I'm sorry, but you do not appear to know a damned thing about the RSAREF license. You were forced to use RSAREF under US Patent law, because that was the only way RSA would give you a license to their patent for free; that does not impact the fact that the RSAREF copyright license did not allow you to modify it. Yet it still benefitted from open peer review, because there was a security hole in RSAREF that was found via peer review.
One important factor of choosing GPL/BSD Operating Systems is that I don't have to pay for a software 'license' to VIEW the code in the first place, and then I don't have to pay for a license to USE the software.
I once again refer you to PGP and RSAREF, neither of which required you to pay for a license to view their code, thus not impairing auditing at all.
Your question is essentially (and probably unintentionally) FUD... the reason OpenBSD is good at what it does in the first place is becuase its *OPEN*BSD.
There are a lot of reasons OpenBSD is good; one is that it is completely unencumbered, so you can put it into an embedded security product. But then, you can put BSD4.4Lite2 in an embedded security product too, but I don't see people doing that. Another reason OpenBSD is good is that the development team does code audits themselves, rather than relying on external parties to bother with it; this process has found more security holes in a few years than third-party audits have revealed in decades. A third reason it is good because you can audit the code yourself.
Two of these benefits have clear impact on security, and two of these benefits do not require open source licensing to be true. Neither the presence of a team doing audits, nor the ability to audit yourself, require open source licensing.
Let me spell it out real nice and easy: the ability to do audits is one of the most important factors in making a piece of software secure. Open source licensing provides that ability, but it is not unique in providing that ability. It is sufficient but not necessary.
It's probably an inappropriate question only because it is too specific, imho.
Are you familiar with the old RSAREF license (relevant before the RSA patent expired)? That wasn't OSS. Taken a look at the PGP license? Again, not OSS. "OSS vs. closed source" is not an issue relevant to the security community.
Here's another question: would you prefer an open source derivative of Red Hat, or a closed source derivative of OpenBSD, as your preferred secure server system? What if you could see the OpenBSD code (particularly diffs), just weren't licensed to build and run it? Is the open-source nature of the product the only factor in deciding?
Look at Opera's history of competing with bundled browsers:
Opera never complained to Microsoft about bundling IE.
Opera never complained to Apple about bundling IE.
Opera even felt that going up against a bundled browser in an extremely marginal operating system (OS/2) was an acceptable money-making proposition.
Opera is even up for competing against free software browsers in Linux.
But, somehow, a beta product is going to cause Opera to leave the Mac market?
Please. The real reason is buried in the article:
"It's not a platform where we've earned a lot of money," said Tetzchner.
Opera simply wasn't successful on the Mac like it was on Windows. I personally suspect that that's because Opera didn't make the switch to the Mac UI very effectively (hard to describe, but it just felt weird relative to using it in Windows), and didn't have the same feel for performance programming in MacOS, making Opera feel sluggish (unlike its trademark lightning-fast performance in Windows).
I think Opera just can't compete on the Mac, knows this, and has made a token "hell why not" offer to Apple to stick around for big money to have an excuse to leave.
TiVo also has a model that controls AT&T's digital receivers over serial. Unfortunately, my digital cable (and maybe all) aren't completely digital yet, so I still have too many channels coming in analog for the completely integrated "store the digital stream on disk rather than digitize an analog image" style TiVo to work yet.
shrug On the system where Chimera text boxes pause on each character, OmniWeb finishes rendering at about the same time as Chimera (I tested the Opera beta too, just for reference, and it finished after both OmniWeb and Chimera). However, OmniWeb isn't as fast or good at progressive rendering.
I didn't test IE, because trying to handle loading a web page in four browsers at once was just too much- couldn't monitor more than three.
As for "displays better" - it's more than that. Good fonts by default, better layout, more closely-coupled rendering and displaying engines.
