There is a big difference here. FreeBSD and Linux both have nearly identical APIs. A few minor #ifdefs to account for Linux peculiarities and you can write pure POSIX/X11 code and be assured that it will run on both.
Porting from Linux to FreeBSD, or vice versa, is as simple as a recompile. Porting from Windows 3.1 to OS/2 3.0 was several magnitudes more difficult. A Unix development firm can trivially support all known unices natively. A Windows development firm had to do major work to support OS/2 natively.
non-usable UI (compared to what CivCTP or Civ 2 had... hey, let's make it more user-friendly by removing all "complexity", like menus...)
I'm not a Windows user, so I thought that was standard:-)
Seriously, the UI really sucks, and the manual doesn't help much. I talked with a friend who have won two games (where does he get the time!) and he talked about saving the game and then retiring just so he could see what his score was. Then I mentioned the "h" button hidden on the unit display and he went gaga. Then he told me about the capital diplomacy menu having different options than the advisor diplomacy menu (which has different options than the "d" button diplomacy menu hidden on the unit display).
Just pretend Civ III is Myst III. There's a hidden control panel there somewhere, and you job is to find it.
Myst III: Exile. A sequel to Myst I, with a Riven (Myst II) feel, and an actual plot! Graphics are cool (realistic panoramic view). Brad Dourif is your thoroughly wicked and entertaining antagonist.
Civilization III: Addicting. Addicting. Addicting. Just two more turns and I'll go to bed. Okay, just two more turns and I'll turn off the alarm clock telling me to get up. A major evolutionary (not revolutionary) update to Civ II. Diplomacy actually works. Conquering the world is damn near impossible.
Civ III is cool! It's also addictive as hell. Don't plan to do anything else the first week you have it. I toyed with it a bit just before Thanksgiving vacation, then started a real game last monday. Duh! I haven't done anything else! My bills are backing up, so that I take them in to work so I can do them because once I get home my brain shuts off and I play Civ III until I fall asleep. I expect to finish the game tomorrow night, then I can finally get some real work done:-)
"Their purpose is not to bind the user to the will of the licensor."
That's exactly what the purpose of all licenses are, including open source licenses.
Funny, I didn't see anything in the FSF's four freedoms about binding people to the wills of other. That antithetical to freedom. (it's also antithetical to contract law, which is why unilateral contracts are Evil)
"The first thing a good license should do is grant unconditional permission to use the software."
In that case, there is no such thing as a good license. Everything should be released into the public domain.
Copyright law has already given the user the right to use the program. No ifs, ands or buts. Since so many commercial licenses say "by using this software you agree to...", I felt it necessary, if redundant, to explicitly assure the user that they can use the software no matter what else the license says. After all, I even know of one person who holds the belief that you may not use GPLd software unless you agree with the philosophical preamble in its entirety.
So the user should have permissions, but those permissions should not be dependent on accepting the license, but the license should inform the user of their permissions. What exactly are you trying to say?
Hmmm, you must be a lawyer, as no one else has so much trouble parsing standard English. Let me restate. You give the user a set of permissions. You then let the user know what these permissions are. You do not keep them secret for the user to guess. Let me give an example: "you may distribute this post to anyone." There. I granted a permission to all readers of this post, and also informed that of it.
"A permission may have conditions attached to it."
Didn't you just say: "The first thing a good license should do is grant unconditional permission to use the software" just five sentences up in this same paragraph?
Yes, I did say that. But double check my post anyway. I had more than one criteria. The right to *use* the software unconditionally is the first of my criteria. Other rights, such as distribution, modification, etc., may be conditional.
Contract law says "different parties have different rights, resources, and privileges, and they can negotiate exchanges of those rights, resources, and privileges under these rules."
Cool. Too bad it doesn't apply to most Open Source licenses. The operative phrase in your quote is "negotiate exchanges". Since that's how I always understood contract law, we must be in agreement on something! But I don't understand where the negotiation or the exchange comes in when I download the Linux kernel and start distributing it. I have negotiated nothing! I have given nothing back to Linus and Friends!
