There already are such standards. Why do you think we can run Linux on a PC at all? Because most things are always the same. This is caused by the fact that PC manufacturers today still maintain compatability with the IBM PC of 1980. The DOS from that PC will still run on today's systems.
Look at USB too. There are some generic device types and they usually just work. More often on Linux even than on Windows.
System-independent BIOS chips on boards also exist.
In all, a lot has been done, but there always remain devices that don't fall into these categories and for which there are no drivers. They especially exist in the low-end segment, where the driver tends to provide more of the device functionality (to reduce cost in the device itself) so reverse-engineering is much more difficult, and the manufacturer is not interested in Linux investments.
>As soon as companies got their own digital switches, spoofing was possible.
As soon as phone companies only cared about money and not about quality of service, it became possible. Having your own switch that sends data to the phone company does not mean the phone company has to accept and forward it without checking. It only does so when the phone company does not care.
Apparently the phone system in your country is a big mess.
When I try to send a wrong caller ID on my ISDN line or in fact on any ISDN-2 or ISDN-30 line in this country, it gets replaced by the line's main number automatically. Sending spoofed numbers this way is impossible.
This happens at the local telephone company's switch. Telephone companies can get around this, but businesses cannot.
Unfortunately is is too easy to become a telephone company if you want to, but that is another story.
You may need to read the original article. This was not about hiding your caller ID, it was about inserting a number of your own choice when making a call.
A subscriber cannot do that, but a telecom company can. The whole idea behind caller ID is that telecom companies are reliable and won't do that. Just as they are supposed to hide your number when you tell them they should. (there is no built-in guarantee for that either, it completely depends on de trustworthyness of the telecom provider)
Do you really believe the size of an application (as it is on disk, in memory, or as a download) has any significance to its market share???
I hope not. This is absolutely not relevant to the average computer user. They do not even have a grasp of what a typical size of an application or datafile is. How large is a 2-page letter, a digital camera photograph, a 3 minute MP3 file, a 1 hour movie, a word processor, a browser? They don't have the foggiest notion and send them as attachments to their (hot)mail without bothering.
No, if you want to explain market share of browsers you will have to look at something other than size, really!
We run Windows 2000 SP4 with all fixes, and about 10 applications (Mozilla, OpenOffice, ICA Client, SAP GUI, Acrobat Reader, JAVA runtime, etc plus MS Office on a few systems). There are even some old systems with 3.2 GB disks and they have room to spare.
But it does not matter, replace the 4 GB with 8 GB if you like and you still see a lot of free space on a business workstation with the smallest disk available today. All data is of course stored on servers.
>Why shouldn't their lucrative, unethical trade move to massively parallel processing on that network to generate these "expensive" messages at our expense, rather than theirs?
What do you mean 'our' expense? I do not count the virus infected PCs of clueless users as 'my' expense, and I would rather like to see that the individual home user gets heavily punished by having a spam relay on their machine (by performance problems). A system with micropayments would probably have the same problem, but it would cost those users real money. Is that your better alternative?
>We're faced with the choice of evolving, or wallowing in the muck we've got now.
It has been like that for nearly a decade, and the Internet mail community has hardly made any moves towards a new mailsystem. I mean a really new mailsystem, not some blocklist or sender ID add-on to the existing broken system.
My real fear is that Microsoft will come up with such a system, will take 99% of the users with it like they did with Internet Explorer, and then lockout all opensource, free software, competing OS manufacturers etc from the world's e-mail system.
Fortunately we have only a single domain. Access control lists should not be a problem, but they should be included (they often aren't in programs intended for file transfer rather than backup).
If I remember correctly, RAR can do almost what we need except for the requirement to store the different parts on different machines (= in different destination directories). And except it does not do on-the-fly PAR2 generation but leaves that to a separate utility that again reads all the files.
And of course, to use it in real life there would need to be some database to store what backup on what date was stored on which workstations and under what filename. Plus a mechanism to purge old files to make room for new ones.
Indeed, this can be expanded and extended almost indefinately, but we are really looking at a simple solution for a small company (with more creative employees than money).
Micropayments do not need to involve money. That avoids a lot of complications and legal stuff.
Solutions have been proposed to create stamps that require some computer power to make, but are easy to verify. Someone wanting to send a mail message can easily spend 10-60 seconds of CPU power on that (e.g. while it is being typed) but a spammer will not be able to afford that.
