It would have been great if the designers of Star Trek did a series on the fall of Star Fleet. It would be creative, dark and interesting change for the Star Trek universe. It will never happen sadly.
How can a software company profit from a free software (only) solution?
If I invest money (engineering staff) and time into developing a solid application, and a group of hackers decide to branch the software into another competeing software and release it for free, what can I do? Sure you may say a similar event occured between Microsoft and Netscape, but the difference is Microsoft developed their software independantly they did not use Netscape software.
I often hear that open software companies benefit from the many-eyes form of debugging, but can that replace an expert quality assurance staff dedicated to your product. What about a testing for a software on a newly released hardware product?
I think you are making the wrong argument. No one is forcing you to use elightenment or any other desktop. I don't believe there is any way KDE/GNOME can compete agaist microsoft who dedicates 100% of its time to figuring out what the other 99% of people want. They also have the resources to provide dynamic content to their customers through there alliances with large companies.
Example: you can open up microsoft money and get live stock quotes as well as business news or advice from msn money. You cannot do this from gnucash because the maintainers cannot supply the information to you. That is a problem you cannot solve with code alone!
I always wanted to know why we should compete on the desktop? I don't think there is a need for grandma to be using *BSD or linux. The whole concept of the desktop PC is a dying concept anyway. That concept is being replaced by a decentralized computing environment (cell phones, pdas, tivos, etc). This is where the *nix community should strike providing backend support and embedded support.
Even microsoft is strategically positioning itself away from the desktop market (dot net, xbox, etc). By the time linux makes a crack in the desktop environment the desktop environment will not even be an issue worth fighting for.
Having a diverse environment of GUIs is great for open source because it give engineers plenty of options for implementing *nux on an console. i.e: If I wanted to implement a tivo like device on the linux kernel, I have tons of GUI environments to choose from each with a different functionality.
I respectfully disagree. The original Star Trek and Star Trek TNG transcended the Sci Fi stigma. I know a lot of non geeky people who watch Star Trek like it is just another show. You might have forgotten that the first star trek movie and first contact we smash hits. Too bad Bablyon 5 could not get past its scifi stigma.
I thought intermmediate format was jitted at run time. Then the results are cached for when you run that code again. The theory is that although you take a perfomance penalty in the beginning the translater or "jitter thingy" would optimize the code to your hardware specs and when that section of code is called again it will have the optimized code ready.
I read that translation of the code is more efficient then translation at installation because people do not use every nook and cranny of the software. Why spend time translating the whole thing? I could be wrong about this.
As a linux supporter I am suprised we did not come up with this innovation first are we sleeping on the job? We can't afford to be trail blazers. I should just shut up and code.
One thing people fail to talk about is stability with respect to scaleability. UNIX/Linux is extremely stable when you have small components interacting with each other. When UNIX/Linux apps start to scale things start to get messy. Currently many linux applications have problems with interoperability with each other. KDE apps work in KDE space and GNOME apps work in GNOME space.
People often complain about how unstable Internet Explorer is, but when you compare it to Konquer you start to see similarities in the frequency of crashes between the two. The problem is while they both crash at the same rate, Microsoft's IE has more features and displays more sophisticated webpages. Similar comparisons can be made about MS Office vs. Star Office. While they are both relatively stable Star Office is a piss poor (featurewise) when compared to MS Office.
The moral of my story is that Microsoft scales better than Linux. This is not to bash Linux in anyway, since I am an avid user/developer of Linux software. The problem is that while I believe the Open Source philosophy is extremely important to our developement as a human race, Microsoft's (Closed Source philosophy) is delivering new and sophisticated technology while we UNIX/Linux coders are playing copy cat.
I must be gullible,since it took me until section 3 (What do you mean I can't restore my files?) and a quick glance at my calendar to figure what the hell was going on.
Not to be an asshole but by definition a professional is traditionally educated and certified in the field they are working in.
"profession n. an occupation requiring advanced academic training, as medicine, law, etc.[webster]"
This is not to say that a person from a liberal arts cannot be a great programmer or computer theorist.