If you look at the typical contract, you will see certain attributes. First, both parties are aware of each other. Second, negotiation of terms is possible even if the negotiation does occur. Third, both parties receive something of benefit. Finally there is an explicit agreement. None of these attributes are present when I download and start legally distributing the Linux kernel. Linus and Friends are not aware of me, or of the fact that I possess a legal copy of the kernel. And it is not possible to negiate terms because a line of communication has not been established (although that communication could be initiated by me). I receive benefit from the kernel, but Linus and Friends receive nothing from me, not even the satisfaction of knowing that I am even using it. Finally, there is no explicit agreement. No signature, no handshake, no verbal "I agree", no clickthrough, no filling out of registration cards, nothing. A transaction of sorts has occured, but there is no contract.
First off, licenses aren't written for end users. Yes, they're purportedly intended to inform a user of their privileges, but the true audience of a license is a judge and a pack of lawyers.
It depends on the license. For the MS EULA, you are certainly correct. But software written by developers that expect/invite/encourage participation in the development process should have a license written for developers. Of course lawyers and judges will be one audience for the license, but they will not be the primary audience.
The next thing is that all licenses are based in contract law. There is no room in copyright law for ranting permissions beyond those explicitly enumerated (and irrevocable) in copyright law.
Here's a sample license: "You may freely copy, distribute, modify, translate, or otherwise transmit and transform this software without restriction." Just how does this qualify as a contract? Where is the agreement? Where is the consideration? Are you saying that the above license is not valid?
...to provide a framework under which two parties can clearly enumerate an exchange of permissions or other benefits.
What exchange? In the case of the Linux (as an example) I have given nothing to Linus Torvalds. No money. No pledges of royalties. No promises that I will ever contribute anything back to the project. I'm not grantingd him any permissions or benefits. Heck, Linus and I have never even met, so how can we possibly exchange anything!
Once that's in place, the author can then impose a license (read: contract) on the user for the software.
Contracts are never "imposed." They must be agreed to voluntarily by both parties.
No, this obligation is a condition to the permission to distribute. The GPL does not take away any rights given to me under copyright law. I can shout to the world "I disagree" and ignore the GPL at my leisure as I give my friend my sole copy of the software under the aegis of first sale. I cannot, under some pedantic interpretations, keep a copy myself after doing so, or give it to more than one person.
And as an aside, I did not necessarily consider the GPL to be a "good" license by my criteria. It is very borderline. The entirety of the license is be based on copyright law, and clause 5 even says the same. Yet clause 5 is attempting to place it under contract law. I can only assume that this is for the purpose of satisfying the overly pedantic lawyers working with the FSF. There's no way you can read the list of four freedoms of the FSF and conclude that RMS thinks software should be distributed under contract.
After seeing half a million OSS licenses, I have concluded that the vast majority of them just don't get it. I'm not talking about the four "freedoms" of the FSF, but rather the freedom of the user not to be insulted by the licensing. Lawyers may love confusing, convoluted and non-parsable legalese, but the users do not.
(The following is my opinion, so please read it as such. When I refer to a "good" open source license, I am making a qualitative assessment, and not trying to set up criteria for any approval process but my own.)
The purpose of open source licenses are to grant the user a broad set of permissions and rights over and above those granted by copyright law. Their purpose is not to bind the user to the will of the licensor. A good open source license must be based on copyright law, not contract law.
The first thing a good license should do is grant unconditional permission to use the software. This should be so basic it to not be worth mentioning, but you would be surprised as some of the licensed submitted. Additionally, the use of the software should not be trigger for anything else. We don't want any EULA's here, thank you. The second thing a good license should do is clearly inform the user of their permissions. These permissions must not be predicated upon acceptance of any agreement. A permission may have conditions attached to it. If there is anything you wish the user NOT to do, make it a condition. Next the license should have a warranty disclaimer, to assure the user that they will not be sued if they contribute stuff to the project. You may (and should if you're a commercial project) include a real warranty as a separate legal document.
Notice that I haven't included anything about what you require the user to do. No blanket obligations. That's on purpose. Open Source and Free Software are NOT about making people do things. It is okay to make an obligation be a condition to a permission. It is not okay to make an obligation be a condition to the entire license. Remember, this is about what the user can and cannot do.
