As somebody who has worked as a contractor developing software for startups for the better part of a decade, I'm here to tell you that you get what you negotiate. Renegotiating the contract you are working in the light of success is very bad form. Live and learn.
When I read the headline, my head parsed "Smithsonian celebrates 150 Years of COBOL"... but I guess that's just because when it's COBOL, it only feels like 150 years.
I've been in the situation where a manager in a crunch period really slowed the whole thing down because they were demanding explanations of every check-in. I've also had the experience of having a technical manager save the team no end of hassle by running interference and buffering us from the political realities even higher up the chain in crunch periods; in those cases the manager was technical enough to just let us get on with it.
Gerald Sussman and Harold Abelson's MIT course "The Structure and Interpretation of Computer Programs" is the best beginner's course I've ever seen anywhere. It starts at the beginning and ends up with advanced subjects like closures, and functional composition, building every concept in small pieces with clear examples. Plus the material is freely available, and there's video of Sussman giving the course to a bunch of hilariously dressed HP engineers in the eighties some time.
It seems to me that the obvious answer to this problem is to establish a 'parts bin' of privacy policy components that are guaranteed to be compatible with each other, and use a publishing mechanism similar to the creative commons licensing site, so the most common ones become known quantities for average users.
there's no free lunch. everybody wants a lamborghini for the cost of an impala. good, fast or cheap -- pick two. nobody puts symmetrix/shark blah blah blah out of business with duct tape and bailing wire. Jesus. if you're all so bored, give up the keyboards and get real jobs.:)
anyhow, the heads over at archive.org spun out a company to develop storage systems closely matching the brief you just laid down. Check this out: Petabox
you can buy the nodes and their stuff is proven in the archive.org infrastructure.:) not free, but then again, not as ludicrously expensive as the EMC/Hitachi/IBM/NetApp alternatives.
I just built a mail system that was intended to scale out to about 4 million users. We replaced over 70 windows machines running iMail with a small array of linux boxes behind load balancers. Mail is largely IO bound, so you pretty much have to get a kickass storage system. I specified fibrechannel connections to a generic SATA array with a lot of spindles. Because you're going to get a lot of delivery and read access concurrency, use Maildir as the spool storage. Select a good fast filesystem. We used Reiser for historical reasons, but XFS would also work well. We used Qmail as the transport with ldap patches and courier-imap(ssl) as the imap server, which practically gives you the webmail component for free. Spend your money on the SAN backend, and give each mail node a lot of ram. Apart from the Maildir requirement, you could build out a like architecture with postfix as well. To date, over 2 million accounts have been transitioned to the new system, and those 70 windows machines have been retasked as webservers. Here's the machine breakdown:
*note, if you have a requirement to save virus prone boxes from commiting sepuku over foreign agents coming in on the mail channel, you'll have to scale the smtp requirements nearly arbitrarily to account for processing each message for malignant components at that stage. Personally, I think it's kind of distasteful to use Free (speech) systems to make windows secure, as it's sort of defeatist, but that's my political baggage. YMMV.;)
... and the likely hassles the community is set to face in the near-term future, I think it would be a good idea to give the money to the Electronic Frontier Foundation, and have them administer the Linux Fund for the betterment of free software in general.
The article alludes to the historical fragmentation of the unix market back in the eighties, paving the way for windows to rule the small computer, and compares that to the emergence of 2.7... I don't understand. 2.1 branched from 2.0, 2.3 from 2.2, 2.5 from 2.4.. do we see a pattern here? Did the 'journalist' just miss the nuance in difference between a devel tree branching and a code system forking? I dunno. mountains. molehills. More stuff that will have to be explained the next time I propose a linux platform for a client problem. It's a drag man.
I've had a Spire bag since 1999, and while I have dragged it all over the world with me, it has outlived three laptops. They cost a lot, but their medium bag is comfortable to be worn all day long, and they're tough tough. Mine has literally weathered typhoons, survived long motorcycle trips, and protected my machines through all of it. Best of all, they don't look like computer bags, so they don't tend to attract the wrong kind of attention when you're travelling.
One of the lovely things about a mac is that it uses a real (openboot) firmware to manage bootstrapping, pretty much like a sparc. if your alias to the boot partition of your OSX install is munged, just give the hardware address to that partition in openboot, and you're good to go.
This is a great idea. I was thinking of porting knoppix to the ppc for some time while my life dispatched the regular inerrupt requestors... and in the meantime somebody delivered pretty much exactly what I wanted. =)
those mirrors are getting slaughtered... would somebody (gentoo? gatech?) put up a bittorrent tracker for those iso's?
I believe the DUNs are generated based on ip allocation justifications submitted to ARIN. If an ISP isn't working within it's IP allocation plan, then I guess, yes, that is possible.