Companies tend to higher untrained "professionals" when trained computer expert supply is low and demand for products are high. When the economy shifts downward they then start looking at the dictionary definition of professional
As a electrical/computer engineer and a soon to be CS graduate student, I will try to expain this as best I can.
Computer engineering has its roots in electrical engineering. Electrical engineers do not only learn about circuit analysis they learn about microwave transmission, waveguides, transmission lines, optics, analog/digital communications, neural networks, etc.
Since computer technology has its extreme set of complications such as VLSI design, embedded system design, digital networks, microprocessor design, etc, the computer engineering field was created to face these specialized challenges.
Computer Scientists on the other hand tend to learn software engineering principles, theory, programming languages, and there math coures tend to be less intensive than an electrical engineers by default. This by no means implies that computer scientists do not or cannot take hardcore math classes.
Computer engineers do take computer science classes such as computer architecture, assembly, object oriented programming.
If you equally love hardware and software pick computer engineering. Although I believe there is more long term security in the traditonal fields of CS and EE.
If this still does not help, you can do what I did, which is take them all:>
Yes sony sold 2,200,000 playstation 2 with 4,000,000 games across 35 titles, but that does not make hurt the playstation per se. Sony's philosophy is to saturate the market with games and let the buyers sort it out. Nintendo is a little more subtle, they create great games that sell alot. Both of these styles work well. I will be definately buying myself a playstation 2 because I believe the programmers who are complaining now will adapt when they get used to the technology.
IMHO Nintendo learned its lesson with respect to cartrige based games and will do extremely well with the dolphin. Square, Konami, Namco, etc are already on board for the playstation 2 and I trust they will figure the kinks out by the time the Cube comes out.
My only fear is that the playstation 2 might be overextending itself. I like the playstation because I could just stick in games and not worry about it. I did not have to deal with hardware problems like the ones games on PC have. But it will be nice to have a DVD player for my TV.
Yes, nuclear power is a safe form of power in the near-term.
The critical issue you miss is the effects of nuclear power on the long-term. Containment of nuclear wastes is an ongoing problem for us now and the future. Now I am no fan of coal either, but don't be foolish enough to think that nuclear waste is a miracle energy.
You can't do sound on a CPU because the output power requirements would be too high. Same thing with video. You will always need boards to perform the dirty work. The CPU can speed things up by taking some of the slack off of the devices to speed things up.
Also if they were magically able to put all of this functionality into a chip they would not call it a CPU.
If you could afford a good sound and video card, I don't see any good reason for doing this!
Laptops users don't quite have that option! SWAR is a tad helpful in that respect.
I do agree with the idea that on chip 3D acceleration would not help too much on a PC with the latest 3D card.
I don't think that streaming media is accelerated by the video card. That would be helpful to PC users. I guess Intel is waiting util broadband cable access hits almost every home and streaming media hits it big.
You can do jumps by adding the program counter. You can also do conditional branches with NAND and and ADD. Risc is beautiful!
Re:Dying for PCs - but PCs are not everything!
on
Vanishing Game Genres
·
· Score: 1
Good luck to you! I played Mr. Do on the MAME program and it is 10 times more addictive then some of the latest games on the PC.
Re:Dying for PCs - but PCs are not everything!
on
Vanishing Game Genres
·
· Score: 1
You are absolutely correct. I cannot stand the way PC games operate. I heard Deus Ex was a great game so I downloaded the demo to test it out and it was horribly slow, this is with a TNT2/450MHz K6/128MB of ram. What the hell do I need to do to keep up with the computer gaming market! All I ask is that the PC game market create a standard base game system and release standard (OPEN) interfaces that scales well with newer systems [Not DirectX]. Also they should simplify gameplay. Yeah consoles are not very flexible but who cares! All of my console games work! Best yet I don't have to install conosle games either!
I nearly choked when I saw the visuals for Chrono Cross and realised I was playing on a system with 2MB of RAM and 90MHz proc. Many PC game developers do not even spend the time pushing technology to the limit before jumping on to another techonology.
That produces games that are incompatible with older systems, and games that are buggy. At least it's pretty:>.
