Every time when I read articles such as these I wonder: why oh why do Democrats in the USA have such a hard time selling the truth to the public? I mean: the current Bush administration has piled misleading and disputable decision on decision, and the American public seems to feel it is all right. How come? Why aren't the Democrats using these obvious limitations on the freedoms of the American citizens to rally the public so they'll support the Democrats and elect a better government in place which will overturn such decisions like a concentration of media companies?
You can come to two conclusions: 1) The Democrats are also after the same money from these media companies as the Republicans are, which in fact makes the USA's democracy rather dead: there is no real choice for Joe Sixpack, the two parties which matter are NOT serving the interests of the people 2) The Democrats are incapable of fighting Bush effectively. Which also makes the USA democracy rather dead, because the general public doesn't KNOW there is an alternative to 'Bush'. When Bush gets the concetration of media in place, and the holders of these media on his side (which seems to be the case) he controls EVERYTHING and the republicans can stay in power, well... forever.
If the republican party would exist in The Netherlands, Europe, they would get at most 2 seats in the 150 seat parlement, roughly guessed. Not because we're all 'stinking liberals', but because we tolerate less a government that thinks of big $$$ first and the interest of the public second.
(To the USA citizens: as a European I see you as a group of people who thought that a president who nailed his intern with cigars should be impeached and a president who started a very expensive war under false intelligence in a time where jobs dissapear very quickly should stay in his office and should stay popular. Think about that for a second.)
What if he links with a BINARY library he has found on his machine and came with the OS on the machine (i.e. the api offered by the OS) ? Developers like linking with binaries since that shortens compile time.
Now, he links with a binary but has to embed a license to his own sourcecode because he links with a library, however he doesn't include any sourcecode from the binary (say, he uses win32's LoadLibrary()).
_THAT_'s the viral part. He calls into a library and because he does that he has to make his own code which he owns copyright of open because someone else who doesn't own the copyright of that work said so.
I'm 100% sure this will not hold in court in The Netherlands or Germany.
Your complaints about C++ are perhaps legit, they have nothing to do with.NET. If your bosses are too stupid to allow you to work with a C++ compiler that works like you want to, that's a problem at your employer, not with MS or.NET
I'm happy C# is available and.NET matures over time, so I can ditch the VB/VC++/COM crap for the n-tier apps I develop all day.
With NT, MS released 3 bad SP's: 2, 4 and 6.:). On WIn2k though, sp2 was very good, sp3 was crap. It has to be so that SP4 is great;) (it works ok on the testserver that's routing my internet conne**NO CARRIER**
Am I the only one who thinks the shown application windows could be a lot smaller when the controls were positioned more elegantly and efficiently (screenspace wise) and without a lot of space between them?
Perhaps people who are not used to operate a computer with gui screens are confused with a lot of information on their screen but this way you need a lot of dialogs to display all the information simply because the dialogs can't contain a lot of info: most space on the dialog is simply not used. OR, apple is hiding a lot of info so you don't need a lot of dialogs because the total amount of information to show is not a hell of a lot.
Either way, the usage of screenspace like how its done in these screenshots is IMHO a waste of pixels. It can be done a lot more efficient. (look at the process viewer. A lot of space is wasted below the list by a total inefficient ordering of controls, so less space is reserved for the real data: the process list).
Early X11 applications also had this example of IMHO bad interface design.
Please, answer these questions: - WHO has seen both the SCO sourcecode and the linux code? - WHO has seen the contracts IBM signed with SCO? - WHO has seen the patents SCO has related to Unix? - WHO has read and understood the complete class action suit text filed by SCO?
If you've answered on all questions "I have", please step forward and enlighten us with your vision on the matter. If you HAVEN'T answered on all questions "I have", please, stop hammering texts and start thinking about a similar case when a given company X is suspect of a GPL-violation.
Re:Where Sun Excells
on
Sun's Last Stand
·
· Score: 2, Insightful
The reason Apple and Sun hardware/software combinations are superiour in stability, is due to the fact that they are made to support each other, unlike in a windows enviroment, where you have a mix and match of hardware, and software drivers that bring in many inknowns sometimes.... which also counts for Linux. You are saying that Linux is unstable too?
Also, it's nice that Sun has very high end servers that keep running when you have to change cpu's or memory, but everybody knows that Sun can't exist by solely these clients. It's the bulk that makes or breaks you and Sun clearly looses that bulk to cheap systems with cheap OS-es because these systems are ALSO very good. Win2k server or Win2k3 and a dell server will hardly crash, a dell box with Linux or a preconfigured linux server from IBM will also be rocksolid. Why buy sun in those situations?
