Some have simply given up on trying to force the general-case IT policy to be useful. They "solve" the usability problem for their specific case by ignoring IT and using outside tools over which IT has no control.
"IT goon: Business comes first. We're here to support your business!"
"Me Great! We build software systems based on open source, so our developers need access to github."
"IT goon: Sorry, that's a file sharing site. Using it is against policy."
'Me: You said business first. Our business uses open source software from github."
"IT goon: That's right! Business first! Just no file sharing sites."
I didn't "lose" the job any more than I "lose" a defective computer when I throw it in the trash. Indeed it would be very hard to consider it a loss when six months later I was earning $10k more per year.
Nor did I put myself in any legal jeopardy. I'll spare you the lengthy analysis.
Best way to handle the problem? Burning bridges rarely is. But sometimes it has a moral righteousness that's hard to defy.
I wrote a memo laying out all the issues in layman's terms and proposing solutions. Then I gave it to my boss. A little while later with no further movement on the problem, I quit.
A year passed and the system was hacked. Publicly. Embarrassingly. Folks here on Slashdot asked what the sysadmins could possibly have been thinking. So, I published a copy of the memo I had written.
The technology is ready to retire. The impediment is regulatory -- without FCC oversight, delivery of last-mile infrastructure becomes thoroughly anticompetitive, a process which has repeated itself over and over again this past half century. POTS and twisted pair has been the last vestige of deregulation in the sector, to the detriment of the public and MUCH to the detriment of inventors and small business.
Why suppose there's something underhanded going on? It boils down to ignorance. The managers don't really understand the job and certainly can't articulate it, so they list the things they *can* understand... like the model numbers on the hardware, the programs the last guy used, and so on.
In theory, the patent examiner is educated in and well steeped in the field for which he examines patent applications. Einstein wasn't a patent clerk by chance alone -- the patent office requires folks with a strong technical background.
Wheels is a new concept. Inflatable tires on the wheels is a new concept. Springs on the axle is a new concept. Steerable wheels is a new concept. That just about covers it. Most of the rest of the improvements around the wheel and axle are derivative, unworthy of a patent. They don't advance the state of the art enough to merit giving their creators exclusive control.
That's what patents are about, right? They're supposed to encourage leaps of brilliance by making it practical for their inventors to profit off them. If there's no genius, just plodding iterative improvement, there shouldn't be a patent.
I understand that patents act on inventions, not ideas. That's because (A) ideas are rather nebulous and (B) patents must work. Hard to demonstrate that an idea actually works without building something that implements it. Nevertheless, the patent tends to protect the concepts that went in to it, not the precise implementation. If they only protected the precise implementation they'd be worthless.
JPEG was based on the notion that digital photographs were not pixel perfect to begin with, so why maintain that non-perfection perfectly if not doing so would make high compression rates possible. Acceptably lossy compression. This was a fresh, novel concept. The invention based on it deserved patent protection.
MPEG and MP3 took the then-well-known concept of lossy compression and ground on it until they came up with something that worked well for video and audio. After all, if the concept works for one media type, it should work for others, right? No unexpected originality, no unforeseen change to the state of the art, just grinding. Should not have been patentable.
See? Not terribly hard to look back and evaluate the state of the art in the day.
That's why I used the word "justify" and not the word "prove."
Right now the presumption is that if a half-hearted and formulaic search turns up no prior art then the invention is novel. I want to turn that on its head: the invention is presumed obvious until you explain why it isn't in terms folks in your field agree with. Who are the people in your field and why aren't they half a step behind you?
I've read some of the software patents. I have better than average skill in the art and I couldn't implement the software described from *any* of them. Not. A. Single. One.
I'm strongly in favor of patent reform, but CBM is not the answer. CBM allows a subset of patents to be challenged administratively on a fast-track, without having to go to court. That hurts the patent trolls, but it hurts anybody without a phalanx of lawyers even more.
Real patent reform has three key parts:
1. Fix "obviousness." The courts didn't like the examiner affirmatively finding that something was obvious so turn it around and require the applicant to justify why anyone of average skill seeking the same result would not have found the same method. Require the examiner to affirmatively find that it isn't obvious. No justification = no patent.
If anybody asked to do X would have tried your approach and X itself doesn't supply the genius either then no patent should be granted. Nor should a minor tweak on something you or somebody else already invented receive a patent. There are too many "routine inventions" receiving patents.
