The Patriot Act doesn't really do anything new. It takes the previous provisions against racketeering, and extends them to suspected terrorists - with the additional safeguard of requiring a judges approval.
But while I think that liberal talk about "power grabs" is way overheated, I am concerned about the Patriot act - even as a conservative. Why? Because long before the Patriot act, the racketeering laws were used against abortion protestors. If a protestor can be labelled a "racketeer", and deprived of legal rights, what's to prevent someone who disapproves of homosexual behaviour from being labelled a terrorist? Another problem is that racketeering is something that takes place on an ongoing basis and everyone knows it - the police are simply trying to determine the culprits. Terrorism is something that the culprits are planning for the future - often with no plans to live beyond the deed. Trying to track them down is uncomfortably close to a department of pre-crime.
However, while I would rather for the present that things like racketeering laws and Patriot acts were not around to be abused, I think the root of the problem is courts exceeding their authority and redefining words, whether they are inventing novel definitions for "racketeering" and "marriage" or writing law from the bench respecting the establishment of religion.
We need more conservative judges - who restrict themselves to interpreting existing law rather than creating new law.
How is adding metadata to each picture file to categorize your vacation pictures any less laborious than placing the vaction pictures into their own directory?
It isn't. The file names are metadata. Links and Symlinks let you have multiple "metadata" entries. If directories represent categories, then you can link a picture into as many categories as applicable.
In terms of power, metadata support is equivalent to support for links. In fact, metatdata could also be encoded into long file names - but that could get pretty ugly. For instance, my company uses a homebrew filesystem where filenames can contain null chars. The normally visible part of a filename is terminated by a null char, and followed by arbitrary metadata in a conventional format - typically a database field table.
So there is no real need for special metadata support. Anything stored as metadata can be equivalently stored using some filename or directory bundle convention. The important thing is to define common conventions.
With the unix approach, in the worst case, with thousands of competing conventions, you can still backup and restore with your favorite tar or cpio like utility. If you go the special metadata route, on the other hand, you have to have specialized backup and restore utilities. This is a great feature for M$ (yet another way to lock you into their platform), but a huge drawback for open source.
Right now SPF is being done with a large library that gets linked in.
I can only speak for sendmail, but none of the sendmail solutions link in any libraries to sendmail. Every solution uses a sendmail milter. The milter process is started independently of sendmail, and does not required access to any files or privileged network ports - so it runs as an isolated user id. The only I/O required by an SPF milter is TCP for the milter protocol, DNS queries, and perhaps a log file. No patches to sendmail are required.
You may be confusing SPF with SRS, which cannot be fully implemented in sendmail via a milter. However, SRS can be securely implemented for sendmail via a socket map in the latest versions. A patch to support socket maps in older sendmail versions is available - or you can use the (horribly innefficient but workable) solution of a program map.
At that time in human development, "history" amounted to what we might call "mythology".
And this is different today, how? Our culture is loaded with myths of cosmic origin (the scientic guess work doesn't change the lack of direct observation and mythic style of presentation), national origin (George Washington chopping down the cherry tree and crossing the Delaware), and story telling (Superman, Star Wars, Tolkein).
This is greatly misunderstood- but his Dialogues were PLAYS.
This is so true. Not only of Plato, but of the Bible and any other literature both ancient and modern. When the director of the movie "The Day After Tomorrow" was criticized for the implausibility of an ice age developing in three days, he replied, "I had only 2 hours, and the movie is fiction, not a documentary."
It is worth noting that a story may be both myth and historical reality. A story is mythic because of the way is it told, not because it is untrue. Thus, although you might believe the story of evolution to be historically true, it is nevertheless usually told in a mythic style. "Millions of years ago, the earth was covered with a reducing atmosphere and a complex solution of dissolved chemicals - the prebiotic soup. One day..." Similarly, I believe that the story of George Washington crossing the Delaware is historical (but not the story about the cherry tree). But both are mythic stories.
Now having made point about understanding literature in light of its intended style, let me say that a popular style today is "historical fiction". In historical fiction, the background events and significant actions of well known characters are expected to be historical, whereas the actions of other characters and day to day actions of well known characters are fictional - although consistent with the historical background.
In the same vein, many of Shakespeare's plays were the historical fiction of the day, and it is not unreasonable to use them as a source for what was generally known at the time about Richard III and other historical characters. Similarly, Plato's stories about Socrates are usually considered to be either historical fiction or "based on a true story" - as opposed to pure fiction like Star Wars.
