"but I'm sure I'm not alone in saying that I had never even heard of SpamAssassin until just now via/. "
You must be new around here or filtering out interesting articles... It's been around for ages, I've been using it now myself for nearly 18 months or more in varying guises.
I guess the latter would make it rather hard for them to force us to change the name for the thing, too.
I don't think I've ever worked in a company that uses only one language for everything, anyway. More to the point, none of the developers I've ever met have been one-horse wonders, either. And no surprise: to be a developer is to know how to *program*, which is a language-independent ability. We use all combinations of C, java, PHP, Tcl, and C# - and I was mildly pleased when one of the developers said he understood the point of s-expressions too, to say nothing of me being the in-office Ruby and Perl guy.
And never mind the linguistic inaccuracies (why didn't the author check each occurrence of `scheme' by hand?!) but what about jobs that are really vapourware to capture more CVs than the competing agencies? Don't they target the big names (VB, VC) more than anything else?
I'm tempted to agree, although in my case I watched it most when Maggie Philbin was on the team, ie when I was about 8-15 or something. Still got the impression it was going downhill; it served its purpose in getting one small eejit hooked on things scientific, and then it all got dumbed-down. Bah.
It's not as though _Horizon_ is any temptation either - the reason we adult types have given up watching it is because of the dumbed-down drama - it even has "drama" in its caption. We don't need drama. We don't need the historical perspective, although it can be one means to present an achievement. I said it at Uni, and I'll say it again: I don't give a constipated monkey whether Einstein *liked* his results, *show* me them and I'll work it out for myself.
I'm not surprised people complained. I *keep on* saying it, but shifting the workload that spammers should be doing onto each and every innocent mailer is morally unjustifiable. It's not as though TMDA is any different and it's been around a long while.
And this is to say nothing of the dangers of sending a spam as though it were from one half of a mail2news gateway, of course. Or any other impersonated-sender scenario...
Choose spamassassin at SMTP injection point. Choose exim-4.10+exiscan. Reject mails with too high a score before they get anywhere near you. Use trap addresses and block them in the To: header, auto-reporting them in Envelope-To:. Whatever. Just don't multiply spams potentially to innocent parties!
a) RPM already has at least two ways of being upgraded dynamically - urpmi and apt-get. It just needs a consistent well-maintained high-quality upstream pool-set
b) Debian supports RPM packages just fine.
c) The standards (specifically, the LSB) say nothing about requiring RPM to be the system's native package-managing system.
d) Debian already strive for LSB-compliancy, at least where it makes sense. This is why we've had/etc/init.d/ since the get-go while RH have been messing around with this "/etc/rc.d" abomination which then needs legacy support on the assumption that there are idiots out there who can't cope with RH correcting their previous mistakes.
"the system directory on Windows 2000 is \WINNT"
Well, only *if* you install it that way.
And one for thought: which is more important, adhering to a standard for the sake of it, or knowing what you're doing? (A specific example of the latter: given an IP#, I expect you to be able to trace through DNAT, netstat -p or similar and through/proc, to tell me where in the filesystem the httpd is located that's responsible for a given webserver. If you can't debug that, you ain't gettin' root on my boxes.)
Nor to me, for a different reason: what about those of us with CFLAGS= set to various strange funky optimizations in Gentoo? What about the Ports system in FreeBSD, similarly?
This thing does not have the potential to spread to all distributions or all unixen.
What about historical storage? Are they really proposing to store an md5sum for/bin/*/usr/bin/* for all packages for all distributions for all releases, or when do older things get purged?
Seems mad to me. Would be better off staying with AIDE instead, IMO.
"So let's return to the *original* definition of spam"
What the one that applies only to Usenet? That's fine by me; the real problem with email is UBE anyway - *not* just UCE, although that's pretty foul crap anyway, but anything unsolicited and sent in bulk (read: not targetted at you specifically). This includes crazy charity/religious-headcase mails as well.
It also means that a mailshot from a company to "just say hi" to someone is still undesirable.
Besides which, it's still the ultimate in cheek to assume that someone's email connection is free - it's *NOT*, dammit, so any senders-of-UCE are stealing resources in order to try and sell you something. I'm expecting to change bank *again* for exactly this reason.
