $359 each, he will earn at least $590,000. I say at least, because at some point he is going to encounter a situation where the next one he sells costs less than $300, perhaps because he had to discard one fewer unit per hundred units, or because the volume of plastic he bought to make the cases cost less at the new volume.
The question becomes is there sufficient demand for 10000 units rather than 1000 units.
Of course that's my opinion. Support or maintain your own as you see fit.
I submit that there are four catagories of software professionals who actively work on code, and whom may be considered programers, one of which may be considered an engineer.
I have read earlier notes talking about programers, scientists, and engineers, which seems to miss the mark as well.
The first group ranges from the script kiddy through the technical school trained programmer. These are people who may, or may not understand the concepts behind good software design, but generally do not create their own tools, or the libraries of functions they use. This is neither good nor bad. Most could not tell you wether they use a top-down, or bottom-up development methodology.
The next group are the people who have developed dynamic libraries of their own software that they use to solve software problems, and who have ported that library of functions across multiple development languages as they need them. They do understand their own libraries, but may mistakenly believe that their functions are better than the equivalent functions built into the tools that they use. Sometimes they are right, but they have no evidence of that. They also make some of their own tools. These people may or may not know how to build good test cases for the software they develop. These programing professionals are generally very good at what they do. However I would not consider them to be Engineers.
Next is the Research and Development group. These people are often working on the very edge of software development. These are the first people who understand how to use mathematics to prove or disprove the validity of the functions they develop or use. This is akin to one of the expectations of an engineer, that structural engineers do their best to overdesigning the structures they work on compared to the expected stresses those structures will withstand. At the same time these people are generally not working in a field that seeks engineering certifications. You will find these people working in university post graduate programs, and some larger software development houses building the tools that the rest of us use.
The group that I would consider closest to the Engineer title are those programers who take the time to review the libraries they use, are capable of pulling the source code for the software tools they use, and repair/replace functions that almost work, feeding back those fixes to the people developing those tools. These people will stand behind the softare they develop, and are able to explain situations where their software will break.
Software development is both an Art and a Science. A programmer who "codes with the flow" is more likely to be an Artist than a Scientist. Someone who uses analytical tools to determine what needs to be written, and what can be re-used is more likely to be a scientist than an artist. There is room for both, in fact for most companies there is more of a demand for the Artist than the Scientist. Artists are less likely to look at a problem and say that it will take x months to write this, but are more likely to work the extra hours per week to complete the project in x-n months.
Lexar and SanDisk have announced high capacity CF media in 4 and 8 Gigabyte capacities. You may not be able to fit an entire Debian distribution into one, but a functioning distribution is a different matter. The entire Knoppix image could be booted uncompressed off of one of these.
My own laptop will not run directly off car power, as it requires 18vdc input. This is not uncommon on new laptops.
That said, components designed for desktop pcs use anything from + and - 12 volts for hard drives, through +/- 5 volts for most logic. There are other voltage levels as well to support cpu and memory in a way that will keep them from burning up.
The biggest problem is that you would need to come up with a way to regulate the various voltage levels you feed to the motherboard as the load for those voltages varies. This is far from impossible, however is perhaps not best done by tapping off a cigaret lighter socket.
Using off the shelf voltage regulators is not always helpful, as their primary solution is to convert excess voltage into heat. If you feed it 12-17 volts, and have an output requirement of 5 volts, you are converting 7-9 volts directly to heat. That may provide a usable voltage, but you are also going to generate a lot of heat.
One option is to use the old form of this, run an inverter off the 12 volts, get a multi-tap winding transformer and pull off the voltages you need then bridge rectify those voltages, pass them through a filter and a capacitor to stabalize them, etc. The problem here is that off the shelf inverters are going to be more reliable, and standard atx-pc power supplies are off the shelf replacement items that will provide you with all the correct voltages if you match it to the inverter. (Don't try to use a 500W power supply with a 150W inverter.)
Of course remember the maxim "Ask the experts, they will tell you what can't be done, and why. Then go do it anyway."
