What does windows have to do with these new Sun CPU's? nothing at all. So why refering to them then? Ain't this editor's note a bit too much offtopic and a bit too much of a 'biased' seed to control the discussions?
Be fair. The new sparc is an awesome CPU. it will be combined with extrodinairy cool hardware in the new sun servers, and will run a kickass operating system, solaris 8. No need for offtopic rants and raves about microsoft and their products. --
Here in the Netherlands we have the most stable and best running economy of europe, with an aweful lot of jobs open for high tech people, and what's nice is that 70% of the working force is working at offices, not in factories. So high tech IT specialists are wanted _always_ and are _important_:).
--
true, however the software patent situation is in Europe, not the US. I'm not sure if all european countries have the same policy as the situation I described, which is the situation of The Netherlands, but I'd be suprised if a lof of EU countries were handling patents differently.
I recently filed for a patent for a total different thing (not software related) so I had to dive into this crap:) It's indeed sad the US passes almost all 'patents' that seem to look smart but are very obvious. --
So, once you've done all that work writing the code, do you want to get sued for patent infringement?
*That* is what software patents are about. Patenting things that a lot of people are bound to implement and 'invent' independent of the patent holder, not even knowing someone has done it before
No. When you want to get a patent for anything, you have to provide evidence to the comittee that your idea/thing is compliant to the rules the comittee has setup. For example, an obvious idea that can come up in every skilled person's head during his work will never be granted a patent. When you want a 20 year patent, you have to prove with examples and materials that your idea is indeed unique and worth the 20 year patent. This is NOT easy (as in: easy to convice the patentcomittee). So don't worry someone patents a forloop that searches a single linked list:). It's more about patents on CLEVER IMPLEMENTATIONS of algoritms that for example do real time LOD on meshes for example, stuff that is now patented using not the sourcecode but ideadescriptions.
Software patents are about clever implementations of an algoritm, not the algoritm itself. That is already patentable. Because, as I said, the rules are strict, it's hard to proof that your 'clever' implementation is unique. A good example of a clever implementation is Dijkstra's semaphoreless concurrency program. --
MSDN is a funny thing. I personally didn't find it all that helpful, for OpenGL or DX. That's why I turned to books and guides on the web.
Weird. The MSDN library has by far the best OpenGL function descriptionset available in digital format. There is _NO_ other source out there which describes every damn glFunction better than the MSDN library. (and don't mention MESA sourcecode or SGI's websites, expecially SGI's resources for programmers are not that good when it comes to plain and simple fast reference material for functions and constants.)
It's not important, but my Linux box does have different hardware. I don't know if you've used Fire GL/PRO, but it doesn't work properly with DX; now, is that DX's fault for not supporting Fire GL properly, or Fire GL's fault for not supporting DX properly? I guess you'd say Fire GL's
I don't know to which problems you refer to but the FireGL Pro had a permedia 2, which lacks several blending modes in hardware. Resulting in a lack of features in OpenGL too (try play q3a. Nice squares around the blended lights, heh!;)). Thats perfectly acceptable, but dont point towards an api lacking a feature the hardware doesn't provide in the first place. For d3d, you probably know, because you've worked so many years with it, that the manufacturer has to write a decent D3D api HAL driver, so the HAL can call the hardware properly. If this driver sux, it's end of story. Because 3dlabs is a pure OpenGL shop, I'm not suprised. --
MS has regular tours around developers offices and talks with them about DX, what they want in it, and how. I doubt SGI does that =)
True, but we, OpenGL programmers, have the very *cough* efficient *cough* ARB board for that;). I also have to say that OpenGL 1.2 is already finished for months, as is the sample implementation MS needs to produce a decent Opengl32.dll. However, MS doesn't ship it yet. Kinda odd I'd say. So the competition isn't really played that fair: win32 systems, the usual gaming platform, still suffers from an old opengl32 dll (OpenGL 1.1) vs a topnotch d3d api. No wonder no-one chooses OpenGL to develop with unless it's the api you know best. --
Carmack chose OpenGL in a time when D3D wasn't up to par. I'm sure if he had to make the same decision TODAY perhaps he'd have made a different one, especially because D3D8 is more powerful than Ogl 1.1 is (unfortunately).
Unreal has an OGL driver but it's very inmature, basicly because the internal engine core isnt able to spit out the primitives in a way OGL can render them at the fastest speed.