We have records from a time when people lived to be almost 1000 (Genesis). The men are recorded as fathering children through their 70s. This implies that the women had the same number of eggs and roughly the same childbearing years as today. In any event, child bearing was confined to the first century or less of those 1000 years.
As an additional data point, consider Sarah, Abraham's wife. Although after the flood, and lifespans were rapidly declining, they still lived hundreds of years. Sarah was a "looker" despite being past childbearing age. (Yes she was barren - but the text specifically mentions them ready to give up hope because she was also past normal childbearing age). So much so that two kings tried to add her to their harem.
So based on that evidence (which some may not accept as historical), it looks as if the number of children born per couple would stay about the same. Even with widespread promiscuity, the number of children born per female would stay about the same as if they were all married. (And VDs would help keep the numbers down.)
a couple of years ago, Sun was telling us we'd all be running on glorified VT100 terminals.
I'm typing this on a glorifed vt100. A NiC (thinknic.com - now defunct), as a matter of fact, running LTSP. I have 4 of them (LTSP terminals - the rest are recycled Pentiums) connected to a $400 Dell 400SC. You can't get much more economical than that. It's true that Sun didn't get any of my miniscule pie, but if they weren't bent on evil, I'd be looking at Sun hardware to run hundreds of glorified vt100s for business customers (as soon as I can convince the customer that their computer doesn't need Windows to run).
I don't know the admin password, our systems guy is a real sourpuss and won't tell us what it is:(
If you have a systems guy, you are probably not running Lindows. The point is, for the Lindows market - where the user and and admin are the same - user security won't help. A typical home user or Mom and Pop business user simply has no idea of when it is appropriate to supply the admin password.
The only solution to security for the Mom and Pop user (other than basic common sense learned in the school of hard knocks) is something like Paladium, where a 3rd party company fills the role of the "systems guy".
If only there were a way to guarantee a choice of who provides the security service (and control which programs you are allowed to install), the Trusted Computing Architecture could be a boon, rather than a strategic tool for World Domination by M$. Switching security providers for a TCA machine should be as easy as switching name servers for an internet domain. (Although a reinstall might be required.)
I've been drooling over the 400SC since we started getting some in our office to set up for customers. It really is whisper quiet. Unfortunately, I bought the 500SC for home - which sounds like a 747 taking off. (Slight exaggeration.) I've only had the 500SC for a year and a half, so it's too early to retire it. If I could spend $25 to quiet the thing, I'd be happy.
Caution on the 400SC with SATA: it won't boot from 3rd party IDE cards (such as you'd use for mirroring).
My hosting business started as a web collective. Everybody was gonna pay for his percentage of the server, and we were going to be a non-profit. For art! For software!
I'm not sure what your web commune had to do with Open Source. Nothing about open source prevents or discourages you from charging for goods and services. You'll find just as many freeloaders eager to take "free beer" from the closed source camp.
The alternative to OSS is Trade Guilds. OSS is a necessary foundation for a free market in the software industry. What Microsoft is promiting is the Trade Guild model - where all the information needed to create and use the underlying technology is carefully guarded and available only to guild members (those who work for Microsoft).
Outsiders might be able write a few scripts to customize things (after registering them with the Guild so that the Trusted Computing Architecture will load them). But the only source for the software infrastructure is the Software Guild, known as Microsoft in the early 21st century. The Theives Guild might offer some alternatives, but with its own costs and risks.
Worse, of course, is the collateral damage. How will I be able to send mail using my own business' domain, as I do today, when it is going out via an ISP server? My "from" address is an alias, not a real sender, and I use it to send via more than one ISP, depending on where I am. SPF seems to make this a lot harder, thereby forcing more people to put their ISPs' name in the From: field, rather than their own. Since email is not portable, a user's address is lost when they change ISPs, or when their ISP changes names (mediaone->attbi->comcast). Personal domains (forwarded via a service like mydomain) solve this. Will SPF kill mydomain?
Your objection mentions a multitude of configuration scenarios. I will address a few:
A home user with a personal domain simply lists their ISP as a valid sender in SPF.
A travelling salesman with a laptop should use a VPN or SMTP AUTH or webmail.
A travelling salesman using anonymous PCs wherever he finds them is crazy. But if he insists, he should use webmail. That is probably less of a risk than entering an SMTP AUTH user and password.