2. A person of average skill in the art should be able to implement the technology from the contents in the patent. Start rejecting packets where that isn't true. Vague or stilted language in the application = no patent.
3. Patent duration should be from application, not from the grant. Effective protection starts with the application. You can't sue anybody until after the grant, but no one dares use the tech unless they're sure the patent won't be granted. That's been abused by delaying the final grant for years or even a decade.
Contracts of adhesion (unilateral contracts) are generally enforceable but they are "interpreted against the drafter" meaning that any ambiguity is interpreted in favor of the customer.
Click-through contracts are less likely to be enforceable than something bearing a physical signature. Add a little unconscionability and no court in the land would uphold that contract. If there even was a contract.
The magic word you're looking for, though, is Libel. These jokers deliberately published a false statement of fact to the credit reporting agencies with the intention of damaging the individual's reputation. That's a cha-ching if you take 'em to court.
However, this part of the story doesn't quite ring true for me. The credit reporting agencies don't like to accept reports without an SSN. Too high a risk they get applied to the wrong person. So how did folks paid via paypal get enough information to attach a complaint to the person's credit report? Maybe I just don't know enough about how the reporting agencies work but for darn sure there's nothing on my credit report from anyone who didn't have my SSN.
High temperature fuel cells. Right. You can indeed eliminate the platinum catalyst if you're willing to run the fuel cells at the better part of the temperature of molten lava.
And they'll slag those hard drives for you when you're ready to dispose of them too.
Of course, you can't really do that inside the cooled computer cabinet, so you don't actually eliminate power distribution within the data center that way.
Iridium data rates are 0.0024 mbps, up and down. On the plus side,they give you that data rate everywhere in the world.
You get 10 to 20 ms to the satellite orbiting 500ish miles away. To actually talk to anything on the ground, your signal is relayed to other satellites, down to Arizona and then across the Internet. If you gang a bunch of channels together to get a dialup-grade data rate (20ish channels yields the equivalent of a 56k modem), you can probably come in at half the latency of a geostationary satellite. Still pretty high though.
Correct, round trip packet time through a geostationary satallite link is a minimum of around 0.5 seconds.
There are other options that don't travel as far (e.g. Iridium is in LEO, 500 miles or so up) but AFAIK none are designed to Internet service at usable bandwidths.
If you're on call, it's appropriate to receive a notification in any situation. That's what it means to be on call, and a lot of young professionals are.
If the notification requires response, it's then appropriate to excuse yourself from the meeting. Just like you'd excuse yourself to hold a person to person side conversation while someone else was presenting.
You're right: I don't know enough about C++ templates to tell you that you're wrong.
I do know enough about the compiler behavior, though, to tell you that when the compiler generates warnings it tracks them back to a file and line number that the warning is about. Which means that once you've checked the warnings which originate from that file you could reasonably tell the compiler to stop reporting warnings just for that file. Without having to sacrifice useful warnings fro the rest of your code.
That mis-states what I asked for. I don't expect the compiler to tell me about optimizations made to memcpy on the assumption that in memcpy(a,b,4), a!=b.
On the other hand, if I write:
a=b; memcpy (a,b,20);
I want a warning when the optimizer decides it can delete the memcpy! Odds are I didn't mean for a to equal b. Silently eliminating the memcpy just makes my bug that much worse.
A classic example of this would be a for loop to set elements in a float array to zero. On most systems it can be replaced by a single, optimized memset
There's an unsubtle difference between transforming portions of my code and removing portions of my code. It's the latter that I want to see a warning about.
I read the article but I'm not following the author's point.
int a = 0x12345678; short *b = (short *)&a; b[1] = 0;
What's wrong with that? You take the location of a and assign it to a pointer to a short. b[0] now contains the first two bytes that comprise a and b[1] now contains the second two bytes that comprise a. Which index contains 0x1234 and which one contains 0x5678 (before b[1]=0 sets it to zero) depends on thee endianness of your machine but that's beside the point. It's very clear what this statement should do: act on the bytes which comprise the 32 bit integer and do it 16 bits at a time.
Slashdot Mirror
You understand github is an internet site while git is a piece of software, right?
Some have simply given up on trying to force the general-case IT policy to be useful. They "solve" the usability problem for their specific case by ignoring IT and using outside tools over which IT has no control.
"IT goon: Business comes first. We're here to support your business!"
