Download the source from here (http://www.zdnet.com/eweek/stories/general/0,1101 1,2646732,00.html) for further insight into what's killing their JSP performance... they're not doing any database connection pooling at all! The connection is released after every request, so they have to renegotiate the connection before each page can render.
The ODBC COM object used in ASP, and (I think) the default PHP database connectors, both provide connection pooling. This is a massive component of the runtime performance that really needs to be addressed. There's plenty of good JDBC connection pooling code available for free; hopefully Sun is listening and will make it part of the java.sql package eventually.
Re:Breeding population
on
TigerCloning
·
· Score: 1
The lack of genetic diversity is definitely bad, but it is not a death sentence... current theory is that cheetahs survived a genetic bottleneck that may have been as small as just one pregnant female.
You will get a very strong founders effect from a sample this small, of course... meaning that any problems in the genotype will be magnified and exaggerated. But if there aren't any lethal recessives, you could establish a breeding population, albeit one with persistent problems such as vulnerability to disease.
I agree with the other replies that point out the fact that the MS argument rests on copyright law. Here's 2 cents on the trade secret discussion...
The formal definition of a trade secret is 1) any information, including any formula, pattern, compilation, program, device, method, technique, or process, that 2) provides a business with a competitive advantage, 3) is not generally known by a company's current or potential competitors, and cannot be readily discovered by them through legitimate means, and 4) is the subject of reasonable efforts to maintain its secrecy.
Now, in general, #3 has been interpreted to include reverse engineering. If the trade secret cannot be ascertained from the product (the formula for Coke is a common example) you can release the product without losing trade secret status. (this probably means that a reverse engineer of the protocol would invalidate the trade secret status right away).
However, and more germane to this discussion... "even when trade secrets become public in violation of the owner's rights, once they are public, they cease to be trade secrets." Damages may be available, but winning them requires victory at trial. (The Entrepreneur's Guide to Business Law, p. 437)
So it looks like the trade secret is out of the bag. The copyright issues are really the only ones that are relevant, long-term.
Slashdot Mirror
Download the source from here (http://www.zdnet.com/eweek/stories/general/0,1101 1,2646732,00.html) for further insight into what's killing their JSP performance... they're not doing any database connection pooling at all! The connection is released after every request, so they have to renegotiate the connection before each page can render.
The ODBC COM object used in ASP, and (I think) the default PHP database connectors, both provide connection pooling. This is a massive component of the runtime performance that really needs to be addressed. There's plenty of good JDBC connection pooling code available for free; hopefully Sun is listening and will make it part of the java.sql package eventually.
You will get a very strong founders effect from a sample this small, of course... meaning that any problems in the genotype will be magnified and exaggerated. But if there aren't any lethal recessives, you could establish a breeding population, albeit one with persistent problems such as vulnerability to disease.
(For more on the cheetah, see this natural history and this genetics discussion)
The formal definition of a trade secret is
1) any information, including any formula, pattern, compilation, program, device, method, technique, or process, that
2) provides a business with a competitive advantage,
3) is not generally known by a company's current or potential competitors, and cannot be readily discovered by them through legitimate means, and
4) is the subject of reasonable efforts to maintain its secrecy.
Now, in general, #3 has been interpreted to include reverse engineering. If the trade secret cannot be ascertained from the product (the formula for Coke is a common example) you can release the product without losing trade secret status. (this probably means that a reverse engineer of the protocol would invalidate the trade secret status right away).
However, and more germane to this discussion... "even when trade secrets become public in violation of the owner's rights, once they are public, they cease to be trade secrets." Damages may be available, but winning them requires victory at trial. ( The Entrepreneur's Guide to Business Law, p. 437 )
So it looks like the trade secret is out of the bag. The copyright issues are really the only ones that are relevant, long-term.