As a programmer and electrical engineer I have to disagree with your generalization. I think the problem is a sign of a systemmic problem with American computer games. Many games focus on complex and precise details, but fails to deliver good gameplay. When I come home from work I don't want to study a 200 page manual on how to get started, and then another 20 pages on how to get it configured correctly (I do that enough at work already). I want to just want to pop the game in press start and have a fun time.
The Japanese may not be the most innovative when it comes to games but they create games that are fun and enjoyable. I would prefer to play a well crafted Playstation game at 320x200 resolution that garunteed to work than a 1024x768 resolution PC game with the latest technology that lacks in gameplay.
Play Chrono Cross and you'll see what I am talking about!
I do let gameplay slide for some games like Nethack because they are extremely fun and interesting when you get into it.
Yes your correct, ignorance about the distribution model has burned alot of potential linux users. The "crazy" distribution model is a why linux is here today. This is why you can find versions of linux in TIVO's, embedded chips, and multiple OSs. This form of ignorance is not only a problem with CEOs and CFOs but with computer professionals, BSD hackers, and even some Linux community members to.
It is hard to prove your point to ignorant people, but at least the community does not take it lying down.
P.S. I would not want you setting up Oracle servers for my company, cause if you don't tell the retailers what version of version of NT you have you might just get the wrong one. Ignorance is Bliss!
I do agree with the fact that there are definately some problems with the Linux community's approach to structured software developement. But I do not agree to your theory that Linux is made by nonprofessionals. There are expert programmers and software engineers working in the Linux community as well as script kiddies and zealots.
We have to make distinctions between the Linux (kernel) and the software that runs on the kernel. Redhat != Linux! And you should also be thankful that alot of the software you are running in BSD was created in Linux and ported over to BSD and an even larger amount of software were developed for both operating systems. You could make BSD as flashy as Linux if you wanted. That aside FreeBSD is an excellent UNIX system that I would recommend anyone to use. I personally don't like Redhat, since I was burned from the older versions.
When arguing about Linux you must take care to discuss the deficiencies in the kernel rather than generalizing about software written for that operating system. With NT on the other hand, I could talk about the GUI and system software because they are tightly integrated into the system. I cannot do that with Linux, because linux is a kernel, and I can make the system look or act anyway I like it to. (TIVO is evidence of my point).
Is it the Linux community's fault or RedHat's fault that you had problems with xconfigurator and linuxconf?
Thats not the point the point was that you cannot have two instructions to handle it. And your right a dotfile would have similar problems to that of a database. My issue is that by raising the configuration issue a to whole new level of abstaction you bring forth more problems than you solve.
I don't think it would be that simple. I don't mean to sound cynical but geeze two instructions! What happens if I run this program over a network, what is the limit of reg_values I can have, what happens if the reg_value is not there, how do I detect registry corruption, what happens if two people try to access the same keys at the same time, could malicious programs delete values of your registry, could registry information be stolen, etc. Even with reusable code things are still complex.
Speaking about deleting registry values how could I remove values with this two instruction set. Can I allocate space for large registry values.
What would happen if I did this:> set_reg_value(user.thisapp.version.property, large_buf[20000]); ?
I guess the/. community prefers a whole level of abstraction greater than the simplicity of the dotfile (gathering from the about of high scoring registry pundits). I guess I don't know what eXtensible Markup Language is, but when I said ascii file format I meant that it is easy to edit in any text editor with only knowledge of the keys and possible data entries. I do agree there should be a common place for configuration data, but I do not nessicarily believe you need a database (registry) to manage them.
I'll give you an example of my fear with the rpm database. I was trying to get all of the depencies I needed for a program, and then I did something no one should ever do, I forced a rpm in. That not only screwed up the database but forced me to reinstall the entire distribution. I do not have a problem with the concept of a registry. The concept is interesting and does have its particular benifits on paper. The practice is what worries me. Dotfiles work and they have for years, if it is not broken why try to fix it.
