How long will this "we need a new e-mail system" go on? The discussion about a new protocol to replace SMTP has gone on for ages, but nothing has happened.
I predict that Microsoft will come up with a new, better secured way of transferring mail messages over the Internet. It will be a closed architecture that requires Windows on all client and server systems. It will take over from e-mail overnight. In about a year's time, you will get more and more comments like "Oh, you still have such and old-fashioned mail address, one with a @ in it?" from most of your mail partners, certainly in business uses of mail...
Why? Because the advocates of open standards only talk about the problems of migrating to a new standard, and don't actually start designing and migrating.
I have always considered this a disadvantage. SCCS had the same problem.
These systems operate with the notion of a background storage system where you checkin and checkout working versions. What I often really want is a system where a number of files that are frequently changed can be separately kept in a versioning system. So, when I checkin something, I do not want the file to disappear.
I wrote a simple script that is run once a day, and puts the current version of a file in CVS when it has changed, but keeps the file available. I always wonder why such a mode of operation is not available.
Who said the writes would have to be synchronous? There may be a maximum throughput, but that does not need to slow down anything (except maybe a system shutdown shortly after a big write)...
Maybe you should hack a bit with the software (drivers). There are only subtle differences between the NE1000 (8bit) and NE2000 (16bit) network cards, and the code to recognize the cards may be misguided by the situation. You probably still can get it to work.
Years ago, I took an NE2000 card, mounted it on a piece of experimenters board, and fitted it into an Atari Mega ST (the ones with the pizzabox case under the monitor, that had a bus slot). After modifying the driver that was in the Linux kernel at that time (1.0 days, I guess) it worked okay. Unfortunately there was no openly available network stack for the machine, so I could only use it with KA9Q NET. So, I had telnet, ftp, smtp etc but no shared drives.
It probably depends on the people and their expectations, plus their flexibility when looking at solutions.
We are running a system like that for over three years now, and our managers are very happy with it. In fact one asked me last friday to replace the Outlook program on his personal (home) system by Mozilla.
They not only like the fact that it is free, they also like that it has never been down except for scheduled maintenance, and that we have never had a virus or trojan.
For scheduling we use Maorong Zou's "Webcalendar". Have a look here: http://www.ma.utexas.edu/~mzou/webCal/index.html
With every migration, there is always the issue of "does the new product implement all 25000 features of the old product". When you start your migration with the assumption "the new product must do everything the old product did plus possibly more" you will not only limit your options, but you will also migrate to more and more complex systems.
Instead, you should look at the requirements of the organisation, and define the properties of the system that will be implemented. Maybe Outlook/Exchange implements them, but that does not mean that it is the only solution to the problem.
Don't try to implement an Outlook/Exchange replacement, but define what your company needs and implement that. Possibly it does not need all the features of Outlook/Exchange and thus they do not need to be present in the "replacement".
We run an IMAP mailserver on Linux, with LDAP address book, and a separate web-based calendaring system. All are accessed from Mozilla on the (Windows) desktops. It works fine. The only thing I would want to be improved is the maintenance of the LDAP address book by nontechnical users.
Well, the problem is (maybe that is not clear from my posting) not that I receive lots of spam, but that I get lots of bounce messages and complaints about spam that I never sent but claims to be from me. Even when I just ignore all bounce mail, it causes an irritating growth of the logfiles and a noticable load of the ADSL link and the system.
Something like SPF will not help me unless all OTHER people implement it.
Re:Spam isn't that much of a problem ...
on
Baffling the Spam Bots
·
· Score: 2, Interesting
Don't count yourself lucky just yet! I used the same method, and my own mailserver with agressive filters, and it worked very well until... a Russian spammer started to send out spam with my mail address as the sender address. He did this via hacked systems (open proxies) so it was not possible to do any blocking. The load of crap that came in was just unbelievable, and all attempts to contact his spamvertized site or their providers just had no result.
In the end the only thing I could do was remove the MX record for the domain. I pointed it to the spamvertized site instead. Hopefully they are happy with their own bounces. Of course I cannot receive any legitimate mail on that address anymore:-(
Indeed the patches were bad. I tried the first one and it caused strange problems. My ISP installed another one and it is even worse: it does not return an error but it simply returns no answer for the wildcarded records.
