Blacklists and Spam Assassin help some, but there are too many false positives
Perhaps not as many as you'd think though - I recently switched from Spamfire (keyword based filter) to POPFile (Bayesian with list of known-to-be-good-senders), and have been very impressed.
It's been running for 10 days, has processed 1108 mails, and made 26 mistakes. Almost all of which were in the first 24 hours - I've been checking my spam folder a couple of times a day, and have had 3 false positives in the last week (all receipts from online orders).
So far it's claiming 97.65% accuracy, with 60% of the mail passing through it being spam.
Norway is not currently part of the EU, although there are close ties between them (e.g., they are in the EEA which brings them somewhat into the internal EU market zone, and they particpate in schemes like Europol).
All the signs are they probably will join eventually, although I believe the last referendum in '94 voted not to (although that was pre-Euro, which may affect how people vote next time).
AFAIK, the beachball first showed up as the wait cursor for MPW (Apple's pre-Mac OS X command line development environment). It started showing up in other software after that.
In terms of the official busy cursor, you're right, it was a wristwatch.
Every day, 80 percent of all incoming mail to Outblaze is rejected as spam and filtered out before Ramasubramanian and his team have to deal with it. Out of the remaining 15 million messages per day that do pass through Outblaze servers
So if 15 million messages is 20% of what they get, they receive 75 million individual messages a day? That seems a little high...
Actually he's English. Born in Surrey apparently, so I guess his accent is fake - Surrey's a long way from Leith...:-)
Although he still sounds Scottish in the thing he's in just now on Sky One ("Mile High"), so I guess it's either part of his image or he lived in Scotland for a while.
Given what OS X is, it's not surprising that it takes some getting used to, despite vaguely looking like Mac OS.
It's gotten a lot better, but the best description I've heard of Mac OS X DPs/10.0 was "it's kind of like a Mac, but a Mac built by people who've only had a Mac described to them over the phone".
There were a number of really quite spurious changes to the UI initially, which probably explains the demand for this kind of book - the change from 9 to X has been more confusing than any OS transition Apple users have ever had to do before, including the move to System 7 (when there was also plenty of grousing to start with).
And deciding what "properly" is, is often the toughest problem...:-) I.e., formal methods are very strong for self-consistency within a system, but weaker when it comes to the boundary between the system and the outside world.
The technique you describe is part of a field known as "Formal Methods".
The term "cleanroom" comes from reverse engineering, where you have team A of engineers write a spec for a competitors product and team B (who've never seen the product) write an implementation. This gives you some degree of legal protection, but does not prove anything about correctness.
Of course, the flaw with formal methods is that they only prove the program is functioning as designed - which is definitely a worthwhile goal, but does not say anything about the correctness of the design itself. E.g., think of the problems with the incorrect mirror for the Hubble Space Telescope - the grinding machine worked perfectly, but the final mirror was still ground to the wrong shape.
Downloading seeds from Apple typically requires you to have an account in their developer program, and to log in to their seed server - you then get a randomly generated user name/password combo for an ftp server that expires in a couple of hours. It would be quite possible to hook into this and to watermark binaries so that you could tell which seed account was used to leak the app.
Or have the app watermark itself on first launch, after prompting for a unique key which gets mailed out to each person in the seed program. Lots of ways you could do it, and I wouldn't be surprised if you start seeing some of them in the future.
We had to introduce a similar scheme for betas of our software, after a leak just before release - which has resulting in end-users mailing support asking for help when their ripped off copy fails to function with the latest data (wasting our time and slowing down support for our customers).
Although the intent behind Habeas isn't really to automate filtering, it's to allow legal action to be taken against spammers.
The headers they append are fragments of a copyrighted/trademarked poem or somesuch: the idea is that eventually enough people will filter based on this that a spammer will "illegally" append them as well.
Remains to be seen how effective it will actually be - I suspect not very, since spammers will only need to bother appending the Habeas marker if 99% of legit mail includes it. Since Habeas charge money to include their mark in business email, most mail will never include it - so spammers will never need to bother.
Why hasn't at least one manufacturer (maybe one that isn't doing as well as HP or Epson) decided to go for the enormous marketing advantage that cheap cartridges would give its products?
Easy - because they all built their business model around "sell the printer cheaply, make it up on ink". They subsidise the design and manufacturing with the revenue they get back from ink sales, so typically the only people interested in re-fillable ink cartridges (or kits to do it) are people who don't sell printers. People who do sell printers would incur a huge cost to switch their business model, and why bother when they're making money the way things stand.