9 - Permission system for both read & write access to the registry
This is nothing you cannot do with dotfile
8 - Versioning, Integrity & Duplication checking included - duplicate entries not allowed! a) Duplicate entries not allowed b) Permission system on how to handle version conflicts
If the structure of the dotfile is written well it will be backwards compatible, for instance if I put a new key and I am running an old version of the program, the dotfile parser can just ignore it or send a error message to the user while using options that it is compatable with. I am afraid of the alternatives the registry gives me. I could also have multiple dotfiles for mutlipule versions of the programs. A nice checksum will take care of integrity.
7 - Ability to delete any desired application's registry entry
vi.whatever
5 - Roll-back features included to return to any previous state
cp.whatever/old_dotfile_archive
4 - Cached for performance benefits (?)
Why would I need a cache for my 1K dot file?
3 - Access granted through system only a) This ensures no 'hidden' entries in registry
less.whatever did I mention dotfiles are in ascii
2 - XML format
Ascii format:>
1 - Meta-Data Only!
Any data I want! I am sure the registry may be great for managing large scale systems, but I don't need a complex database for my xmms. I also don't want to learn the usually complex APIs to implement it in my simple text editor.
Do not get me wrong there is no standard dot file format, there is really no standard way of parsing a dot file either. My major gripes with registries would be major gripes that go for any large database, in order for your code to be portable everyone must support that database. Dot files are truely independant because they are written in ascii. Yes I do admit when I do a dir -al it there are 100 dotfiles some of which I don't know what program they belong to.
Security-wise access to dotfiles can be handled just like any other file. You don't need a separate security abstaction to handle it. I have had problems in some cases with GUI dotfile managers like network configuration tools, but if written correctly as a series of keys and entrys it would be very hard to f*ck a dotfile parser up. Well written comments and/or documentation help the user understand whats in a dotfile and be able to configure them.
IMHO there is a fine line between configuration files and well structured dotfiles. For example crontab has a great structure that can be easily edited by humans or by a parser, while the sysV init or BSD are a set configuration scripts which is not that easy to parse by computer and written by hand. Most programmers I think will have a lot more of an easy time programming for dotfiles rather than for the registry. And its great to know just what information your program is saving about you.
Slashdot Mirror
It would have been great if the designers of Star Trek did a series on the fall of Star Fleet. It would be creative, dark and interesting change for the Star Trek universe. It will never happen sadly.
How can a software company profit from a free software (only) solution?
If I invest money (engineering staff) and time into developing a solid application, and a group of hackers decide to branch the software into another competeing software and release it for free, what can I do? Sure you may say a similar event occured between Microsoft and Netscape, but the difference is Microsoft developed their software independantly they did not use Netscape software.
I often hear that open software companies benefit from the many-eyes form of debugging, but can that replace an expert quality assurance staff dedicated to your product. What about a testing for a software on a newly released hardware product?
Am I crazy or what?
I think you are making the wrong argument. No one is forcing you to use elightenment or any other desktop. I don't believe there is any way KDE/GNOME can compete agaist microsoft who dedicates 100% of its time to figuring out what the other 99% of people want. They also have the resources to provide dynamic content to their customers through there alliances with large companies.
Example: you can open up microsoft money and get live stock quotes as well as business news or advice from msn money. You cannot do this from gnucash because the maintainers cannot supply the information to you. That is a problem you cannot solve with code alone!
I always wanted to know why we should compete on the desktop? I don't think there is a need for grandma to be using *BSD or linux. The whole concept of the desktop PC is a dying concept anyway. That concept is being replaced by a decentralized computing environment (cell phones, pdas, tivos, etc). This is where the *nix community should strike providing backend support and embedded support.
Even microsoft is strategically positioning itself away from the desktop market (dot net, xbox, etc). By the time linux makes a crack in the desktop environment the desktop environment will not even be an issue worth fighting for.
Having a diverse environment of GUIs is great for open source because it give engineers plenty of options for implementing *nux on an console. i.e: If I wanted to implement a tivo like device on the linux kernel, I have tons of GUI environments to choose from each with a different functionality.
I respectfully disagree. The original Star Trek and Star Trek TNG transcended the Sci Fi stigma. I know a lot of non geeky people who watch Star Trek like it is just another show. You might have forgotten that the first star trek movie and first contact we smash hits. Too bad Bablyon 5 could not get past its scifi stigma.