Still, I love my Chimera, and I rank IE second as it's the most mature, feature-complete, accurate browser on the mac today. Still, there's no tabs, no popup-blocking and no speed. So the honor of being my favorite browser goes to Chimera. I'm using one of the nightly builds, and it's fantastic.
I admit, I've never liked tabbed browsing. I've played with tabs some (even a tabbed window manager in X), just doesn't grab me. I don't really think it's as innovative and purely great as some people think (not to say that it isn't innovative, just that it's not that innovative).
OmniWeb trumps IE or Chimera on everything but rendering correctness and tabs; but it does a very credible approximation of correctness, and displays better than anything else I've seen.
How slow is your Mac? Or how the hell fast can you type? Because I type ~65-70 words a minute (fairly fast) and I'm typing this on Chimera 0.60 and I'm not seeing any "pauses" between characters.
500MHz G3. Is that not fast enough to run a frickin' web browser? I remember running (an admittedly much less capable version of) OmniWeb on a 25MHz NeXTStation. How much more processing power does Chimera require? Chimera pauses for each character entered, and it is definitely not a Cocoa text field.
The text field, where OS X actually innovated a lot - actually improved usability over other systems a lot - isn't native. Things like reasonably complete emacs bindings for cursor movement, interface to the spell checker, and so on are things I've come to expect in OmniWeb. Redrawing the entire text box every time I enter a character is not what I expect.
I might rank Chimera and OmniWeb neck and neck, but probably not. Netscape 7 is completely, unabashedly unusable, and doesn't belong in any list of browsers people use.
IE renders more correctly than OmniWeb, but the user experience is sufficiently inferior that I only use it if it correctly handles something OmniWeb fails at (that is, one site a week).
I would love for OmniWeb to render websites more correctly. But, frankly, it does a good enough job. Does that limit web developers? Sure. I feel for them. But expecting users to use crap like Netscape 7 is simply insulting. Expecting me to use stuff like Chimera, which offers a "Cocoa" interface with all non-Cocoa widets for interaction, is also insulting. I'm using OS X for a reason.
Here's a quick test: is posting to Slashdot pleasant in these other browsers? No. Only OmniWeb can spell check my text as I enter it, and only OmniWeb and IE have text entry that is reasonably fast. Why would anyone use a browser that obviously pauses for each character I enter into a text field?
You seem to be saying that as employers don't offer you a guaranteed job, you don't offer them guaranteed service.
Only partially.
Absolutely.. however.. will you work for an employer who has a record of laying off his workers every 12 months or so?
No. But I think the parallel between "laid off workers" and "stopped working there" is a bad one. After all, he didn't say "quit their job every 12 months". What about a company that's extremely upwardly mobile, with people not holding the same positions for long? (that's quite a bit different from annual layoffs) What about an employee who went from contractor to employee for a vendor to employee for a customer?
If the work is obvious contract work, that's not what he is talking about
Excuse me if I take his words at face value, and ignore your attempts to interpret it. He made a pretty clear statement that didn't provide for "if it was contract work."
he's talking about people (and I'm sure we all know people like this) who work for a job for around a year, and then look for the bigger, better deal.
No, he's not, because that can't be measured with the metric he mentioned. Maybe that's what he's trying to winnow out, but that's not all he's cutting out.
Why did you leave your last job? You will be asked this in an interview. If your main reasons seem to be "more money" then it's obvious that as soon as someone comes along offering a bit more cash, you will flee.
Not necessarily. As an example, I left my last job because as much as I loved it, they weren't prepared to offer me enough to support both me and my fiance moving into town. When I get a kid, I probably won't be comfortable working where I am now, much as I love it. Then, and in the future, I will be leaving for "more money," but I'm not a fly-by-night programmer looking for my next buck. In fact, I pick jobs (and work environments) well enough that I expect my main reason for leaving, over time, will be either "they went under" (won't happen at Carnegie Mellon), or "I needed more money."
