I currently work on a project that is CMM Level 4 rated, and we will be going for Level 5 next year. Our software is not what you would call widely used(not for consumers, but for Govt.). The main thing that using the CMM model does for us is in finding software defects.
With the number of reviews and the incremental development lifecycle, we catch most of our bugs way before the software is shipped to the customer. This is a very good thing, because the software is not very easy to upgrade or patch once in the field.
If our customer suddenly asks us to use an older version of the software to test certain features on hardware(An older version which has no extra code, so testing the hardware is easier), we can go and get the exact version that was shipped, apply any necessary fixes, and deliver in almost no time.
The main thing that CMM forces you to use is a process. Everything is based on processes. How to build the software, how to report a defect, how to note the fix in the CM database, how to keep track of versions, how to ship, etc. Everytime you do anything you follow the process. It's annoying at times, but it allows everything to be repeated exactly. By optimizing the process we can "speed" up development time(our ship dates are known years in advance, so it just means that we make deadlines). It also means that when customers are looking for a company to provide them with a solution, they see that if we have created successful projects in the past, then we probably will be able to repeat the success on a new project.
I think that some things in the CMM are very good, but I also don't see them as being able to apply very well to the open source world, primarily because it requires large amounts of time to make sure the processes are correct, time that developers don't want to spend(Thank goodness for managers).
So how does having a digital theatre help show movies filmed in Super 16 blown up to 32mm?
Many 'indie' film makers still shoot in Super 8, 16 and Super 16. The film 'Best in Show' was shot on Super 16 then enlarged to 32(as Super 16 works very well when enlarged to 32mm).
I don't think that digital theatres will replace non-digital theatres will happen for a long time. Most movies are not digital, and they will probably not switch to digital until the quality of digital beats out the quality of film.
I know this is an old discussion, and no one will probably read this, but there is a problem with speed limiters. First, imagine this, your're driving along maxed out at 30 MPH when out of no where a car comes out of nowhere(a driveway) right beside you, you see them, and have less than a second to react. In real life if this happens, you speed up, yup that's right, you speed up over the speed limit and avoid hitting the car. With a speed limiter, then you get hit. I'm sure there are other examples of this, but I'm not thinking of any right now.
There is also a GNU C compatibility mode, which provides the extensions supported by GCC (version 3.0.1), along with various undocumented
features and bugs. The compatibility is good enough that the front end can compile the Linux kernel and utilities. At present, there is no g++
mode, i.e., no way to enable GNU extensions and C++ mode at the same time.
Which/bin/sh are you using?
Is it really ash as most BSD distributions provide, is it really bash which several linux distros put as/bin/sh, or is it really AT&T's real/bin/sh.
The point I'm trying to make is that even though most Unix Systems won't start without a/bin/sh, there are a bunch of drop in replacements for/bin/sh.
If I could replace every instance of MSIE with Mozilla, then this would be a fair comparison.
If something's not compiling that you think should, you end up wading through the mile-long error messages. If it does compile but doesn't work right, you're going to find yourself in the debugger trying to step through some of that crazy obscure STL C++ code to try to figure out what the heck is going wrong. Neither is much fun.
You find that with most libraries. On most large projects one usese the debugger constantly, to find out why the code isn't working "right". So I don't really find digging through the debugger a chore since you then really know what your code is doing.
But when it does compile and run correctly STL is pretty nice!
Agreed.
I'm looking forward to somebody starting over some day and coming up with a language that supports generic programming as well as C++, but which doesn't have the terrible syntax of C++ templates. It must be possible.
People may not like me for this answer, but how about Ada? I've not had any problems with their generics(Similar to templates) on the client side(I'm not writing the generics).
For example:
type Light_Spectrum is range 10.. 900;
package Wavelength_IO is new Ada.Text_IO.Integer_IO(Light_Spectrum);
Defines a new type(Wavelength) which accepts the integers from 10 to 900(inclusive), then instantiates an IO package for it. This allows for basic input and output of the type Wavelength.
NOTE: Wavelength is not an integer, it's stored however the compiler wants to store it.