What you describe is like an rsync server that sends the correct files to the client. That can already be done, you just need someone to setup the rsync server.
But indeed I have always wondered why I need to download an entire 25MB RPM for each and every security fix of some packages. Sometimes the fix is not even in the binary files, but in some config file or the installation script of the RPM!
What you would expect is a small update, that depends on the original RPM (program-1.0) and upgrades it to a fixed program (program-1.0-2 or program-1.1). Installation of the fix RPM should apply the patch and change the RPM database to the new situation, just as if you had removed the old version installed the new version from a big RPM file.
I have discussed the idea of using PAR2 files for backups at work. The situation is this: we have some 200 workstations. These have way too much diskspace, as you cannot buy small disks these days anymore. 4GB would be plenty for a business-use workstation (just the OS and some applications are stored there), but the workstations have 40 or 80 GB disks. This means there is about 8TB of unused diskspace.
The servers have much less diskspace than that, in total. It would be possible to store several backups on the workstation disks. The idea is to make backups as a set of files, distributed over a number of workstations, and protected by PAR2 files (in case a workstation fails or is re-installed). The backup data would of course be encrypted, and not accessible to the workstation user because of access rights.
This could serve as a quickly accessible backup, in addition to the normal tape backups we use. It could also be more reliable (tape backup units are not the most reliable parts in computers).
What is missing, or at least as far as we know, is a tool that allows to store a directory tree, access rights included, into a a number of parts with added PAR2 files, where the destination of each file can be specified separately (e.g. via a control file). We would not want to first copy all server data to a local set of files and then move each of these files to a workstation in a second step, but the files would have to be directly written to the workstations over the network.
I suspect eventually we'll see dns servers implementing a custom record type for SPF to replace the current TXT records, but other than that, you don't really need anything else.
The big problem with SPF right now is the very limited support for TXT records by providers of DNS service. I would really like to experiment with SPF but all three DNS providers I am using at home and at work don't offer the editing of TXT records via the usual web interface.
I am still hoping that more wide use of SPF would lead to the addition of TXT to the usual A, PTR, CNAME and MX repertoire of DNS records available to end users, but I fear that asking for a new record type will be too much.
That was the old approach. It does not really work because people will check the list to conclude that something will NOT work, and there is little incentive to add something that works to a database. Remember you are dependent on voluntary action by someone who owns te hardware and can judge if "it works".
When a devide does NOT work, and you have tried to find drivers and contact the manufacturer, you may be more likely to have found the "incompatability list" in the process, and add your device to that.
It looks like it controls the HVAC using 24V AC switching. Is that still used in the USA?
Over here (Netherlands) any modern heating system uses a digital serial link between the equipment and the thermostat. The thermostat not only switches the heater on and off, but can also set the level (modulation), read back and display status information, set parameters like hot water operation mode etc.
As it does not use straight on/off switching, such a system operates much more smoothly. Rather than cycling the heater with a dutycycle that results in the correct room temperature, it sets the heater to the correct level to keep that temperature.
There exists a standard protocol called OpenTherm. Although it is called "Open" this does not mean that specifications are available for everyone to download:-(
Wannabe firewall reviewers like to whine about "stealthed" and "closed" ports. Usually they don't know what they are talking about. When a review starts to mention how this or that firewall has listed a port as "closed" instead of "stealthed" and claims that this is less optimal, to me this devaluates the entire review.
This depends on your background and on the problem you are looking at. There are Windows problems that I have never been able to completely understand and solve, mostly because of the black-box situation you find yourself in. And there are many tasks that you can solve in a few minutes in Linux, and take hours of RSI-inducing clicking in Windows.
A typical cylinder lock has just as many combinations and works just the same as this type of lock. The pistons are just aranged differently, but the principle is the same.
Cylinder locks look secure, but often can be openened in 30 seconds with two paperclips.
Slashdot Mirror
There already are such standards. Why do you think we can run Linux on a PC at all? Because most things are always the same.
This is caused by the fact that PC manufacturers today still maintain compatability with the IBM PC of 1980. The DOS from that PC will still run on today's systems.
Look at USB too. There are some generic device types and they usually just work. More often on Linux even than on Windows.
System-independent BIOS chips on boards also exist.
In all, a lot has been done, but there always remain devices that don't fall into these categories and for which there are no drivers. They especially exist in the low-end segment, where the driver tends to provide more of the device functionality (to reduce cost in the device itself) so reverse-engineering is much more difficult, and the manufacturer is not interested in Linux investments.