My point being, it's a complicated issue that requires consideration of a number of factors and information coming from a number of sources (former employers, the prospective employee, references...). However, a recruiter sees enough applications that they can't give all this information due consideration for every applicant- they have to have simple, moderately effective, metrics for throwing out 90% of the resumes with very little thought.
So while I'm arguing that this fella's metric isn't perfect, I'd probably do something similar.
Another thing I want is some longevity. If I'm hiring for a full-time position, I view it as a purchase as opposed to a rental.
Hmmm. I assume longevity, then, is something you as a manager can offer your employees, too? Can they feel confident that there won't be layoffs, that the company won't collapse, that there won't be huge accounting scandals that ruin their retirement?
What assurances do I have that such people are not simply hopping from one contract to the next, leaving behind a trail of destruction?
I dunno. Why not ask the applicant? Maybe former employers too. Maybe you can find out whether they did well in those jobs.
Maybe he's hopping from contract to contract, leaving a trail of happy employers- maybe he's worked for a lot of high-risk companies that went under or had to lay him off, and he's looking for something now that's a bit more stable (if you purport to offer it).
I don't think your metric has much value, but I'll be honest- with a big stack of resumes, some winnowing technique must be used. It's hard to argue with one that's worked, and isn't outright unethical.
Unions only increase costs, decrease productivity, and guarantee that the industry will need a government bail-out or protection in 20-30 years.
That's not all they do. They also help deal with employee exploitation, and serve as a collective power to balance out the collective power of a monopoly employer.
They are an answer to a particular problem, namely an unfree job market. If employees can find different jobs, then employers will see value in offering better working conditions. If employers can hire different people, employees will see value in picking what they're good at as their job.
Some people think that "legacy stretching back further than TeX" is not a "win".
These same people might question why you are suggesting that something "used for man page formatting on tty devices" should be used for printed documentation.
Of course, such people clearly don't know the power of groff (for those people: try printing man pages out, and you'll see that it handles paper copy very well), but you're not exactly selling it.
And really I would recommend (La)TeX because I think it's more usable (closer to "what you see is what you mean" than roff); but if someone doesn't want to use TeX, I would recommend looking to modern-day roff users for information on using it: Plan 9.
Granted this is a sort of grey area since some people would argue that you're paying for.Mac instead of for the backup software itself but the backup software is designed to be used with the.Mac service so using it without paying Apple is akin to piracy.
Speaking as someone who shelled out money for.Mac, I don't want my data going on Apple's servers... so this helps me without reducing the money Apple gets. Piracy? Or Fair use?
Of course, I've long been hoping that they'd roll out ".Mac for Corporations", or something, to push OS X Server. Sell you the software they use in.Mac (IMAP, DAV, web publishing, the whole bit), and let you provide the.Mac service to employees/clients/whatever. If it were per-seat, but cheaper than.Mac, they'd make money hand over fist because it was cheaper for end users and cheaper for Apple (and we IT guys would be able to eat the hardware cost pretty easily, too, because we wouldn't have as many clients as Apple).
Cyrus's mailstore system is actually quite different from Maildir, in particular because it doesn't need to play games with user processes (the way read/unread messages are handled in Maildir is handled that way so multiple processes can manipulate messages at the same time, for instance).
Also, most of the abilities you list are simply unavailable via POP; Cyrus is massive overkill for a POP server, and would require even more resources (particularly disk: the users that have 40MB spool files now could probably find themselves with 2GB of mail if you let them... and even the non-abusive users would require more storage for IMAP than for POP).
Incidentally, we use qpopper to handle POP - and quite a few users go over 40MB without killing our (not particularly beefy, and not dedicated mail) servers. I suspect the real problem is that the guy is using uw-imap's POP server - the author of which is notoriously unconcerned with the performance (or lack thereof) of spoolfiles being served over POP. Which is perfectly reasonable - he writes an IMAP server, he should be concerned with IMAP performance, and if he writes a better mailbox format (he has) then he should also concern himself with that and not a 20+ year old format.