This assumes that the failing of one engine has no influence whatsoever on any other engine. This is of course very far from the truth.
When one engine falls off (manufacturers say this cannot happen) it can fall aways sideways (very unlikely) and damage the wing at the same time. The aircraft can spin down uncontrolled, and it could fly into an appartment building.
When you calculate the chances that this all happens at the same time, you get an astronomically small number. But still it happened. Why? because these chances are not independent. So the calculation is not correct.
If you were to put drives from manufacturers 1 and 2 into a RAID-0 array. You would have the worst of both worlds because one drive would be waiting for the other drive two thirds of the time.
This will only be true when the two drives are driven completely synchronously, e.g. with a controller that writes only single sectors and waits for both writes to complete before starting the next. In practice this is rarely done. There are buffers on the controller or in the kernel that allow the two drives to run asynchronously. When reading, the two drives are operated interleaved (like in RAID-0) so their heads are probably in different positions anyway.
(early RAID arrays used spindle sync, but who is still doing that today? now, the drives are just running independently)
Actually, Maxtor are marketing large IDE drives now for what they call "near-line storage". It seems they consider this a step between on-line and off-line storage.
These drives are specified for storage of backups, archives etc where the access frequency is lower than for a fileserver.
SCART is a 21-pin connector that has all those signals on it. You can make or buy a conversion cable between SCART and S-video, composite video, component video, or whatever you need.
Recently I have been looking for an A/V amplifier to be used as a home theatre system (combined with a screen, speakers, DVD player, DM7000 etc).
Unfortunately there are not many that offer SCART connectors. IMHO they are the best solution. Who wants to use those flimsy RCA connectors or even worse: S-video? I want many RGB video connections.
It seems the US video world has invented something incompatible with the industry standard RGB: component video. Probably to force the customer to buy new stuff. The DM7000 can output YPrPb but why was this invented instead of RGB that everyone already used?
Even then, an A/V amplifier has at best two component video inputs and one such output. And often it only switches the input (component, S-video or baseband video) to the same type of output. Of course what you would want is to have different types of input and only a single (RGB or component video) output to send to the screen.
Why is A/V equipment built in such an impractical way? It would be good if US-oriented manufacturers (I think there are no US manufacturers, all this stuff comes from the far east) look at SCART and desirable operating characteristics of their equipment. I (and I presume other customers) want to just plug in the cables and have perfect quality picture, without having to fight with 30 RCA plugs and live with funny limitations in the equipment.
It is not at all a DVR, it is a satellite (DVB-S) receiver with recording functionality. It cannot record anything from external sources.
It has access to program guides transmitted in DVB, but most program guides cover only the current/next program and sometimes a couple of days. When there is a program guide, it can record an item from it. But only when you can select the item at the time you program the recording (there is no "record all items that have star trek in the title" function. of course you can add it!)
There is a flexible recurring recording programming system (record on selected days of the week at a specified time) but it does not use the EPG.
Well, if you run a mailserver with maildir storage format, you are interested in the performance of directories with 5000 small files. You want to quickly do readdir operations, quickly open many of the files and read some data from them, etc.
When you run a fileserver and don't want to explain to your users that it is not a good idea to have 100.000 files in a single directory as a storage format for filled-in entry forms, you want that situation to be handled well by the filesystem.
When a nightly backup operation needs to read the tree that includes that directory and write it out to a directory on a different disk (and remove the copy of a few days before), it should be able to do that without spending ridiculous amounts of time or excessively wearing out the diskdrives.
Those are not hypothetical situations, those are situations that I encounter in real life.
Slashdot Mirror
The difference with roaming profiles is in the "World Wide Web" thingy. Roaming profiles roam to and from an SMB server, like Windows NT or SAMBA.
However, the old Netscape 4.x program did have this roaming profiles over World Wide Web capability.
It required a customization kit to be activated.
They are working towards a monopoly on services. That is even worse than a monopoly on servers would be.