At this point, I would suspect it is mostly market forces. While revenue for LCD flat pannel displays in the next year will probably cross that of CRT monitors, that still means that CRT sales will be a significantly larger volume of sales.
Additionally from what I have seen, most LCD flat pannel displays support vga/15-pin inputs as well as DVI, yet I don't know of a CRT that accepts the DVI input. As a result, an interest in selling _more_ of these units will be served by supporting the 15-pin analog interface for some time to come.
I have seen MBs with headers for DVI without the physical DVI interface on them. I would suspect that this will be an option some of the itx manufactures will be taking going forward as well.
One last thought, I would not be surprised if the designers were protecting themselves as well. With a DVI output in addition to the TV and 15-pin analog, I am reasonably sure that some marketing droid would say "multi-head support" and someone would buy into the hype, discover the error and publicly lambast the manufacturer. One or two mistakes like this, and you don't live it down.
Just my opnions by the way, I have been wrong before.
One of the other problems that PB included was that these components could not be disabled. From what I have read and seen, the Bios on these boards should support peripheral replacement, so if you really want to replace the 16550afn uart driven/dev/ttyS0 interface with a pci card based 8250 uart driven/dev/ttyS0 port, You can.
PB was also not the only company who put components on the MB. Compaq and IBM also have systems built with video, serial, printer, and drive management built into the motherboard. It is a mixed blessing. You can also find full sized atx MBs with just about everything you need built into the port package around the keyboard port.
One major positive for this concept, is that in some of the suggested uses for these boards, the fact that the stuff is soldered down means that in high vibration environments (cars, boats, backpacks, bikes, whatever) you are less likely to have a system failure as the interface slips out of the pci slot.
In my own opinion, CF memory has two handy features that make it a reasonable purchase.
First of all, my Z runs with a wifi card in the CF slot most of the time. However I have several situations where that is of little or no use. My low-power cf card does not support Kismet, so I can not use it to scan for open nets. For the most part I do not need the additional battery drain either.
When I do not need the wifi card, I have found that having a couple of hours of music on a 128M cf memory card can help while away longer bus rides, without affecting how much application space is available on my SD card.
Also I have several CF interfaces for my computers, being able to read/write the CF to transfer stuff when I don't want to use a wifi card is also handy.
As noted in the article sd/mmc cards cost more and tend to have lower capacity than CF cards.
As a last note, there are bluetooth interfaces built to run in sd/mmc slots, which may be of interest to people using blutooth enabled network, phone, printer, or even headset devices. In that case the memory expansion opportunity would be the CF slot.
Then again, that's just my opnion. You will have to make up your own mind.
Yes, everything runs as root. No need for a root kit, etc.
There is an app for th Z that gives you a login screen, I have not tested it, so I do not know if it provides user-id with different privledges than root.
This is functionally a single user multi-tasking device. It happens to run a multi-user capable OS, however that does not require that that feature be used.
Does this create potential problems? Sure. However the system files can not be removed by the user, though certain configuration files can. Even so, if you really care, you can reset the platform to it's original (or last flash rom) state by completely draining the battery. Make regular backups to cf memory.
All of the discrete components of the computer at my house may on their own function with or without the active participation of the other components of the system at my house. On their own, they do not perform all the the tasks that I can accomplish as they are constructed at this time.
This means that the single mac address that my internet service provider observes when viewing my network connection is fundamentally representative of all of the devices that are considered a part of, or peers of that device. I happen to own a cluster computer doing independent processing of a variety of functions.
I may own a functioning 10 node cluster, however the number of nodes in the cluster is variable, and the participants are all acting at my behest.
As the nodes are not all running the same operating system, nor are they running some form of failover, or even common management system, this cluster is anything but a beowolf cluster. Further if I sum the discreete processing power of the components of my cluster, it does not approach the processing power of a single system built with the newest fastest processors, but at the same time I can afford it, and it works.
While it may not have the "latest and newest" of all packages in the distribution, I would hesitate to call that something negative.