I was going to post my.procmailrc entry here but slashdot censored it saying "too many junk characters". So you'll have to work it out for yourself based on Content-Type headers matching
More to the point, you never want to filter based on Received: headers, unless you can safely say that e.g. *no-one* in Korea is ever going to want to contact you. Otherwise, grab the IP# listings from IANA and see what netblocks are assigned to APNIC and score them down in your mail processing rules.
For what it's worth, I've had a lot of success using Bayesian filters to identify dodgy-charset mails - both ifile and bogofilter do a great job.
Sending emails back to spammers is for brainless cretins - it serves only to clutter up your mail queue and risks offending innocent impersonated senders or having your email address confirmed as valid for spam.
And sending automated emails back to legitimate senders is downright *immoral* - making everyone do the work that a spammer *should* be doing to get through to you is indefensible.
And I've seen a case recently where this TMDA thing was so misconfigured that it sent an mail back to a mailing list saying there was an unrecognized sender address, and of course that mailing list was half of the gnu.emacs.help mail2news gateway, so the message appeared on the newsgroup for *all* to see. Talk about efficiently multiplying spam.
Now for something useful. Use one of the Bayesian filters, seeing as they're all the rage and get about 97-98% spam matched correctly, coupled with SpamAssassin as a fall-back for the remaining 2% cases, and you'll have far less of a problem. Now incorporate those filters in your MTA so that the whole body is checked for spammishness before being "accepted for delivery" and you'll have the best solution of them all: bounce the mail at injection-point and be done.
South London (Croydon) was mostly clouded over especially in the south and just above the northern horizon. However, I saw two between 0245 and 0310 this morning before going back to bed.
Why do these things have to happen on a week-day?!;)
Here's an idea. How about seeing it, not as "censoring" (which is generally bad, except that the word can have a fairly wide definition), but as choosing to separate folks according to what suits them best?
I think the main problem is that Mum and Dad are still involved in making sure kids stay within that domain - what's to stop some brat typing `slashdot.org' into the browser location bar?!
OTOH, there's no real problem with saying "this helps make a clear area appropriate for a given approximate age-range", and expecting folks to make contacts and get used to clicking around in there, then migrating out into the real world. In fact, if it means said real world can be left intact with fewer stupid rules then all to the good.
True, although most people would consider it nice to have a picture-format as well as a text-delivery system all openly and commonly available, so...
While we're passing by, I saw `DjVU' touted as a possible successor to JPEG; however, that also has explicit "nice" (pro-open-source) patenting surrounding it, which I still think is a dodgy idea.
So how *do* I get pictures in web-pages? Is it only PNG now and stuff the older browsers?
That doesn't have to happen. All the spammers need to do is carry on pushing crap out and someone's bound to get it - drop a block of varied nice-words in the bottom of the mail (start "talking about" PHP, java, diet-coke and stuff, and you'll be going up in my estimation) and the damage potential will increase. And in the process, more bandwidth will be consumed (which is where the real work-load is inflicted on the 'Net, not and the end-recipient) and the poor eejits out there *not* using these filters will still be getting crap.
That's also not the point. It might not cost them 10x as much to send 10x as much spam if that becomes the case. Not to mention, someone's bound to *try* circumventing any sufficiently common anti-spam system.
(And Bayesian filters are no panacea, nor are they invulnerable, btw; I still get the occasional junk-mail through here, despite my better efforts. I even had one over the weekend that fell back to SpamAssassin and scored over 27 but bogofilter considered clean - simply base64-encoded all its body, and I don't know that bogofilter either handles that well, or if it does, it hadn't been well enough trained against that sort of mail.)
The best way is still a moderately aggressive "just say NO!", not a passive acceptance, IMO. Spam should always be nuked at source.