Actually, if one were so inclined, IMAP makes a better POP than POP3 - just disable the ability to create new folders, and use a better mailbox format (mbx, Maildir,...).
In several of the flights, actually. When they stopped caring about giving the pilots problems, the cell phones seemed to work OK. I also suspect, but don't know for sure, that at the times the cell phones were being used the planes were a bit lower than normal.
By all-in-one I mean I want a Digital Camera/Cell Phone/Pager/mp3 player/PDA with wireless networking all in one no bigger than palm-sized package. Yes, I know it will cost a lot of money, but I don't see it as an impossibility.
I see. So you want to trust some software vendor to not introduce bugs in trying to get all that crap to work together? You're not worried about the extent to which they'll have to integrate it all, so that if one tiny little problem crops up your only real option is to buy a new one?
I have a friend with a Kyocera 6035 phone/PDA - just a phone and PDA, no camera, no MP3 player, no wireless networking (other than dial up) - and oops, the digitizer has issues. Now he can't use it as a PDA, the screen is too custom for him to find replacement parts for... he just has to wait until the next model comes out, and his model gets cheap enough to buy a second one.
And now you want them to integrate a bunch of other stuff that most vendors still get wrong all by itself?
No thanks. I've become a believer in Bluetooth, myself. I don't need to have my phone be part of my PDA, I certainly don't need something as self-contained as an mp3 player to be attached to a number pad or a touch-sensitive screen. Only the camera might need to be attached to the phone, certainly nothing else.
because the file doesn't get removed until Mac OS 9 has had a clean shut down.
Do I have to go through this again?
OS X has a dirty shutdown.
OS 9 will pop up Disk First Aid at next OS 9 boot, UNLESS
I run fsck manually in OS X.
Now, if OS 9 and OS X have different methods for marking a partition dirty, why would a dirty shutdown in OS X affect OS 9? Why would fsck - which is the "OS X" way of dealing with dirty filesystems - also affect OS 9's opinion of the dirtiness of the filesystem?
OS X can make OS 9 think that the filesystem is dirty, and that the filesystem is clean. I don't see how this in any way suggests that OS X uses a different method than OS 9 for determining whether the filesystem is clean. This leads to the conclusion that there is no boot-time fsck in OS X.
Truth is, looking at the man page for autodiskmount, which uses hfs.util to handle fscking, it looks like things got a little complicated and somewhere along the line, fscking stopped happening.
Green? (also, given all that, I would recommend running fsck after a dirty shutdown yourself and seeing whether fsck thinks OS X already fixed all the problems:-)
OSX uses a different method of detecting an improper shutdown to OS9.
Really? Think carefully before you answer. Why does running fsck in OS X convince OS 9 that the filesystem is not dirty, then?
when OSX shuts down it doesn't make the drive clean from OS9's perspective.
Really? Again, think carefully.
When a filesystem is mounted read-write, it's also tagged as dirty. Part of cleanly unmounting it is tagging it clean. So if they really use different methods of tagging it dirty, then OS 9 would not be aware of improper shutdowns in OS X, and booting into OS 9 after an improper shutdown would not call up Disk First Aid. But my whole problem is booting into OS 9 after an improper shutdown in OS X (if I haven't run fsck since that improper shutdown).
Soooo, if you're in OS9, set your startup disk to OSX and force a reboot, OSX will probably boot fine, no fsck etc, and when you restart to 9, you'll get the Disk First Aid.
Not if I run fsck on the filesystem while I'm in OS X.
The only conclusion I can draw is that OS X does not, in fact, run fsck on filesystems after a dirty shutdown.
Simple questions: if what you say is true, then why doesn't the automatic fsck that you say happens at boot remove this file? Why does running fsck manually report that the filesystem was modified, if it already ran when it booted? Why does running fsck guarantee that I won't have to check the filesystem in OS 9 (which is orders of magnitude slower)?
OSX fsck's the file systems itself while it boots up.