Here's part of the declaration of Ada.Text_IO.Integer_IO.
generic
type Num is range <>;
package Integer_Io is
Default_Width : Field:= Num'Width;
This says that type Num is some range that we don't know yet, that's it, then just use Num where ever you would use the Actual type.
I like it, but I'm really tired so this might be garbled somewhat.
Washington State(at least when I used to live there) had different sales taxes in Pierce and King counties. I think Pierce was 7.25 and King was 8, but I may not be remembering correctly.
Of course that's why you just go down to OR to buy anything:)
I have only one problem with point number 3 in your argument. I think that you should have the option to license software, or to purchase it. If I am a company, would I rather buy 40 copies of application A, or buy a site license and install the software on each new PC I need it on?
On the other hand, it is easier for me to buy 2 copies of some game to play on multiple computers at once then it is for me to license those two copies.
I don't think that there is a very slow rate of adoption in the corporate setting for Windows XP. Most corporate settings are just now getting around to upgrading to Win2k.
I held a job supporting Cigna Healthcare's last big upgrade, which was about a year after Win2k, and they were upgrading all client PC's to NT4. My current company still use NT4 on the PCs, Solaris 2.7(and we just upgraded to this late 2001) on most of the suns, and HP11i on the HP-UX machines.
So corporate customers generally wait a long time to upgrade, to allow stability issues to be fixed, and to allow the applications that they use to catch up to the operating system.
Ah, but his argument(I think) was that even after a small number of people started getting arrested for drug possesion, most users did not quit.
I don't really know if 90% would stop using out of fear as many people think that "It can't happen to me", "They'll never catch/find/identify me", or "I'm above the law". Those people need to be arrested themselves before changing their behavior.
Anyway, I don't think anyone has any hard facts to support either of these arguments.
We allow corporations to rip off huge quantities of code and call it "the Windows (R) network stack"
Parts of the Win network(TCP/IP) stack are based off of the BSD network(TCP/IP) stack, but then so is every network(TCP/IP) stack, as it was first implemented by BSD(Way back when, but I don't remember the date). This is due to the BSD license, and most BSD people have no problem with that. Also, Microsoft admits to including code by the CSRG at Berkely, it's listed in the release notes(i.e. This software includes portions written by the CSRG at UCB(something like that). So that point has nothing to do with the rest of your argument.
Also, it is wrong to make unauthorised copies of Microsoft's code, they have the _copyright_ on it, because they wrote it. You still have the ability to go and get the BSD TCP/IP stack and fix it yourself.
The requirement's don't say that the libraries must be free. I'd assume that if they were looking at C# and VB.Net that cost is not an issue in this case.
Yeah, this is posting to an old discussion, but the FreeBSD ultrasparc port is very close to self hosting. As the port already boots multi-user on PCI based sun4u systems. Also there is a PowerPC port of FreeBSD(don't know the status).
The problem is that even "safe" Java has had security problems. Not relating to the language itself necessarily, but relating to browser/platform implementations.
So until the languge/CLR mature enough, then there will be more problems with an insecure language.
Also, note that most early Java security problems were found because sun encouraged people to find them, and then Sun would fix the problems. Microsoft doesn't want people to find and disclose bugs in it's software, so it may take longer to mature security wise.
--xPhase
P.S. pardon any spelling errors, i'm tired.
Hi we sent email to sourceforge members, and linux kernel mailing list members, and got a lots of numbers.
First, the response rate from Sourceforge was 34.2%, which is not representative of most of sourceforge. How many of those contacted were in active development? How many alpha projects are there on sourceforge that have had little development, and will continue with little development?
Second, there was only a 2.4% response by the lkml members. Would this 2.4% be the active members of the list, or mainly the people that don't participate?
Third, this only accounts for people on sourceforge projects or linux, which may not be representative of all open source projects. Yes it would be very difficult to survey more members, but how do we know that sourceforge members and linux hackers are not different than other projects?
