I repeat the question. Bear in mind that the app doesn't have anything coded into it that says where that call is found. It just has an entry that says "I need to call RenderHTML()". The loader then locates the RenderHTML() function in the visible libraries and resolves the references to it. If it's in MSHTML.DLL or MOZHTML.DLL, the loader will still resolve it and the app won't know the difference. That's the whole point of modular programming and libraries and COM/COM+/etc.. It's what lets you upgrade from DirectX 6 to DirectX 8 without having to throw all your DirectX 6.x games away and buy new copies. If it works for DirectX, why wouldn't it work for IBrowser?
Ummm... What's the difference to an app making a RenderHTML( &window, &HTMLObject ) call whether the RenderHTML() implementation is in the OS or a DLL? In fact, in Windows, it is in a DLL even when it's in the OS, so the only thing that would change is which DLL it's in.
Which is the point Gates doesn't want to admit to, because as soon as he does the whole "everything must be integrated into Windows or it won't work" argument explodes and his main method of fending off competitors evaporates.
...is that he can't take the code without honoring the owner's wishes as far as payment goes. With every other license he can find a way to get the code without having to shell out anything significant, but with the GPL he can't get out of paying back in the same coin he received: code.
In your setup where I have to register and wait for a username/password to be e-mailed to me before I can download the software, the usual reason I never download it is that I had to wait for the username/password to be e-mailed to me. There's always some lag, from a few minutes to a day, and even if it's just a few minutes I can't sit there fetching and refetching my mail until the message arrives. So instead I go off to do other things until one of my hourly mail runs picks up the message and I check my inbox and notice it. The problem is I usually notice it in the middle of the workday, and I've got Yet Another Bug to fix right now, so I file the message in a folder to get to later. By the time I get free time, I've forgotten about it. I'll notice it when I do my daily check of things I have to deal with, but if it's not high-priority that I get that piece of software up and running (and eval software is rarely high priority) then I'll put it off again, and probably forget about it again. Usually by the time I get around to going back for the download, the username has expired and I can't get at the software. At which point I shrug and go on to one of the fifteen other jobs in my inbox.
If you want people to get the software, make it downloadable now. Mail them a key to activate it if you have to, but let them get the package when they're at your site and ready to get it.
Oh, and as for thinking that people who put in bogus e-mail addresses aren't good prospects, let me give you a clue: anybody who has any idea what they're doing these days never gives out their real e-mail address to anybody until they know who they're dealing with and are ready to buy from them. You're filing your most technically competent prospects, the ones who'll be making the recommendations to the CTOs about which products will actually do what they want, in the "bad prospect" file.
I hate to say this but, with the Hammer's Slammers books, David Drake is probably one of the better-known writers in sci-fi circles. He's not quite up in Issac Asimov's league, but he's probably in the top 20% nonetheless. And he's seeing better sales of his books that are in the BFL than the ones available in more restricted electronic forms. By your theory he should be hurt by the free availability, but it appears not. He may not be helped as much as Eric (he's a bigger name than Eric, but is only making as much in electronic royalties as Eric), but he's still making 100 times as much in royalties on BFL electronic books as on non-BFL electronic books.
One correction to the blurb I'd like to make: the used-book sales aren't royaltyless. No royalty is paid to the author on this sale because the author has already received the royalty on that copy of the book the first time it was sold (as a new book). The Guild is complaining not that the author isn't being paid, but that they aren't being paid multiple times for a single copy sold.
I think that's one of the points: when all those VS.NET developers get told by their bosses "35 million AOL users can't access our site. Fix it.", they'll wind up having to either abandon VS.NET (because they can't fix the problem while staying in it) or telling their bosses "Sorry, you'll just have to write off a third of the Internet as incompatible with our site.". 3 guesses how the boss will react to the second option. IMHO this would be a Good Thing.
I had something better. The only Yahoo e-mail address I have was one from a Pepsi promo campaign. I don't need it anymore. I took the attitude of "I'm sorry, if I didn't change those settings then any changes are invalid and on your heads be the consequences.", and added a rule to my.procmailrc to EXITCODE = 77 anything coming in to that address. All that address is good for now is clogging up the sender's mail system with messages that can't be delivered.