If you go through the list of release candidates you will find that over a month ago Mandrake locked the collection of software that would be included with 9.1.
When it comes down to it, Mandrake is a Publisher, and the people who select what software to include in the product that is going to be distributed are editors. You or I may have written the worlds best Linux aplication yesterday, completely bug free, and a Windows killer. You or I may have made binary packages for every known distribution of Linux, and source packages for those distributions that only accept source code, last night. The software may be completely compelling, may make Linux itself so easy to use that Bill Gates would close up Microsoft and convert to Linux.
Even so, Mandrake and Red Hat would be remiss if they tried to include the software in their release distributions for the next couple of weeks. Not because the software was no good, or buggy, or anything like that. It would be because they realize that their customers expect them to test software and do their best to insure that there are no negative interactions between it and other software included in their distributions.
As has been noted, there is a method of doing just that. Don't forget, you can also take that network.img boot floppy image and use it as the bootable image on a CD that you burn yourself.
Using this method you can install across the network on a PC that does not have a floppy disk drive.
I don't personally do this, and will not claim that it is either recomended, or not. Your milage may vary.
One of the reasons that release candidates and beta releases exist are to get copies of the software that a company will release into the hands of reporters and reviewers. The expectation is that there will not be significant changes between the softare being reviewed and what ends up in customer's hands.
Strange as it may seem, a lot of people put beta software into production environments. Then again, I understand there are people who use "Windows", knowing that the software is considered to have some rather well known holes, and the company that maintains it has been known to re-introduce holes when patching others.
While the software developers recomend against using beta, or even release candidate software in a production environment, if your production environment needs something that is in that release, you may decide to run with it, and work around the bugs.
One example of a feature that may be enough to move to RC3 of Mandrake 9.1 is the fact that 9.0 would not properly support some tablets, yet 9.1 RC3 does. In my case 9.0 would not support a Wacom Graphire 2, but both rc2 and rc3 of 9.1 have.
If you are a graphic artist, I suspect that this would be a deciding point.
Then again, that's just my opinion. I may be wrong.
speculation would be that the xscale processor supports operational modes that the strongArm processor does not (shutting down parts of the processor that are not in use, idle sleep, etc) and the fact that the two processors run at very different speeds, which affects anything else running on the processor buffers.
There are several servers available for the Z, including but not limited to ftp, apache, and samba. With both ftp and samba, since you are effectively root, there is noting preventing joe random hacker from downloading, editing and uploading your/etc/ files and making the device do lots of things you would otherwise not be planning on.
Who needs a root-kit if the device does everything as root?
Just a recomendation on what cf 802.11b card to look at...
I have used both the SMC 802.11b cf card and the Socket Low Power 802.11b cards. The Socket definately draws less power, and my experience is that it has an equal range.
The down side of the Socket is that none of the built in wireless apps, nor any of the wireless apps (other than the text mode wireless tools) know how to communicate with the card, nor do they recognize it as a wireless network card.
The driver for the card is on the http://www.zaurus.com/feed server as spectrum24drivers (link is untested, set up your ipkg installer to get the drivers via your usb connection, or pull them down, get them on a cf or sd card and install from there.)
One of the local Office Depot stores is where I got my Socket card this week, There is a $30 mail in rebate, bringing the price down from 149 to 119 after the rebate.
My experience with the SMC card was I would get about half an hour of live time with the card installed. With the Socket card I am getting much closer to 2 hours, if not 3. Considering that I only get about 4 hours of live time without any network interface, I am fairly impressed.
If you drain the battery with this card, while sitting on the crapper, you really do have more problems than you thought.
Slashdot Mirror
So businesses and govornments are going to use software that anyone can see the source code for. Does anyone else see this as a security risk?
Of course knowing exactly what software a government agency uses poses a potential security risk. At the same time this gives people who are monitoring security risks a list of the contacts that it makes a lot of sense to notify when a vulnerability in that code comes up.