I thought intermmediate format was jitted at run time. Then the results are cached for when you run that code again. The theory is that although you take a perfomance penalty in the beginning the translater or "jitter thingy" would optimize the code to your hardware specs and when that section of code is called again it will have the optimized code ready.
I read that translation of the code is more efficient then translation at installation because people do not use every nook and cranny of the software. Why spend time translating the whole thing? I could be wrong about this.
As a linux supporter I am suprised we did not come up with this innovation first are we sleeping on the job? We can't afford to be trail blazers. I should just shut up and code.
One thing people fail to talk about is stability with respect to scaleability. UNIX/Linux is extremely stable when you have small components interacting with each other. When UNIX/Linux apps start to scale things start to get messy. Currently many linux applications have problems with interoperability with each other. KDE apps work in KDE space and GNOME apps work in GNOME space.
People often complain about how unstable Internet Explorer is, but when you compare it to Konquer you start to see similarities in the frequency of crashes between the two. The problem is while they both crash at the same rate, Microsoft's IE has more features and displays more sophisticated webpages. Similar comparisons can be made about MS Office vs. Star Office. While they are both relatively stable Star Office is a piss poor (featurewise) when compared to MS Office.
The moral of my story is that Microsoft scales better than Linux. This is not to bash Linux in anyway, since I am an avid user/developer of Linux software. The problem is that while I believe the Open Source philosophy is extremely important to our developement as a human race, Microsoft's (Closed Source philosophy) is delivering new and sophisticated technology while we UNIX/Linux coders are playing copy cat.
I must be gullible,since it took me until section 3 (What do you mean I can't restore my files?) and a quick glance at my calendar to figure what the hell was going on.
You guyz!!!
Not to be an asshole but by definition a professional is traditionally educated and certified in the field they are working in.
"profession n. an occupation requiring advanced academic training, as medicine, law, etc.[webster]"
This is not to say that a person from a liberal arts cannot be a great programmer or computer theorist.
Companies tend to higher untrained "professionals" when trained computer expert supply is low and demand for products are high. When the economy shifts downward they then start looking at the dictionary definition of professional
As a electrical/computer engineer and a soon to be CS graduate student, I will try to expain this as best I can.
:>
Computer engineering has its roots in electrical engineering. Electrical engineers do not only learn about circuit analysis they learn about microwave transmission, waveguides, transmission lines, optics, analog/digital communications, neural networks, etc.
Since computer technology has its extreme set of complications such as VLSI design, embedded system design, digital networks, microprocessor design, etc, the computer engineering field was created to face these specialized challenges.
Computer Scientists on the other hand tend to learn software engineering principles, theory, programming languages, and there math coures tend to be less intensive than an electrical engineers by default. This by no means implies that computer scientists do not or cannot take hardcore math classes.
Computer engineers do take computer science classes such as computer architecture, assembly, object oriented programming.
To sum it up:
Electrical Engineers: Physics, computer architecture, math, minimal programming.
Computer Engineers: Computer architecture, firmware design, math, low-level programming.
Computer Scientists: Software engineering, programming languages, theory, math.
If you equally love hardware and software pick computer engineering. Although I believe there is more long term security in the traditonal fields of CS and EE.
If this still does not help, you can do what I did, which is take them all
Who said buying a how without furniture was a bad thing?
Yes sony sold 2,200,000 playstation 2 with 4,000,000 games across 35 titles, but that does not make hurt the playstation per se. Sony's philosophy is to saturate the market with games and let the buyers sort it out. Nintendo is a little more subtle, they create great games that sell alot. Both of these styles work well. I will be definately buying myself a playstation 2 because I believe the programmers who are complaining now will adapt when they get used to the technology.
IMHO Nintendo learned its lesson with respect to cartrige based games and will do extremely well with the dolphin. Square, Konami, Namco, etc are already on board for the playstation 2 and I trust they will figure the kinks out by the time the Cube comes out.
My only fear is that the playstation 2 might be overextending itself. I like the playstation because I could just stick in games and not worry about it. I did not have to deal with hardware problems like the ones games on PC have. But it will be nice to have a DVD player for my TV.