Software licenses as contracts was an invention of the proprietary software industry. There was a time not that long ago when copyright law as very vague as to the status of software. So the industry decided to use contract law instead, and created licenses that had such bizarre phrases in them as "by reading this sentence you agree to the following obligations...". That's bullshit and Open Source and Free Software should have nothing to do with such rubbish.
There's wasn't much discussion on the KRL. Don't really know why. I read it through several times, couldn't grok it, so never bothered commenting. Remember, license-discuss isn't composed of attorneys, so a license that doesn't parse to laymen English won't get much discussion.
Well, you heard of "weak copyleft?" Well this is "stong unrestricted." You get more permissions than the MIT/BSD license (really), but the license agreement must be retained in all distributions. This is different from the MIT/BSD licenses in that they require the license to be included, but not necessarily applied to, any copies or derivations.
That's why this is good news. The licenses rejected were among the most bizarre, redundant or useless licenses yet submitted to OSI. Perhaps that's why they made the news.
Not to name names, but one license had a restrictive patent grant that only applied to GPL'ed operating systems.
And what a bizarre license that was (not to name names). It was essentially the BSD license word for word, with the aforementioned patent grant. Yet you couldn't legally use the software on a BSD licensed operating system.
Another was more of a rant than a license.
A delicious rant to be sure. I quite enjoyed it, despite its wrongheadedness. It could not be approved of course, since it explicitly denied its own validity.
The one license that did pass was the Python Software Foundation License.
Whoohoo! In this age of a million open source licenses, it's nice to see that a sensible license that fills a gap in open source gets approved while the frivolous crap gets flushed.
Okay, my reply was a copout. I apologize. I asked a question and you gave a sincere answer.
I wanted to know what made a GUI "user friendly" and you answered basically "readability". Thank you. I will agree that readability is a big factor. I've seen some GUI installers that weren't readable, but a GUI done right can be more readable than a console program done right.
Corel LinuxOS: Crashed every time it tried to probe the video card, even though the XFree86 documentation clearly says in unambiguous terms not to probe my video card. I found now way around it, and no obvious textmode installer to fall back to. It's the only Linux distro I have never been able to install. It's was also cited as a model for user friendliness before even the newbies gave up on it. (it wasn't an obsolete card)
SuSE Linux: On a friends computer: His card did not have a standard VGA mode (seriously) but that was the mode that the installer wanted to use. On the bright side, at least it was easy to get to a textmode installer. (it wasn't an obsolete card)
Installers have to be, by definition, a one-size-fits-all program. Unfortunately there is no such thing as a one-size-fits-all graphics standard. The PC video display industry is anything BUT standard. As my friend's computer demonstrated, even the VGA display standard is not standard. The only display that every PC can be assured of having is a textmode console.
I'm not prejudiced against graphics. I love the GUI. But I love reliability more. I have learned through painful experience that setting up XFree86 by hand is more reliable than trusting the installer to do it automatically. Go ahead and make your super-duper works-on-anything GUI installer, but keep a textmode installer around for those that require it.
VGA compatible hardware may be a common denominator standard, but that doesn't mean VGA hardware is installed on on your rackmount.
All you need is a console. It could be a Matrox G450+, a VGA display, a Hercules card, an even older MDA card, or even a connection to the serial port. I would recommend the latter as opposed to swapping cards in and out, or course:-)
Past girlfriends, some roomates, and relatives of mine agree that once you have to tab between things to install something the install stops and they look for something else.
We're talking about unix here. If someone can't handle using the tab key, then perhaps, just perhaps, *BSD is not for them.
I've never seen an OS that was truly one-size-fits-all, and I don't ever expect to.
It also gives you more room for onscreen instructions and help and so on.
Hmmm, everytime I've tried a linux GUI installer I get some stoopid 640x480 screen. Doesn't help the readability much.
The big problem with (FreeBSD|Linux) textmode installs that I've seen is not that they are text, but that they are crappy applications.
You mean like that Mandrake installer in a purple and yellow color scheme that makes choosing packages a nightmare?
Re:though the suggestions might be usefull...
on
Homepage Usability
·
· Score: 2
do you make a seperate page, for form incompatible browsers?