Quake engines use OpenGL, but are also strictly written with OpenGL in mind. Unreal is not. And it's not just a coincidense Valve abandoned the quake/quake2 engine and rewrote their extensions to those engines as a complete new d3d powered engine for their next game. --
Also, the point of a HAL is to abstract the hardware away, not to provide a way for you to stick 50,000 case statements in your code in case some particular feature is supported or not. That's a stupid way to design an API.
Ah, I hear someone talking who programs 3D software for a lot of cards every day, plus includes support for sound for a lot of cards!
What you describe with the case statements is how OpenGL programmers handle different codepaths in their apps because some extensions are supported and others are not. Direct3D just gives you the possibility to execute a feature, even if it's not supported by hardware. OGL doesn't have that.
No switch-case crap needed. Also, for soundcards and other hardware specific crap (like joysticks/mice!) you don't want to have to write 2 or more different codepaths to do 1 single thing. With a HAL you can do that.
But of course your remark fits perfectly with a supporter of a monolitic, hardware specific kernel, who simply ignores the fact that microkernels with HAL's are more flexible and easy to use --
The domainname you 'have' is in fact a rented name. You own the right to rent that name. No-one can also rent that name, because you have the exclusive right. As in this case, it's not about 'who OWNS sex.com' (because the ICANN does) but who owns the RIGHT to rent it from ICANN. If you obtain the right to rent the name, say, slashdot.org, no-one can do something about it, only the organisation who lets out the names, the ICANN (or their representatives, NSI and others).
See it as an apartment. If you rent an apartment from a certain organisation, you live there. but if someone else obtains a legal contract to rent that same apartment and you thereby loose the right to rent it, you have to move.
The ONLY RIGHT thing that should be done here is that the organisation who owns the domain, and thus lets out the domains to others, fixes the rent contracts so the legally owner of the RIGHT to rent sex.com is given back that right.
IMHO, the judge should have ruled that the ICANN or NSI have to fix the records so sex.com is back to the original subscriber. --
When you read those articles, for example the SQLserver 7 security how to here, with good tips on securing databases inside SQLserver, ODBC links to databases etc etc, you'll learn that SEVERAL TIMES you're advised to give the 'sa' account a secure password (that is: a password difficult to guess) and NEVER USE the 'sa' account again, only in case of trouble. You're adviced to setup accounts in NT and to use these inside SQLserver, and how to use NT security over SQLserver security (thus, using NT accounts instead of SQLserver accounts, like the 'sa' account, over thrusted pipes.)
I simply don't understand why MS has to be blamed for typical misbehaviour of end-users. If an end user doesn't want to read the articles online or doesn't want to understand the issues concerning security and internet when installing and setting up corporate systems (we're not talking a deskop system here), why is it suddenly the vendor's problem? "Yes, dear RedHat helpdesk guy, I did rm -rf/* when I was logged in as root, why is it MY fault that everything is gone?".
MS won't license you the sourcecode if you're going to use that sourcecode to create a competitive product to MS' own stuff. WINE is a competitive product, because it makes win9x or any other win32 compatible client OS unnecessary. --
of course in theory you can build the stuff with win32 functions.
For example, the standard explorer shell extensions for adding icons to the tray (notify icons) don't use any 'hidden' APIs
That's not what I was refering to. The SH* functions are more and more used as 'normal' win32 functions. If you don't know better and reading the MSDN, you'd understand these ARE win32 functions, which they ain't. (I made that mistake).
I wasn't refering to the 'shell' extensions like 'command prompt here' stuff when you right click in explorer. I was referring to stuff like the webpage formatting of a dir in win2k explorer. Not something you can just add.
Win32 is the layer on top of the NT core and win9x core libs. Normally apps talk to win32, instead of all the api's beneath it (you know, the so called 'secret api's). IE uses win32 to do stuff but also ADDS stuff to win32 (namely the shell extensions, SH* functions). This means it uses layers below win32, layers build with the normally hidden api's.
If you want to port IE to another platform, you have to know what the functions do that are used by IE. The win32 functions are documented in the MSDN, but the NT / win9x core api calls aint. So you need the sourcecode.
The browser will work fine without the shell extensions, it's just the shell extensions that make it has to use the lower level api calls. And because IE is part of the Shell of NT/win9x/win2000, it's called 'part of the OS', but NT runs great without it. But if you want to use the SH* extensionfunctions in your code, you NEED IE installed. (the SH* functions create nice dirtree's for example in controls)
I totally agree with you. Within a few months I'm going to open up the sourcecode for DemoGL, which is win32 focussed. I truelly will FORK an eventually port to another platform. IMHO this is better for the quality of the code: platform specific (or philosophy specific) features can then evolve in a forked codebase and won't mess up other codebases with endless #ifdef crap.