But I can run X-Windows without a window manager. In fact there's very good reasons for doing so. For example, a cash register needs to run the register application and only the register application, the cashier shouldn't be able to close or move windows or get at the desktop except as the register application allows them. Ie., they need a system without a window manager or desktop, just the application running in a full-screen window.
If you replace the IBrowser implementation completely, then what IE's implementation used underneath the hood doesn't matter anymore. Why should a Mozilla-based browser call IE's rendering and manipulation routines, after all? It's got it's own.
So Microsoft isn't using their own recommended method for displaying HTML: use COM to create an IBrowser object and pass what you need displayed to it? That's what they want everyone else to do, and if they follow their own recommendations then you can drop any browser or other app that implements the IBrowser interface into the system and it'll get used just fine. At least as long as what's being displayed is standard HTML and not MS-proprietary non-HTML stuff.
One major difference, though. If I go to Fiori's every day, the staff there will get to know me and my habits. If I go to Giovanni's, the staff at Fiori's don't know what I ordered there, and it's none of their business. If I want them to know, I'll tell them. My objection to the way businesses want the Internet to work is that it's the equivalent of Fiori's and Giovanni's not just sharing my habits without my permission but the local Restaurant Association snagging the information from both of them to target advertising at me without my permission to do so. If I had wanted Fiori's or the RA to know what I ordered at Giovanni's, I'd've told them. I didn't, so I probably don't and I don't appreciate their spying on me behind my back.
And if I'm going to the Kelly Blue Book site, I probably don't want ads for particular cars. I want information on the cars I ask for. This is where most businesses get tripped up, trying to guess what the customer wants instead of just asking him and then giving him what he asked for. I don't appreciate stores where the staff are constantly hovering behind me interrupting while I'm looking, and I don't appreciate it on-line either.
It's not that software engineers don't want to remove errors from their code. It's that their business-type managers don't feel there's a business case for removing the errors, since that would delay the product and cost money and won't have any noticeable effect on sales.
The Brilliant client gets executable code downloaded from the Brilliant servers and download of the code is under the control of the servers, not the client. If someone got control of the Brilliant servers they could download code to your machine that either used your access or exploited a security hole to gain admin access and completely compromise your machine. It could then set up a server like Back Orifice and wait for orders.
Scenarios like that are one reason I refuse to install software that does things under the control of someone else's servers. I can control my machine and what I do, I can't control their servers and what they do, and if I don't have control I have no way of insuring that nothing happens that breaks security.
Actually it would be significant. It's an indicator of people's behavior. A relatively constant session time would indicate people are most likely browsing, following links and wandering around in the on-line equivalent of window-shopping at the mall. A steadily decreasing session time would indicate that people are instead using search engines, bookmarks or some other method of locating what they need/want, going there, doing what they need done and then leaving. The first would be indicative of entertainment or casual use, the second is more indicative of purposeful, goal-directed use, with all the implications for things like advertising the different entails.
I think you miss the point: Microsoft won't negotiate on that point. You either take a contract to ship Windows on every PC or you won't have a contract. With the retail price of Windows being greater than their profit margin on a box, this means they've got to increase the price of their systems noticeably (10% or so) above the prices of their competitors or lose money on each sale. You can't do either of those and stay in business.
For cache checking, yes you'll need to check URI in the request, Host header if present and other things along with the original destination IP address. Any cache should tag the cached data with the IP address of the server as well, since it's not guaranteed that two servers with the same name will return the same data if they've different IP addresses (in fact under round-robin DNS it's possible for two servers to be required to return different information).
I was speaking, though, of how the cache determines which server to talk to when it needs to talk to the destination server. At that point a transparent proxy should only connect to the actual IP address used by the client. Anything else leads to legal configurations which the proxy will break.