As a comparison point, when code red, nimda, and slammer came out, was there any kind of list of agencies or businesses that should be notified of that fact? I suspect that well over 90% of the agencies and businesses were made aware of the existence of the problem by their own systems responding unusually.
-Rusty
I think it will take a few more, and even more impressive, events, but perhaps we are looking at the end of the Crazy Years?
-Rusty
I don't know, I think I would rather have the local topless bar use some of their tallent to deliver kegs to my house, and draw it for me.
-Rusty
The clasic example of an "incompatible monitor" would be one for which the refresh rate was not supported by the video card. Usually this would only be a problem on 21" or larger fixed frequency monitors, but I don't know what monitor the history of this thread is refering to.
Another way that a monitor could be incompatible is if it responds to the pnp enquiry with information indicating that it is monitor XYZ1504, for which the detection software has not compariable reference for. Either it would be too new, or just have too small of a market share to have an entry in the monitor database.
Just a couple of ideas.
-Rusty
Another recomendation. If you are going to have one or more printers, that you would like two or more computers in your house to use, get a network print server for your printer.
I have used multi-port print servers, and printer network cards, and have found both to be acceptable for most tasks. I have even seen them built into wi-fi APs so you may be able to use a single device to provide both services. (If you do I recomend that you do not use that same device as a broadband router, unless the documentation gives you a nice easy way to prevent or restrict remote users from printing to you printer.)
I am currently using a print server with one parallel and two usb ports, with a usb-parallel adapter to support two parallel port printers. (already had the usb-parallel adapter.) I went this route because it was less expensive (even if I were to buy two usb-parallel adapters to go with it) than the three parallel port print servers. It's small, sits in the office with the printers, is low power (meaning that I am not leaving a computer powered up to make these printers available) and is a workable target for every operating system I am using as an lpd and iss print server. It even supports AppleTalk if I have the printer drivers for my Macs.
Why do I bring this up? Once you install a printer on the network, (other than possibly some win-printers) you pretty much never have to worry about what printer port mode (std, ecp, epp) the printer supports or allows.
Caveat, while you are welcome to test this on the Lexmark Zx1 and Zx2 printer lines, I would recomend making sure that you can return the print server, or the printer at your option. These are WinPrinters and while they may work with Linux on a parallel port, I have not had good luck with them on print servers of any type. Fourtunately the low end of these are generally disposable as the replacement ink cartridge for them is more expensive than the printer itself is. (Z11, Z12, Z32 as examples, the Z5x set start to get a bit more expensive.)
Then again, all this is just my opinion. Make your own choices based uopn your own opinions and observations.
-Rusty
More accuratly the small population of idiots that are part of the entire population of the military is neither controlling, or particularly idiotic.
They also tend to be self limiting in that they are the ones that end up becoming targets first, which helps the remainder of the troops figure out where the enemy happens to be. As a result the population also tends to become very controled.
-Rusty
Sure we can. But what fun would that be? I suspect that some poor soul would think that hell had frozen over.
I use several different OSs. I have yet to find anything that Windows does better over the long haul. Over the next couple of years it will even loose it's current edge in gaming.
Then again that's puerly my opinion. Maintain and support your own, even if I happen to think they are wrong.
On topic, OpenZaurus, on top of either th 5500 or the 5600 supports all of Debian. This means that you can install anything that has been compiled for the processor included. Or install and compile from source
Well, back to Snow Crash.
-Rusty
So what if someone's parents never wed. That's no reason to equate them to Microsoft executive.
um, don't you mean "bring them down to the level of a Microsoft executive."?
-Rusty
Do you mean that you will believe it when Microsoft releases "OpenWindows (TM)" and reports that all the doors have been secured?
-Rusty
... when the people building and selling these devices realize that they can make more money overall by selling more at a lower margin than they can make by selling fewer at a higher margin.
If the monitor costs $300 to manufacture, and the maker sells 1000 at $500 each, he earns $200,000. If he sells 10000 at $359 each, he will earn at least $590,000. I say at least, because at some point he is going to encounter a situation where the next one he sells costs less than $300, perhaps because he had to discard one fewer unit per hundred units, or because the volume of plastic he bought to make the cases cost less at the new volume.