Unfortunately, allowing the clueful few who wouldn't buy or respond to spam anyway (meaning mozilla users) to filter mail easier does nothing in the way of telling spammers to get knotted. That's where integrating the spam-check at the MTA receipt stage would come in handy; you don't have to passively accept crap, nor do you have to waste bandwidth sending complete bounce emails (potentially to innocent victims), but rather you drop things at injection. Fine by me:)
There's a patch and spiel about an exim module and filter that hooks into SpamAssassin at http://marc.merlins.org/linux/exim/sa.html
Agreed, although some say this apparently gets harder with increasing antiquity;)
I would say there's another thing to be on the look-out for: abstraction. Someone who "learns Pascal" or "learns Fortran" or "learns C" might be said to be missing out on the general "knows how to program" aspect, which is far more employable. Likewise "knows web-design" is missing out on the fun of the open w3c standards and tagged-up text-processing tools that go with XML.
Ask yourself what aspects of IT will be around in 10yrs' time. Windoze? No way. Linux? Probably not in any form similar to present. Cisco? Likewise, mutation expected. It's not worthwhile investing time and money for certification in any proprietary single technology. I'd say two key things that'll still be around are an understanding of SECURITY and of ETHICS. Read up on those - get to know security for what it is, and how best to apply it to a couple of current OSs, and work through the moral mazes of a few ethical questions, and you'll set yourself up for a longer career.
[catty comment] It's Harry Potter - it's unwatchable anyway!;8)
"Everyone assumes Valenti and Rosen are right: that piracy is damaging the film and music businesses."
Agreed. I don't like seeing "illegal" copies, or premature leaks of media, but when it comes to previewing one or two tracks off an album in advance of buying the CD, I'm not going to waste half my lunchtime walking to $supermarket and listening to a few select noises when I can download someone else's copy of those tracks - and then I'll decide whether to go out and buy it or not (and delete the crap as need be).
This makes it apparent that there are at least two factors in the "does it cost us money?" equation - those folks who are freeloaders and will never buy a CD of even a proportion of the tracks they have, versus those who'll use it as a means of exploration and actively add more to the companies' profits.
I see it as simplistic to take either position - "it loses us money" or "it adds to the revenue!"; there's a mixture of both, and I've yet to see a calculation of the ratio of both phenomena.
While I'm passing, I'm of the opinion that milking that extra potential profit-margin (the freeloaders - who also exist in non-Internet-related realms) is pretty much the definition of greed.
It also seems increasingly apparent to me that the movie (and record) companies are simply trying to enforce an old marketing model - "you must get your crap here", rather than embracing an opportunity to expand to new horizons (FFS - think of the patent capabilities of helping invent a new distribution mechanism!). That's positively asking for rip-offs, IMO.
I *want* a "PC" - whatever the heck that might mean these days - but most importantly, I want Architecture. I'm sick of buying Intel, and I don't want to buy into the intel concept by getting AMD instead, either. x86 can go whistle, I want ppc or alpha. And I want to run NetBSD on it, too, as forcing linux to run on various bits of kit (amiga A500+, anyone? Psion 5MX?) has lost its appeal with me. I have intel with Gentoo, FreeBSD and OpenBSD atm; need to complete the set!
So who else do we know who does G4-800 chips? One answer: Apple. And how much do they cost? about 2x-4x as much as this new amiga effort. So by sacrificing apple's proprietary mobo and peripherals, I fulfil my desire for "Architecture" and slice the cost right down. That's looking pretty peachy to me.
Now, the really worrying and annoying thing is that clause about needing some "enabler" to get OS4 to run on it. I read it like this: they want me to pay for something (always dodgy) that helps them fight me. Erm, yeah, right, time to open-source it and sell CDs with added manuals or other value for $20. *Then* I'd be interested in the OS.
"when I set up servers that do a lot of logging (firewall, web, ids, etc...) and I give a big/var partition (10+ GB) how little is filled up after several months."
Some sites' webserver logfiles are of the order a gig a day. Fortunately, apache CLF is well compressible, so an archival CD only needs burnt every 3-4 days.
Some sites' traffic in reporting a handful of servers' firewall hits back to a loghost contribute 4K/s of syslog traffic alone.
Do we have any stats for our great esteemed slashdot to compare?
"I just want to USE my computer no have to worry about all the moving parts under the hood."