By "old" standard, I don't mean the ones the cavemen wrought, but merely something older than last week. If my browser doesn't support the established standards, then that is MY problem, and not yours.
Let's say I take my form incompatible browser to your site. I could either end up with a lot of garbage, in which case I can only conclude that my browser sucks. Or you can throw up a page saying my browser sucks, in which cas I can only conclude that you're an asshole.
As far as browsers go, having used IE 5.0 around the time of the nimda worm... I have suggested in my project that we FLATLY regect anyone who doesn't have THE most up to date web browser, AS A PUBLIC SERVICE!
I certainly empathize with this viewpoint. However, when it comes to the web, I am not a developer, I am a user. When I come to a website that says "go away we don't want you", I can only assume that the webmaster is a) rude, b) way too rich to want business, or c) all of the above. The better solution is to simply say "IE 5.0 has a problem, please upgrade to the fix".
Imagine you drive into a service station to get some gas. You know your car uses gas because you put some in last week. However this fuel pump has a square nozzle and your gas tank has a round hole. When you try to jam the square nozzle in up flashes a sign that says "you drive a Dodge so go away!"
Whaa whaa.... I can't see any of the cool internet sites with my C64 anymore!
What pisses me off, and what you apparently can't understand, is that my browser is compliant with this weeks standards, but that you choose not to support it because it wasn't made by Microsoft or Netscape.
The installer can still ask the same questions, but in a more user friendly manner.
Why does a "more user friendly" installer have to be a GUI? What is there about a GUI that makes things easier? I've asked this question before in other forums, but I've never gotten a straight answer.
To be sure, there are many advantages to a GUI, but I don't see where "user friendly" has anything to do with it.
People are like that. Remember that recent article about the open source OSX developer who was calling it quits? He linked to a bunch of emails trying to show that he was being reasonable when in fact it showed him to be a complete ass.
Everything you make has those five steps. Decide what beer, choose ingredients, brew, taste, store in cellar. A lot a people got into Extreme Programming because they thought it would excuse them from some of those onerous steps. Then they discovered that they're not so onerous after all.
Slashdot Mirror
There is a big difference here. FreeBSD and Linux both have nearly identical APIs. A few minor #ifdefs to account for Linux peculiarities and you can write pure POSIX/X11 code and be assured that it will run on both.
Porting from Linux to FreeBSD, or vice versa, is as simple as a recompile. Porting from Windows 3.1 to OS/2 3.0 was several magnitudes more difficult. A Unix development firm can trivially support all known unices natively. A Windows development firm had to do major work to support OS/2 natively.
non-usable UI (compared to what CivCTP or Civ 2 had... hey, let's make it more user-friendly by removing all "complexity", like menus...)
:-)
I'm not a Windows user, so I thought that was standard
Seriously, the UI really sucks, and the manual doesn't help much. I talked with a friend who have won two games (where does he get the time!) and he talked about saving the game and then retiring just so he could see what his score was. Then I mentioned the "h" button hidden on the unit display and he went gaga. Then he told me about the capital diplomacy menu having different options than the advisor diplomacy menu (which has different options than the "d" button diplomacy menu hidden on the unit display).
Just pretend Civ III is Myst III. There's a hidden control panel there somewhere, and you job is to find it.
Myst III: Exile. A sequel to Myst I, with a Riven (Myst II) feel, and an actual plot! Graphics are cool (realistic panoramic view). Brad Dourif is your thoroughly wicked and entertaining antagonist.
Civilization III: Addicting. Addicting. Addicting. Just two more turns and I'll go to bed. Okay, just two more turns and I'll turn off the alarm clock telling me to get up. A major evolutionary (not revolutionary) update to Civ II. Diplomacy actually works. Conquering the world is damn near impossible.