How long will this "we need a new e-mail system" go on? The discussion about a new protocol to replace SMTP has gone on for ages, but nothing has happened.
I predict that Microsoft will come up with a new, better secured way of transferring mail messages over the Internet. It will be a closed architecture that requires Windows on all client and server systems. It will take over from e-mail overnight. In about a year's time, you will get more and more comments like "Oh, you still have such and old-fashioned mail address, one with a @ in it?" from most of your mail partners, certainly in business uses of mail...
Why? Because the advocates of open standards only talk about the problems of migrating to a new standard, and don't actually start designing and migrating.
I have always considered this a disadvantage. SCCS had the same problem.
These systems operate with the notion of a background storage system where you checkin and checkout working versions.
What I often really want is a system where a number of files that are frequently changed can be separately kept in a versioning system. So, when I checkin something, I do not want the file to disappear.
I wrote a simple script that is run once a day, and puts the current version of a file in CVS when it has changed, but keeps the file available. I always wonder why such a mode of operation is not available.
Who said the writes would have to be synchronous?
There may be a maximum throughput, but that does not need to slow down anything (except maybe a system shutdown shortly after a big write)...
Maybe you should hack a bit with the software (drivers).
There are only subtle differences between the NE1000 (8bit) and NE2000 (16bit) network cards, and the code to recognize the cards may be misguided by the situation. You probably still can get it to work.
Years ago, I took an NE2000 card, mounted it on a piece of experimenters board, and fitted it into an Atari Mega ST (the ones with the pizzabox case under the monitor, that had a bus slot).
After modifying the driver that was in the Linux kernel at that time (1.0 days, I guess) it worked okay.
Unfortunately there was no openly available network stack for the machine, so I could only use it with KA9Q NET. So, I had telnet, ftp, smtp etc but no shared drives.
It probably depends on the people and their expectations, plus their flexibility when looking at solutions.
We are running a system like that for over three years now, and our managers are very happy with it. In fact one asked me last friday to replace the Outlook program on his personal (home) system by Mozilla.
They not only like the fact that it is free, they also like that it has never been down except for scheduled maintenance, and that we have never had a virus or trojan.
For scheduling we use Maorong Zou's "Webcalendar".
Have a look here: http://www.ma.utexas.edu/~mzou/webCal/index.html
With every migration, there is always the issue of "does the new product implement all 25000 features of the old product".
When you start your migration with the assumption "the new product must do everything the old product did plus possibly more" you will not only limit your options, but you will also migrate to more and more complex systems.
Instead, you should look at the requirements of the organisation, and define the properties of the system that will be implemented. Maybe Outlook/Exchange implements them, but that does not mean that it is the only solution to the problem.
Don't try to implement an Outlook/Exchange replacement, but define what your company needs and implement that. Possibly it does not need all the features of Outlook/Exchange and thus they do not need to be present in the "replacement".
We run an IMAP mailserver on Linux, with LDAP address book, and a separate web-based calendaring system. All are accessed from Mozilla on the (Windows) desktops. It works fine.
The only thing I would want to be improved is the maintenance of the LDAP address book by nontechnical users.
The reboot is only required on Windows...
Well, the problem is (maybe that is not clear from my posting) not that I receive lots of spam, but that I get lots of bounce messages and complaints about spam that I never sent but claims to be from me.
Even when I just ignore all bounce mail, it causes an irritating growth of the logfiles and a noticable load of the ADSL link and the system.
Something like SPF will not help me unless all OTHER people implement it.
Don't count yourself lucky just yet!
:-(
I used the same method, and my own mailserver with agressive filters, and it worked very well until... a Russian spammer started to send out spam with my mail address as the sender address. He did this via hacked systems (open proxies) so it was not possible to do any blocking.
The load of crap that came in was just unbelievable, and all attempts to contact his spamvertized site or their providers just had no result.
In the end the only thing I could do was remove the MX record for the domain. I pointed it to the spamvertized site instead. Hopefully they are happy with their own bounces.
Of course I cannot receive any legitimate mail on that address anymore
According to the ITRON page linked, it is an operating system specification, not an operating system.
That would make it a competitor of Posix, instead of Linux.