SPF does not prevent you from putting whatever you want into the From: header. (Yahoo domain keys addresses that.) SPF only authenticates the envelope sender. The envelope sender is used for bounces, return receipts, and the like.
If you can't authenticate from the field in any way, you could always publish SPF that lists your home system and leaves everything else neutral. Since anyone, including me or a spammer can currently send email claiming to be from you, that accurately reflects the situation.
I run folding@home and seti@home as Linux services via sysvinit. They run with
nice -20
and are mostly unnoticed - except that the load average never goes below 1.0 (for one project running) or 2.0 (for two projects running). Then, I have to tune lots of other daemons that stop working when the LA goes above some level - typically defaulting to 1.0. For instance, sendmail will only queue mail unless you tune QUEUE_LA and REFUSE_LA.
I would like to have explicit support for background processes like this. They should be treated like an idle loop - never run unless there is nothing else to do, and not reported in the load average.
We used SCO for several years in the 1980's, and were very dissatisfied with their support in general. The answer to any problem was "Reformat the disk and reinstall Xenix to see if the problem persists." Naturally, this was never a viable option for a production system. A better solution was to replace the buggy SCO software with the fledgling GNU software. If it didn't work much better already, it could be fixed. I bought several cartridge tape distros from GNU to support them.
When ESDI disks came out, we thought it would be a good idea to try and get better support for the new technology. So we signed up for the $1200/yr premium support plan. That kind of money should at least get us past the "reformat your disk" nonsense.
We got our first ESDI system, and booted the latest Xenix install with ESDI support from diskette. Everything went smoothly until it got to the part where you format and partition the disk. Two thirds of the way through the formatting, it found a bad sector. No biggie, these were common and just added to the bad block map in those days. However, it kept finding the *same* bad sector over and over - ad infinitum.
So we called our premium tech support - confident that now we had a problem that they couldn't possibly blow off with "reformat the disk", since that was exactly what we were trying to do. Not to mention the big bucks we were paying. I explained the problem, and to my horror and consternation, the guy said, "Reformat your disk and reinstall Xenix." I completely lost it, and told him he was a complete idiot and needed a new career. He told me I needed to calm down and follow instructions if I wanted his help. I told him what he could do with his help. The boss gave me a long lecture on the relative number of flies caught with honey versus vinegar - however, that was the last SCO system we ever bought.
The bms.py application is the SPF checking milter (with lots of other features). The RedHat RPM provides a sysvinit start/stop/status script, defanged mail cleanup cron script, and puts log files in/var/log/milter (along with the config - to be fixed). If you are building your own package from source (as you would with FreeBSD), just run bms.py with python and add to sendmail.mc per the README. milter.cfg documentation is comments in the sample milter.cfg. Free free to email me directly from my address on the web page (stuart). I do not have FreeBSD, but will be glad to include BSD specific packaging/setup docs. There is an openbsd port - but I don't know if that applies to freebsd.
A C spfmilter has been discusses on the spf-discuss list. Check the archives. However, I am using Python Milter to check SPF. It is in use by several sites that process 10K+ messages/hr. I like having something that is so easy to tweak. At one point, I was processing 40K+ messages/day - and the 400Mhz 686 was still loafing. I don't think CPU is an issue with either Perl or Python for this application (I don't recommend pure Python for ray tracing:-). Python is pretty stingy with memory beyond the shared memory for the interpreter and modules.
There is no startup cost because the milter process is persistent. Each connection is handled by its own thread. If your objection to Perl has to do with syntax, then I am with you. Otherwise, Perl has similar issues.
Exokernels reinvent IBMs VM system. "An Exokernel securely multiplexes the raw hardware, and application libraries directly implement traditional OS functions." This does not mean that applications must now include their own drivers for every possible hardware they might use. It means that drivers can now be packaged as shared libraries in user space rather than as kernel modules.
To summarize, let me call the part that securely multiplexes hardware the "kernel".
Monolithic makes drivers share address space with the "kernel".
With Microkernel, "kernel", drivers, filesystems, applications, etc each get their own address spaces.
Exokernel makes drivers share address space with applications. (Hopefully, filesystems get their own process and address space.)
As you can see, as soon as you start partitioning applications into separate processes for security and robustness, the distinction between Exokernel and Microkernel becomes rather vague. The advantage of the Exokernel or VM approach is that you get the flexibility of keeping things like filesystems in a separate process for security and robustness, and things like video drivers in the same address space for performance. You might even have an X server as a separate process, but still allow full screen mode games that directly call the driver libraries for performance.