It is unsurprising that AOL has taken this step... most postmasters I know utilise a DUN list of some sort, and one of the most prominently represented dial nets in any DUN worth it's footprint in ram is AOL's. I'd much rather see large mail transports like AOL's refuse smtp from dial nets than have raw smtp blocked at the ISP's network, which is what happens on many large dial networks now... preventing things like privacy services from working correctly without the further complication of tunneling the traffic out of the ISP's dial network. sympatico.ca in Canada blocks raw SMTP, and has done for ages. A fact that made me insane on a semi-regular basis.
It is a real shame that the antisocial behaviour manifested on the network in the form of spam has caused us to break our own rules out of a real practical necessity. I'm still not convinced it isn't possible to deal with this kind of detritus by technical means. I think a dynamic BGP blackhole triggered by pseudo bayesian criteria could really cause the businesses and organisations involved with open relays and problem customers to deal directly with the problem in a timely way. It would certainly limit the collateral damage of breaking RFCs on a global scale, just to deal with a few pathologically antisocial money fixated anal retentive ass bandits who insist on flushing their shite into the public networks of the world.
One day...
on
Fun With Wine
·
· Score: 3, Interesting
This kind of virtualisation stress test is interesting, but largely academic. I'm still waiting for the day when it is less hassle to load the (very few) windows applications I need under wine than it is to reboot my workstation to deal with those tasks under windows. Screwing around with wine to get it to load even small windows applications is one of the most frustrating things I can think of in association with *nix systems. I hear good things about the transgaming stuff, but it obviously hasn't made it back into the main branch of the wine tree. The promise of wine has been hanging out there for a lot of years now; I'm just wondering if perhaps they're trying to build a glass house on quicksand.
There was discussion in this thread about whether Revenue Canada runs their
own court; there is an interesting true story of one guy's fight with Revenue
Canada in the courts available at this site
. In short, the government tries to ellicit information from the guy using
force and threats of imprisonment, he challenges in court, and when he gets
before the judge, the judge claims that he is not in fact an agent of the
Law, but an agent of RevCan. Worth a read.
well, that's partly what I was talking about. The kind of flesh-eating lawyers wielded like linebackers in the software/content business these days are more than formidable -- they're completely unsurmountable by individuals and small corporations.When is the chilling effect going to sink all the way back to the engineering decisions made at the design stage of product development for these kinds of appliances?
... watching how commodity appliances based on free software evolve and mutate beyond the intentions of their creators. While this is obviously good for the consumer, this kind of activity seems to be broadly percieved as threatening to the producers of these toys. I wonder if the real threat to free software systems in embedded roles is the building legal backlash; eventually the costs of defending engineering choices is going to offset the advantages of choosing free systems to begin with and purely pragmatic people (accountants) will decide that it's not worth the effort and time. Is this what groups like the MPAA, RIAA and the proprietary software houses are counting on?
sure, but in that circumstance any correctly configured mail server would reject mail coming from a dial pool anyhow, so the "openness" of the port is completely moot. The only people you would be able to send mail to would be people on other dial access connections. The conversation is ludicrous.
Fully 50% of the SPAM I used to see originated from the dialpools of the gargantuan ISPs of the world. UUnet, Earthlink, PSInet etc. Since I have converted all of the mail servers I administer to utilise the MAPS-DUL the spam traversing my network has been massively cut down. None of these servers will accept smtp traffic from inside dial-pool netblocks. the RBL is very effective against the cretins who run open exchangers. I would like to note that since implementing the MAPS-DUL, my stress level has also signifigantly dropped as I no longer have to endure the silly autoresponders that these large ISPs have working the abuse@ , postmaster@, and root@ email addresses. I'm positive that the majority of abuse mail is piped to/dev/nul.
ssh tunnels to a shell outside the employer
network. They can't monitor what they can't
read, and they can't seize what you don't have
in your possession.
-f
Hybrid cars don't have to be slow. Chrysler developed a race car called the Patriot that used a high-output alternator and a carbon-fiber flywheel, spun up by a turbine to drive a 600hp electric motor which transmited power via a manual 6spd transmission. The car reclaimed braking energy, putting it back into the flywheel, and had a terminal velocity of over 200mph. It also had a range of over 1500miles. Wired had an article on the car early in the mag's run.. just searched and found this: Patriot
You seem to be labouring under the opinion that we live in a rational universe, and as products of this universe are in point of fact inherrently rational ourselves. Can you support that assumption?
Slashdot Mirror
As somebody who has worked as a contractor developing software for startups for the better part of a decade, I'm here to tell you that you get what you negotiate. Renegotiating the contract you are working in the light of success is very bad form. Live and learn.