Yes, nuclear power is a safe form of power in the near-term.
The critical issue you miss is the effects of nuclear power on the long-term. Containment of nuclear wastes is an ongoing problem for us now and the future. Now I am no fan of coal either, but don't be foolish enough to think that nuclear waste is a miracle energy.
You can't do sound on a CPU because the output power requirements would be too high. Same thing with video. You will always need boards to perform the dirty work. The CPU can speed things up by taking some of the slack off of the devices to speed things up.
Also if they were magically able to put all of this functionality into a chip they would not call it a CPU.
If you could afford a good sound and video card, I don't see any good reason for doing this!
Laptops users don't quite have that option! SWAR is a tad helpful in that respect.
I do agree with the idea that on chip 3D acceleration would not help too much on a PC with the latest 3D card.
I don't think that streaming media is accelerated by the video card. That would be helpful to PC users. I guess Intel is waiting util broadband cable access hits almost every home and streaming media hits it big.
You can do jumps by adding the program counter. You can also do conditional branches with NAND and and ADD. Risc is beautiful!
Good luck to you! I played Mr. Do on the MAME program and it is 10 times more addictive then some of the latest games on the PC.
You are absolutely correct. I cannot stand the way PC games operate. I heard Deus Ex was a great game so I downloaded the demo to test it out and it was horribly slow, this is with a TNT2/450MHz K6/128MB of ram. What the hell do I need to do to keep up with the computer gaming market! All I ask is that the PC game market create a standard base game system and release standard (OPEN) interfaces that scales well with newer systems [Not DirectX]. Also they should simplify gameplay. Yeah consoles are not very flexible but who cares! All of my console games work! Best yet I don't have to install conosle games either!
:>.
I nearly choked when I saw the visuals for Chrono Cross and realised I was playing on a system with 2MB of RAM and 90MHz proc. Many PC game developers do not even spend the time pushing technology to the limit before jumping on to another techonology.
That produces games that are incompatible with older systems, and games that are buggy. At least it's pretty
As a programmer and electrical engineer I have to disagree with your generalization. I think the problem is a sign of a systemmic problem with American computer games. Many games focus on complex and precise details, but fails to deliver good gameplay. When I come home from work I don't want to study a 200 page manual on how to get started, and then another 20 pages on how to get it configured correctly (I do that enough at work already). I want to just want to pop the game in press start and have a fun time.
The Japanese may not be the most innovative when it comes to games but they create games that are fun and enjoyable. I would prefer to play a well crafted Playstation game at 320x200 resolution that garunteed to work than a 1024x768 resolution PC game with the latest technology that lacks in gameplay.
Play Chrono Cross and you'll see what I am talking about!
I do let gameplay slide for some games like Nethack because they are extremely fun and interesting when you get into it.
Yes your correct, ignorance about the distribution model has burned alot of potential linux users. The "crazy" distribution model is a why linux is here today. This is why you can find versions of linux in TIVO's, embedded chips, and multiple OSs. This form of ignorance is not only a problem with CEOs and CFOs but with computer professionals, BSD hackers, and even some Linux community members to.
It is hard to prove your point to ignorant people, but at least the community does not take it lying down.
P.S. I would not want you setting up Oracle servers for my company, cause if you don't tell the retailers what version of version of NT you have you might just get the wrong one. Ignorance is Bliss!
I do agree with the fact that there are definately some problems with the Linux community's approach to structured software developement. But I do not agree to your theory that Linux is made by nonprofessionals. There are expert programmers and software engineers working in the Linux community as well as script kiddies and zealots.
We have to make distinctions between the Linux (kernel) and the software that runs on the kernel. Redhat != Linux! And you should also be thankful that alot of the software you are running in BSD was created in Linux and ported over to BSD and an even larger amount of software were developed for both operating systems. You could make BSD as flashy as Linux if you wanted. That aside FreeBSD is an excellent UNIX system that I would recommend anyone to use. I personally don't like Redhat, since I was burned from the older versions.