IBM's VM was never that popular in its raw "Exokernel" mode with drivers in application space. However, it is still hugely popular as a way to run multiple Operating Systems as the "applications". Your mainframe can securely run multiple instances of S/390 Linux and traditional mainframe systems together.
And you point out what SPF tries to downplay: The ultimate purpose of ?all is to be treated as spam. Therefore, any "solution" that suggests use of ?all is disingenuous.
The ultimate purpose of SPF is for all email senders to be authenticated so that there is no more envelope header forgery. The ?all solution is not disingenuous - it provides a way to implement as much of SPF as you can now without breaking anything. SPF provides flexible options, and a gradual phased adoption plan so that email providers can choose from a number of strategies and take as long as they need to implement the ultimate goal of sender authentication.
for any large ISP, managing the notification and inclusion of hundreds of thousands if not millions of individual forwards is a logistical nightmare, and it's the height of hubris to suggest that the solution is just "notify the recipient MTA of the forwards".
Of course. And that is why AOL does not currently check SPF. And why when they do, they will not reject mail solely on the basis of SPF.
Senders that use SPF + SES (Signed Envelope Sender) make it possible for a large ISP recipient to reject forged mail even without the cooperation of forwarders.
Currently, AOL does not check SPF for incoming mail. Supposing for the sake of argument that they did, they would need to provide a way for users to authorize non-SRS forwarders (it won't hurt to redundantly authorize an SRS forwarder). If the student then fails to authorize foo.edu, then the student is incompentent.
However, since AOL is designed for people that are general incompetent when it comes to computers, (because they have other interests), then AOL might wisely choose not to reject any mail based solely on SPF.
While foo.edu didn't do anything wrong, they also could make life easier for the student by implementing SRS for mail that they forward. I install my pysrs package for sendmail. Enabling SRS is a single line in sendmail.mc.
This would effectively be saying, "I'm probably spam; please filter me."
"~all", known as 'softfail' means "I'm probably spam; please filter me." It is intended for a transition period. "?all", known as 'neutral' is equivalent to not publishing SPF records at all. It allows you to validate mail sent from machines you control while saying nothing about machines you don't control. Whether you have no SPF record at all, or whether you send from a machine that hits the "?all", the result for a receiving MTA that checks SPF is:
Received-SPF: neutral
If it ever gets to the point where a neutral result means "I'm probably spam", then SPF will have been a smashing success.
BTW, with regard to SMTP AUTH and/or SSL, if mainstream mail clients like Outlook or Mozilla or pine are not available where you need to send mail, then
you might consider using web mail for sending validated email from remote locations.
You don't need to currently because AOL doesn't check SPF. (They only publish it.) Should they decide to begin checking SPF, then hopefully they are competent enough to do it correctly - so you still don't need to. If not, then their incompetence in handling email will likely lose some mail - SPF is not required for incompetence to have that effect.
In general, forwarders are selected by the email recipient and handling them correctly when implementing SPF is the responsibility of the email recipient. It is easy for the recipient to "punt" as a first step to implementing SPF and simply add Received-SPF headers without actually rejecting anything. The Received-SPF headers are remarkably effective at aiding content filters to do their job.
With SPF, you don't just know where the zombie resides, you know which domain authorizes the zombie to send mail. You know who registered the domain, and hence is responsible for authorizing the zombie. Unauthorized zombies can be ignored. Because AOL publishes SPF, I am ignoring most zombie mail now.
Without SPF, you have no idea who is responsible for the zombie machine (unless you are RIAA or the ISP). Furthermore, the vast majority of zombies aren't authorized by *any* domain - their owners are computer clueless AOL or Yahoo users.
If SPF were widely implemented, Mom and Pop (and their Zombie masters) would not be able to send mail directly from their Windoze box without first registering a domain and publishing an SPF record to authorize it (or asking their ISP to authorize their box).
I am dismayed at how often this misunderstanding has been repeated here.
If the receiver does not check SPF, then no mail is rejected and forwarding is not broken.
If the receiver does check SPF, but doesn't use any forwarders, then forwarding is not broken.
If the receiver does check SPF, but uses only forwarders that implement SRS, then forwarding is not broken.
If the receiver does check SPF and uses a non-SRS forwarder, but uses a whitelist to avoid rejecting mail from that forwarder, then forwarding is not broken.