Fourth, there listing of open source principles(slide 8) is only representative of a specific group of open source developers. The intellectual property ideas("Free speech, not free beer", and copyleft) don't apply to people not using the gpl. This may mis-represent the people who participated in the survey.
This survey is much more useful if, instead of claiming to represent all open source developers, it admitted to being primarily about linux/gpl developers. Or if more information was given about what projects(activity, license, and activity by developer) the participants of the sourceforge survey were involved in.
The survey is interesting if looked at in relation only to the sourceforge community, but is not able to be applied to all open source developers.
I'm not commenting on whether or not someone should post this information on a website.
I would like to mention that material on how to make bombs has been in circulation since before the internet, and it will be in circulation even if all the sites with HOW-TO's get taken down.
Also, how many Afghani terrorists have internet accounts? I mean other than John Katz's friend?
Whether or not I agree with the information being available on-line, I do not feel that it puts our freedom in any sort of jeopardy.
Does anyone here know if FreeBSD is committed to adopting NetBSD's rc system by some specified release? 4.5? 5.0?
According to JKH, there is no definate plan for the integration, so it may never happen. There is currently an rc system used by FreeBSD, and it(as well as NetBSD's) are completely different ideas than the way SysV init works. There are some that don't like the way init works.
I believe 5.0 will bring FreeBSD's kernel up to rough parity with the Linux kernel.
Many will agree that in some areas FreeBSD's kernel is superior to the linux kernel, and in some areas it is lacking. I don't think that there will ever be a rough parity between the two. I think that FreeBSD will always be better at some things while linux will be better at others.
Slashdot Mirror
I currently work on a project that is CMM Level 4 rated, and we will be going for Level 5 next year. Our software is not what you would call widely used(not for consumers, but for Govt.). The main thing that using the CMM model does for us is in finding software defects.
With the number of reviews and the incremental development lifecycle, we catch most of our bugs way before the software is shipped to the customer. This is a very good thing, because the software is not very easy to upgrade or patch once in the field.
If our customer suddenly asks us to use an older version of the software to test certain features on hardware(An older version which has no extra code, so testing the hardware is easier), we can go and get the exact version that was shipped, apply any necessary fixes, and deliver in almost no time.
The main thing that CMM forces you to use is a process. Everything is based on processes. How to build the software, how to report a defect, how to note the fix in the CM database, how to keep track of versions, how to ship, etc. Everytime you do anything you follow the process. It's annoying at times, but it allows everything to be repeated exactly. By optimizing the process we can "speed" up development time(our ship dates are known years in advance, so it just means that we make deadlines). It also means that when customers are looking for a company to provide them with a solution, they see that if we have created successful projects in the past, then we probably will be able to repeat the success on a new project.
I think that some things in the CMM are very good, but I also don't see them as being able to apply very well to the open source world, primarily because it requires large amounts of time to make sure the processes are correct, time that developers don't want to spend(Thank goodness for managers).
YMMV,
xPhase
So how does having a digital theatre help show movies filmed in Super 16 blown up to 32mm?
Many 'indie' film makers still shoot in Super 8, 16 and Super 16. The film 'Best in Show' was shot on Super 16 then enlarged to 32(as Super 16 works very well when enlarged to 32mm).
I don't think that digital theatres will replace non-digital theatres will happen for a long time. Most movies are not digital, and they will probably not switch to digital until the quality of digital beats out the quality of film.
--xPhase
I know this is an old discussion, and no one will probably read this, but there is a problem with speed limiters. First, imagine this, your're driving along maxed out at 30 MPH when out of no where a car comes out of nowhere(a driveway) right beside you, you see them, and have less than a second to react. In real life if this happens, you speed up, yup that's right, you speed up over the speed limit and avoid hitting the car. With a speed limiter, then you get hit. I'm sure there are other examples of this, but I'm not thinking of any right now.
--xPhase
Yes, I use it here at work. I even find that some of its features are better than Rational's Ada compiler.
For info see: Ada Core Technologies
They are a private company that has done loads of work on gnat(GNU Ada compiler), and yes, the compiler is available for free, with source.