Civ III is cool! It's also addictive as hell. Don't plan to do anything else the first week you have it. I toyed with it a bit just before Thanksgiving vacation, then started a real game last monday. Duh! I haven't done anything else! My bills are backing up, so that I take them in to work so I can do them because once I get home my brain shuts off and I play Civ III until I fall asleep. I expect to finish the game tomorrow night, then I can finally get some real work done :-)
Funny, I didn't see anything in the FSF's four freedoms about binding people to the wills of other. That antithetical to freedom. (it's also antithetical to contract law, which is why unilateral contracts are Evil)
Copyright law has already given the user the right to use the program. No ifs, ands or buts. Since so many commercial licenses say "by using this software you agree to...", I felt it necessary, if redundant, to explicitly assure the user that they can use the software no matter what else the license says. After all, I even know of one person who holds the belief that you may not use GPLd software unless you agree with the philosophical preamble in its entirety.
Hmmm, you must be a lawyer, as no one else has so much trouble parsing standard English. Let me restate. You give the user a set of permissions. You then let the user know what these permissions are. You do not keep them secret for the user to guess. Let me give an example: "you may distribute this post to anyone." There. I granted a permission to all readers of this post, and also informed that of it.
Yes, I did say that. But double check my post anyway. I had more than one criteria. The right to *use* the software unconditionally is the first of my criteria. Other rights, such as distribution, modification, etc., may be conditional.
Contract law says "different parties have different rights, resources, and privileges, and they can negotiate exchanges of those rights, resources, and privileges under these rules."
Cool. Too bad it doesn't apply to most Open Source licenses. The operative phrase in your quote is "negotiate exchanges". Since that's how I always understood contract law, we must be in agreement on something! But I don't understand where the negotiation or the exchange comes in when I download the Linux kernel and start distributing it. I have negotiated nothing! I have given nothing back to Linus and Friends!
If you look at the typical contract, you will see certain attributes. First, both parties are aware of each other. Second, negotiation of terms is possible even if the negotiation does occur. Third, both parties receive something of benefit. Finally there is an explicit agreement. None of these attributes are present when I download and start legally distributing the Linux kernel. Linus and Friends are not aware of me, or of the fact that I possess a legal copy of the kernel. And it is not possible to negiate terms because a line of communication has not been established (although that communication could be initiated by me). I receive benefit from the kernel, but Linus and Friends receive nothing from me, not even the satisfaction of knowing that I am even using it. Finally, there is no explicit agreement. No signature, no handshake, no verbal "I agree", no clickthrough, no filling out of registration cards, nothing. A transaction of sorts has occured, but there is no contract.
My principle motivation in being involved in Linux is to weaken Microsoft.
As an aside...
What happens when Microsoft becomes weakened? Will you subsequently abandon Linux development?
First off, licenses aren't written for end users. Yes, they're purportedly intended to inform a user of their privileges, but the true audience of a license is a judge and a pack of lawyers.
...to provide a framework under which two parties can clearly enumerate an exchange of permissions or other benefits.
It depends on the license. For the MS EULA, you are certainly correct. But software written by developers that expect/invite/encourage participation in the development process should have a license written for developers. Of course lawyers and judges will be one audience for the license, but they will not be the primary audience.
The next thing is that all licenses are based in contract law. There is no room in copyright law for ranting permissions beyond those explicitly enumerated (and irrevocable) in copyright law.
Here's a sample license: "You may freely copy, distribute, modify, translate, or otherwise transmit and transform this software without restriction." Just how does this qualify as a contract? Where is the agreement? Where is the consideration? Are you saying that the above license is not valid?
What exchange? In the case of the Linux (as an example) I have given nothing to Linus Torvalds. No money. No pledges of royalties. No promises that I will ever contribute anything back to the project. I'm not grantingd him any permissions or benefits. Heck, Linus and I have never even met, so how can we possibly exchange anything!
Once that's in place, the author can then impose a license (read: contract) on the user for the software.
Contracts are never "imposed." They must be agreed to voluntarily by both parties.
No, this obligation is a condition to the permission to distribute. The GPL does not take away any rights given to me under copyright law. I can shout to the world "I disagree" and ignore the GPL at my leisure as I give my friend my sole copy of the software under the aegis of first sale. I cannot, under some pedantic interpretations, keep a copy myself after doing so, or give it to more than one person.