It's possible a 3rd party could come out with a compatible print cartridge, but this is why more and more companies are moving to embedding electronics in the cartridge itself - then they get access to legislation like the DMCA to prevent competition, since it's not just a physical interface.
If it hasn't happened to inkjets, what does that mean? Incredibly effective price-fixing strategies in the inkjet industry?
It doesn't mean anything - inkjets can get away with their lock-in model since a cheap and easy substitute isn't widely available. I guess you could try ink from pens, but the nozzle on an inkjet is far more prone to blockage than something like a fountain pen nib.
I'm not sure refuelling from drinking alcohol will ever take off - all of the devices which are going to use fuel cells will also take mains power, and a much better emergency power source would be something like a hand-powered generator from freeplay.
Given that fuel cells are going to be a pretty niche item to start with, I doubt they're going to go to the extra expense of making them user-servicable and reusable rather than disposable.
Apple's look and feel lawsuit against Microsoft failed. Was that because Apple had copied it's look and feel from the Xerox PARC?
No, it's because Apple's lawyers didn't do a good job of drawing up the contract they signed with MS. They didn't realise how much "look and feel" they were actually allowing MS to use, and contested MS' intepretation of the contract. The courts agreed with MS - nothing to do with PARC, just a typical "this means X/no, it means Y" legal situation.
Actually, yes he was. We were sitting side by side with our Apple laptops. I was preparing for a talk at a conference and he was writing code.
Hmm, he's got more patience than me then...:-)
I tried doing Mac+PC development on a Mac using VPC and VC++, but found it was incredibly slow (as in a 30 minute build on the Mac would be several hours on the PC). Perhaps I just had something misconfigured or something, but we eventually gave in and got a bare-bones PC for doing builds on.
Re:Big advantages for developers
on
Virtual PC 6 Review
·
· Score: 2, Informative
I highly doubt they were using the Mac version for development - they were much more likely to be using the Windows version of VPC.
This will give you near-native speed since the instruction set doesn't need to be interpreted, but you can still have multiple virtual machines (which definitely makes life easier: particularly for QA, where you can run multiple versions of Windows for testing and just fetch a fresh disk image when things get broken).
This whole virtual-x86-machine-on-x86 is exactly why MS bought them: the fact that there's a Mac version was pretty incidental to them I imagine - what they're really after is the virtual server market.
If some overachiever finds a bug in code you are responsible for at 2:30 AM Sunday morning, you get paged.
I find that hard to believe - unless you're on the brink of shipping a product, in which case you're probably up early, what possible good can bringing you in 6 hours early do?
The rules of efficiency have been rewritten, it seems to be drifting towards not how good your program is, but how fast you can crank it out.
That's just a reflection of the times - programmer time is now much more costly than machine time, which wasn't always the case (quite the reverse in fact).
Look at how popular scripting languages are nowadays - they're popular because they're expressive enough to let you solve problems quickly, even though they can easily be orders of magnitude slower to execute than a hard-coded equivalent.
Have you actually tried TiPaint.com? According to several of the reader reports at MacInTouch, they're a scam.
the exterior of the machine, being metal, is prone to denting.
Hardly something you can fault them for - build it out of plastic and you're more likely to crack it than make a dent. Not much you can do other than take care: they're computers, they're always going to be delicate.
what Apple really needs to do is come out with a G4 iBook
That would be the 12" PowerBook. It's pretty obvious that this was going to be "the G4 iBook", but they probably couldn't get the cost down for the first rev - but that's what it is.
Neither does Virtual PC - an older version used to support Glide (3Dfx's API) and map that onto RAVE (Apple's API), but this was dropped and never really followed through on (e.g., a D3D->OpenGL translator would have been the logical step, but they presumably felt it wouldn't be very heavily used by their typical customer).
QuickDraw 3D was the thing that was replaced with OpenGL - QuickDraw is the 2D drawing API that dates back to the original Mac, and is the equivalent of GDI.
It has been superceded with Quartz in Mac OS X, but it's still fully supported.
Slashdot Mirror
Alan! You know how I feel about you taking the Lord's name in vain. You've been spending too much time with Michael again.
Blacklists and Spam Assassin help some, but there are too many false positives
Perhaps not as many as you'd think though - I recently switched from Spamfire (keyword based filter) to POPFile (Bayesian with list of known-to-be-good-senders), and have been very impressed.