--xPhase
Supported C++ and C Language Features
This page also says:
--xPhase
Which /bin/sh are you using?
Is it really ash as most BSD distributions provide, is it really bash which several linux distros put as /bin/sh, or is it really AT&T's real /bin/sh.
The point I'm trying to make is that even though most Unix Systems won't start without a /bin/sh, there are a bunch of drop in replacements for /bin/sh.
If I could replace every instance of MSIE with Mozilla, then this would be a fair comparison.
--xPhase
You find that with most libraries. On most large projects one usese the debugger constantly, to find out why the code isn't working "right". So I don't really find digging through the debugger a chore since you then really know what your code is doing.
But when it does compile and run correctly STL is pretty nice!
Agreed.
I'm looking forward to somebody starting over some day and coming up with a language that supports generic programming as well as C++, but which doesn't have the terrible syntax of C++ templates. It must be possible.
People may not like me for this answer, but how about Ada? I've not had any problems with their generics(Similar to templates) on the client side(I'm not writing the generics).
For example: .. 900;
type Light_Spectrum is range 10
package Wavelength_IO is new Ada.Text_IO.Integer_IO(Light_Spectrum);
Defines a new type(Wavelength) which accepts the integers from 10 to 900(inclusive), then instantiates an IO package for it. This allows for basic input and output of the type Wavelength.
NOTE: Wavelength is not an integer, it's stored however the compiler wants to store it.
Here's part of the declaration of Ada.Text_IO.Integer_IO.
generic := Num'Width;
type Num is range <>;
package Integer_Io is
Default_Width : Field
This says that type Num is some range that we don't know yet, that's it, then just use Num where ever you would use the Actual type.
I like it, but I'm really tired so this might be garbled somewhat.
Washington State(at least when I used to live there) had different sales taxes in Pierce and King counties. I think Pierce was 7.25 and King was 8, but I may not be remembering correctly.
:)
Of course that's why you just go down to OR to buy anything
--xPhase
I have only one problem with point number 3 in your argument. I think that you should have the option to license software, or to purchase it. If I am a company, would I rather buy 40 copies of application A, or buy a site license and install the software on each new PC I need it on?
On the other hand, it is easier for me to buy 2 copies of some game to play on multiple computers at once then it is for me to license those two copies.
--xPhase
Why couldn't you just have an app like:
int main(){
return 0;
}
?
Then you wouldn't have any windows pop-up at all?
As far as I know Wasabi seems to be doing fine.
They even work in the embedded systems area.
No, I haven't seen any financial reports, but I've not heard of any trouble either.
--xPhase
I don't think that there is a very slow rate of adoption in the corporate setting for Windows XP. Most corporate settings are just now getting around to upgrading to Win2k.
I held a job supporting Cigna Healthcare's last big upgrade, which was about a year after Win2k, and they were upgrading all client PC's to NT4. My current company still use NT4 on the PCs, Solaris 2.7(and we just upgraded to this late 2001) on most of the suns, and HP11i on the HP-UX machines.
So corporate customers generally wait a long time to upgrade, to allow stability issues to be fixed, and to allow the applications that they use to catch up to the operating system.
--xPhase
Sorry for not making a huge long rambling post, but you really should check out the Risks Digest
--xPhase
Actually I think that it is illegal for Video Rental companies to sell information about you rental habits. At least in the US.
See: Here
--xPhase
Ah, but his argument(I think) was that even after a small number of people started getting arrested for drug possesion, most users did not quit.
I don't really know if 90% would stop using out of fear as many people think that "It can't happen to me", "They'll never catch/find/identify me", or "I'm above the law". Those people need to be arrested themselves before changing their behavior.
Anyway, I don't think anyone has any hard facts to support either of these arguments.
--xPhase
We allow corporations to rip off huge quantities of code and call it "the Windows (R) network stack"
Parts of the Win network(TCP/IP) stack are based off of the BSD network(TCP/IP) stack, but then so is every network(TCP/IP) stack, as it was first implemented by BSD(Way back when, but I don't remember the date). This is due to the BSD license, and most BSD people have no problem with that. Also, Microsoft admits to including code by the CSRG at Berkely, it's listed in the release notes(i.e. This software includes portions written by the CSRG at UCB(something like that). So that point has nothing to do with the rest of your argument.