Oh, and if even a short period of downtime costs you a HELL of a lot of money, you won't buy Sun, you will buy stratus systems or tandem systems (HP has given them another name, same hw). Sun looses also in that area. It's a loose/loose situation for Sun. At the uni, I loved the sun boxes and SunOS, but today I can hardly imagine using one when the alternatives (win2k/xp/win2k3 or linux/bsd) are cheaper and as good as Sun's hw.
About a month ago my mailserver started to receive a lot of hotmail connections from the range 65.54.*.*., guess what the bay range servers inside hotmail.com. I contacted abuse@hotmail.com, tried a few times to convince the drone at the other end that my mailserver was receiving a connection from a hotmail server every 20 seconds, but they didn't understand it. I mailed mailserver logs, explanations, links to threads about this on usenet, no clue. After a while I simply blocked all hotmail servers from my server. It's really weird that they have people on the abuse staff that do not understand what 'abuse' means or how to prevent it.
A week ago I removed the block to check if things had changed. To my suprise, no connection since. Apparantly MS has solved this problem finally (that is: installed the WebDAV patch that is what, 2 months old?).
Yes, the patches themselves. People don't install them because they break critical production software which must not be broken. That critical production software NEEDS a patch, f.e. it has a security hole, or runs on top of an OS that has a security hole. THerefor it IS already broken and thus needs patching. THere is NO excuse for not patching your software, like there is also no excuse for having security holes in your software.
Your E=mc^2 idea is neat but 'c' also relies on time, which means that the equation relies on time, and thus doesn't solve the 'it relies on timemeasurement' problem.
I haven't tested this (I don't own an FX nor an ati card with the balls to run it) but on the VE3D forums someone mentioned this to make Dawn naked:
Well for you sickies looking for the nude patch, no need. Just change the fairy.exe file name to one of two things. Quake3.exe will make her nude, and 3dmark03.exe will make her nude and wingless.
On some other forum (here) they talk about renaming some file to get rid of her erm.. cloths/leaves.:) Enjoy.
I agree with you that the parent should probably read some more books, but Sun also should be looking more to their worst enemy Microsoft how easy system administration CAN be done. I'm not saying Win2k is easy to admin when you set up a large forest with AD, but applying patches etc, it's darn easy. Sun (and other companies as well) should learn that even professionals do not LIKE it when their job is HARD because the 'tools' they have to work with are very low level and effectually do not help the professional a lot. Tools should help the user, not work against him/her.
Ok, apple invests in software too, but their hardware is also designed mostly by IBM/Motorola, nvidia and gee... Intel (PCI/AGP). Dell sells hardware THEY designed, using common parts, but THEY designed it. Apple does the same.
However, WHY would anyone buy sun? They only company I can think of is... Oracle. IBM can't even buy it without severy interference by the anti-trust people in the government, Dell will not buy it since it will not add anything to their portfolio: their high end systems are designed around Microsoft's software, not unix.
However, Oracle's database is mostly run on Sun hardware, or better: when you buy a Sun E* server and decide to run a database on it, 10 to 1 it will be Oracle. It does make sence in a way for Oracle to buy Sun so they can sell the complete package to high end customers: hardware AND software.
And I mean: really useful. Plus your application has to be of a quality that is only found in commercial applications. Then you have a combination in your hands that is attractive to an enormous amount of people: quality software that is useful for a cheap price: free. There are a lot of open source projects that do not meet one or more of these requirements and are therefor not succesful. A lot of developers think they have made a great tool, but do not understand that the tool they made is only useful to them because they understand it, i.e.: the quality is not up to par, and an average user will not be able to use it as it is ment, which will result in a request from the user for another tool which seems to do the same: the user will look elsewhere.
I've made 2 open source tools: DemoGL and LLBLGen. DemoGL was a library for win32 which let you develop OpenGL effects easily in C++. It was not that useful to a lot of people, since most people didn't need it and got things started with a few tutorials from NeHe. LLBLGen on the other hand is very succesful, it's a DAL generator for sqlserver/.NET/C#/VB.NET. The reason why it's so succesful (according to the vast amount of reactions I received from users) is that it's simple to use, yet does what the user wants and costs nothing. It comes with the source so if you want to adjust it you even can do that. (This is what I understand most users found about the tool).
The fun thing is: LLBLGen was a pet project, a project to learn C#:) it wasn't ment to be released as a worldwide succes, and DemoGL was, but failed.
