I don't know how exactly a $7000 3D modeling/animation/rendering package is a replacment for flash?
Mental ray (the renderer integrated into XSI) can't even output to SWF, or any vector format for that matter, except for postscript from its contour shaders, which are damn rarely used.
Novell has not thought about any of this stuff at all.
As I said to the other poster: The BDB code may be covered by patents which Oracle could choose to exercise, thereby preventing the BDB code from being used despite its liberal license.
Or, you know, PostgreSQL (http://www.postgresql.org/) which is a far better database than MySQL by most measures, and much closer to Oracle in terms of features and performance (and even PL/SQL compatibility to a limited extent in the form of PL/PGSQL).
This seems like it fits with their other purchases if their strategy is to kill of the commercial incarnation of MySQL. First the InnoDB purchase threatened MySQL's commercial business being the primary transaction based backend, and now BDB too is threatened.
Can MySQL license the code (and any patents covering it) to continue commercial MySQL sales/support?
Actually Linux is far more sophisticated than that.
Linux can efficiently handle NUMA and hotplugging CPUs at the same time. Memory locality and extremely scalable TLB management also help make it one of the fastest and most scalable systems out there. Linux can handle 128 CPUs out of the box as it were (vanilla kernel), and with patches from SGI up to 512 CPUs. I think only Solaris/SPARC can beat it with such high-end configurations.
I used to use cygwin heavily (not just for perl), but with recent versions, it has gotten very very slow. For many of my scripts, ActivePerl is more than 100% faster than cygwin.
Another thing is that cygwin has a number of bugs when working with sockets (e.g. select() uses 100% CPU).
For much better results, use Coverity on your code regularly.
Re:Balkanization
on
Demise of C++?
·
· Score: 4, Informative
In many cases, a good C++ compiler will produce better code if the C sources are C++ clean, due to the extra type-safety in C++ the compiler is safely allowed to make more assumptions leading to better optimized code.
That work will still be done (perhaps not by the same people though) - the vanilla kernel goes to 128 already. 512 is not too far off.
AMD's next generation CPUs will essentially be a bunch of Opterons with a new generation of hypertransport to interconnect them. This will give commodity clusters machines with 16 or 32 CPUs, then scalability work will accelerate in the OSes.
512 is impressive, but not too difficult to attain given the right resources.
I just ran diskeeper's analysis on one of the machines, and the average number of fragments per file is 1.02. This is on a machine with 3 harddisks, totalling 760GB of disk space (479GB of it is used).
The most fragmented file is 4.2GB and has 12 fragments.
Because they don't charge $50000 per CPU per year, and thus can afford fewer people working full-time on the code.
Even with full support, there are plenty of companies who will support it and the competition keeps the prices low... unlike Oracles.
Oracle is great if you really NEED that level of scalability, and those that do, already use Oracle or DB2 as no OSS solution meets their needs. But if you are not handling 100 million transactions per hour or working with terabyte datasets, then PG is just fine, and costs NOTHING.
My laptop does not "peg out", nor have I ever lost data due to hardware problems.
Reiser4 is already better at almost everything (when it finally hits the disk, the latency is higher than some other filesystems).
Granted, my workflow is very different from a "normal" user, but it works amazingly well for me. Reiser4 has saves me literally hours everyday when working with huge numbers of small files. A task for which I used Berkeley DB in the past. Not only is reiser4 more than twice as fast, it rarely touches the disk. It does use a bit more space though.
These kind of changes can only be made with changes to the driver model.
They can't make it impossible to do this kind of thing on 32-bit versions of Windows (without breaking A LOT of drivers and programs), but on all 64-bit Windows versions this is already impossible.
MS does not inconvenience its users to defrag. I have been running several Windows machines for many years (still running Win2K on some of them), and I have honestly never defragged them.
Two are running Battlefield 2 servers 24x7, one is running a web server and a database server to aggregate statistics on players. All have been in service doing this kind of thing for over two years.
I am not the GP poster, but I regularly use this feature in my applications. It is in fact very useful, to the extent that I wish it were available on Linux (if nothing else, than for the sake of Samba).
This is probably a good way to learn Oracle's nuances and help you on your way to become a DBA for Oracles bigger stuff.
This is not in response to OSS (since they target two distinct markets: MySQL low-end, Oracle/DB2 high-end). Rather, this is likely in response to the MSDE (and the SQL Server 2005 version thereof), and the free DB2 products.
This version is licensed for commercial use (unlike all the other free stuff oracle gives away), so they are probably hoping people will develop for this and eventually migrate up to the higher-end products offered.
MS Access is "supposed" to be the interface to it. Access is actually a very nice interface and handles all of the SQL Server (2000) features extremely well, from simple views to complex stored procedures and macros.
I haven't yet had a chance to play around with SQL Server 2005, but I understand that the entire.NET framework and runtime has been deeply integrated with it, and as such, all.NET languages can be used to create first class database objects like stored procedures, and even custom data types.
Slashdot Mirror
How exactly is XSI a replacement for Flash?
I don't know how exactly a $7000 3D modeling/animation/rendering package is a replacment for flash?
Mental ray (the renderer integrated into XSI) can't even output to SWF, or any vector format for that matter, except for postscript from its contour shaders, which are damn rarely used.
Novell has not thought about any of this stuff at all.
I think he/she was referring to http://it.slashdot.org/article.pl?sid=05/11/18/065 6204&tid=102&tid=221&tid=218
There is no shortage of companies supporting PostgreSQL: http://www.postgresql.org/support/professional_sup port_northamerica
Besides, who would you trust not to disappear next week, MySQL AB, or say, Sun?