"Me Great! We build software systems based on open source, so our developers need access to github."
"IT goon: Sorry, that's a file sharing site. Using it is against policy."
'Me: You said business first. Our business uses open source software from github."
"IT goon: That's right! Business first! Just no file sharing sites."
They aren't paid more because of their caution.
I didn't "lose" the job any more than I "lose" a defective computer when I throw it in the trash. Indeed it would be very hard to consider it a loss when six months later I was earning $10k more per year.
Nor did I put myself in any legal jeopardy. I'll spare you the lengthy analysis.
Best way to handle the problem? Burning bridges rarely is. But sometimes it has a moral righteousness that's hard to defy.
I wrote a memo laying out all the issues in layman's terms and proposing solutions. Then I gave it to my boss. A little while later with no further movement on the problem, I quit.
A year passed and the system was hacked. Publicly. Embarrassingly. Folks here on Slashdot asked what the sysadmins could possibly have been thinking. So, I published a copy of the memo I had written.
Your mileage may vary.
The technology is ready to retire. The impediment is regulatory -- without FCC oversight, delivery of last-mile infrastructure becomes thoroughly anticompetitive, a process which has repeated itself over and over again this past half century. POTS and twisted pair has been the last vestige of deregulation in the sector, to the detriment of the public and MUCH to the detriment of inventors and small business.
Why suppose there's something underhanded going on? It boils down to ignorance. The managers don't really understand the job and certainly can't articulate it, so they list the things they *can* understand... like the model numbers on the hardware, the programs the last guy used, and so on.
Yep. And several great free shards to play on these days too.
In theory, the patent examiner is educated in and well steeped in the field for which he examines patent applications. Einstein wasn't a patent clerk by chance alone -- the patent office requires folks with a strong technical background.
Wheels is a new concept. Inflatable tires on the wheels is a new concept. Springs on the axle is a new concept. Steerable wheels is a new concept. That just about covers it. Most of the rest of the improvements around the wheel and axle are derivative, unworthy of a patent. They don't advance the state of the art enough to merit giving their creators exclusive control.
That's what patents are about, right? They're supposed to encourage leaps of brilliance by making it practical for their inventors to profit off them. If there's no genius, just plodding iterative improvement, there shouldn't be a patent.
I understand that patents act on inventions, not ideas. That's because (A) ideas are rather nebulous and (B) patents must work. Hard to demonstrate that an idea actually works without building something that implements it. Nevertheless, the patent tends to protect the concepts that went in to it, not the precise implementation. If they only protected the precise implementation they'd be worthless.
JPEG was based on the notion that digital photographs were not pixel perfect to begin with, so why maintain that non-perfection perfectly if not doing so would make high compression rates possible. Acceptably lossy compression. This was a fresh, novel concept. The invention based on it deserved patent protection.
MPEG and MP3 took the then-well-known concept of lossy compression and ground on it until they came up with something that worked well for video and audio. After all, if the concept works for one media type, it should work for others, right? No unexpected originality, no unforeseen change to the state of the art, just grinding. Should not have been patentable.
See? Not terribly hard to look back and evaluate the state of the art in the day.
That's why I used the word "justify" and not the word "prove."
Right now the presumption is that if a half-hearted and formulaic search turns up no prior art then the invention is novel. I want to turn that on its head: the invention is presumed obvious until you explain why it isn't in terms folks in your field agree with. Who are the people in your field and why aren't they half a step behind you?
I've read some of the software patents. I have better than average skill in the art and I couldn't implement the software described from *any* of them. Not. A. Single. One.
I'm strongly in favor of patent reform, but CBM is not the answer. CBM allows a subset of patents to be challenged administratively on a fast-track, without having to go to court. That hurts the patent trolls, but it hurts anybody without a phalanx of lawyers even more.
Real patent reform has three key parts:
1. Fix "obviousness." The courts didn't like the examiner affirmatively finding that something was obvious so turn it around and require the applicant to justify why anyone of average skill seeking the same result would not have found the same method. Require the examiner to affirmatively find that it isn't obvious. No justification = no patent.
If anybody asked to do X would have tried your approach and X itself doesn't supply the genius either then no patent should be granted. Nor should a minor tweak on something you or somebody else already invented receive a patent. There are too many "routine inventions" receiving patents.