Indeed the patches were bad. I tried the first one and it caused strange problems.
My ISP installed another one and it is even worse: it does not return an error but it simply returns no answer for the wildcarded records.
about:config
This assumes that the failing of one engine has no influence whatsoever on any other engine.
This is of course very far from the truth.
When one engine falls off (manufacturers say this cannot happen) it can fall aways sideways (very unlikely) and damage the wing at the same time.
The aircraft can spin down uncontrolled, and it could fly into an appartment building.
When you calculate the chances that this all happens at the same time, you get an astronomically small number. But still it happened.
Why? because these chances are not independent. So the calculation is not correct.
If you were to put drives from manufacturers 1 and 2 into a RAID-0 array. You would have the worst of both worlds because one drive would be waiting for the other drive two thirds of the time.
This will only be true when the two drives are driven completely synchronously, e.g. with a controller that writes only single sectors and waits for both writes to complete before starting the next.
In practice this is rarely done. There are buffers on the controller or in the kernel that allow the two drives to run asynchronously. When reading, the two drives are operated interleaved (like in RAID-0) so their heads are probably in different positions anyway.
(early RAID arrays used spindle sync, but who is still doing that today? now, the drives are just running independently)
Actually, Maxtor are marketing large IDE drives now for what they call "near-line storage". It seems they consider this a step between on-line and off-line storage.
s /maxline_data_sheet.pdf
These drives are specified for storage of backups, archives etc where the access frequency is lower than for a fileserver.
Check this: http://www.maxtor.com/en/documentation/data_sheet
Buy only one? I think you should always buy drives in pairs and put them in RAID-1 configuration...
Linux is a multi-user operating system.
SCART is a 21-pin connector that has all those signals on it. You can make or buy a conversion cable between SCART and S-video, composite video, component video, or whatever you need.
Recently I have been looking for an A/V amplifier to be used as a home theatre system (combined with a screen, speakers, DVD player, DM7000 etc).
Unfortunately there are not many that offer SCART connectors. IMHO they are the best solution. Who wants to use those flimsy RCA connectors or even worse: S-video? I want many RGB video connections.
It seems the US video world has invented something incompatible with the industry standard RGB: component video. Probably to force the customer to buy new stuff. The DM7000 can output YPrPb but why was this invented instead of RGB that everyone already used?
Even then, an A/V amplifier has at best two component video inputs and one such output. And often it only switches the input (component, S-video or baseband video) to the same type of output. Of course what you would want is to have different types of input and only a single (RGB or component video) output to send to the screen.
Why is A/V equipment built in such an impractical way? It would be good if US-oriented manufacturers (I think there are no US manufacturers, all this stuff comes from the far east) look at SCART and desirable operating characteristics of their equipment. I (and I presume other customers) want to just plug in the cables and have perfect quality picture, without having to fight with 30 RCA plugs and live with funny limitations in the equipment.
The latest firmware does NTSC.
It is not at all a DVR, it is a satellite (DVB-S) receiver with recording functionality.
It cannot record anything from external sources.
It has access to program guides transmitted in DVB, but most program guides cover only the current/next program and sometimes a couple of days. When there is a program guide, it can record an item from it. But only when you can select the item at the time you program the recording (there is no "record all items that have star trek in the title" function. of course you can add it!)
There is a flexible recurring recording programming system (record on selected days of the week at a specified time) but it does not use the EPG.
Or fuel cells... a charge of H2 and O2 in a small box should be fun as well.
Well, if you run a mailserver with maildir storage format, you are interested in the performance of directories with 5000 small files.
You want to quickly do readdir operations, quickly open many of the files and read some data from them, etc.
When you run a fileserver and don't want to explain to your users that it is not a good idea to have 100.000 files in a single directory as a storage format for filled-in entry forms, you want that situation to be handled well by the filesystem.
When a nightly backup operation needs to read the tree that includes that directory and write it out to a directory on a different disk (and remove the copy of a few days before), it should be able to do that without spending ridiculous amounts of time or excessively wearing out the diskdrives.
Those are not hypothetical situations, those are situations that I encounter in real life.