As I said to the other poster: The BDB code may be covered by patents which Oracle could choose to exercise, thereby preventing the BDB code from being used despite its liberal license.
Are you sure the BDB code is not covered by patents now owned by Oracle? What happens if they should choose to exercise them?
Or, you know, PostgreSQL (http://www.postgresql.org/) which is a far better database than MySQL by most measures, and much closer to Oracle in terms of features and performance (and even PL/SQL compatibility to a limited extent in the form of PL/PGSQL).
This seems like it fits with their other purchases if their strategy is to kill of the commercial incarnation of MySQL. First the InnoDB purchase threatened MySQL's commercial business being the primary transaction based backend, and now BDB too is threatened.
Can MySQL license the code (and any patents covering it) to continue commercial MySQL sales/support?
Actually Linux is far more sophisticated than that.
Linux can efficiently handle NUMA and hotplugging CPUs at the same time. Memory locality and extremely scalable TLB management also help make it one of the fastest and most scalable systems out there. Linux can handle 128 CPUs out of the box as it were (vanilla kernel), and with patches from SGI up to 512 CPUs. I think only Solaris/SPARC can beat it with such high-end configurations.
If there's a non-browser, non-mozilla XUL app out there, I haven't heard of it.
Activestate Komodo.
That said, your point still stands... there are almost no non-mozilla apps built on XUL.
I used to use cygwin heavily (not just for perl), but with recent versions, it has gotten very very slow. For many of my scripts, ActivePerl is more than 100% faster than cygwin.
Another thing is that cygwin has a number of bugs when working with sockets (e.g. select() uses 100% CPU).
Splint is absolutely useless with C++.
For much better results, use Coverity on your code regularly.
In many cases, a good C++ compiler will produce better code if the C sources are C++ clean, due to the extra type-safety in C++ the compiler is safely allowed to make more assumptions leading to better optimized code.
It is the single best open source project around.
I am going to be modded troll for this I am sure, but I have to say it anyway.
How many OSes have ever been written from scratch?
I can think of only 3, none of which has even 0.1% market share. In fact, Plan9 is the only one of them alive.
What is the big deal with bashing microsoft for copying ideas from people?
Isn't that what OSS is built around, copying good ideas?
That work will still be done (perhaps not by the same people though) - the vanilla kernel goes to 128 already. 512 is not too far off.
AMD's next generation CPUs will essentially be a bunch of Opterons with a new generation of hypertransport to interconnect them. This will give commodity clusters machines with 16 or 32 CPUs, then scalability work will accelerate in the OSes.
512 is impressive, but not too difficult to attain given the right resources.
See my reply my post. I was making a joke... slashdot happened to eat the sarcasm tags around it :)
Sorry, there should have some been sarcasm tags there, but /. ate them (dont flame me) :)
I just ran diskeeper's analysis on one of the machines, and the average number of fragments per file is 1.02. This is on a machine with 3 harddisks, totalling 760GB of disk space (479GB of it is used).
The most fragmented file is 4.2GB and has 12 fragments.
This machine does NOT need defragging.
Because they don't charge $50000 per CPU per year, and thus can afford fewer people working full-time on the code.
Even with full support, there are plenty of companies who will support it and the competition keeps the prices low... unlike Oracles.
Oracle is great if you really NEED that level of scalability, and those that do, already use Oracle or DB2 as no OSS solution meets their needs. But if you are not handling 100 million transactions per hour or working with terabyte datasets, then PG is just fine, and costs NOTHING.
My laptop does not "peg out", nor have I ever lost data due to hardware problems.
Reiser4 is already better at almost everything (when it finally hits the disk, the latency is higher than some other filesystems).
Granted, my workflow is very different from a "normal" user, but it works amazingly well for me. Reiser4 has saves me literally hours everyday when working with huge numbers of small files. A task for which I used Berkeley DB in the past. Not only is reiser4 more than twice as fast, it rarely touches the disk. It does use a bit more space though.
These kind of changes can only be made with changes to the driver model.
They can't make it impossible to do this kind of thing on 32-bit versions of Windows (without breaking A LOT of drivers and programs), but on all 64-bit Windows versions this is already impossible.
MS does not inconvenience its users to defrag. I have been running several Windows machines for many years (still running Win2K on some of them), and I have honestly never defragged them.
Two are running Battlefield 2 servers 24x7, one is running a web server and a database server to aggregate statistics on players. All have been in service doing this kind of thing for over two years.
Never defragged, still running as good as new.
I am not the GP poster, but I regularly use this feature in my applications. It is in fact very useful, to the extent that I wish it were available on Linux (if nothing else, than for the sake of Samba).
This is probably a good way to learn Oracle's nuances and help you on your way to become a DBA for Oracles bigger stuff.
This is not in response to OSS (since they target two distinct markets: MySQL low-end, Oracle/DB2 high-end). Rather, this is likely in response to the MSDE (and the SQL Server 2005 version thereof), and the free DB2 products.
This version is licensed for commercial use (unlike all the other free stuff oracle gives away), so they are probably hoping people will develop for this and eventually migrate up to the higher-end products offered.
MS Access is "supposed" to be the interface to it. Access is actually a very nice interface and handles all of the SQL Server (2000) features extremely well, from simple views to complex stored procedures and macros.
.NET framework and runtime has been deeply integrated with it, and as such, all .NET languages can be used to create first class database objects like stored procedures, and even custom data types.
I haven't yet had a chance to play around with SQL Server 2005, but I understand that the entire