Probably users like the fact that Skype does not need gateways or other kinds of servers to work.
A single company in control over the source code is one thing, but worse is a single company in control over all the "servers", like with MSN.
Ah, they should just come up with a new version that does not require a connection.
The sendmail support is from Sendmail, Inc. not from the open-source sendmail at sendmail.org
>As soon as companies got their own digital switches, spoofing was possible.
As soon as phone companies only cared about money and not about quality of service, it became possible.
Having your own switch that sends data to the phone company does not mean the phone company has to accept and forward it without checking. It only does so when the phone company does not care.
Apparently the phone system in your country is a big mess.
When I try to send a wrong caller ID on my ISDN line or in fact on any ISDN-2 or ISDN-30 line in this country, it gets replaced by the line's main number automatically. Sending spoofed numbers this way is impossible.
This happens at the local telephone company's switch. Telephone companies can get around this, but businesses cannot.
Unfortunately is is too easy to become a telephone company if you want to, but that is another story.
When your phone company's switch accepts that number when it is not one of the numbers you have subscribed to, it is seriously misconfigured.
Sending the main switchboard number instead of a direct dialling number should be no problem, but sending someone else's number should be disallowed.
>I thought that caller ID was done through the phone company and people couldn't alter it.
That is true. But if you setup a small phone company and let people make calls through your systems, you can do whatever you like.
Of course there should be some authority that will immediately disconnect such subversive phone companies from the net.
You may need to read the original article.
This was not about hiding your caller ID, it was about inserting a number of your own choice when making a call.
A subscriber cannot do that, but a telecom company can. The whole idea behind caller ID is that telecom companies are reliable and won't do that. Just as they are supposed to hide your number when you tell them they should.
(there is no built-in guarantee for that either, it completely depends on de trustworthyness of the telecom provider)
It has NOT been fixed.
It will be fixed in a future release.
I run the newest 1.7.2 release of Mozilla and it still has this problem.
Do you really believe the size of an application (as it is on disk, in memory, or as a download) has any significance to its market share???
I hope not. This is absolutely not relevant to the average computer user. They do not even have a grasp of what a typical size of an application or datafile is. How large is a 2-page letter, a digital camera photograph, a 3 minute MP3 file, a 1 hour movie, a word processor, a browser? They don't have the foggiest notion and send them as attachments to their (hot)mail without bothering.
No, if you want to explain market share of browsers you will have to look at something other than size, really!
Yes, I have this problem, no fix.
This issue has been there for a long time, and it amazes me too that this does not get fixed.
FYI, you do not need to reload, you can press "back" and "forward", that will fix it too with less traffic and less waiting (but two clicks).
We run Windows 2000 SP4 with all fixes, and about 10 applications (Mozilla, OpenOffice, ICA Client, SAP GUI, Acrobat Reader, JAVA runtime, etc plus MS Office on a few systems).
There are even some old systems with 3.2 GB disks and they have room to spare.
But it does not matter, replace the 4 GB with 8 GB if you like and you still see a lot of free space on a business workstation with the smallest disk available today. All data is of course stored on servers.
>Why shouldn't their lucrative, unethical trade move to massively parallel processing on that network to generate these "expensive" messages at our expense, rather than theirs?
What do you mean 'our' expense? I do not count the virus infected PCs of clueless users as 'my' expense, and I would rather like to see that the individual home user gets heavily punished by having a spam relay on their machine (by performance problems). A system with micropayments would probably have the same problem, but it would cost those users real money. Is that your better alternative?
>We're faced with the choice of evolving, or wallowing in the muck we've got now.
It has been like that for nearly a decade, and the Internet mail community has hardly made any moves towards a new mailsystem. I mean a really new mailsystem, not some blocklist or sender ID add-on to the existing broken system.
My real fear is that Microsoft will come up with such a system, will take 99% of the users with it like they did with Internet Explorer, and then lockout all opensource, free software, competing OS manufacturers etc from the world's e-mail system.
Fortunately we have only a single domain. Access control lists should not be a problem, but they should be included (they often aren't in programs intended for file transfer rather than backup).
If I remember correctly, RAR can do almost what we need except for the requirement to store the different parts on different machines (= in different destination directories). And except it does not do on-the-fly PAR2 generation but leaves that to a separate utility that again reads all the files.
And of course, to use it in real life there would need to be some database to store what backup on what date was stored on which workstations and under what filename. Plus a mechanism to purge old files to make room for new ones.