When I read the headline, my head parsed "Smithsonian celebrates 150 Years of COBOL" ... but I guess that's just because when it's COBOL, it only feels like 150 years.
Build better bugs, experience more fruitful testing.
I've been in the situation where a manager in a crunch period really slowed the whole thing down because they were demanding explanations of every check-in. I've also had the experience of having a technical manager save the team no end of hassle by running interference and buffering us from the political realities even higher up the chain in crunch periods; in those cases the manager was technical enough to just let us get on with it.
Gerald Sussman and Harold Abelson's MIT course "The Structure and Interpretation of Computer Programs" is the best beginner's course I've ever seen anywhere. It starts at the beginning and ends up with advanced subjects like closures, and functional composition, building every concept in small pieces with clear examples. Plus the material is freely available, and there's video of Sussman giving the course to a bunch of hilariously dressed HP engineers in the eighties some time.
It seems to me that the obvious answer to this problem is to establish a 'parts bin' of privacy policy components that are guaranteed to be compatible with each other, and use a publishing mechanism similar to the creative commons licensing site, so the most common ones become known quantities for average users.
there's no free lunch. everybody wants a lamborghini for the cost of an impala. good, fast or cheap -- pick two. nobody puts symmetrix/shark blah blah blah out of business with duct tape and bailing wire. Jesus. if you're all so bored, give up the keyboards and get real jobs. :)
:) not free, but then again, not as ludicrously expensive as the EMC/Hitachi/IBM/NetApp alternatives.
anyhow, the heads over at archive.org spun out a company to develop storage systems closely matching the brief you just laid down. Check this out: Petabox
you can buy the nodes and their stuff is proven in the archive.org infrastructure.
I just built a mail system that was intended to scale out to about 4 million users. We replaced over 70 windows machines running iMail with a small array of linux boxes behind load balancers. Mail is largely IO bound, so you pretty much have to get a kickass storage system. I specified fibrechannel connections to a generic SATA array with a lot of spindles. Because you're going to get a lot of delivery and read access concurrency, use Maildir as the spool storage. Select a good fast filesystem. We used Reiser for historical reasons, but XFS would also work well. We used Qmail as the transport with ldap patches and courier-imap(ssl) as the imap server, which practically gives you the webmail component for free. Spend your money on the SAN backend, and give each mail node a lot of ram. Apart from the Maildir requirement, you could build out a like architecture with postfix as well. To date, over 2 million accounts have been transitioned to the new system, and those 70 windows machines have been retasked as webservers. Here's the machine breakdown:
;)
2 * dual xeon/1G ram/small boot drive -- SMTP
6 * dual xeon/4G ram/small boot drive -- imap/pop
2 * dual xeon/4G ram/internal raid -- LDAP
4 Terrabyte SATA Raid backend connected by fibrechannel cards
*note, if you have a requirement to save virus prone boxes from commiting sepuku over foreign agents coming in on the mail channel, you'll have to scale the smtp requirements nearly arbitrarily to account for processing each message for malignant components at that stage. Personally, I think it's kind of distasteful to use Free (speech) systems to make windows secure, as it's sort of defeatist, but that's my political baggage. YMMV.
... and the likely hassles the community is set to face in the near-term future, I think it would be a good idea to give the money to the Electronic Frontier Foundation, and have them administer the Linux Fund for the betterment of
free software in general.
The article alludes to the historical fragmentation of the unix market back in the eighties, paving the way for windows to rule the small computer, and compares that to the emergence of 2.7... I don't understand. 2.1 branched from 2.0, 2.3 from 2.2, 2.5 from 2.4.. do we see a pattern here? Did the 'journalist' just miss the nuance in difference between a devel tree branching and a code system forking? I dunno. mountains. molehills. More stuff that will have to be explained the next time I propose a linux platform for a client problem. It's a drag man.
I've had a Spire bag since 1999, and while I have dragged it all over the world with me, it has outlived three laptops. They cost a lot, but their medium bag is comfortable to be worn all day long, and they're tough tough. Mine has literally weathered typhoons, survived long motorcycle trips, and protected my machines through all of it. Best of all, they don't look like computer bags, so they don't tend to attract the wrong kind of attention when you're travelling.
One of the lovely things about a mac is that it uses a real (openboot) firmware to manage bootstrapping, pretty much like a sparc. if your alias to the boot partition of your OSX install is munged, just give the hardware address to that partition in openboot, and you're good to go.
Use bootx to boot oldworld macs. The details of this proceedure are well documented in a few different places. google it.
This is a great idea. I was thinking of porting knoppix to the ppc for some time while my life dispatched the regular inerrupt requestors... and in the meantime somebody delivered pretty much exactly what I wanted. =)
those mirrors are getting slaughtered... would somebody (gentoo? gatech?) put up a bittorrent tracker for those iso's?