I don't think so. Why don't I think so? Because when I boot into OS 9, it thinks the filesystem is dirty if I boot it without running fsck in OS X before booting into OS 9. Why does OS 9 think the partition needs checking if OS X already ran a fsck?
I've actually been quite unhappy with OS X's blase attitude toward fsck - every time I boot into OS 9, I run fsck first "just in case." That's pretty lame (fortunately, I boot into OS 9 about once a month).
I know someone who switched from OS X to XP and liked it. It wasn't any particular feature they liked in XP or disliked in OS X, it was just what he felt was the incongruous mix of Mac and Unix, and what he felt was the excessively raw nature of OS X (he did use 10.1, but not 10.2 - but then, I've not used 10.2).
I can't say I understand his position, much less agree, but there it is. He's a a bit of a Unix geek and does most of his work on Macs (as dictated to him - he's a grad student), so it's not like he was planning on liking it.
OTOH, his TiBook was stolen, and his current laptop is in much less danger of being the target of thievery - it's just so danged ugly. Maybe that's what he prefers?:-P
Slashdot Mirror
I'm sorry, but you do not appear to know a damned thing about the RSAREF license. You were forced to use RSAREF under US Patent law, because that was the only way RSA would give you a license to their patent for free; that does not impact the fact that the RSAREF copyright license did not allow you to modify it. Yet it still benefitted from open peer review, because there was a security hole in RSAREF that was found via peer review.
I once again refer you to PGP and RSAREF, neither of which required you to pay for a license to view their code, thus not impairing auditing at all.
There are a lot of reasons OpenBSD is good; one is that it is completely unencumbered, so you can put it into an embedded security product. But then, you can put BSD4.4Lite2 in an embedded security product too, but I don't see people doing that. Another reason OpenBSD is good is that the development team does code audits themselves, rather than relying on external parties to bother with it; this process has found more security holes in a few years than third-party audits have revealed in decades. A third reason it is good because you can audit the code yourself.
Two of these benefits have clear impact on security, and two of these benefits do not require open source licensing to be true. Neither the presence of a team doing audits, nor the ability to audit yourself, require open source licensing.
Let me spell it out real nice and easy: the ability to do audits is one of the most important factors in making a piece of software secure. Open source licensing provides that ability, but it is not unique in providing that ability. It is sufficient but not necessary.
Are you familiar with the old RSAREF license (relevant before the RSA patent expired)? That wasn't OSS. Taken a look at the PGP license? Again, not OSS. "OSS vs. closed source" is not an issue relevant to the security community.
Here's another question: would you prefer an open source derivative of Red Hat, or a closed source derivative of OpenBSD, as your preferred secure server system? What if you could see the OpenBSD code (particularly diffs), just weren't licensed to build and run it? Is the open-source nature of the product the only factor in deciding?
Look at Opera's history of competing with bundled browsers:
But, somehow, a beta product is going to cause Opera to leave the Mac market?
Please. The real reason is buried in the article:
Opera simply wasn't successful on the Mac like it was on Windows. I personally suspect that that's because Opera didn't make the switch to the Mac UI very effectively (hard to describe, but it just felt weird relative to using it in Windows), and didn't have the same feel for performance programming in MacOS, making Opera feel sluggish (unlike its trademark lightning-fast performance in Windows).I think Opera just can't compete on the Mac, knows this, and has made a token "hell why not" offer to Apple to stick around for big money to have an excuse to leave.
Heh. Take a look.
don't geeks go in for, ah, technical details?
Where was the technical detail in saying "Darwin's VM system has to take into account different memory usage patterns"?
(I enjoyed the article, I guess, but "geeky"?)
TiVo also has a model that controls AT&T's digital receivers over serial. Unfortunately, my digital cable (and maybe all) aren't completely digital yet, so I still have too many channels coming in analog for the completely integrated "store the digital stream on disk rather than digitize an analog image" style TiVo to work yet.
shrug On the system where Chimera text boxes pause on each character, OmniWeb finishes rendering at about the same time as Chimera (I tested the Opera beta too, just for reference, and it finished after both OmniWeb and Chimera). However, OmniWeb isn't as fast or good at progressive rendering.