I was reading those lines, copying it over to post it, and saw your posting, so posting it again would be really redundant, but it is indeed pretty funny, and that poor Ari is receiving his 100 penis enlargement offers a day really soon I guess;)
The "NT kernel series" sucks when you try to port Unix-style thread or process per client model server software to it, because of the process limit I discussed and the VMS-like heaviweight processes. The ideal # of concurrent executing threads on 2K3 is one per processor, SQL Server and Exchange are modeled on this. Kernels do not suck. Kernels, if properly written (but not properly written kernels are hardly ever used in OS-es, except the win9x 'kernels'), do what was designed up-front. They have specs. If you want to use something, you read the specs and see if what you want to use can perform what you want it to perform so the usage is succesful. If you want to use an NT kernel as a monolithic UNIX kernel, and try to schedule processes instead of threads, it won't work efficiently. NT-based OS-es do not use shared memory for their processes. That's why they use threads, because these can use shared memory. When you want to use processes on an NT-kernel and you see that the performance is poor, it's not the fault of the kernel, it's the fault of the programmer who sux big time and doesn't know a thing about what he/she is doing.
Also your remark about 'VMS-like Heavyweight processes' is of the same quality: an NT-kernel based OS works differently than UNIX. You also do not eat soup with a fork, do you? even when the fork performs brilliantly when eating potatoes!
One thread per processor is the optimum? Whoa:) Tell me, how would it be possible to execute say 4 threads simultaneously on a processor (without HT) ? SQLServer and Exchange are modelled on the 'thread-scheduling' model, pure and simple. SQLServer's kernel (yes it has a kernel too) even uses NT-fibers, a part of the OS which can boost threads (and other threads are suffering on this). If Win2k3 has better thread-scheduling and less process-scheduling, SQLServer and Exchange will benefit from this, but also ALL threaded applications will benefit from this (can you say: IIS? aspnet_wp.exe ?)
Windows server performance is top notch, the kernels are tuned excellently, and with each server release they get better and better. I also found your resources-remark rather amusing. You are refering to the handle-count in each process. So you think it is a good thing, a process will open (2^32)-1 objects and thus has that much handles open? I think that's a bad thing. An open handle means you have an open resource, and are keeping it open. Not a lot of resources qualify for that, most resources get opened, used and are closed right after they are used. That's good programming practise.
I'm very happy for you that you think your monolithic, hardware-specific kernel is the way of the future. I also hope that you WON'T understand that how a kernel works internally is not that important, it's how the OS it is part of runs the software YOU want to run and use. If you WILL understand this, you will regret your swap. Until then, enjoy the ride, while it lasts.
MS DTC (Microsoft Distributed Transaction Coordinator) manages the transactions which span more processes, across machines. You can see this as a P2P program, it communicates with the other MS DTC instances on other machines if parts of a distributed transaction are executed there. This means that a transaction started in a client application which rolls through tiers down to the sqlserver machine, both on the client and the server a MS DTC instance is running and coordinating the part of the transaction running on their boxes. It coordinates commiting the distributed transaction acros all the servers enlisted in the transaction.
If the client started the transaction on the server by an RPC, only on the server the MS DTC is started and only on the server there is a COM+ transaction, not on the client. However, I don't think that was the case. I think, and if I read your post correctly, you assumed this too, that on the clients software was running which initiated the transaction locally and which was rolled down to the server and back and which was coordinated by the transaction coordinator tuxedo or COM+ (which uses DTC). So client-server communication IS necessary in both situations.
So to me the end result counts: what kind of setup is faster (because both need client - server communication) and the SQLServer setup is very fast for the money it costs.
In Iraq US soldiers are putting their life on the line because they think they fight for 'freedom'. However by the day that is less and less true. Where will it end? In the situation where the New Iraq will 'liberate' the US from their occupiers so the people can live in 'freedom'?
All this talk about electronical gadgets... millions of people in Iraq suffer every damn day because there is no water or too less water or only dirty water and not enough food.
And the US of A can only talk about cellphones. I don't know but then you really are out of touch with reality. Give the iraqi people food and water and let them rebuild basic infrastructure first. This will cost a few years. After that the debates about the unnecessary gadgets can begin.
World + Models should be rendered in 1 pass.
on
Cheating Online Gamers
·
· Score: 4, Informative
Now, the worlds in most FPS games are rendered first, then the models and other entities are rendered, using clipping/depth buffer info of the world. A lot of engines use 2 different render routines to do this: the world is mostly static and uses a different routine than the model renderer.
THe result is, that when you 'patch' the world renderer so that it f.e. renders wireframes instead of solid polys (in OpenGL based engines this is 'not that hard', you just change the value passed to glBegin()) the models still are rendered solid, plus because most renderers for models rely on the depthbuffer filled during the world rendering, the models close to corners are fully rendered, since the depthbuffer is empty. So you can easily 'see' the models close to corners. If you also 'patch' the model renderer for not doing world clipping, you will see ALL models rendered in your window.