Indeed, this can be expanded and extended almost indefinately, but we are really looking at a simple solution for a small company (with more creative employees than money).
Micropayments do not need to involve money.
That avoids a lot of complications and legal stuff.
Solutions have been proposed to create stamps that require some computer power to make, but are easy to verify. Someone wanting to send a mail message can easily spend 10-60 seconds of CPU power on that (e.g. while it is being typed) but a spammer will not be able to afford that.
What you describe is like an rsync server that sends the correct files to the client. That can already be done, you just need someone to setup the rsync server.
But indeed I have always wondered why I need to download an entire 25MB RPM for each and every security fix of some packages. Sometimes the fix is not even in the binary files, but in some config file or the installation script of the RPM!
What you would expect is a small update, that depends on the original RPM (program-1.0) and upgrades it to a fixed program (program-1.0-2 or program-1.1). Installation of the fix RPM should apply the patch and change the RPM database to the new situation, just as if you had removed the old version installed the new version from a big RPM file.
I have discussed the idea of using PAR2 files for backups at work.
The situation is this: we have some 200 workstations. These have way too much diskspace, as you cannot buy small disks these days anymore. 4GB would be plenty for a business-use workstation (just the OS and some applications are stored there), but the workstations have 40 or 80 GB disks.
This means there is about 8TB of unused diskspace.
The servers have much less diskspace than that, in total. It would be possible to store several backups on the workstation disks.
The idea is to make backups as a set of files, distributed over a number of workstations, and protected by PAR2 files (in case a workstation fails or is re-installed). The backup data would of course be encrypted, and not accessible to the workstation user because of access rights.
This could serve as a quickly accessible backup, in addition to the normal tape backups we use. It could also be more reliable (tape backup units are not the most reliable parts in computers).
What is missing, or at least as far as we know, is a tool that allows to store a directory tree, access rights included, into a a number of parts with added PAR2 files, where the destination of each file can be specified separately (e.g. via a control file). We would not want to first copy all server data to a local set of files and then move each of these files to a workstation in a second step, but the files would have to be directly written to the workstations over the network.
Anyone seen something like that?
I suspect eventually we'll see dns servers implementing a custom record type for SPF to replace the current TXT records, but other than that, you don't really need anything else.
The big problem with SPF right now is the very limited support for TXT records by providers of DNS service.
I would really like to experiment with SPF but all three DNS providers I am using at home and at work don't offer the editing of TXT records via the usual web interface.
I am still hoping that more wide use of SPF would lead to the addition of TXT to the usual A, PTR, CNAME and MX repertoire of DNS records available to end users, but I fear that asking for a new record type will be too much.
It seems that in the USA, the word Moore is filtered. Can we draw conclusions from that?
That was the old approach.
It does not really work because people will check the list to conclude that something will NOT work, and there is little incentive to add something that works to a database. Remember you are dependent on voluntary action by someone who owns te hardware and can judge if "it works".
When a devide does NOT work, and you have tried to find drivers and contact the manufacturer, you may be more likely to have found the "incompatability list" in the process, and add your device to that.
It looks like it controls the HVAC using 24V AC switching. Is that still used in the USA?
:-(
Over here (Netherlands) any modern heating system uses a digital serial link between the equipment and the thermostat. The thermostat not only switches the heater on and off, but can also set the level (modulation), read back and display status information, set parameters like hot water operation mode etc.
As it does not use straight on/off switching, such a system operates much more smoothly. Rather than cycling the heater with a dutycycle that results in the correct room temperature, it sets the heater to the correct level to keep that temperature.
There exists a standard protocol called OpenTherm. Although it is called "Open" this does not mean that specifications are available for everyone to download
Wannabe firewall reviewers like to whine about "stealthed" and "closed" ports. Usually they don't know what they are talking about. When a review starts to mention how this or that firewall has listed a port as "closed" instead of "stealthed" and claims that this is less optimal, to me this devaluates the entire review.
This depends on your background and on the problem you are looking at.
There are Windows problems that I have never been able to completely understand and solve, mostly because of the black-box situation you find yourself in.
And there are many tasks that you can solve in a few minutes in Linux, and take hours of RSI-inducing clicking in Windows.
A typical cylinder lock has just as many combinations and works just the same as this type of lock. The pistons are just aranged differently, but the principle is the same.
Cylinder locks look secure, but often can be openened in 30 seconds with two paperclips.