So how do you distinguish between bits you "use" and are obviously therefore the only worthwhile bits for the rest of the world, and bits that we happen to like to use, or to choose between?
And why do you install tcsh if you're not going to use it? Do you have no idea what you're doing when you install a machine, and somehow that's the upstream distribution's fault for providing things that you have to choose to use or not, and you just can't be arsed making a choice yourself?
Stay with Windows. Please. For the sake of the kittens...
Dual feedback loops. Every mail that matches spam gets fed back into the system so both the is-spam wordlist AND the is-good wordlists become more "concentrated" over time. Ifile does this, bogofilter does this with some wangling in procmail,...
That way, if someone sends something that's still mostly spam (one or two words in common with spam, enough to tip the balance) then all the neutral words will be tarnished as well.
Re:How dense can you get?
on
One of Many
·
· Score: 2
Right, yeah, it's not hard to convert the g/cm^3 thing into proper units - but my point is that 1cm^3 is easily visualized, and it weighing half a bag of sugar is totally easy to imagine - so I'm with the other respondents who've pointed out the missing "^" - 10^94 would be far more like it - especially if there's something like 10^80 -odd particles in the universe, for example.
How dense can you get?
on
One of Many
·
· Score: 3
The second article, `inflation for beginners' says:
the density was not infinite but "only" some 1094 grams per cubic centimetre. These are the absolute limits on size and density allowed by quantum physics.
On that picture, the first puzzle is how anything that dense could ever expand -- it would have an enormously strong gravitational field, turning it into a black hole and snuffing it out of existence (back into the singularity) as soon as it was born.
Erm, could someone explain to me just how dense 1094g/cm^3 really is? I'm trying to picture a bag of sugar and a small cube of steel here... and I'm thinking maybe there's a scaling problem somwhere...
Slashdot Mirror
"but I'm sure I'm not alone in saying that I had never even heard of SpamAssassin until just now via /. "
You must be new around here or filtering out interesting articles... It's been around for ages, I've been using it now myself for nearly 18 months or more in varying guises.
I guess the latter would make it rather hard for them to force us to change the name for the thing, too.
"When you cousider that the universe could be infinatly old,"
Not in this reality, it couldn't be infin*i*tely old.
Ever since I was at school over 10 years ago it's been "between 10-20 billion years, most likely 16ish". That's why I also wonder what's changed.
I don't think I've ever worked in a company that uses only one language for everything, anyway. More to the point, none of the developers I've ever met have been one-horse wonders, either. And no surprise: to be a developer is to know how to *program*, which is a language-independent ability. We use all combinations of C, java, PHP, Tcl, and C# - and I was mildly pleased when one of the developers said he understood the point of s-expressions too, to say nothing of me being the in-office Ruby and Perl guy.
And never mind the linguistic inaccuracies (why didn't the author check each occurrence of `scheme' by hand?!) but what about jobs that are really vapourware to capture more CVs than the competing agencies? Don't they target the big names (VB, VC) more than anything else?
I'm tempted to agree, although in my case I watched it most when Maggie Philbin was on the team, ie when I was about 8-15 or something.
Still got the impression it was going downhill; it served its purpose in getting one small eejit hooked on things scientific, and then it all got dumbed-down. Bah.
It's not as though _Horizon_ is any temptation either - the reason we adult types have given up watching it is because of the dumbed-down drama - it even has "drama" in its caption. We don't need drama. We don't need the historical perspective, although it can be one means to present an achievement.
I said it at Uni, and I'll say it again: I don't give a constipated monkey whether Einstein *liked* his results, *show* me them and I'll work it out for myself.
RIP TW, then.
I'm not surprised people complained. I *keep on* saying it, but shifting the workload that spammers should be doing onto each and every innocent mailer is morally unjustifiable. It's not as though TMDA is any different and it's been around a long while.
And this is to say nothing of the dangers of sending a spam as though it were from one half of a mail2news gateway, of course. Or any other impersonated-sender scenario...
Choose spamassassin at SMTP injection point. Choose exim-4.10+exiscan. Reject mails with too high a score before they get anywhere near you. Use trap addresses and block them in the To: header, auto-reporting them in Envelope-To:. Whatever. Just don't multiply spams potentially to innocent parties!