The question becomes is there sufficient demand for 10000 units rather than 1000 units.
Of course that's my opinion. Support or maintain your own as you see fit.
-Rusty
I submit that there are four catagories of software professionals who actively work on code, and whom may be considered programers, one of which may be considered an engineer.
I have read earlier notes talking about programers, scientists, and engineers, which seems to miss the mark as well.
The first group ranges from the script kiddy through the technical school trained programmer. These are people who may, or may not understand the concepts behind good software design, but generally do not create their own tools, or the libraries of functions they use. This is neither good nor bad. Most could not tell you wether they use a top-down, or bottom-up development methodology.
The next group are the people who have developed dynamic libraries of their own software that they use to solve software problems, and who have ported that library of functions across multiple development languages as they need them. They do understand their own libraries, but may mistakenly believe that their functions are better than the equivalent functions built into the tools that they use. Sometimes they are right, but they have no evidence of that. They also make some of their own tools. These people may or may not know how to build good test cases for the software they develop. These programing professionals are generally very good at what they do. However I would not consider them to be Engineers.
Next is the Research and Development group. These people are often working on the very edge of software development. These are the first people who understand how to use mathematics to prove or disprove the validity of the functions they develop or use. This is akin to one of the expectations of an engineer, that structural engineers do their best to overdesigning the structures they work on compared to the expected stresses those structures will withstand. At the same time these people are generally not working in a field that seeks engineering certifications. You will find these people working in university post graduate programs, and some larger software development houses building the tools that the rest of us use.
The group that I would consider closest to the Engineer title are those programers who take the time to review the libraries they use, are capable of pulling the source code for the software tools they use, and repair/replace functions that almost work, feeding back those fixes to the people developing those tools. These people will stand behind the softare they develop, and are able to explain situations where their software will break.
Software development is both an Art and a Science. A programmer who "codes with the flow" is more likely to be an Artist than a Scientist. Someone who uses analytical tools to determine what needs to be written, and what can be re-used is more likely to be a scientist than an artist. There is room for both, in fact for most companies there is more of a demand for the Artist than the Scientist. Artists are less likely to look at a problem and say that it will take x months to write this, but are more likely to work the extra hours per week to complete the project in x-n months.
Then again that's just my opinion.
-Rusty
Lexar and SanDisk have announced high capacity CF media in 4 and 8 Gigabyte capacities. You may not be able to fit an entire Debian distribution into one, but a functioning distribution is a different matter. The entire Knoppix image could be booted uncompressed off of one of these.
-Rusty
My own laptop will not run directly off car power, as it requires 18vdc input. This is not uncommon on new laptops.
That said, components designed for desktop pcs use anything from + and - 12 volts for hard drives, through +/- 5 volts for most logic. There are other voltage levels as well to support cpu and memory in a way that will keep them from burning up.
The biggest problem is that you would need to come up with a way to regulate the various voltage levels you feed to the motherboard as the load for those voltages varies. This is far from impossible, however is perhaps not best done by tapping off a cigaret lighter socket.
Using off the shelf voltage regulators is not always helpful, as their primary solution is to convert excess voltage into heat. If you feed it 12-17 volts, and have an output requirement of 5 volts, you are converting 7-9 volts directly to heat. That may provide a usable voltage, but you are also going to generate a lot of heat.
One option is to use the old form of this, run an inverter off the 12 volts, get a multi-tap winding transformer and pull off the voltages you need then bridge rectify those voltages, pass them through a filter and a capacitor to stabalize them, etc. The problem here is that off the shelf inverters are going to be more reliable, and standard atx-pc power supplies are off the shelf replacement items that will provide you with all the correct voltages if you match it to the inverter. (Don't try to use a 500W power supply with a 150W inverter.)
Of course remember the maxim "Ask the experts, they will tell you what can't be done, and why. Then go do it anyway."