It's been running for 10 days, has processed 1108 mails, and made 26 mistakes. Almost all of which were in the first 24 hours - I've been checking my spam folder a couple of times a day, and have had 3 false positives in the last week (all receipts from online orders).
So far it's claiming 97.65% accuracy, with 60% of the mail passing through it being spam.
Norway is not currently part of the EU, although there are close ties between them (e.g., they are in the EEA which brings them somewhat into the internal EU market zone, and they particpate in schemes like Europol).
All the signs are they probably will join eventually, although I believe the last referendum in '94 voted not to (although that was pre-Euro, which may affect how people vote next time).
AFAIK, the beachball first showed up as the wait cursor for MPW (Apple's pre-Mac OS X command line development environment). It started showing up in other software after that.
In terms of the official busy cursor, you're right, it was a wristwatch.
Every day, 80 percent of all incoming mail to Outblaze is rejected as spam and filtered out before Ramasubramanian and his team have to deal with it. Out of the remaining 15 million messages per day that do pass through Outblaze servers
So if 15 million messages is 20% of what they get, they receive 75 million individual messages a day? That seems a little high...
Actually he's English. Born in Surrey apparently, so I guess his accent is fake - Surrey's a long way from Leith... :-)
Although he still sounds Scottish in the thing he's in just now on Sky One ("Mile High"), so I guess it's either part of his image or he lived in Scotland for a while.
Given what OS X is, it's not surprising that it takes some getting used to, despite vaguely looking like Mac OS.
It's gotten a lot better, but the best description I've heard of Mac OS X DPs/10.0 was "it's kind of like a Mac, but a Mac built by people who've only had a Mac described to them over the phone".
There were a number of really quite spurious changes to the UI initially, which probably explains the demand for this kind of book - the change from 9 to X has been more confusing than any OS transition Apple users have ever had to do before, including the move to System 7 (when there was also plenty of grousing to start with).
the function of the code must be defined properly
:-) I.e., formal methods are very strong for self-consistency within a system, but weaker when it comes to the boundary between the system and the outside world.
And deciding what "properly" is, is often the toughest problem...
The technique you describe is part of a field known as "Formal Methods".
The term "cleanroom" comes from reverse engineering, where you have team A of engineers write a spec for a competitors product and team B (who've never seen the product) write an implementation. This gives you some degree of legal protection, but does not prove anything about correctness.
Of course, the flaw with formal methods is that they only prove the program is functioning as designed - which is definitely a worthwhile goal, but does not say anything about the correctness of the design itself. E.g., think of the problems with the incorrect mirror for the Hubble Space Telescope - the grinding machine worked perfectly, but the final mirror was still ground to the wrong shape.
Downloading seeds from Apple typically requires you to have an account in their developer program, and to log in to their seed server - you then get a randomly generated user name/password combo for an ftp server that expires in a couple of hours. It would be quite possible to hook into this and to watermark binaries so that you could tell which seed account was used to leak the app.
Or have the app watermark itself on first launch, after prompting for a unique key which gets mailed out to each person in the seed program. Lots of ways you could do it, and I wouldn't be surprised if you start seeing some of them in the future.
We had to introduce a similar scheme for betas of our software, after a leak just before release - which has resulting in end-users mailing support asking for help when their ripped off copy fails to function with the latest data (wasting our time and slowing down support for our customers).
Although the intent behind Habeas isn't really to automate filtering, it's to allow legal action to be taken against spammers.
The headers they append are fragments of a copyrighted/trademarked poem or somesuch: the idea is that eventually enough people will filter based on this that a spammer will "illegally" append them as well.
Remains to be seen how effective it will actually be - I suspect not very, since spammers will only need to bother appending the Habeas marker if 99% of legit mail includes it. Since Habeas charge money to include their mark in business email, most mail will never include it - so spammers will never need to bother.
Actually, G4s perform better than P4s on RC5 due to their faster hardware rotate support - not AltiVec.
(And you meant "I hope they all die in traffic accidents". Plural on the last word.)
No, just one really big accident.
Why hasn't at least one manufacturer (maybe one that isn't doing as well as HP or Epson) decided to go for the enormous marketing advantage that cheap cartridges would give its products?