This can't be done if the world + models use a single render routine, i.e.: model polygons and world polygons are packed together as THE set of polygons to render, then the single render routine will eat these single pack of polys to render. If you patch the routine for wireframing, you will see the models also wireframed, if you patch out the world clipping, you will get the complete world in your window, not what you want.
I think in future game engines there will be a merger between world + model polygon sets, because worlds are more and more modfyable in game by the player, which in the end requires that the modifyable parts are 'models' too. However games based on the current crop of quake * engines will keep on suffering from this.
"We" do not have troops there..
on
Strike on Iraq
·
· Score: 2, Interesting
I'm not a citizen of the US nor UK. The 'our' in 'support our troops' is not for everybody a given since 'their' troops are not there at the moment (thankfully).
People who volunteered to be in an army should face the consequences when their commanders think they should start playing cowboys and indians in some desert. Why should I support those people? Because they are 'fighting for freedom' ? 'Giving their lives for our lives' ? Freedom is degrading all over the world especially in the USA. If the fight for freedom should be held somewhere it should be in the USA and against its own government, not somewhere in the Middle East.
I'm all for freedom and peace for the iraqi people, as I also am for freedom and peace for the people in the occupied palestine territories (Israel occupies them for 30 years now, despite a UN resolution (242) which says Israel should retreat (resolution is 30 (!) years old)), as I am also for freedom for all those African citizens who suffer from war day after day for decades in a row.
However I'm against hypocrisy and a single war against Iraq solely to 'bring freedom' while ignoring all those other countries where people suffer day after day (for a much longer period!). This war is wrong, the consequences will be hard for a lot of people, the deaths of possible thousands of civilians will be caused by US lead troops.
Now, tell me, why should I support these 'freedom fighters' in killing innocent people?
You claim a lot of things, name percentages etc, but where did YOU get those figures and how did YOU test the products you mention to back up YOUR claims?
I don't see any link in your article. (like the one about linux versions running faster (running WHAT faster? An endless loop?) on the same hardware than windows).
You can argue Postgres, but I've never run into a case where I couldn't work around those features that haven't been implemented in MySQL yet. Really? Now, how did you implement relational integrity? You added a lot of queries to perform the relational integrity? You can't! You can't add relational integrity by code ON TOP of a database system. Especially not in a multi-user environment. Especially with a database system which doesn't support transactions. So you can first test with an sql statement, and based on that result, perform some actions, but those 2 things are not 1 atomic unit. So just after the test-statement is done, another user could be messing with your data. Oh, you are not going to tell me you lock primary key rows when inserting a foreign key, are you?
The one thing I can't stand is when someone suggests: "I can't afford Oracle, so lets' go with a MSSQL database." That's like, I can't afford a space shuttle, and a ferarri isn't good enough for me, so I'm going to buy this million dollar llama instead because 1000 marketing agents can't be wrong, right?" It has all the same feautres as Oracle, it's just that the features in Oracle WORK. Whoa, sherlock! You are saying that features in MSSQL don't work but they do in Oracle? Please, can you give me a list of them, so I can use that list in my daily work, so I can avoid features that are not working in MSSQL.
MSSQL is worlds fastest database AND supports true ACID compliance, stored procedures, views, partitioned views etc. etc. It also costs a hell of a lot less than Oracle. Oracle does have more features though. But... why wouldn't it be a reasonable choice? Because the features don't work but they do in Oracle?
Have you ever worked with a true database management system? I have the fear you haven't
All nice, but MySQL doesn't offer a lot on the area of relational integrity (constraints, triggers). Also it doesn't support subqueries so you are 'joining' a lot.
The 4 CRUD actions are not the total picture you need, for even a small database driven webapplication. And if price is a problem, postgresql is far better: it offers relatonal integrity, sports subqueries and more. MySQL is _NEVER_ the best option. NEVER.
Grow up, child! Even if it was sarcastic or in a wicked way ment to be 'funny', it's too pathetic for words. Microsoft is like any other company which wants to make money. Employees who work at Microsoft, do that because they get paid for what they do there, like people at Sun or IBM (or Red Hat).
Just because MS mistreats some of its customers, doesn't mean the individual employee there is a bad person, or worse: stupid, because he felt for the 'power of the dark side'. For once, keep marketingpoop and real life separated.