2. A person of average skill in the art should be able to implement the technology from the contents in the patent. Start rejecting packets where that isn't true. Vague or stilted language in the application = no patent.
3. Patent duration should be from application, not from the grant. Effective protection starts with the application. You can't sue anybody until after the grant, but no one dares use the tech unless they're sure the patent won't be granted. That's been abused by delaying the final grant for years or even a decade.
Contracts of adhesion (unilateral contracts) are generally enforceable but they are "interpreted against the drafter" meaning that any ambiguity is interpreted in favor of the customer.
Click-through contracts are less likely to be enforceable than something bearing a physical signature. Add a little unconscionability and no court in the land would uphold that contract. If there even was a contract.
The magic word you're looking for, though, is Libel. These jokers deliberately published a false statement of fact to the credit reporting agencies with the intention of damaging the individual's reputation. That's a cha-ching if you take 'em to court.
However, this part of the story doesn't quite ring true for me. The credit reporting agencies don't like to accept reports without an SSN. Too high a risk they get applied to the wrong person. So how did folks paid via paypal get enough information to attach a complaint to the person's credit report? Maybe I just don't know enough about how the reporting agencies work but for darn sure there's nothing on my credit report from anyone who didn't have my SSN.
High temperature fuel cells. Right. You can indeed eliminate the platinum catalyst if you're willing to run the fuel cells at the better part of the temperature of molten lava.
And they'll slag those hard drives for you when you're ready to dispose of them too.
Of course, you can't really do that inside the cooled computer cabinet, so you don't actually eliminate power distribution within the data center that way.
Iridium data rates are 0.0024 mbps, up and down. On the plus side,they give you that data rate everywhere in the world.
You get 10 to 20 ms to the satellite orbiting 500ish miles away. To actually talk to anything on the ground, your signal is relayed to other satellites, down to Arizona and then across the Internet. If you gang a bunch of channels together to get a dialup-grade data rate (20ish channels yields the equivalent of a 56k modem), you can probably come in at half the latency of a geostationary satellite. Still pretty high though.
Correct, round trip packet time through a geostationary satallite link is a minimum of around 0.5 seconds.
There are other options that don't travel as far (e.g. Iridium is in LEO, 500 miles or so up) but AFAIK none are designed to Internet service at usable bandwidths.
So, I can run lots of wire or I can run... enough gas pipes that errors are inevitable? When those servers blow up they'll really blow up!
Plus as I'm sure you know the platinum used as a catalyst in fuel cells has no major cost to it either.
Microsoft. Full of Win. As usual.
If you're on call, it's appropriate to receive a notification in any situation. That's what it means to be on call, and a lot of young professionals are.
If the notification requires response, it's then appropriate to excuse yourself from the meeting. Just like you'd excuse yourself to hold a person to person side conversation while someone else was presenting.
If I compile it with debugging symbols (in addition to optimization) the debugger somehow manages to find its way back.
You're right: I don't know enough about C++ templates to tell you that you're wrong.
I do know enough about the compiler behavior, though, to tell you that when the compiler generates warnings it tracks them back to a file and line number that the warning is about. Which means that once you've checked the warnings which originate from that file you could reasonably tell the compiler to stop reporting warnings just for that file. Without having to sacrifice useful warnings fro the rest of your code.
That mis-states what I asked for. I don't expect the compiler to tell me about optimizations made to memcpy on the assumption that in memcpy(a,b,4), a!=b.
On the other hand, if I write:
a=b;
memcpy (a,b,20);
I want a warning when the optimizer decides it can delete the memcpy! Odds are I didn't mean for a to equal b. Silently eliminating the memcpy just makes my bug that much worse.
A classic example of this would be a for loop to set elements in a float array to zero. On most systems it can be replaced by a single, optimized memset
There's an unsubtle difference between transforming portions of my code and removing portions of my code. It's the latter that I want to see a warning about.
I read the article but I'm not following the author's point.
int a = 0x12345678;
short *b = (short *)&a;
b[1] = 0;
What's wrong with that? You take the location of a and assign it to a pointer to a short. b[0] now contains the first two bytes that comprise a and b[1] now contains the second two bytes that comprise a. Which index contains 0x1234 and which one contains 0x5678 (before b[1]=0 sets it to zero) depends on thee endianness of your machine but that's beside the point. It's very clear what this statement should do: act on the bytes which comprise the 32 bit integer and do it 16 bits at a time.