If a codebase is the BASIS for different systems, evolving along different philosophy lines (for example: (theoretically spoken) if interbase A will have it's focus more on distributed computing and if interbase B will be more focussed on central computing with large (real large!) databases than the fork between A and B FROM the basis codebase is a good one IMHO. No cluttered code in either of the codebases to exclude functionality from both philosophies.)
IMHO more OSS projects should fork (or branch, whatever you want to call it) their codebase to clean up the code. The more #ifdef SPECIFIC_DEFINE there are included in a sourcefile, the more it gets unreadable and unmaintainable. --
Please forgive me if I don't understand the whole situation, but if a piece of code is licensed with 2 licenses, and I as a programmer use that piece of code, which license should I obey? because: if one of the licenses forbids me to do something but the other doesn't, there is a conflict in the general rules how a piece of code can be used. Also, if one of the licenses forces the rights on modifications of the code back to the original programmers of the code (as in the MPL) and the other license doesn't force you to do that, why should a programmer suddenly feel he has to do that in the new situation? because the GPL is ALSO a license on THE SAME CODE.
To me it sounds as if the codebase is branched and will go on as 2 different codebases with 2 different licenses. Not that practical if you ask me, especially when the product derived from the codebase desperately needs to be released to gain back marketshare. --
It's perhaps wise to check out what your common 'enemy':) has cooked up in a (old but still good read) document about the next generation GDI (win32's gui rendering subsystem). Perhaps some things can be included in the rendering system currently in the works for XFree.
According to that book (you know, the bible) some 'god' created life on earth etc etc. When the theory about life is coming from another planet is true, does that also imply we have solid proof that christians believe in a god that doesn't exists?
It's not giving out tools to exploit a security hole with a single click on a button. That's something else. Security holes still can be described and mentioned at sites like bugtraq, but when they're not coming with detailed descriptions and 'run this tool to exploit this hole' executables, is this bad?
No! ofcourse not! You see: even if a vendor of software, let's take Sun or Microsoft, checks bugtraq at the same time as a legion of scriptmorons do, they can't avoid a possible hell on earth when the bugtraq posting comes with a plain tool to exploit the security hole right away!
There is simply no _TIME_ for the vendor (or linux distributor) to patch. It then also takes a while before the majority of products around the globe are updated, so during that time, the scriptkiddie or similar individual has a weapon at hand, and only has to press a button to exploit the bug.
This discussion should be focussed on the area between: TOTAL nondisclosure of ANY securityhole whatsoever vs TOTAL disclosure of ALL securityholes and every securityhole should come with FULL exploit tools and exploit how-to's.
IMHO too many people are focussed on one of the ends of the possible solution area. Don't. both edges are not helpful. --
Thing is, this is all going to get fixed. Several companies are working on the printing problem. Once they all screw it up and present 15 different conflicting standards, some group of free programmers will get pissed off enough to write one from scratch. X could go away as well.
Pal, really... stop it:) you're killing me.
Don't you get it? people don't want to WAIT for shit to get fixed if they can get it at the store TODAY, only it has a box with a 'microsoft' logo on it. It's fun to work on tools that will bring utopia on earth ONCE they're finished. but for productivity, for getting things done, stuff like: 'it is going to get fixed, and if the fixes don't work we all get mad and work together as ONE! Oh yeah baby! (orgasmic cry)' won't work. Trust me. perhaps in 'my own private hobby world', but not in 'the Real World(tm)'. I'm sorry. --
Slashdot Mirror
Thankfully there is also a crapstory about Microsoft on the frontpage, so the 'hack'news won't catch anyone's attention.
--
What does windows have to do with these new Sun CPU's? nothing at all. So why refering to them then? Ain't this editor's note a bit too much offtopic and a bit too much of a 'biased' seed to control the discussions?
Be fair. The new sparc is an awesome CPU. it will be combined with extrodinairy cool hardware in the new sun servers, and will run a kickass operating system, solaris 8. No need for offtopic rants and raves about microsoft and their products.
--
Brillant wizards don't rant about something bad; they do something about it. period.
--
I already see this slogan slammed on every billboard in town :)
--
Here in the Netherlands we have the most stable and best running economy of europe, with an aweful lot of jobs open for high tech people, and what's nice is that 70% of the working force is working at offices, not in factories. So high tech IT specialists are wanted _always_ and are _important_ :).
--
I recently filed for a patent for a total different thing (not software related) so I had to dive into this crap :) It's indeed sad the US passes almost all 'patents' that seem to look smart but are very obvious.