-Rusty
At this point, I would suspect it is mostly market forces. While revenue for LCD flat pannel displays in the next year will probably cross that of CRT monitors, that still means that CRT sales will be a significantly larger volume of sales.
Additionally from what I have seen, most LCD flat pannel displays support vga/15-pin inputs as well as DVI, yet I don't know of a CRT that accepts the DVI input. As a result, an interest in selling _more_ of these units will be served by supporting the 15-pin analog interface for some time to come.
I have seen MBs with headers for DVI without the physical DVI interface on them. I would suspect that this will be an option some of the itx manufactures will be taking going forward as well.
One last thought, I would not be surprised if the designers were protecting themselves as well. With a DVI output in addition to the TV and 15-pin analog, I am reasonably sure that some marketing droid would say "multi-head support" and someone would buy into the hype, discover the error and publicly lambast the manufacturer. One or two mistakes like this, and you don't live it down.
Just my opnions by the way, I have been wrong before.
-Rusty
One of the other problems that PB included was that these components could not be disabled. From what I have read and seen, the Bios on these boards should support peripheral replacement, so if you really want to replace the 16550afn uart driven /dev/ttyS0 interface with a pci card based 8250 uart driven /dev/ttyS0 port, You can.
PB was also not the only company who put components on the MB. Compaq and IBM also have systems built with video, serial, printer, and drive management built into the motherboard. It is a mixed blessing. You can also find full sized atx MBs with just about everything you need built into the port package around the keyboard port.
One major positive for this concept, is that in some of the suggested uses for these boards, the fact that the stuff is soldered down means that in high vibration environments (cars, boats, backpacks, bikes, whatever) you are less likely to have a system failure as the interface slips out of the pci slot.
Then again that's just my opinion.
-Rusty
In my own opinion, CF memory has two handy features that make it a reasonable purchase.
First of all, my Z runs with a wifi card in the CF slot most of the time. However I have several situations where that is of little or no use. My low-power cf card does not support Kismet, so I can not use it to scan for open nets. For the most part I do not need the additional battery drain either.
When I do not need the wifi card, I have found that having a couple of hours of music on a 128M cf memory card can help while away longer bus rides, without affecting how much application space is available on my SD card.
Also I have several CF interfaces for my computers, being able to read/write the CF to transfer stuff when I don't want to use a wifi card is also handy.
As noted in the article sd/mmc cards cost more and tend to have lower capacity than CF cards.
As a last note, there are bluetooth interfaces built to run in sd/mmc slots, which may be of interest to people using blutooth enabled network, phone, printer, or even headset devices. In that case the memory expansion opportunity would be the CF slot.
Then again, that's just my opnion. You will have to make up your own mind.
-Rusty
Yes, everything runs as root. No need for a root kit, etc.
There is an app for th Z that gives you a login screen, I have not tested it, so I do not know if it provides user-id with different privledges than root.
This is functionally a single user multi-tasking device. It happens to run a multi-user capable OS, however that does not require that that feature be used.
Does this create potential problems? Sure. However the system files can not be removed by the user, though certain configuration files can. Even so, if you really care, you can reset the platform to it's original (or last flash rom) state by completely draining the battery. Make regular backups to cf memory.
-Rusty
All of the discrete components of the computer at my house may on their own function with or without the active participation of the other components of the system at my house. On their own, they do not perform all the the tasks that I can accomplish as they are constructed at this time.
This means that the single mac address that my internet service provider observes when viewing my network connection is fundamentally representative of all of the devices that are considered a part of, or peers of that device. I happen to own a cluster computer doing independent processing of a variety of functions.
I may own a functioning 10 node cluster, however the number of nodes in the cluster is variable, and the participants are all acting at my behest.
As the nodes are not all running the same operating system, nor are they running some form of failover, or even common management system, this cluster is anything but a beowolf cluster. Further if I sum the discreete processing power of the components of my cluster, it does not approach the processing power of a single system built with the newest fastest processors, but at the same time I can afford it, and it works.