What's the problem?
/etc/init.d/ since the get-go while RH have been messing around with this "/etc/rc.d" abomination which then needs legacy support on the assumption that there are idiots out there who can't cope with RH correcting their previous mistakes.
/proc, to tell me where in the filesystem the httpd is located that's responsible for a given webserver. If you can't debug that, you ain't gettin' root on my boxes.)
a) RPM already has at least two ways of being upgraded dynamically - urpmi and apt-get. It just needs a consistent well-maintained high-quality upstream pool-set
b) Debian supports RPM packages just fine.
c) The standards (specifically, the LSB) say nothing about requiring RPM to be the system's native package-managing system.
d) Debian already strive for LSB-compliancy, at least where it makes sense. This is why we've had
"the system directory on Windows 2000 is \WINNT"
Well, only *if* you install it that way.
And one for thought: which is more important, adhering to a standard for the sake of it, or knowing what you're doing? (A specific example of the latter: given an IP#, I expect you to be able to trace through DNAT, netstat -p or similar and through
"Doesn't seem overly useful to me...."
/bin/* /usr/bin/* for all packages for all distributions for all releases, or when do older things get purged?
Nor to me, for a different reason: what about those of us with CFLAGS= set to various strange funky optimizations in Gentoo? What about the Ports system in FreeBSD, similarly?
This thing does not have the potential to spread to all distributions or all unixen.
What about historical storage? Are they really proposing to store an md5sum for
Seems mad to me. Would be better off staying with AIDE instead, IMO.
"So let's return to the *original* definition of spam"
:)
What the one that applies only to Usenet? That's fine by me; the real problem with email is UBE anyway - *not* just UCE, although that's pretty foul crap anyway, but anything unsolicited and sent in bulk (read: not targetted at you specifically). This includes crazy charity/religious-headcase mails as well.
It also means that a mailshot from a company to "just say hi" to someone is still undesirable.
Besides which, it's still the ultimate in cheek to assume that someone's email connection is free - it's *NOT*, dammit, so any senders-of-UCE are stealing resources in order to try and sell you something.
I'm expecting to change bank *again* for exactly this reason.
Rewrite the definition: stamp out UBE!
Filter on Charset headers instead?
.procmailrc entry here but slashdot censored it saying "too many junk characters". So you'll have to work it out for yourself based on Content-Type headers matching
k r) "?
I was going to post my
charset="?(big5|ks_c_5601-1987|iso-2022-jp|euc-
instead.
More to the point, you never want to filter based on Received: headers, unless you can safely say that e.g. *no-one* in Korea is ever going to want to contact you. Otherwise, grab the IP# listings from IANA and see what netblocks are assigned to APNIC and score them down in your mail processing rules.
For what it's worth, I've had a lot of success using Bayesian filters to identify dodgy-charset mails - both ifile and bogofilter do a great job.
I've said it before, and I'll say it again.
Sending emails back to spammers is for brainless cretins - it serves only to clutter up your mail queue and risks offending innocent impersonated senders or having your email address confirmed as valid for spam.
And sending automated emails back to legitimate senders is downright *immoral* - making everyone do the work that a spammer *should* be doing to get through to you is indefensible.
And I've seen a case recently where this TMDA thing was so misconfigured that it sent an mail back to a mailing list saying there was an unrecognized sender address, and of course that mailing list was half of the gnu.emacs.help mail2news gateway, so the message appeared on the newsgroup for *all* to see. Talk about efficiently multiplying spam.
Now for something useful. Use one of the Bayesian filters, seeing as they're all the rage and get about 97-98% spam matched correctly, coupled with SpamAssassin as a fall-back for the remaining 2% cases, and you'll have far less of a problem.
Now incorporate those filters in your MTA so that the whole body is checked for spammishness before being "accepted for delivery" and you'll have the best solution of them all: bounce the mail at injection-point and be done.
South London (Croydon) was mostly clouded over especially in the south and just above the northern horizon. However, I saw two between 0245 and 0310 this morning before going back to bed.