--
we weren't talking about the software patent granted in the US, were we? We're talking Europe here. US != europe. (sorry to shock you)
--
*That* is what software patents are about. Patenting things that a lot of people are bound to implement and 'invent' independent of the patent holder, not even knowing someone has done it before
No. When you want to get a patent for anything, you have to provide evidence to the comittee that your idea/thing is compliant to the rules the comittee has setup. For example, an obvious idea that can come up in every skilled person's head during his work will never be granted a patent. When you want a 20 year patent, you have to prove with examples and materials that your idea is indeed unique and worth the 20 year patent. This is NOT easy (as in: easy to convice the patentcomittee). So don't worry someone patents a forloop that searches a single linked list :). It's more about patents on CLEVER IMPLEMENTATIONS of algoritms that for example do real time LOD on meshes for example, stuff that is now patented using not the sourcecode but ideadescriptions.
Software patents are about clever implementations of an algoritm, not the algoritm itself. That is already patentable. Because, as I said, the rules are strict, it's hard to proof that your 'clever' implementation is unique. A good example of a clever implementation is Dijkstra's semaphoreless concurrency program.
--
It's not important, but my Linux box does have different hardware. I don't know if you've used Fire GL/PRO, but it doesn't work properly with DX; now, is that DX's fault for not supporting Fire GL properly, or Fire GL's fault for not supporting DX properly? I guess you'd say Fire GL's
I don't know to which problems you refer to but the FireGL Pro had a permedia 2, which lacks several blending modes in hardware. Resulting in a lack of features in OpenGL too (try play q3a. Nice squares around the blended lights, heh! ;)). Thats perfectly acceptable, but dont point towards an api lacking a feature the hardware doesn't provide in the first place. For d3d, you probably know, because you've worked so many years with it, that the manufacturer has to write a decent D3D api HAL driver, so the HAL can call the hardware properly. If this driver sux, it's end of story. Because 3dlabs is a pure OpenGL shop, I'm not suprised.
--
True, but we, OpenGL programmers, have the very *cough* efficient *cough* ARB board for that ;). I also have to say that OpenGL 1.2 is already finished for months, as is the sample implementation MS needs to produce a decent Opengl32.dll. However, MS doesn't ship it yet. Kinda odd I'd say. So the competition isn't really played that fair: win32 systems, the usual gaming platform, still suffers from an old opengl32 dll (OpenGL 1.1) vs a topnotch d3d api. No wonder no-one chooses OpenGL to develop with unless it's the api you know best.
--
Unreal has an OGL driver but it's very inmature, basicly because the internal engine core isnt able to spit out the primitives in a way OGL can render them at the fastest speed.
Quake engines use OpenGL, but are also strictly written with OpenGL in mind. Unreal is not. And it's not just a coincidense Valve abandoned the quake/quake2 engine and rewrote their extensions to those engines as a complete new d3d powered engine for their next game.
--
Ah, I hear someone talking who programs 3D software for a lot of cards every day, plus includes support for sound for a lot of cards!
What you describe with the case statements is how OpenGL programmers handle different codepaths in their apps because some extensions are supported and others are not. Direct3D just gives you the possibility to execute a feature, even if it's not supported by hardware. OGL doesn't have that.
No switch-case crap needed. Also, for soundcards and other hardware specific crap (like joysticks/mice!) you don't want to have to write 2 or more different codepaths to do 1 single thing. With a HAL you can do that.
But of course your remark fits perfectly with a supporter of a monolitic, hardware specific kernel, who simply ignores the fact that microkernels with HAL's are more flexible and easy to use
--
And it worked flawlessly! ;))
O.
--
See it as an apartment. If you rent an apartment from a certain organisation, you live there. but if someone else obtains a legal contract to rent that same apartment and you thereby loose the right to rent it, you have to move.
The ONLY RIGHT thing that should be done here is that the organisation who owns the domain, and thus lets out the domains to others, fixes the rent contracts so the legally owner of the RIGHT to rent sex.com is given back that right.
IMHO, the judge should have ruled that the ICANN or NSI have to fix the records so sex.com is back to the original subscriber.
--
http://www.microsoft.com/tec hnet/security/database.asp
When you read those articles, for example the SQLserver 7 security how to here, with good tips on securing databases inside SQLserver, ODBC links to databases etc etc, you'll learn that SEVERAL TIMES you're advised to give the 'sa' account a secure password (that is: a password difficult to guess) and NEVER USE the 'sa' account again, only in case of trouble. You're adviced to setup accounts in NT and to use these inside SQLserver, and how to use NT security over SQLserver security (thus, using NT accounts instead of SQLserver accounts, like the 'sa' account, over thrusted pipes.)