If the receiver check SPF and uses a non-SRS forwarder, but configures their MTA to reject mail from that forwarder, then their incompetence will result in rejected mail. How is this the fault of SPF?
This is not a problem with SPF. The easiest solution is to simply leave the default for your SPF record as "?all" - which says "I don't know" for any of the hotel sites you might send from. The better solution is to use SMTP AUTH or SMTP over SSL to relay your mail from the hotel through the home office.
SPF tells you if the envelope sender of the mail is forged. Nothing tells you if the mail as a whole is forged (except perhaps PGP/GPG/S-MIME if you are careful with your private key password). Someone could have sat down at your PC while you were at lunch and forgot to lock your screen.
Slashdot Mirror
But while I think that liberal talk about "power grabs" is way overheated, I am concerned about the Patriot act - even as a conservative. Why? Because long before the Patriot act, the racketeering laws were used against abortion protestors. If a protestor can be labelled a "racketeer", and deprived of legal rights, what's to prevent someone who disapproves of homosexual behaviour from being labelled a terrorist? Another problem is that racketeering is something that takes place on an ongoing basis and everyone knows it - the police are simply trying to determine the culprits. Terrorism is something that the culprits are planning for the future - often with no plans to live beyond the deed. Trying to track them down is uncomfortably close to a department of pre-crime.
However, while I would rather for the present that things like racketeering laws and Patriot acts were not around to be abused, I think the root of the problem is courts exceeding their authority and redefining words, whether they are inventing novel definitions for "racketeering" and "marriage" or writing law from the bench respecting the establishment of religion.
We need more conservative judges - who restrict themselves to interpreting existing law rather than creating new law.
It isn't. The file names are metadata. Links and Symlinks let you have multiple "metadata" entries. If directories represent categories, then you can link a picture into as many categories as applicable.
In terms of power, metadata support is equivalent to support for links. In fact, metatdata could also be encoded into long file names - but that could get pretty ugly. For instance, my company uses a homebrew filesystem where filenames can contain null chars. The normally visible part of a filename is terminated by a null char, and followed by arbitrary metadata in a conventional format - typically a database field table.
So there is no real need for special metadata support. Anything stored as metadata can be equivalently stored using some filename or directory bundle convention. The important thing is to define common conventions.
With the unix approach, in the worst case, with thousands of competing conventions, you can still backup and restore with your favorite tar or cpio like utility. If you go the special metadata route, on the other hand, you have to have specialized backup and restore utilities. This is a great feature for M$ (yet another way to lock you into their platform), but a huge drawback for open source.
I can only speak for sendmail, but none of the sendmail solutions link in any libraries to sendmail. Every solution uses a sendmail milter. The milter process is started independently of sendmail, and does not required access to any files or privileged network ports - so it runs as an isolated user id. The only I/O required by an SPF milter is TCP for the milter protocol, DNS queries, and perhaps a log file. No patches to sendmail are required.
You may be confusing SPF with SRS, which cannot be fully implemented in sendmail via a milter. However, SRS can be securely implemented for sendmail via a socket map in the latest versions. A patch to support socket maps in older sendmail versions is available - or you can use the (horribly innefficient but workable) solution of a program map.
And this is different today, how? Our culture is loaded with myths of cosmic origin (the scientic guess work doesn't change the lack of direct observation and mythic style of presentation), national origin (George Washington chopping down the cherry tree and crossing the Delaware), and story telling (Superman, Star Wars, Tolkein).
This is greatly misunderstood- but his Dialogues were PLAYS.
This is so true. Not only of Plato, but of the Bible and any other literature both ancient and modern. When the director of the movie "The Day After Tomorrow" was criticized for the implausibility of an ice age developing in three days, he replied, "I had only 2 hours, and the movie is fiction, not a documentary."
It is worth noting that a story may be both myth and historical reality. A story is mythic because of the way is it told, not because it is untrue. Thus, although you might believe the story of evolution to be historically true, it is nevertheless usually told in a mythic style. "Millions of years ago, the earth was covered with a reducing atmosphere and a complex solution of dissolved chemicals - the prebiotic soup. One day ..." Similarly, I believe that the story of George Washington crossing the Delaware is historical (but not the story about the cherry tree). But both are mythic stories.
Now having made point about understanding literature in light of its intended style, let me say that a popular style today is "historical fiction". In historical fiction, the background events and significant actions of well known characters are expected to be historical, whereas the actions of other characters and day to day actions of well known characters are fictional - although consistent with the historical background.
