The author of the article presents CAFC as if it were uniformly a bad experience. This is not exactly the case. Prior to CAFC the state of patent litigation was very bad - judges with no experience in patent law were making arbitrary decisions that arbitrary and inconsistent, venue shopping was the order of the day, many really valid patents and important were being overturned, etc.
While CAFC has resulted in some quite bad decisions, it has at least brought some consistency. Now that there is this consitency it is up to Congress to correct the patent law to exclude or refine the scope of patent law.
If you have a problem, who do you turn to? A vendor who you paid a lot of money to for support, or a mailing list that may or may not get back to you? Most businesses won't accept that kind of uncertainty.
That's the reason for the existence on companies like Red Hat. You get the source code, the informal support AND the ability to get paid support and accontability. The best of all worlds.
If you have architectural problems with a C++ application that are affecting programmer productivity, it seems to me that the answer is to fix those problems through a refactoring process. This process can be controlled so it is relatively non-invasive and low impact. Not as sexy and harder to sell to management, but it is the way to go.
Switching languages won't help long-term. While a rewrite to a different language is one way to gain the refactoring you are after, it comes with a lot of undesirable side effects. Worse yet it doesn't instill the culture that software engineering really should include continuous refactoring as part of the development process ifo order to keep the hedges trimmed. So inevetably as time goes on that clean new design will accumulate cruft. Unlike some pundants state, source code does rust.
Switching from C++ to Java may marginally improve programmer productivity because memory management is easier in Java, but it seems I bet that it will take about 3 years before you reach the level of expertise needed to start seeing that benefit.
"the general consensus among Java devs curious enough to check out another language is that Ruby is far better"
I'd like to see some backup for that statement. I'm a Java dev that has taken the time to check out both Ruby and Python, and have reached the conclusion "don't quite your day job".
Say, I am curious as to what laptop you owned that had more than 4GB of RAM. It musta have been some nice machine. Or maybe you didn't realize that B stands for bytes, not bits.
*sheesh* I guess I was expecting too much for Slashdot.
4 GB dual core laptops are starting to appear. I would think that a pair of HP nc1440's could be set up to give all the power you need and still be about 10 lbs.
First, the fiber network that was laid out during the.com boom globally by companies like global crossing currently contains a lot of dark fiber. So that part is cheap.
The routers / switches / head ends / last mile ARE NOT CHEAP. Verizon is laying FIOS and it is taking them years at a cost of many billions.
Do you have any idea how hard it is to wire in Manhattan? The existing infrastructure is deep and old, and the cost of opening a street is huge. Costs are much lower in the burbs.
There are places in the US where you can get fiber to the home (Verizon FIOS); and even Cable with DOCSIS 1.1 is capable of the bit rates the Okinawan author is describing. My cable provider (Optimum Online) offers a 2nd tier service at $65 / mo that is 30 Mbs down. They are talking about a 3rd tier that will be 100 Mbps which I assume will be DOCSIS 2.0.
384 kbps is a flea bite. We are talking on demand delivery of 1080p HD and NOT at some artifact-laden compression rate either. Think more like 20 MBps, i.e. 50 times your broadcast rate.
The point is not the technical ability to create the infrastructure, what is required is well known. The question is who is going to pay for it. Personally I have no problem using the existing models - in some areas of the country it already seems to be working. CableVision Boost is 30MB/sec, Verizon FIOS is similar. A little work on the QOS and you are there. The end user might have to pay $60-70 per month for a pipe that robust, but I think it is still better than having a tiered service.
The fiber was aid to prevent having to dig up the streets additional times when capacity needs increased. As fiber that is all it is - there are no switches, routers, management systems, last mile cabling, etc. etc.
All of the ancilliaries are far more expensive than the used glass fiber itself.
The articles that your search results point to generally contain warnings that text replacement techniques cause accessability problems. And as far as dynamic image creation, I dare you to try that on a server that is under any kind of significant load.
Thanks for the article. It sounded like the authors felt there wasn't enough evidence yet to make a firm conclusion yet because the excess seems to be decreasing with time (or improved study methods?).
Slashdot Mirror
The author of the article presents CAFC as if it were uniformly a bad experience. This is not exactly the case. Prior to CAFC the state of patent litigation was very bad - judges with no experience in patent law were making arbitrary decisions that arbitrary and inconsistent, venue shopping was the order of the day, many really valid patents and important were being overturned, etc.
While CAFC has resulted in some quite bad decisions, it has at least brought some consistency. Now that there is this consitency it is up to Congress to correct the patent law to exclude or refine the scope of patent law.
If you have a problem, who do you turn to? A vendor who you paid a lot of money to for support, or a mailing list that may or may not get back to you? Most businesses won't accept that kind of uncertainty.