Also, it is wrong to make unauthorised copies of Microsoft's code, they have the _copyright_ on it, because they wrote it. You still have the ability to go and get the BSD TCP/IP stack and fix it yourself.
blah
--xPhase
The requirement's don't say that the libraries must be free. I'd assume that if they were looking at C# and VB.Net that cost is not an issue in this case.
--xPhase
You can do cross platform GUI building using the QT library. It supports all major platforms.
--xPhase
Yeah, this is posting to an old discussion, but the FreeBSD ultrasparc port is very close to self hosting. As the port already boots multi-user on PCI based sun4u systems. Also there is a PowerPC port of FreeBSD(don't know the status).
The problem is that even "safe" Java has had security problems. Not relating to the language itself necessarily, but relating to browser/platform implementations.
See the Risks Digest:
17.39
17.83
18.18
and there are many more listed in the archives.
So until the languge/CLR mature enough, then there will be more problems with an insecure language.
Also, note that most early Java security problems were found because sun encouraged people to find them, and then Sun would fix the problems. Microsoft doesn't want people to find and disclose bugs in it's software, so it may take longer to mature security wise.
--xPhase
P.S. pardon any spelling errors, i'm tired.
Agreed, but this is 1/3 of 1648 developers of sourceforge. What % of the total number of developers on sourceforge is this?
--xPhase
This study is meaningless.
Hi we sent email to sourceforge members, and linux kernel mailing list members, and got a lots of numbers.
First, the response rate from Sourceforge was 34.2%, which is not representative of most of sourceforge. How many of those contacted were in active development? How many alpha projects are there on sourceforge that have had little development, and will continue with little development?
Second, there was only a 2.4% response by the lkml members. Would this 2.4% be the active members of the list, or mainly the people that don't participate?
Third, this only accounts for people on sourceforge projects or linux, which may not be representative of all open source projects. Yes it would be very difficult to survey more members, but how do we know that sourceforge members and linux hackers are not different than other projects?
Fourth, there listing of open source principles(slide 8) is only representative of a specific group of open source developers. The intellectual property ideas("Free speech, not free beer", and copyleft) don't apply to people not using the gpl. This may mis-represent the people who participated in the survey.
This survey is much more useful if, instead of claiming to represent all open source developers, it admitted to being primarily about linux/gpl developers. Or if more information was given about what projects(activity, license, and activity by developer) the participants of the sourceforge survey were involved in.
The survey is interesting if looked at in relation only to the sourceforge community, but is not able to be applied to all open source developers.
--xPhase
I saw a poster at some gas station the other night, apparently they have sprint pre-paid internet access available.
You just buy the card, then run with it.
Ugh!
--xPhase
I'm not commenting on whether or not someone should post this information on a website.
I would like to mention that material on how to make bombs has been in circulation since before the internet, and it will be in circulation even if all the sites with HOW-TO's get taken down.
Also, how many Afghani terrorists have internet accounts? I mean other than John Katz's friend?
Whether or not I agree with the information being available on-line, I do not feel that it puts our freedom in any sort of jeopardy.
--xPhase
Does anyone here know if FreeBSD is committed to adopting NetBSD's rc system by some specified release? 4.5? 5.0?
According to JKH, there is no definate plan for the integration, so it may never happen. There is currently an rc system used by FreeBSD, and it(as well as NetBSD's) are completely different ideas than the way SysV init works. There are some that don't like the way init works.
I believe 5.0 will bring FreeBSD's kernel up to rough parity with the Linux kernel.
Many will agree that in some areas FreeBSD's kernel is superior to the linux kernel, and in some areas it is lacking. I don't think that there will ever be a rough parity between the two. I think that FreeBSD will always be better at some things while linux will be better at others.
blah
--xPhase