I didn't test IE, because trying to handle loading a web page in four browsers at once was just too much- couldn't monitor more than three.
As for "displays better" - it's more than that. Good fonts by default, better layout, more closely-coupled rendering and displaying engines.
Erik K. Veland wrote:
I admit, I've never liked tabbed browsing. I've played with tabs some (even a tabbed window manager in X), just doesn't grab me. I don't really think it's as innovative and purely great as some people think (not to say that it isn't innovative, just that it's not that innovative).
OmniWeb trumps IE or Chimera on everything but rendering correctness and tabs; but it does a very credible approximation of correctness, and displays better than anything else I've seen.
In short, have you tried OmniWeb yet? :-P
TiMac wrote:
500MHz G3. Is that not fast enough to run a frickin' web browser? I remember running (an admittedly much less capable version of) OmniWeb on a 25MHz NeXTStation. How much more processing power does Chimera require? Chimera pauses for each character entered, and it is definitely not a Cocoa text field.
The text field, where OS X actually innovated a lot - actually improved usability over other systems a lot - isn't native. Things like reasonably complete emacs bindings for cursor movement, interface to the spell checker, and so on are things I've come to expect in OmniWeb. Redrawing the entire text box every time I enter a character is not what I expect.
Spoken like someone who hasn't used any of them.
I might rank Chimera and OmniWeb neck and neck, but probably not. Netscape 7 is completely, unabashedly unusable, and doesn't belong in any list of browsers people use.
IE renders more correctly than OmniWeb, but the user experience is sufficiently inferior that I only use it if it correctly handles something OmniWeb fails at (that is, one site a week).
I would love for OmniWeb to render websites more correctly. But, frankly, it does a good enough job. Does that limit web developers? Sure. I feel for them. But expecting users to use crap like Netscape 7 is simply insulting. Expecting me to use stuff like Chimera, which offers a "Cocoa" interface with all non-Cocoa widets for interaction, is also insulting. I'm using OS X for a reason.
Here's a quick test: is posting to Slashdot pleasant in these other browsers? No. Only OmniWeb can spell check my text as I enter it, and only OmniWeb and IE have text entry that is reasonably fast. Why would anyone use a browser that obviously pauses for each character I enter into a text field?
mindstrm wrote:
Only partially.
No. But I think the parallel between "laid off workers" and "stopped working there" is a bad one. After all, he didn't say "quit their job every 12 months". What about a company that's extremely upwardly mobile, with people not holding the same positions for long? (that's quite a bit different from annual layoffs) What about an employee who went from contractor to employee for a vendor to employee for a customer?
Excuse me if I take his words at face value, and ignore your attempts to interpret it. He made a pretty clear statement that didn't provide for "if it was contract work."
No, he's not, because that can't be measured with the metric he mentioned. Maybe that's what he's trying to winnow out, but that's not all he's cutting out.
Not necessarily. As an example, I left my last job because as much as I loved it, they weren't prepared to offer me enough to support both me and my fiance moving into town. When I get a kid, I probably won't be comfortable working where I am now, much as I love it. Then, and in the future, I will be leaving for "more money," but I'm not a fly-by-night programmer looking for my next buck. In fact, I pick jobs (and work environments) well enough that I expect my main reason for leaving, over time, will be either "they went under" (won't happen at Carnegie Mellon), or "I needed more money."
My point being, it's a complicated issue that requires consideration of a number of factors and information coming from a number of sources (former employers, the prospective employee, references...). However, a recruiter sees enough applications that they can't give all this information due consideration for every applicant- they have to have simple, moderately effective, metrics for throwing out 90% of the resumes with very little thought.