;)
Why do these things have to happen on a week-day?!
Here's an idea. How about seeing it, not as "censoring" (which is generally bad, except that the word can have a fairly wide definition), but as choosing to separate folks according to what suits them best?
I think the main problem is that Mum and Dad are still involved in making sure kids stay within that domain - what's to stop some brat typing `slashdot.org' into the browser location bar?!
OTOH, there's no real problem with saying "this helps make a clear area appropriate for a given approximate age-range", and expecting folks to make contacts and get used to clicking around in there, then migrating out into the real world.
In fact, if it means said real world can be left intact with fewer stupid rules then all to the good.
> Neither JPG or GIF are W3C specifications.
...
:(
True, although most people would consider it nice to have a picture-format as well as a text-delivery system all openly and commonly available, so
While we're passing by, I saw `DjVU' touted as a possible successor to JPEG; however, that also has explicit "nice" (pro-open-source) patenting surrounding it, which I still think is a dodgy idea.
So how *do* I get pictures in web-pages? Is it only PNG now and stuff the older browsers?
Certainly feels a little bit retro, to me
That doesn't have to happen. All the spammers need to do is carry on pushing crap out and someone's bound to get it - drop a block of varied nice-words in the bottom of the mail (start "talking about" PHP, java, diet-coke and stuff, and you'll be going up in my estimation) and the damage potential will increase.
And in the process, more bandwidth will be consumed (which is where the real work-load is inflicted on the 'Net, not and the end-recipient) and the poor eejits out there *not* using these filters will still be getting crap.
That's also not the point. It might not cost them 10x as much to send 10x as much spam if that becomes the case. Not to mention, someone's bound to *try* circumventing any sufficiently common anti-spam system.
(And Bayesian filters are no panacea, nor are they invulnerable, btw; I still get the occasional junk-mail through here, despite my better efforts. I even had one over the weekend that fell back to SpamAssassin and scored over 27 but bogofilter considered clean - simply base64-encoded all its body, and I don't know that bogofilter either handles that well, or if it does, it hadn't been well enough trained against that sort of mail.)
The best way is still a moderately aggressive "just say NO!", not a passive acceptance, IMO. Spam should always be nuked at source.
Unfortunately, allowing the clueful few who wouldn't buy or respond to spam anyway (meaning mozilla users) to filter mail easier does nothing in the way of telling spammers to get knotted. That's where integrating the spam-check at the MTA receipt stage would come in handy; you don't have to passively accept crap, nor do you have to waste bandwidth sending complete bounce emails (potentially to innocent victims), but rather you drop things at injection. Fine by me :)
There's a patch and spiel about an exim module and filter that hooks into SpamAssassin at http://marc.merlins.org/linux/exim/sa.html
HTH,
Agreed, although some say this apparently gets harder with increasing antiquity ;)
I would say there's another thing to be on the look-out for: abstraction. Someone who "learns Pascal" or "learns Fortran" or "learns C" might be said to be missing out on the general "knows how to program" aspect, which is far more employable. Likewise "knows web-design" is missing out on the fun of the open w3c standards and tagged-up text-processing tools that go with XML.
Ask yourself what aspects of IT will be around in 10yrs' time.
Windoze? No way. Linux? Probably not in any form similar to present. Cisco? Likewise, mutation expected.
It's not worthwhile investing time and money for certification in any proprietary single technology. I'd say two key things that'll still be around are an understanding of SECURITY and of ETHICS. Read up on those - get to know security for what it is, and how best to apply it to a couple of current OSs, and work through the moral mazes of a few ethical questions, and you'll set yourself up for a longer career.
$0.02. HTH.
"It's almost unwatchable."
;8)
[catty comment] It's Harry Potter - it's unwatchable anyway!
"Everyone assumes Valenti and Rosen are right: that piracy is damaging the film and music businesses."
Agreed. I don't like seeing "illegal" copies, or premature leaks of media, but when it comes to previewing one or two tracks off an album in advance of buying the CD, I'm not going to waste half my lunchtime walking to $supermarket and listening to a few select noises when I can download someone else's copy of those tracks - and then I'll decide whether to go out and buy it or not (and delete the crap as need be).