I believe the DUNs are generated based on ip allocation justifications submitted to ARIN. If an ISP isn't working within it's IP allocation plan, then I guess, yes, that is possible.
It is unsurprising that AOL has taken this step... most postmasters I know utilise a DUN list of some sort, and one of the most prominently represented dial nets in any DUN worth it's footprint in ram is AOL's. I'd much rather see large mail transports like AOL's refuse smtp from dial nets than have raw smtp blocked at the ISP's network, which is what happens on many large dial networks now... preventing things like privacy services from working correctly without the further complication of tunneling the traffic out of the ISP's dial network. sympatico.ca in Canada blocks raw SMTP, and has done for ages. A fact that made me insane on a semi-regular basis.
It is a real shame that the antisocial behaviour manifested on the network in the form of spam has caused us to break our own rules out of a real practical necessity. I'm still not convinced it isn't possible to deal with this kind of detritus by technical means. I think a dynamic BGP blackhole triggered by pseudo bayesian criteria could really cause the businesses and organisations involved with open relays and problem customers to deal directly with the problem in a timely way. It would certainly limit the collateral damage of breaking RFCs on a global scale, just to deal with a few pathologically antisocial money fixated anal retentive ass bandits who insist on flushing their shite into the public networks of the world.
This kind of virtualisation stress test is interesting, but largely academic. I'm still waiting for the day when it is less hassle to load the (very few) windows applications I need under wine than it is to reboot my workstation to deal with those tasks under windows. Screwing around with wine to get it to load even small windows applications is one of the most frustrating things I can think of in association with *nix systems. I hear good things about the transgaming stuff, but it obviously hasn't made it back into the main branch of the wine tree. The promise of wine has been hanging out there for a lot of years now; I'm just wondering if perhaps they're trying to build a glass house on quicksand.
There was discussion in this thread about whether Revenue Canada runs their
own court; there is an interesting true story of one guy's fight with Revenue
Canada in the courts available at this
site
. In short, the government tries to ellicit information from the guy using
force and threats of imprisonment, he challenges in court, and when he gets
before the judge, the judge claims that he is not in fact an agent of the
Law, but an agent of RevCan. Worth a read.
well, that's partly what I was talking about. The kind of flesh-eating lawyers wielded like linebackers in the software/content business these days are more than formidable -- they're completely unsurmountable by individuals and small corporations.When is the chilling effect going to sink all the way back to the engineering decisions made at the design stage of product development for these kinds of appliances?
... watching how commodity appliances based on free software evolve and mutate beyond the intentions of their creators. While this is obviously good for the consumer, this kind of activity seems to be broadly percieved as threatening to the producers of these toys. I wonder if the real threat to free software systems in embedded roles is the building legal backlash; eventually the costs of defending engineering choices is going to offset the advantages of choosing free systems to begin with and purely pragmatic people (accountants) will decide that it's not worth the effort and time. Is this what groups like the MPAA, RIAA and the proprietary software houses are counting on?
It already runs X. Developers have been exporting
the display to their desktop terminals all along.
sure, but in that circumstance any correctly configured mail server would reject mail coming from a dial pool anyhow, so the "openness" of the port is completely moot. The only people you would be able to send mail to would be people on other dial access connections. The conversation is ludicrous.
/dev/nul.
Fully 50% of the SPAM I used to see originated from the dialpools of the gargantuan ISPs of the world. UUnet, Earthlink, PSInet etc. Since I have converted all of the mail servers I administer to utilise the MAPS-DUL the spam traversing my network has been massively cut down. None of these servers will accept smtp traffic from inside dial-pool netblocks. the RBL is very effective against the cretins who run open exchangers. I would like to note that since implementing the MAPS-DUL, my stress level has also signifigantly dropped as I no longer have to endure the silly autoresponders that these large ISPs have working the abuse@ , postmaster@, and root@ email addresses. I'm positive that the majority of abuse mail is piped to
ssh tunnels to a shell outside the employer network. They can't monitor what they can't read, and they can't seize what you don't have in your possession. -f
Hybrid cars don't have to be slow. Chrysler developed a race car called the Patriot that used a high-output alternator and a carbon-fiber flywheel, spun up by a turbine to drive a 600hp electric motor which transmited power via a manual 6spd transmission. The car reclaimed braking energy, putting it back into the flywheel, and had a terminal velocity of over 200mph. It also had a range of over 1500miles. Wired had an article on the car early in the mag's run.. just searched and found this: Patriot
You seem to be labouring under the opinion that we live in a rational universe, and as products of this universe are in point of fact inherrently rational ourselves. Can you support that assumption?