-Rusty
While it may not have the "latest and newest" of all packages in the distribution, I would hesitate to call that something negative.
If you go through the list of release candidates you will find that over a month ago Mandrake locked the collection of software that would be included with 9.1.
When it comes down to it, Mandrake is a Publisher, and the people who select what software to include in the product that is going to be distributed are editors. You or I may have written the worlds best Linux aplication yesterday, completely bug free, and a Windows killer. You or I may have made binary packages for every known distribution of Linux, and source packages for those distributions that only accept source code, last night. The software may be completely compelling, may make Linux itself so easy to use that Bill Gates would close up Microsoft and convert to Linux.
Even so, Mandrake and Red Hat would be remiss if they tried to include the software in their release distributions for the next couple of weeks. Not because the software was no good, or buggy, or anything like that. It would be because they realize that their customers expect them to test software and do their best to insure that there are no negative interactions between it and other software included in their distributions.
Then again, perhaps that's just my opinion...
-Rusty
As has been noted, there is a method of doing just that. Don't forget, you can also take that network.img boot floppy image and use it as the bootable image on a CD that you burn yourself.
Using this method you can install across the network on a PC that does not have a floppy disk drive.
I don't personally do this, and will not claim that it is either recomended, or not. Your milage may vary.
-Rusty
One of the reasons that release candidates and beta releases exist are to get copies of the software that a company will release into the hands of reporters and reviewers. The expectation is that there will not be significant changes between the softare being reviewed and what ends up in customer's hands.
Strange as it may seem, a lot of people put beta software into production environments. Then again, I understand there are people who use "Windows", knowing that the software is considered to have some rather well known holes, and the company that maintains it has been known to re-introduce holes when patching others.
While the software developers recomend against using beta, or even release candidate software in a production environment, if your production environment needs something that is in that release, you may decide to run with it, and work around the bugs.
One example of a feature that may be enough to move to RC3 of Mandrake 9.1 is the fact that 9.0 would not properly support some tablets, yet 9.1 RC3 does. In my case 9.0 would not support a Wacom Graphire 2, but both rc2 and rc3 of 9.1 have.
If you are a graphic artist, I suspect that this would be a deciding point.
Then again, that's just my opinion. I may be wrong.
-Rusty
speculation would be that the xscale processor supports operational modes that the strongArm processor does not (shutting down parts of the processor that are not in use, idle sleep, etc) and the fact that the two processors run at very different speeds, which affects anything else running on the processor buffers.
-Rusty
From what I understand, the SL-5600 uses Opera6 as the default browser.
I have a SL-5500 which comes with Opera 5 as it's default browser.
Downloadable browsers include Konquer and Links-ssl.
-Rusty
There are several servers available for the Z, including but not limited to ftp, apache, and samba. With both ftp and samba, since you are effectively root, there is noting preventing joe random hacker from downloading, editing and uploading your /etc/ files and making the device do lots of things you would otherwise not be planning on.
Who needs a root-kit if the device does everything as root?
-Rusty
Just a recomendation on what cf 802.11b card to look at...
I have used both the SMC 802.11b cf card and the Socket Low Power 802.11b cards. The Socket definately draws less power, and my experience is that it has an equal range.
The down side of the Socket is that none of the built in wireless apps, nor any of the wireless apps (other than the text mode wireless tools) know how to communicate with the card, nor do they recognize it as a wireless network card.
The driver for the card is on the http://www.zaurus.com/feed server as spectrum24drivers (link is untested, set up your ipkg installer to get the drivers via your usb connection, or pull them down, get them on a cf or sd card and install from there.)
One of the local Office Depot stores is where I got my Socket card this week, There is a $30 mail in rebate, bringing the price down from 149 to 119 after the rebate.
My experience with the SMC card was I would get about half an hour of live time with the card installed. With the Socket card I am getting much closer to 2 hours, if not 3. Considering that I only get about 4 hours of live time without any network interface, I am fairly impressed.
If you drain the battery with this card, while sitting on the crapper, you really do have more problems than you thought.
-Rusty