You should read "Proudly serving my corporate masters" by Adam Barr ( I believe he even is a slashdotter). Then you will understand that interns at Microsoft are not picked up at MacDonalds, but recruited at the finest universities and should pass a tough selection program. No wonder as a company they are treated as normal human beings: the best people know they are the best and will only work for... the best, ('best' can be different for a lot of people) so Microsoft will do everything they can to get them on board (like IBM, Sun and other companies will do too).
Slashdot Mirror
Every time when I read articles such as these I wonder: why oh why do Democrats in the USA have such a hard time selling the truth to the public? I mean: the current Bush administration has piled misleading and disputable decision on decision, and the American public seems to feel it is all right. How come? Why aren't the Democrats using these obvious limitations on the freedoms of the American citizens to rally the public so they'll support the Democrats and elect a better government in place which will overturn such decisions like a concentration of media companies?
You can come to two conclusions:
1) The Democrats are also after the same money from these media companies as the Republicans are, which in fact makes the USA's democracy rather dead: there is no real choice for Joe Sixpack, the two parties which matter are NOT serving the interests of the people
2) The Democrats are incapable of fighting Bush effectively. Which also makes the USA democracy rather dead, because the general public doesn't KNOW there is an alternative to 'Bush'. When Bush gets the concetration of media in place, and the holders of these media on his side (which seems to be the case) he controls EVERYTHING and the republicans can stay in power, well... forever.
If the republican party would exist in The Netherlands, Europe, they would get at most 2 seats in the 150 seat parlement, roughly guessed. Not because we're all 'stinking liberals', but because we tolerate less a government that thinks of big $$$ first and the interest of the public second.
(To the USA citizens: as a European I see you as a group of people who thought that a president who nailed his intern with cigars should be impeached and a president who started a very expensive war under false intelligence in a time where jobs dissapear very quickly should stay in his office and should stay popular. Think about that for a second.)
What if he links with a BINARY library he has found on his machine and came with the OS on the machine (i.e. the api offered by the OS) ? Developers like linking with binaries since that shortens compile time.
Now, he links with a binary but has to embed a license to his own sourcecode because he links with a library, however he doesn't include any sourcecode from the binary (say, he uses win32's LoadLibrary()).
_THAT_'s the viral part. He calls into a library and because he does that he has to make his own code which he owns copyright of open because someone else who doesn't own the copyright of that work said so.
I'm 100% sure this will not hold in court in The Netherlands or Germany.
Your complaints about C++ are perhaps legit, they have nothing to do with .NET. If your bosses are too stupid to allow you to work with a C++ compiler that works like you want to, that's a problem at your employer, not with MS or .NET
.NET matures over time, so I can ditch the VB/VC++/COM crap for the n-tier apps I develop all day.
I'm happy C# is available and
With NT, MS released 3 bad SP's: 2, 4 and 6. :). On WIn2k though, sp2 was very good, sp3 was crap. It has to be so that SP4 is great ;) (it works ok on the testserver that's routing my internet conne**NO CARRIER**
Am I the only one who thinks the shown application windows could be a lot smaller when the controls were positioned more elegantly and efficiently (screenspace wise) and without a lot of space between them?
Perhaps people who are not used to operate a computer with gui screens are confused with a lot of information on their screen but this way you need a lot of dialogs to display all the information simply because the dialogs can't contain a lot of info: most space on the dialog is simply not used. OR, apple is hiding a lot of info so you don't need a lot of dialogs because the total amount of information to show is not a hell of a lot.
Either way, the usage of screenspace like how its done in these screenshots is IMHO a waste of pixels. It can be done a lot more efficient. (look at the process viewer. A lot of space is wasted below the list by a total inefficient ordering of controls, so less space is reserved for the real data: the process list).
Early X11 applications also had this example of IMHO bad interface design.
Please, answer these questions:
- WHO has seen both the SCO sourcecode and the linux code?
- WHO has seen the contracts IBM signed with SCO?
- WHO has seen the patents SCO has related to Unix?
- WHO has read and understood the complete class action suit text filed by SCO?
If you've answered on all questions "I have", please step forward and enlighten us with your vision on the matter. If you HAVEN'T answered on all questions "I have", please, stop hammering texts and start thinking about a similar case when a given company X is suspect of a GPL-violation.
The reason Apple and Sun hardware/software combinations are superiour in stability, is due to the fact that they are made to support each other, unlike in a windows enviroment, where you have a mix and match of hardware, and software drivers that bring in many inknowns sometimes.
Also, it's nice that Sun has very high end servers that keep running when you have to change cpu's or memory, but everybody knows that Sun can't exist by solely these clients. It's the bulk that makes or breaks you and Sun clearly looses that bulk to cheap systems with cheap OS-es because these systems are ALSO very good. Win2k server or Win2k3 and a dell server will hardly crash, a dell box with Linux or a preconfigured linux server from IBM will also be rocksolid. Why buy sun in those situations?