This makes it apparent that there are at least two factors in the "does it cost us money?" equation - those folks who are freeloaders and will never buy a CD of even a proportion of the tracks they have, versus those who'll use it as a means of exploration and actively add more to the companies' profits.
I see it as simplistic to take either position - "it loses us money" or "it adds to the revenue!"; there's a mixture of both, and I've yet to see a calculation of the ratio of both phenomena.
While I'm passing, I'm of the opinion that milking that extra potential profit-margin (the freeloaders - who also exist in non-Internet-related realms) is pretty much the definition of greed.
It also seems increasingly apparent to me that the movie (and record) companies are simply trying to enforce an old marketing model - "you must get your crap here", rather than embracing an opportunity to expand to new horizons (FFS - think of the patent capabilities of helping invent a new distribution mechanism!). That's positively asking for rip-offs, IMO.
Has anyone actually found a source tarball anywhere?
:8(
What's with the sourceforge.net page saying the project "has not released any files"?
If you download the tangle module in CVS, you get a bunch of C++ that looks like it's missing rather a lot...
Has anyone built this from source yet?
Methinks you have it exactly the wrong way up.
I *want* a "PC" - whatever the heck that might mean these days - but most importantly, I want Architecture. I'm sick of buying Intel, and I don't want to buy into the intel concept by getting AMD instead, either. x86 can go whistle, I want ppc or alpha. And I want to run NetBSD on it, too, as forcing linux to run on various bits of kit (amiga A500+, anyone? Psion 5MX?) has lost its appeal with me. I have intel with Gentoo, FreeBSD and OpenBSD atm; need to complete the set!
So who else do we know who does G4-800 chips? One answer: Apple. And how much do they cost? about 2x-4x as much as this new amiga effort. So by sacrificing apple's proprietary mobo and peripherals, I fulfil my desire for "Architecture" and slice the cost right down. That's looking pretty peachy to me.
Now, the really worrying and annoying thing is that clause about needing some "enabler" to get OS4 to run on it. I read it like this: they want me to pay for something (always dodgy) that helps them fight me. Erm, yeah, right, time to open-source it and sell CDs with added manuals or other value for $20. *Then* I'd be interested in the OS.
"when I set up servers that do a lot of logging (firewall, web, ids, etc...) and I give a big /var partition (10+ GB) how little is filled up after several months."
Some sites' webserver logfiles are of the order a gig a day. Fortunately, apache CLF is well compressible, so an archival CD only needs burnt every 3-4 days.
Some sites' traffic in reporting a handful of servers' firewall hits back to a loghost contribute 4K/s of syslog traffic alone.
Do we have any stats for our great esteemed slashdot to compare?
"I just want to USE my computer no have to worry about all the moving parts under the hood."
So how do you distinguish between bits you "use" and are obviously therefore the only worthwhile bits for the rest of the world, and bits that we happen to like to use, or to choose between?
And why do you install tcsh if you're not going to use it? Do you have no idea what you're doing when you install a machine, and somehow that's the upstream distribution's fault for providing things that you have to choose to use or not, and you just can't be arsed making a choice yourself?
Stay with Windows. Please. For the sake of the kittens...
Dual feedback loops. Every mail that matches spam gets fed back into the system so both the is-spam wordlist AND the is-good wordlists become more "concentrated" over time. ...
Ifile does this, bogofilter does this with some wangling in procmail,
That way, if someone sends something that's still mostly spam (one or two words in common with spam, enough to tip the balance) then all the neutral words will be tarnished as well.
Right, yeah, it's not hard to convert the g/cm^3 thing into proper units - but my point is that 1cm^3 is easily visualized, and it weighing half a bag of sugar is totally easy to imagine - so I'm with the other respondents who've pointed out the missing "^" - 10^94 would be far more like it - especially if there's something like 10^80 -odd particles in the universe, for example.
Erm, could someone explain to me just how dense 1094g/cm^3 really is? I'm trying to picture a bag of sugar and a small cube of steel here... and I'm thinking maybe there's a scaling problem somwhere...