Actually the ISP is broken. The customer opens a connection to IP address A.B.C.D, but the ISP is delivering a connection to IP address M.N.O.P instead. It shouldn't matter which DNS roots the user is using, or whether they're using DNS for that matter (I access some of my company's servers using/etc/hosts entries because their specific names aren't in DNS, for example). The ISP should be looking at the destination address in the IP packets and nothing else when doing a transparent proxy. If it does that then it behaves correctly for all customers no matter what they're using for name resolution. If it doesn't, then it breaks completely legal setups. The ISP is broken.
In California at least the employment contract does not trump the law on this matter. California law places limits on what work the company may lay claim to, and explicitly limits contract terms to no more than what the law allows. I can dig the exact wording up out of my files, but it amounts to saying the company can claim work that:
is done on company time.
is assigned to you by the company to do, whether on company time or not.
falls within the normal type of work you would do for the company (which is read fairly narrowly) and which the company, when notified of it's existence, elects to use.
If you work in California, at least, your employer is not permitted to claim anything and everything you create.
Simple: notarization and all that stuff aren't required when buying or selling a house. A simple oral agreement is all that's required. All the rest of the notarization and escrow and such is there not as a requirement of the sale, but as insurance by each party that the other will have a really hard time denying the terms later on. It's hard for either party to deny they signed a contract, for example, when the other guy has a notarized contract with the denying party's signature on it. Same for escrow, it's just a way of insuring that when you hand over the deed or the check for a couple of hundred grand the other party won't take it and then refuse to hand over their part of the transaction. None of it's legally required, technically, it's a framework that allows you to trust that what's been agreed to will happen even when you don't trust the other party completely.
Jon, sanity check here. Electric lights aren't high-tech, they're near a century old now. Pick a few dozen of your friends and ask them how lights work. I'll bet that at least half of them give an answer along the lines of "You flip a switch and the magic electricity flows and the bulb lights.". IOW they know how to work a light but not anything about how and why it actually lights up. They know the electric grid and generating plants work, but for all they know about how and why they work the whole process of getting electricity to a light bulb and making it glow might as well be a magic spell. It might not be obvious on the surface, since probably you understand it, but if you go probing you'll find that most people don't.
Now, if after a century we have a large percentage of people who don't understand the stuff that makes something as simple and ubiquitous as a light bulb work, why should we expect any significant percentage to understand what's behind something as complex and relatively new as a computer? That's what seperates the tech-savvy from the rest: we're the relatively small percentage who're actually interested enough to learn the whys and wherefores and whats behind using the technology. We know how the wires are connected behind the walls, how generating plants and electric grids work and what the switch actually does to the electrons, not just that flipping the switch makes the magic bulb glow. We've actually ferreted out why things work, not just how to make them work.
That's also why the gap won't go away: you can't teach curiousity. You can encourage those that have it to use it, but if they don't have it you can't force-feed it to them. Trying just annoys them, and they start asking why they need to know all this technical stuff just to use X.
This is also, IMHO, why the average savvy person's choice of reading is so different from the mainstream: we like things that make us think, the mainstream doesn't.
Counter-point: if AOL announces the change-over, no web site out there can afford to not work with Mozilla when the change occurs. You simply can't afford to tell 30% of your user-base "Sorry, we don't want your business.". If they're under the gun, I suspect the web designers will scramble to accomodate this large chunk of their user-base. It's not like they don't have the tools available, after all.
Re:A lot of use of the term 'preferred form' here
on
Abusing the GPL?
·
· Score: 2
It's not the preferred form for compiling, though. What they compile is irrelevant, the GPL specifies the preferred form for making modifications to the code. Even if they built for distribution from the obfuscated code, unless that's actually the code their programmers work on it's not in preferred form per the GPL. The GPL also says preferred form, not preferred format, and form implies things in addition to just, say, text format.
Read the section of the GPL that mentions "preferred form of the work for making modifications to it". Deliberate obfuscation of the code with the intent of making it useless for the purposes of modification could easily be construed as an attempt to violate this part of the license, and I'm sure the FSF, after recent court decisions, would be happy to discuss the point before a judge.