Oh, and if even a short period of downtime costs you a HELL of a lot of money, you won't buy Sun, you will buy stratus systems or tandem systems (HP has given them another name, same hw). Sun looses also in that area. It's a loose/loose situation for Sun. At the uni, I loved the sun boxes and SunOS, but today I can hardly imagine using one when the alternatives (win2k/xp/win2k3 or linux/bsd) are cheaper and as good as Sun's hw.
About a month ago my mailserver started to receive a lot of hotmail connections from the range 65.54.*.*., guess what the bay range servers inside hotmail.com. I contacted abuse@hotmail.com, tried a few times to convince the drone at the other end that my mailserver was receiving a connection from a hotmail server every 20 seconds, but they didn't understand it. I mailed mailserver logs, explanations, links to threads about this on usenet, no clue. After a while I simply blocked all hotmail servers from my server. It's really weird that they have people on the abuse staff that do not understand what 'abuse' means or how to prevent it.
A week ago I removed the block to check if things had changed. To my suprise, no connection since. Apparantly MS has solved this problem finally (that is: installed the WebDAV patch that is what, 2 months old?).
Yes, the patches themselves. People don't install them because they break critical production software which must not be broken.
That critical production software NEEDS a patch, f.e. it has a security hole, or runs on top of an OS that has a security hole. THerefor it IS already broken and thus needs patching. THere is NO excuse for not patching your software, like there is also no excuse for having security holes in your software.
Your E=mc^2 idea is neat but 'c' also relies on time, which means that the equation relies on time, and thus doesn't solve the 'it relies on timemeasurement' problem.
On some other forum (here) they talk about renaming some file to get rid of her erm.. cloths/leaves.
I agree with you that the parent should probably read some more books, but Sun also should be looking more to their worst enemy Microsoft how easy system administration CAN be done. I'm not saying Win2k is easy to admin when you set up a large forest with AD, but applying patches etc, it's darn easy. Sun (and other companies as well) should learn that even professionals do not LIKE it when their job is HARD because the 'tools' they have to work with are very low level and effectually do not help the professional a lot. Tools should help the user, not work against him/her.
Ok, apple invests in software too, but their hardware is also designed mostly by IBM/Motorola, nvidia and gee... Intel (PCI/AGP). Dell sells hardware THEY designed, using common parts, but THEY designed it. Apple does the same.
However, WHY would anyone buy sun? They only company I can think of is... Oracle. IBM can't even buy it without severy interference by the anti-trust people in the government, Dell will not buy it since it will not add anything to their portfolio: their high end systems are designed around Microsoft's software, not unix.
However, Oracle's database is mostly run on Sun hardware, or better: when you buy a Sun E* server and decide to run a database on it, 10 to 1 it will be Oracle. It does make sence in a way for Oracle to buy Sun so they can sell the complete package to high end customers: hardware AND software.
And I mean: really useful. Plus your application has to be of a quality that is only found in commercial applications. Then you have a combination in your hands that is attractive to an enormous amount of people: quality software that is useful for a cheap price: free. There are a lot of open source projects that do not meet one or more of these requirements and are therefor not succesful. A lot of developers think they have made a great tool, but do not understand that the tool they made is only useful to them because they understand it, i.e.: the quality is not up to par, and an average user will not be able to use it as it is ment, which will result in a request from the user for another tool which seems to do the same: the user will look elsewhere.
:) it wasn't ment to be released as a worldwide succes, and DemoGL was, but failed.
I've made 2 open source tools: DemoGL and LLBLGen. DemoGL was a library for win32 which let you develop OpenGL effects easily in C++. It was not that useful to a lot of people, since most people didn't need it and got things started with a few tutorials from NeHe. LLBLGen on the other hand is very succesful, it's a DAL generator for sqlserver/.NET/C#/VB.NET. The reason why it's so succesful (according to the vast amount of reactions I received from users) is that it's simple to use, yet does what the user wants and costs nothing. It comes with the source so if you want to adjust it you even can do that. (This is what I understand most users found about the tool).
The fun thing is: LLBLGen was a pet project, a project to learn C#
I was reading those lines, copying it over to post it, and saw your posting, so posting it again would be really redundant, but it is indeed pretty funny, and that poor Ari is receiving his 100 penis enlargement offers a day really soon I guess ;)
The "NT kernel series" sucks when you try to port Unix-style thread or process per client model server software to it, because of the process limit I discussed and the VMS-like heaviweight processes. The ideal # of concurrent executing threads on 2K3 is one per processor, SQL Server and Exchange are modeled on this.