In the same vein, many of Shakespeare's plays were the historical fiction of the day, and it is not unreasonable to use them as a source for what was generally known at the time about Richard III and other historical characters. Similarly, Plato's stories about Socrates are usually considered to be either historical fiction or "based on a true story" - as opposed to pure fiction like Star Wars.
As an additional data point, consider Sarah, Abraham's wife. Although after the flood, and lifespans were rapidly declining, they still lived hundreds of years. Sarah was a "looker" despite being past childbearing age. (Yes she was barren - but the text specifically mentions them ready to give up hope because she was also past normal childbearing age). So much so that two kings tried to add her to their harem.
So based on that evidence (which some may not accept as historical), it looks as if the number of children born per couple would stay about the same. Even with widespread promiscuity, the number of children born per female would stay about the same as if they were all married. (And VDs would help keep the numbers down.)
I'm typing this on a glorifed vt100. A NiC (thinknic.com - now defunct), as a matter of fact, running LTSP. I have 4 of them (LTSP terminals - the rest are recycled Pentiums) connected to a $400 Dell 400SC. You can't get much more economical than that. It's true that Sun didn't get any of my miniscule pie, but if they weren't bent on evil, I'd be looking at Sun hardware to run hundreds of glorified vt100s for business customers (as soon as I can convince the customer that their computer doesn't need Windows to run).
If you have a systems guy, you are probably not running Lindows. The point is, for the Lindows market - where the user and and admin are the same - user security won't help. A typical home user or Mom and Pop business user simply has no idea of when it is appropriate to supply the admin password.
The only solution to security for the Mom and Pop user (other than basic common sense learned in the school of hard knocks) is something like Paladium, where a 3rd party company fills the role of the "systems guy".
If only there were a way to guarantee a choice of who provides the security service (and control which programs you are allowed to install), the Trusted Computing Architecture could be a boon, rather than a strategic tool for World Domination by M$. Switching security providers for a TCA machine should be as easy as switching name servers for an internet domain. (Although a reinstall might be required.)
Dear accounting,
I have pictures of your daughter! Install the enclosed RPM to view them. Enter the admin password when prompted.
Caution on the 400SC with SATA: it won't boot from 3rd party IDE cards (such as you'd use for mirroring).
I'm not sure what your web commune had to do with Open Source. Nothing about open source prevents or discourages you from charging for goods and services. You'll find just as many freeloaders eager to take "free beer" from the closed source camp.
The alternative to OSS is Trade Guilds. OSS is a necessary foundation for a free market in the software industry. What Microsoft is promiting is the Trade Guild model - where all the information needed to create and use the underlying technology is carefully guarded and available only to guild members (those who work for Microsoft).
Outsiders might be able write a few scripts to customize things (after registering them with the Guild so that the Trusted Computing Architecture will load them). But the only source for the software infrastructure is the Software Guild, known as Microsoft in the early 21st century. The Theives Guild might offer some alternatives, but with its own costs and risks.
Your objection mentions a multitude of configuration scenarios. I will address a few:
I would like to have explicit support for background processes like this. They should be treated like an idle loop - never run unless there is nothing else to do, and not reported in the load average.
When ESDI disks came out, we thought it would be a good idea to try and get better support for the new technology. So we signed up for the $1200/yr premium support plan. That kind of money should at least get us past the "reformat your disk" nonsense.
We got our first ESDI system, and booted the latest Xenix install with ESDI support from diskette. Everything went smoothly until it got to the part where you format and partition the disk. Two thirds of the way through the formatting, it found a bad sector. No biggie, these were common and just added to the bad block map in those days. However, it kept finding the *same* bad sector over and over - ad infinitum.
So we called our premium tech support - confident that now we had a problem that they couldn't possibly blow off with "reformat the disk", since that was exactly what we were trying to do. Not to mention the big bucks we were paying. I explained the problem, and to my horror and consternation, the guy said, "Reformat your disk and reinstall Xenix." I completely lost it, and told him he was a complete idiot and needed a new career. He told me I needed to calm down and follow instructions if I wanted his help. I told him what he could do with his help. The boss gave me a long lecture on the relative number of flies caught with honey versus vinegar - however, that was the last SCO system we ever bought.