And as an aside, I did not necessarily consider the GPL to be a "good" license by my criteria. It is very borderline. The entirety of the license is be based on copyright law, and clause 5 even says the same. Yet clause 5 is attempting to place it under contract law. I can only assume that this is for the purpose of satisfying the overly pedantic lawyers working with the FSF. There's no way you can read the list of four freedoms of the FSF and conclude that RMS thinks software should be distributed under contract.
After seeing half a million OSS licenses, I have concluded that the vast majority of them just don't get it. I'm not talking about the four "freedoms" of the FSF, but rather the freedom of the user not to be insulted by the licensing. Lawyers may love confusing, convoluted and non-parsable legalese, but the users do not.
(The following is my opinion, so please read it as such. When I refer to a "good" open source license, I am making a qualitative assessment, and not trying to set up criteria for any approval process but my own.)
The purpose of open source licenses are to grant the user a broad set of permissions and rights over and above those granted by copyright law. Their purpose is not to bind the user to the will of the licensor. A good open source license must be based on copyright law, not contract law.
The first thing a good license should do is grant unconditional permission to use the software. This should be so basic it to not be worth mentioning, but you would be surprised as some of the licensed submitted. Additionally, the use of the software should not be trigger for anything else. We don't want any EULA's here, thank you. The second thing a good license should do is clearly inform the user of their permissions. These permissions must not be predicated upon acceptance of any agreement. A permission may have conditions attached to it. If there is anything you wish the user NOT to do, make it a condition. Next the license should have a warranty disclaimer, to assure the user that they will not be sued if they contribute stuff to the project. You may (and should if you're a commercial project) include a real warranty as a separate legal document.
Notice that I haven't included anything about what you require the user to do. No blanket obligations. That's on purpose. Open Source and Free Software are NOT about making people do things. It is okay to make an obligation be a condition to a permission. It is not okay to make an obligation be a condition to the entire license. Remember, this is about what the user can and cannot do.
Software licenses as contracts was an invention of the proprietary software industry. There was a time not that long ago when copyright law as very vague as to the status of software. So the industry decided to use contract law instead, and created licenses that had such bizarre phrases in them as "by reading this sentence you agree to the following obligations...". That's bullshit and Open Source and Free Software should have nothing to do with such rubbish.
There's wasn't much discussion on the KRL. Don't really know why. I read it through several times, couldn't grok it, so never bothered commenting. Remember, license-discuss isn't composed of attorneys, so a license that doesn't parse to laymen English won't get much discussion.
I've got to ask, this _is_ ((eggs+cheese)+sausage+banana bread), right? Not (eggs+cheese+sausage+banana bread), which is how I initally read it.
I think it was ((eggs+cheese)+((sausage+banana)+bread)), or in other words, your normal eggs and cheese as a side to a banana sausage sandwich.
Well, you heard of "weak copyleft?" Well this is "stong unrestricted." You get more permissions than the MIT/BSD license (really), but the license agreement must be retained in all distributions. This is different from the MIT/BSD licenses in that they require the license to be included, but not necessarily applied to, any copies or derivations.
Open source needs less licences not more..
That's why this is good news. The licenses rejected were among the most bizarre, redundant or useless licenses yet submitted to OSI. Perhaps that's why they made the news.
Not to name names, but one license had a restrictive patent grant that only applied to GPL'ed operating systems.
And what a bizarre license that was (not to name names). It was essentially the BSD license word for word, with the aforementioned patent grant. Yet you couldn't legally use the software on a BSD licensed operating system.
Another was more of a rant than a license.
A delicious rant to be sure. I quite enjoyed it, despite its wrongheadedness. It could not be approved of course, since it explicitly denied its own validity.
The one license that did pass was the Python Software Foundation License.
Whoohoo! In this age of a million open source licenses, it's nice to see that a sensible license that fills a gap in open source gets approved while the frivolous crap gets flushed.
Okay, my reply was a copout. I apologize. I asked a question and you gave a sincere answer.
I wanted to know what made a GUI "user friendly" and you answered basically "readability". Thank you. I will agree that readability is a big factor. I've seen some GUI installers that weren't readable, but a GUI done right can be more readable than a console program done right.