:) Tell me, how would it be possible to execute say 4 threads simultaneously on a processor (without HT) ? SQLServer and Exchange are modelled on the 'thread-scheduling' model, pure and simple. SQLServer's kernel (yes it has a kernel too) even uses NT-fibers, a part of the OS which can boost threads (and other threads are suffering on this). If Win2k3 has better thread-scheduling and less process-scheduling, SQLServer and Exchange will benefit from this, but also ALL threaded applications will benefit from this (can you say: IIS? aspnet_wp.exe ?)
Kernels do not suck. Kernels, if properly written (but not properly written kernels are hardly ever used in OS-es, except the win9x 'kernels'), do what was designed up-front. They have specs. If you want to use something, you read the specs and see if what you want to use can perform what you want it to perform so the usage is succesful. If you want to use an NT kernel as a monolithic UNIX kernel, and try to schedule processes instead of threads, it won't work efficiently. NT-based OS-es do not use shared memory for their processes. That's why they use threads, because these can use shared memory. When you want to use processes on an NT-kernel and you see that the performance is poor, it's not the fault of the kernel, it's the fault of the programmer who sux big time and doesn't know a thing about what he/she is doing.
Also your remark about 'VMS-like Heavyweight processes' is of the same quality: an NT-kernel based OS works differently than UNIX. You also do not eat soup with a fork, do you? even when the fork performs brilliantly when eating potatoes!
One thread per processor is the optimum? Whoa
Windows server performance is top notch, the kernels are tuned excellently, and with each server release they get better and better. I also found your resources-remark rather amusing. You are refering to the handle-count in each process. So you think it is a good thing, a process will open (2^32)-1 objects and thus has that much handles open? I think that's a bad thing. An open handle means you have an open resource, and are keeping it open. Not a lot of resources qualify for that, most resources get opened, used and are closed right after they are used. That's good programming practise.
I'm very happy for you that you think your monolithic, hardware-specific kernel is the way of the future. I also hope that you WON'T understand that how a kernel works internally is not that important, it's how the OS it is part of runs the software YOU want to run and use. If you WILL understand this, you will regret your swap. Until then, enjoy the ride, while it lasts.
MS DTC (Microsoft Distributed Transaction Coordinator) manages the transactions which span more processes, across machines. You can see this as a P2P program, it communicates with the other MS DTC instances on other machines if parts of a distributed transaction are executed there. This means that a transaction started in a client application which rolls through tiers down to the sqlserver machine, both on the client and the server a MS DTC instance is running and coordinating the part of the transaction running on their boxes. It coordinates commiting the distributed transaction acros all the servers enlisted in the transaction.
If the client started the transaction on the server by an RPC, only on the server the MS DTC is started and only on the server there is a COM+ transaction, not on the client. However, I don't think that was the case. I think, and if I read your post correctly, you assumed this too, that on the clients software was running which initiated the transaction locally and which was rolled down to the server and back and which was coordinated by the transaction coordinator tuxedo or COM+ (which uses DTC). So client-server communication IS necessary in both situations.
So to me the end result counts: what kind of setup is faster (because both need client - server communication) and the SQLServer setup is very fast for the money it costs.
In Iraq US soldiers are putting their life on the line because they think they fight for 'freedom'. However by the day that is less and less true. Where will it end? In the situation where the New Iraq will 'liberate' the US from their occupiers so the people can live in 'freedom'?
All this talk about electronical gadgets... millions of people in Iraq suffer every damn day because there is no water or too less water or only dirty water and not enough food.
And the US of A can only talk about cellphones. I don't know but then you really are out of touch with reality. Give the iraqi people food and water and let them rebuild basic infrastructure first. This will cost a few years. After that the debates about the unnecessary gadgets can begin.
Now, the worlds in most FPS games are rendered first, then the models and other entities are rendered, using clipping /depth buffer info of the world. A lot of engines use 2 different render routines to do this: the world is mostly static and uses a different routine than the model renderer.
THe result is, that when you 'patch' the world renderer so that it f.e. renders wireframes instead of solid polys (in OpenGL based engines this is 'not that hard', you just change the value passed to glBegin()) the models still are rendered solid, plus because most renderers for models rely on the depthbuffer filled during the world rendering, the models close to corners are fully rendered, since the depthbuffer is empty. So you can easily 'see' the models close to corners. If you also 'patch' the model renderer for not doing world clipping, you will see ALL models rendered in your window.