That's the reason for the existence on companies like Red Hat. You get the source code, the informal support AND the ability to get paid support and accontability. The best of all worlds.
Yup it is reliably and dependably CRAP.
This whole friggin site is "-1: Flamebait".
I just bought an HP laptop that was FedEx'ed directly from Kunshun China to my door.
Another desperate geek cry for help.
If you have architectural problems with a C++ application that are affecting programmer productivity, it seems to me that the answer is to fix those problems through a refactoring process. This process can be controlled so it is relatively non-invasive and low impact. Not as sexy and harder to sell to management, but it is the way to go.
Switching languages won't help long-term. While a rewrite to a different language is one way to gain the refactoring you are after, it comes with a lot of undesirable side effects. Worse yet it doesn't instill the culture that software engineering really should include continuous refactoring as part of the development process ifo order to keep the hedges trimmed. So inevetably as time goes on that clean new design will accumulate cruft. Unlike some pundants state, source code does rust.
Switching from C++ to Java may marginally improve programmer productivity because memory management is easier in Java, but it seems I bet that it will take about 3 years before you reach the level of expertise needed to start seeing that benefit.
God yes. FC is a PITA because the supported lifetime of a release is so short. CentOS is probably the best free solution available today.
If finalize was good enough you wouldn't have GenericServlet.destroy().
"the general consensus among Java devs curious enough to check out another language is that Ruby is far better"
I'd like to see some backup for that statement. I'm a Java dev that has taken the time to check out both Ruby and Python, and have reached the conclusion "don't quite your day job".
Say, I am curious as to what laptop you owned that had more than 4GB of RAM. It musta have been some nice machine. Or maybe you didn't realize that B stands for bytes, not bits.
*sheesh* I guess I was expecting too much for Slashdot.
4 GB dual core laptops are starting to appear. I would think that a pair of HP nc1440's could be set up to give all the power you need and still be about 10 lbs.
That is nothing. How about Java with destructors and a sizeof method.
A WiFi card and a copy of MS Flight Simulator and YOU, yes YOU are in charge.
BWAHAHAHAHA.
First, the fiber network that was laid out during the .com boom globally by companies like global crossing currently contains a lot of dark fiber. So that part is cheap.
The routers / switches / head ends / last mile ARE NOT CHEAP. Verizon is laying FIOS and it is taking them years at a cost of many billions.
60 GB / month? LOL. Delivery of 1080p HD will consume that in about 10 hours.
Do you have any idea how hard it is to wire in Manhattan? The existing infrastructure is deep and old, and the cost of opening a street is huge. Costs are much lower in the burbs.
There are places in the US where you can get fiber to the home (Verizon FIOS); and even Cable with DOCSIS 1.1 is capable of the bit rates the Okinawan author is describing. My cable provider (Optimum Online) offers a 2nd tier service at $65 / mo that is 30 Mbs down. They are talking about a 3rd tier that will be 100 Mbps which I assume will be DOCSIS 2.0.
384 kbps is a flea bite. We are talking on demand delivery of 1080p HD and NOT at some artifact-laden compression rate either. Think more like 20 MBps, i.e. 50 times your broadcast rate.
The point is not the technical ability to create the infrastructure, what is required is well known. The question is who is going to pay for it. Personally I have no problem using the existing models - in some areas of the country it already seems to be working. CableVision Boost is 30MB/sec, Verizon FIOS is similar. A little work on the QOS and you are there. The end user might have to pay $60-70 per month for a pipe that robust, but I think it is still better than having a tiered service.
The fiber was aid to prevent having to dig up the streets additional times when capacity needs increased. As fiber that is all it is - there are no switches, routers, management systems, last mile cabling, etc. etc.
All of the ancilliaries are far more expensive than the used glass fiber itself.
That's why you uxe multifactor security.
The articles that your search results point to generally contain warnings that text replacement techniques cause accessability problems. And as far as dynamic image creation, I dare you to try that on a server that is under any kind of significant load.
Give it up. Text as images is a WTF.
domestic wiretapping, indefinite detentions and torture, acts of aggression and brinkmanship against sovereign nations, lying to the U.N.
Sounds like the manifesto of pretty much any superpower.
and managing to convince the majority of Americans that this was all incidental
Judging by Bush's approval ratings I don't think your point here works.
Yeah, putting any kind of energy into a cell including heat can affect cell chemistry. Big deal. That is NOT the same as creating a tumor.
Thanks for the article. It sounded like the authors felt there wasn't enough evidence yet to make a firm conclusion yet because the excess seems to be decreasing with time (or improved study methods?).