I simply don't understand why MS has to be blamed for typical misbehaviour of end-users. If an end user doesn't want to read the articles online or doesn't want to understand the issues concerning security and internet when installing and setting up corporate systems (we're not talking a deskop system here), why is it suddenly the vendor's problem? "Yes, dear RedHat helpdesk guy, I did rm -rf /* when I was logged in as root, why is it MY fault that everything is gone?".
--
MS won't license you the sourcecode if you're going to use that sourcecode to create a competitive product to MS' own stuff. WINE is a competitive product, because it makes win9x or any other win32 compatible client OS unnecessary.
--
For example, the standard explorer shell extensions for adding icons to the tray (notify icons) don't use any 'hidden' APIs
That's not what I was refering to. The SH* functions are more and more used as 'normal' win32 functions. If you don't know better and reading the MSDN, you'd understand these ARE win32 functions, which they ain't. (I made that mistake).
I wasn't refering to the 'shell' extensions like 'command prompt here' stuff when you right click in explorer. I was referring to stuff like the webpage formatting of a dir in win2k explorer. Not something you can just add.
--
Win32 is the layer on top of the NT core and win9x core libs. Normally apps talk to win32, instead of all the api's beneath it (you know, the so called 'secret api's). IE uses win32 to do stuff but also ADDS stuff to win32 (namely the shell extensions, SH* functions). This means it uses layers below win32, layers build with the normally hidden api's.
If you want to port IE to another platform, you have to know what the functions do that are used by IE. The win32 functions are documented in the MSDN, but the NT / win9x core api calls aint. So you need the sourcecode.
The browser will work fine without the shell extensions, it's just the shell extensions that make it has to use the lower level api calls. And because IE is part of the Shell of NT/win9x/win2000, it's called 'part of the OS', but NT runs great without it. But if you want to use the SH* extensionfunctions in your code, you NEED IE installed. (the SH* functions create nice dirtree's for example in controls)
--
If a codebase is the BASIS for different systems, evolving along different philosophy lines (for example: (theoretically spoken) if interbase A will have it's focus more on distributed computing and if interbase B will be more focussed on central computing with large (real large!) databases than the fork between A and B FROM the basis codebase is a good one IMHO. No cluttered code in either of the codebases to exclude functionality from both philosophies.)
IMHO more OSS projects should fork (or branch, whatever you want to call it) their codebase to clean up the code. The more #ifdef SPECIFIC_DEFINE there are included in a sourcefile, the more it gets unreadable and unmaintainable.
--
To me it sounds as if the codebase is branched and will go on as 2 different codebases with 2 different licenses. Not that practical if you ask me, especially when the product derived from the codebase desperately needs to be released to gain back marketshare.
--
http://www.microsoft.com/HWDEV/vid eo/GDInext.htm
--
I know adobe tells you different, but here we run pagemaker 6.5 on win2k workstation, and no weird errors or malfunctioning.
--
Not to start a holy war, just a thought.
--
No! ofcourse not! You see: even if a vendor of software, let's take Sun or Microsoft, checks bugtraq at the same time as a legion of scriptmorons do, they can't avoid a possible hell on earth when the bugtraq posting comes with a plain tool to exploit the security hole right away!
There is simply no _TIME_ for the vendor (or linux distributor) to patch. It then also takes a while before the majority of products around the globe are updated, so during that time, the scriptkiddie or similar individual has a weapon at hand, and only has to press a button to exploit the bug.
This discussion should be focussed on the area between: TOTAL nondisclosure of ANY securityhole whatsoever vs TOTAL disclosure of ALL securityholes and every securityhole should come with FULL exploit tools and exploit how-to's.
IMHO too many people are focussed on one of the ends of the possible solution area. Don't. both edges are not helpful.
--
Pal, really... stop it :) you're killing me.
Don't you get it? people don't want to WAIT for shit to get fixed if they can get it at the store TODAY, only it has a box with a 'microsoft' logo on it. It's fun to work on tools that will bring utopia on earth ONCE they're finished. but for productivity, for getting things done, stuff like: 'it is going to get fixed, and if the fixes don't work we all get mad and work together as ONE! Oh yeah baby! (orgasmic cry)' won't work. Trust me. perhaps in 'my own private hobby world', but not in 'the Real World(tm)'. I'm sorry.
--