This can't be done if the world + models use a single render routine, i.e.: model polygons and world polygons are packed together as THE set of polygons to render, then the single render routine will eat these single pack of polys to render. If you patch the routine for wireframing, you will see the models also wireframed, if you patch out the world clipping, you will get the complete world in your window, not what you want.
I think in future game engines there will be a merger between world + model polygon sets, because worlds are more and more modfyable in game by the player, which in the end requires that the modifyable parts are 'models' too. However games based on the current crop of quake * engines will keep on suffering from this.
I'm not a citizen of the US nor UK. The 'our' in 'support our troops' is not for everybody a given since 'their' troops are not there at the moment (thankfully).
People who volunteered to be in an army should face the consequences when their commanders think they should start playing cowboys and indians in some desert. Why should I support those people? Because they are 'fighting for freedom' ? 'Giving their lives for our lives' ? Freedom is degrading all over the world especially in the USA. If the fight for freedom should be held somewhere it should be in the USA and against its own government, not somewhere in the Middle East.
I'm all for freedom and peace for the iraqi people, as I also am for freedom and peace for the people in the occupied palestine territories (Israel occupies them for 30 years now, despite a UN resolution (242) which says Israel should retreat (resolution is 30 (!) years old)), as I am also for freedom for all those African citizens who suffer from war day after day for decades in a row.
However I'm against hypocrisy and a single war against Iraq solely to 'bring freedom' while ignoring all those other countries where people suffer day after day (for a much longer period!). This war is wrong, the consequences will be hard for a lot of people, the deaths of possible thousands of civilians will be caused by US lead troops.
Now, tell me, why should I support these 'freedom fighters' in killing innocent people?
You claim a lot of things, name percentages etc, but where did YOU get those figures and how did YOU test the products you mention to back up YOUR claims?
I don't see any link in your article. (like the one about linux versions running faster (running WHAT faster? An endless loop?) on the same hardware than windows).
You can argue Postgres, but I've never run into a case where I couldn't work around those features that haven't been implemented in MySQL yet.
Really? Now, how did you implement relational integrity? You added a lot of queries to perform the relational integrity? You can't! You can't add relational integrity by code ON TOP of a database system. Especially not in a multi-user environment. Especially with a database system which doesn't support transactions. So you can first test with an sql statement, and based on that result, perform some actions, but those 2 things are not 1 atomic unit. So just after the test-statement is done, another user could be messing with your data. Oh, you are not going to tell me you lock primary key rows when inserting a foreign key, are you?
The one thing I can't stand is when someone suggests: "I can't afford Oracle, so lets' go with a MSSQL database." That's like, I can't afford a space shuttle, and a ferarri isn't good enough for me, so I'm going to buy this million dollar llama instead because 1000 marketing agents can't be wrong, right?" It has all the same feautres as Oracle, it's just that the features in Oracle WORK.
Whoa, sherlock! You are saying that features in MSSQL don't work but they do in Oracle? Please, can you give me a list of them, so I can use that list in my daily work, so I can avoid features that are not working in MSSQL.
MSSQL is worlds fastest database AND supports true ACID compliance, stored procedures, views, partitioned views etc. etc. It also costs a hell of a lot less than Oracle. Oracle does have more features though. But... why wouldn't it be a reasonable choice? Because the features don't work but they do in Oracle?
Have you ever worked with a true database management system? I have the fear you haven't
All nice, but MySQL doesn't offer a lot on the area of relational integrity (constraints, triggers). Also it doesn't support subqueries so you are 'joining' a lot.
The 4 CRUD actions are not the total picture you need, for even a small database driven webapplication. And if price is a problem, postgresql is far better: it offers relatonal integrity, sports subqueries and more. MySQL is _NEVER_ the best option. NEVER.
"You do not know the power of the Dark Side".
Grow up, child! Even if it was sarcastic or in a wicked way ment to be 'funny', it's too pathetic for words. Microsoft is like any other company which wants to make money. Employees who work at Microsoft, do that because they get paid for what they do there, like people at Sun or IBM (or Red Hat).
Just because MS mistreats some of its customers, doesn't mean the individual employee there is a bad person, or worse: stupid, because he felt for the 'power of the dark side'. For once, keep marketingpoop and real life separated.
You should read "Proudly serving my corporate masters" by Adam Barr ( I believe he even is a slashdotter). Then you will understand that interns at Microsoft are not picked up at MacDonalds, but recruited at the finest universities and should pass a tough selection program. No wonder as a company they are treated as normal human beings: the best people know they are the best and will only work for... the best, ('best' can be different for a lot of people) so Microsoft will do everything they can to get them on board (like IBM, Sun and other companies will do too).