So while I'm arguing that this fella's metric isn't perfect, I'd probably do something similar.
dcavanaugh wrote:
Hmmm. I assume longevity, then, is something you as a manager can offer your employees, too? Can they feel confident that there won't be layoffs, that the company won't collapse, that there won't be huge accounting scandals that ruin their retirement?
I dunno. Why not ask the applicant? Maybe former employers too. Maybe you can find out whether they did well in those jobs.
Maybe he's hopping from contract to contract, leaving a trail of happy employers- maybe he's worked for a lot of high-risk companies that went under or had to lay him off, and he's looking for something now that's a bit more stable (if you purport to offer it).
I don't think your metric has much value, but I'll be honest- with a big stack of resumes, some winnowing technique must be used. It's hard to argue with one that's worked, and isn't outright unethical.
zerofoo wrote:
That's not all they do. They also help deal with employee exploitation, and serve as a collective power to balance out the collective power of a monopoly employer.
They are an answer to a particular problem, namely an unfree job market. If employees can find different jobs, then employers will see value in offering better working conditions. If employers can hire different people, employees will see value in picking what they're good at as their job.
Some people think that "legacy stretching back further than TeX" is not a "win".
These same people might question why you are suggesting that something "used for man page formatting on tty devices" should be used for printed documentation.
Of course, such people clearly don't know the power of groff (for those people: try printing man pages out, and you'll see that it handles paper copy very well), but you're not exactly selling it.
And really I would recommend (La)TeX because I think it's more usable (closer to "what you see is what you mean" than roff); but if someone doesn't want to use TeX, I would recommend looking to modern-day roff users for information on using it: Plan 9.
Speaking as someone who shelled out money for .Mac, I don't want my data going on Apple's servers... so this helps me without reducing the money Apple gets. Piracy? Or Fair use?
Of course, I've long been hoping that they'd roll out ".Mac for Corporations", or something, to push OS X Server. Sell you the software they use in .Mac (IMAP, DAV, web publishing, the whole bit), and let you provide the .Mac service to employees/clients/whatever. If it were per-seat, but cheaper than .Mac, they'd make money hand over fist because it was cheaper for end users and cheaper for Apple (and we IT guys would be able to eat the hardware cost pretty easily, too, because we wouldn't have as many clients as Apple).
Mostly right, in a very broad non-technical way.
Cyrus's mailstore system is actually quite different from Maildir, in particular because it doesn't need to play games with user processes (the way read/unread messages are handled in Maildir is handled that way so multiple processes can manipulate messages at the same time, for instance).
Also, most of the abilities you list are simply unavailable via POP; Cyrus is massive overkill for a POP server, and would require even more resources (particularly disk: the users that have 40MB spool files now could probably find themselves with 2GB of mail if you let them... and even the non-abusive users would require more storage for IMAP than for POP).
Incidentally, we use qpopper to handle POP - and quite a few users go over 40MB without killing our (not particularly beefy, and not dedicated mail) servers. I suspect the real problem is that the guy is using uw-imap's POP server - the author of which is notoriously unconcerned with the performance (or lack thereof) of spoolfiles being served over POP. Which is perfectly reasonable - he writes an IMAP server, he should be concerned with IMAP performance, and if he writes a better mailbox format (he has) then he should also concern himself with that and not a 20+ year old format.
Actually, if one were so inclined, IMAP makes a better POP than POP3 - just disable the ability to create new folders, and use a better mailbox format (mbx, Maildir, ...).
I suppose you could call any software that attempts to factor large numbers badly written, on the principle that trying to do it is stupid, but...
The kinds of problems being attempted 20 years ago were much easier than the ones being done now, much less than the ones we still can't do.
In several of the flights, actually. When they stopped caring about giving the pilots problems, the cell phones seemed to work OK. I also suspect, but don't know for sure, that at the times the cell phones were being used the planes were a bit lower than normal.
Blockquoth Apreche:
I see. So you want to trust some software vendor to not introduce bugs in trying to get all that crap to work together? You're not worried about the extent to which they'll have to integrate it all, so that if one tiny little problem crops up your only real option is to buy a new one?