Easy - because they all built their business model around "sell the printer cheaply, make it up on ink". They subsidise the design and manufacturing with the revenue they get back from ink sales, so typically the only people interested in re-fillable ink cartridges (or kits to do it) are people who don't sell printers. People who do sell printers would incur a huge cost to switch their business model, and why bother when they're making money the way things stand.
It's possible a 3rd party could come out with a compatible print cartridge, but this is why more and more companies are moving to embedding electronics in the cartridge itself - then they get access to legislation like the DMCA to prevent competition, since it's not just a physical interface.
If it hasn't happened to inkjets, what does that mean? Incredibly effective price-fixing strategies in the inkjet industry?
It doesn't mean anything - inkjets can get away with their lock-in model since a cheap and easy substitute isn't widely available. I guess you could try ink from pens, but the nozzle on an inkjet is far more prone to blockage than something like a fountain pen nib.
I'm not sure refuelling from drinking alcohol will ever take off - all of the devices which are going to use fuel cells will also take mains power, and a much better emergency power source would be something like a hand-powered generator from freeplay.
Given that fuel cells are going to be a pretty niche item to start with, I doubt they're going to go to the extra expense of making them user-servicable and reusable rather than disposable.
I'm getting old. And I'm only 18...
You're not old - you still feel the need to tell people your age. Young people always do.
Apple's look and feel lawsuit against Microsoft failed. Was that because Apple had copied it's look and feel from the Xerox PARC?
No, it's because Apple's lawyers didn't do a good job of drawing up the contract they signed with MS. They didn't realise how much "look and feel" they were actually allowing MS to use, and contested MS' intepretation of the contract. The courts agreed with MS - nothing to do with PARC, just a typical "this means X/no, it means Y" legal situation.
Actually, yes he was. We were sitting side by side with our Apple laptops. I was preparing for a talk at a conference and he was writing code.
:-)
Hmm, he's got more patience than me then...
I tried doing Mac+PC development on a Mac using VPC and VC++, but found it was incredibly slow (as in a 30 minute build on the Mac would be several hours on the PC). Perhaps I just had something misconfigured or something, but we eventually gave in and got a bare-bones PC for doing builds on.
I highly doubt they were using the Mac version for development - they were much more likely to be using the Windows version of VPC.
This will give you near-native speed since the instruction set doesn't need to be interpreted, but you can still have multiple virtual machines (which definitely makes life easier: particularly for QA, where you can run multiple versions of Windows for testing and just fetch a fresh disk image when things get broken).
This whole virtual-x86-machine-on-x86 is exactly why MS bought them: the fact that there's a Mac version was pretty incidental to them I imagine - what they're really after is the virtual server market.
Actually, he says "Get your ass to Mars!". Quite the potty mouth.
If some overachiever finds a bug in code you are responsible for at 2:30 AM Sunday morning, you get paged.
I find that hard to believe - unless you're on the brink of shipping a product, in which case you're probably up early, what possible good can bringing you in 6 hours early do?
The rules of efficiency have been rewritten, it seems to be drifting towards not how good your program is, but how fast you can crank it out.
That's just a reflection of the times - programmer time is now much more costly than machine time, which wasn't always the case (quite the reverse in fact).
Look at how popular scripting languages are nowadays - they're popular because they're expressive enough to let you solve problems quickly, even though they can easily be orders of magnitude slower to execute than a hard-coded equivalent.
The enamel keeps peeling off [tipaint.com]
Have you actually tried TiPaint.com? According to several of the reader reports at MacInTouch, they're a scam.
the exterior of the machine, being metal, is prone to denting.
Hardly something you can fault them for - build it out of plastic and you're more likely to crack it than make a dent. Not much you can do other than take care: they're computers, they're always going to be delicate.
what Apple really needs to do is come out with a G4 iBook
That would be the 12" PowerBook. It's pretty obvious that this was going to be "the G4 iBook", but they probably couldn't get the cost down for the first rev - but that's what it is.
and it doesn't have accelerated 3D video drivers
Neither does Virtual PC - an older version used to support Glide (3Dfx's API) and map that onto RAVE (Apple's API), but this was dropped and never really followed through on (e.g., a D3D->OpenGL translator would have been the logical step, but they presumably felt it wouldn't be very heavily used by their typical customer).
The reason why QuickDraw and QuickDraw3D
QuickDraw 3D was the thing that was replaced with OpenGL - QuickDraw is the 2D drawing API that dates back to the original Mac, and is the equivalent of GDI.
It has been superceded with Quartz in Mac OS X, but it's still fully supported.