The bms.py application is the SPF checking milter (with lots of other features). The RedHat RPM provides a sysvinit start/stop/status script, defanged mail cleanup cron script, and puts log files in /var/log/milter (along with the config - to be fixed). If you are building your own package from source (as you would with FreeBSD), just run bms.py with python and add to sendmail.mc per the README. milter.cfg documentation is comments in the sample milter.cfg. Free free to email me directly from my address on the web page (stuart). I do not have FreeBSD, but will be glad to include BSD specific packaging/setup docs. There is an openbsd port - but I don't know if that applies to freebsd.
There is no startup cost because the milter process is persistent. Each connection is handled by its own thread. If your objection to Perl has to do with syntax, then I am with you. Otherwise, Perl has similar issues.
To summarize, let me call the part that securely multiplexes hardware the "kernel".
- Monolithic makes drivers share address space with the "kernel".
- With Microkernel, "kernel", drivers, filesystems, applications, etc each get their own address spaces.
- Exokernel makes drivers share address space with applications. (Hopefully, filesystems get their own process and address space.)
As you can see, as soon as you start partitioning applications into separate processes for security and robustness, the distinction between Exokernel and Microkernel becomes rather vague. The advantage of the Exokernel or VM approach is that you get the flexibility of keeping things like filesystems in a separate process for security and robustness, and things like video drivers in the same address space for performance. You might even have an X server as a separate process, but still allow full screen mode games that directly call the driver libraries for performance.IBM's VM was never that popular in its raw "Exokernel" mode with drivers in application space. However, it is still hugely popular as a way to run multiple Operating Systems as the "applications". Your mainframe can securely run multiple instances of S/390 Linux and traditional mainframe systems together.
The ultimate purpose of SPF is for all email senders to be authenticated so that there is no more envelope header forgery. The ?all solution is not disingenuous - it provides a way to implement as much of SPF as you can now without breaking anything. SPF provides flexible options, and a gradual phased adoption plan so that email providers can choose from a number of strategies and take as long as they need to implement the ultimate goal of sender authentication.
Of course. And that is why AOL does not currently check SPF. And why when they do, they will not reject mail solely on the basis of SPF.
Senders that use SPF + SES (Signed Envelope Sender) make it possible for a large ISP recipient to reject forged mail even without the cooperation of forwarders.
However, since AOL is designed for people that are general incompetent when it comes to computers, (because they have other interests), then AOL might wisely choose not to reject any mail based solely on SPF.
While foo.edu didn't do anything wrong, they also could make life easier for the student by implementing SRS for mail that they forward. I install my pysrs package for sendmail. Enabling SRS is a single line in sendmail.mc.
"~all", known as 'softfail' means "I'm probably spam; please filter me." It is intended for a transition period. "?all", known as 'neutral' is equivalent to not publishing SPF records at all. It allows you to validate mail sent from machines you control while saying nothing about machines you don't control. Whether you have no SPF record at all, or whether you send from a machine that hits the "?all", the result for a receiving MTA that checks SPF is:
If it ever gets to the point where a neutral result means "I'm probably spam", then SPF will have been a smashing success.
BTW, with regard to SMTP AUTH and/or SSL, if mainstream mail clients like Outlook or Mozilla or pine are not available where you need to send mail, then you might consider using web mail for sending validated email from remote locations.
In general, forwarders are selected by the email recipient and handling them correctly when implementing SPF is the responsibility of the email recipient. It is easy for the recipient to "punt" as a first step to implementing SPF and simply add Received-SPF headers without actually rejecting anything. The Received-SPF headers are remarkably effective at aiding content filters to do their job.
Without SPF, you have no idea who is responsible for the zombie machine (unless you are RIAA or the ISP). Furthermore, the vast majority of zombies aren't authorized by *any* domain - their owners are computer clueless AOL or Yahoo users.
If SPF were widely implemented, Mom and Pop (and their Zombie masters) would not be able to send mail directly from their Windoze box without first registering a domain and publishing an SPF record to authorize it (or asking their ISP to authorize their box).
This is not a problem with SPF. The easiest solution is to simply leave the default for your SPF record as "?all" - which says "I don't know" for any of the hotel sites you might send from. The better solution is to use SMTP AUTH or SMTP over SSL to relay your mail from the hotel through the home office.
SPF tells you if the envelope sender of the mail is forged. Nothing tells you if the mail as a whole is forged (except perhaps PGP/GPG/S-MIME if you are careful with your private key password). Someone could have sat down at your PC while you were at lunch and forgot to lock your screen.