My GUI Installer Experiences:
Corel LinuxOS: Crashed every time it tried to probe the video card, even though the XFree86 documentation clearly says in unambiguous terms not to probe my video card. I found now way around it, and no obvious textmode installer to fall back to. It's the only Linux distro I have never been able to install. It's was also cited as a model for user friendliness before even the newbies gave up on it. (it wasn't an obsolete card)
SuSE Linux: On a friends computer: His card did not have a standard VGA mode (seriously) but that was the mode that the installer wanted to use. On the bright side, at least it was easy to get to a textmode installer. (it wasn't an obsolete card)
Installers have to be, by definition, a one-size-fits-all program. Unfortunately there is no such thing as a one-size-fits-all graphics standard. The PC video display industry is anything BUT standard. As my friend's computer demonstrated, even the VGA display standard is not standard. The only display that every PC can be assured of having is a textmode console.
I'm not prejudiced against graphics. I love the GUI. But I love reliability more. I have learned through painful experience that setting up XFree86 by hand is more reliable than trusting the installer to do it automatically. Go ahead and make your super-duper works-on-anything GUI installer, but keep a textmode installer around for those that require it.
VGA compatible hardware may be a common denominator standard, but that doesn't mean VGA hardware is installed on on your rackmount.
:-)
All you need is a console. It could be a Matrox G450+, a VGA display, a Hercules card, an even older MDA card, or even a connection to the serial port. I would recommend the latter as opposed to swapping cards in and out, or course
Past girlfriends, some roomates, and relatives of mine agree that once you have to tab between things to install something the install stops and they look for something else.
We're talking about unix here. If someone can't handle using the tab key, then perhaps, just perhaps, *BSD is not for them.
I've never seen an OS that was truly one-size-fits-all, and I don't ever expect to.
It also gives you more room for onscreen instructions and help and so on.
Hmmm, everytime I've tried a linux GUI installer I get some stoopid 640x480 screen. Doesn't help the readability much.
The big problem with (FreeBSD|Linux) textmode installs that I've seen is not that they are text, but that they are crappy applications.
You mean like that Mandrake installer in a purple and yellow color scheme that makes choosing packages a nightmare?
do you make a seperate page, for form incompatible browsers?
By "old" standard, I don't mean the ones the cavemen wrought, but merely something older than last week. If my browser doesn't support the established standards, then that is MY problem, and not yours.
Let's say I take my form incompatible browser to your site. I could either end up with a lot of garbage, in which case I can only conclude that my browser sucks. Or you can throw up a page saying my browser sucks, in which cas I can only conclude that you're an asshole.
As far as browsers go, having used IE 5.0 around the time of the nimda worm... I have suggested in my project that we FLATLY regect anyone who doesn't have THE most up to date web browser, AS A PUBLIC SERVICE!
I certainly empathize with this viewpoint. However, when it comes to the web, I am not a developer, I am a user. When I come to a website that says "go away we don't want you", I can only assume that the webmaster is a) rude, b) way too rich to want business, or c) all of the above. The better solution is to simply say "IE 5.0 has a problem, please upgrade to the fix".
Imagine you drive into a service station to get some gas. You know your car uses gas because you put some in last week. However this fuel pump has a square nozzle and your gas tank has a round hole. When you try to jam the square nozzle in up flashes a sign that says "you drive a Dodge so go away!"
Whaa whaa.... I can't see any of the cool internet sites with my C64 anymore!
What pisses me off, and what you apparently can't understand, is that my browser is compliant with this weeks standards, but that you choose not to support it because it wasn't made by Microsoft or Netscape.
The installer can still ask the same questions, but in a more user friendly manner.
Why does a "more user friendly" installer have to be a GUI? What is there about a GUI that makes things easier? I've asked this question before in other forums, but I've never gotten a straight answer.
To be sure, there are many advantages to a GUI, but I don't see where "user friendly" has anything to do with it.
People are like that. Remember that recent article about the open source OSX developer who was calling it quits? He linked to a bunch of emails trying to show that he was being reasonable when in fact it showed him to be a complete ass.
Everything you make has those five steps. Decide what beer, choose ingredients, brew, taste, store in cellar. A lot a people got into Extreme Programming because they thought it would excuse them from some of those onerous steps. Then they discovered that they're not so onerous after all.
But this topic is about *OpenBSD*.