I have a friend with a Kyocera 6035 phone/PDA - just a phone and PDA, no camera, no MP3 player, no wireless networking (other than dial up) - and oops, the digitizer has issues. Now he can't use it as a PDA, the screen is too custom for him to find replacement parts for... he just has to wait until the next model comes out, and his model gets cheap enough to buy a second one.
And now you want them to integrate a bunch of other stuff that most vendors still get wrong all by itself?
No thanks. I've become a believer in Bluetooth, myself. I don't need to have my phone be part of my PDA, I certainly don't need something as self-contained as an mp3 player to be attached to a number pad or a touch-sensitive screen. Only the camera might need to be attached to the phone, certainly nothing else.
Blockquoth Hes Nikke:
Do I have to go through this again?
- OS X has a dirty shutdown.
- OS 9 will pop up Disk First Aid at next OS 9 boot, UNLESS
- I run fsck manually in OS X.
Now, if OS 9 and OS X have different methods for marking a partition dirty, why would a dirty shutdown in OS X affect OS 9? Why would fsck - which is the "OS X" way of dealing with dirty filesystems - also affect OS 9's opinion of the dirtiness of the filesystem?OS X can make OS 9 think that the filesystem is dirty, and that the filesystem is clean. I don't see how this in any way suggests that OS X uses a different method than OS 9 for determining whether the filesystem is clean. This leads to the conclusion that there is no boot-time fsck in OS X.
Truth is, looking at the man page for autodiskmount, which uses hfs.util to handle fscking, it looks like things got a little complicated and somewhere along the line, fscking stopped happening.
Green? (also, given all that, I would recommend running fsck after a dirty shutdown yourself and seeing whether fsck thinks OS X already fixed all the problems :-)
Blockquoth stux:
Really? Think carefully before you answer. Why does running fsck in OS X convince OS 9 that the filesystem is not dirty, then?
Really? Again, think carefully.
When a filesystem is mounted read-write, it's also tagged as dirty. Part of cleanly unmounting it is tagging it clean. So if they really use different methods of tagging it dirty, then OS 9 would not be aware of improper shutdowns in OS X, and booting into OS 9 after an improper shutdown would not call up Disk First Aid. But my whole problem is booting into OS 9 after an improper shutdown in OS X (if I haven't run fsck since that improper shutdown).
Not if I run fsck on the filesystem while I'm in OS X.
The only conclusion I can draw is that OS X does not, in fact, run fsck on filesystems after a dirty shutdown.
Simple questions: if what you say is true, then why doesn't the automatic fsck that you say happens at boot remove this file? Why does running fsck manually report that the filesystem was modified, if it already ran when it booted? Why does running fsck guarantee that I won't have to check the filesystem in OS 9 (which is orders of magnitude slower)?
Blockquoth gabe:
I don't think so. Why don't I think so? Because when I boot into OS 9, it thinks the filesystem is dirty if I boot it without running fsck in OS X before booting into OS 9. Why does OS 9 think the partition needs checking if OS X already ran a fsck?
I've actually been quite unhappy with OS X's blase attitude toward fsck - every time I boot into OS 9, I run fsck first "just in case." That's pretty lame (fortunately, I boot into OS 9 about once a month).
I know someone who switched from OS X to XP and liked it. It wasn't any particular feature they liked in XP or disliked in OS X, it was just what he felt was the incongruous mix of Mac and Unix, and what he felt was the excessively raw nature of OS X (he did use 10.1, but not 10.2 - but then, I've not used 10.2).
I can't say I understand his position, much less agree, but there it is. He's a a bit of a Unix geek and does most of his work on Macs (as dictated to him - he's a grad student), so it's not like he was planning on liking it.
OTOH, his TiBook was stolen, and his current laptop is in much less danger of being the target of thievery - it's just so danged ugly. Maybe that's what he prefers? :-P