When arguing about Linux you must take care to discuss the deficiencies in the kernel rather than generalizing about software written for that operating system. With NT on the other hand, I could talk about the GUI and system software because they are tightly integrated into the system. I cannot do that with Linux, because linux is a kernel, and I can make the system look or act anyway I like it to. (TIVO is evidence of my point).
Is it the Linux community's fault or RedHat's fault that you had problems with xconfigurator and linuxconf?
Thats not the point the point was that you cannot have two instructions to handle it. And your right a dotfile would have similar problems to that of a database. My issue is that by raising the configuration issue a to whole new level of abstaction you bring forth more problems than you solve.
I don't think it would be that simple.
:>
I don't mean to sound cynical but geeze two instructions! What happens if I run this program over a network, what is the limit of reg_values I can have, what happens if the reg_value is not there, how do I detect registry corruption, what happens if two people try to access the same keys at the same time, could malicious programs delete values of your registry, could registry information be stolen, etc. Even with reusable code things are still complex.
Speaking about deleting registry values how could I remove values with this two instruction set. Can I allocate space for large registry values.
What would happen if I did this
set_reg_value(user.thisapp.version.property, large_buf[20000]); ?
I guess the /. community prefers a whole level of abstraction greater than the simplicity of the dotfile (gathering from the about of high scoring registry pundits). I guess I don't know what eXtensible Markup Language is, but when I said ascii file format I meant that it is easy to edit in any text editor with only knowledge of the keys and possible data entries. I do agree there should be a common place for configuration data, but I do not nessicarily believe you need a database (registry) to manage them.
I'll give you an example of my fear with the rpm database. I was trying to get all of the depencies I needed for a program, and then I did something no one should ever do, I forced a rpm in. That not only screwed up the database but forced me to reinstall the entire distribution.
I do not have a problem with the concept of a registry. The concept is interesting and does have its particular benifits on paper. The practice is what worries me. Dotfiles work and they have for years, if it is not broken why try to fix it.
Fine, I will make my rebuttle in reverse order:
.whatever
.whatever /old_dotfile_archive
.whatever did I mention dotfiles are in ascii
:>
9 - Permission system for both read & write access to the registry
This is nothing you cannot do with dotfile
8 - Versioning, Integrity & Duplication checking included - duplicate entries not allowed! a) Duplicate entries not allowed b) Permission system on how to handle version conflicts
If the structure of the dotfile is written well it will be backwards compatible, for instance if I put a new key and I am running an old version of the program, the dotfile parser can just ignore it or send a error message to the user while using options that it is compatable with. I am afraid of the alternatives the registry gives me. I could also have multiple dotfiles for mutlipule versions of the programs. A nice checksum will take care of integrity.
7 - Ability to delete any desired application's registry entry
vi
5 - Roll-back features included to return to any previous state
cp
4 - Cached for performance benefits (?)
Why would I need a cache for my 1K dot file?
3 - Access granted through system only a) This ensures no 'hidden' entries in registry
less
2 - XML format
Ascii format
1 - Meta-Data Only!
Any data I want! I am sure the registry may be great for managing large scale systems, but I don't need a complex database for my xmms. I also don't want to learn the usually complex APIs to implement it in my simple text editor.
Do not get me wrong there is no standard dot file format, there is really no standard way of parsing a dot file either. My major gripes with registries would be major gripes that go for any large database, in order for your code to be portable everyone must support that database. Dot files are truely independant because they are written in ascii. Yes I do admit when I do a dir -al it there are 100 dotfiles some of which I don't know what program they belong to.
Security-wise access to dotfiles can be handled just like any other file. You don't need a separate security abstaction to handle it. I have had problems in some cases with GUI dotfile managers like network configuration tools, but if written correctly as a series of keys and entrys it would be very hard to f*ck a dotfile parser up. Well written comments and/or documentation help the user understand whats in a dotfile and be able to configure them.
IMHO there is a fine line between configuration files and well structured dotfiles. For example crontab has a great structure that can be easily edited by humans or by a parser, while the sysV init or BSD are a set configuration scripts which is not that easy to parse by computer and written by hand. Most programmers I think will have a lot more of an easy time programming for dotfiles rather than for the registry. And its great to know just what information your program is saving about you.