In the eyes of the recording companies and some musicians, yes you are ripping them off by buying used CDs. However, as Garth Brooks found out, the law doesn't see it that way. When he tried suing to prevent resale of his CDs by used CD stores, the judge told him flat-out that copyright law ended his rights to royalties at first sale. After that, he had no rights to royalties from additional sales of that copy and no legal right to prevent those sales.
Slashdot Mirror
I repeat the question. Bear in mind that the app doesn't have anything coded into it that says where that call is found. It just has an entry that says "I need to call RenderHTML()". The loader then locates the RenderHTML() function in the visible libraries and resolves the references to it. If it's in MSHTML.DLL or MOZHTML.DLL, the loader will still resolve it and the app won't know the difference. That's the whole point of modular programming and libraries and COM/COM+/etc.. It's what lets you upgrade from DirectX 6 to DirectX 8 without having to throw all your DirectX 6.x games away and buy new copies. If it works for DirectX, why wouldn't it work for IBrowser?
Ummm... What's the difference to an app making a RenderHTML( &window, &HTMLObject ) call whether the RenderHTML() implementation is in the OS or a DLL? In fact, in Windows, it is in a DLL even when it's in the OS, so the only thing that would change is which DLL it's in.
Which is the point Gates doesn't want to admit to, because as soon as he does the whole "everything must be integrated into Windows or it won't work" argument explodes and his main method of fending off competitors evaporates.
...is that he can't take the code without honoring the owner's wishes as far as payment goes. With every other license he can find a way to get the code without having to shell out anything significant, but with the GPL he can't get out of paying back in the same coin he received: code.
In your setup where I have to register and wait for a username/password to be e-mailed to me before I can download the software, the usual reason I never download it is that I had to wait for the username/password to be e-mailed to me. There's always some lag, from a few minutes to a day, and even if it's just a few minutes I can't sit there fetching and refetching my mail until the message arrives. So instead I go off to do other things until one of my hourly mail runs picks up the message and I check my inbox and notice it. The problem is I usually notice it in the middle of the workday, and I've got Yet Another Bug to fix right now, so I file the message in a folder to get to later. By the time I get free time, I've forgotten about it. I'll notice it when I do my daily check of things I have to deal with, but if it's not high-priority that I get that piece of software up and running (and eval software is rarely high priority) then I'll put it off again, and probably forget about it again. Usually by the time I get around to going back for the download, the username has expired and I can't get at the software. At which point I shrug and go on to one of the fifteen other jobs in my inbox.
If you want people to get the software, make it downloadable now. Mail them a key to activate it if you have to, but let them get the package when they're at your site and ready to get it.
Oh, and as for thinking that people who put in bogus e-mail addresses aren't good prospects, let me give you a clue: anybody who has any idea what they're doing these days never gives out their real e-mail address to anybody until they know who they're dealing with and are ready to buy from them. You're filing your most technically competent prospects, the ones who'll be making the recommendations to the CTOs about which products will actually do what they want, in the "bad prospect" file.
I hate to say this but, with the Hammer's Slammers books, David Drake is probably one of the better-known writers in sci-fi circles. He's not quite up in Issac Asimov's league, but he's probably in the top 20% nonetheless. And he's seeing better sales of his books that are in the BFL than the ones available in more restricted electronic forms. By your theory he should be hurt by the free availability, but it appears not. He may not be helped as much as Eric (he's a bigger name than Eric, but is only making as much in electronic royalties as Eric), but he's still making 100 times as much in royalties on BFL electronic books as on non-BFL electronic books.
One correction to the blurb I'd like to make: the used-book sales aren't royaltyless. No royalty is paid to the author on this sale because the author has already received the royalty on that copy of the book the first time it was sold (as a new book). The Guild is complaining not that the author isn't being paid, but that they aren't being paid multiple times for a single copy sold.
I think that's one of the points: when all those VS .NET developers get told by their bosses "35 million AOL users can't access our site. Fix it.", they'll wind up having to either abandon VS .NET (because they can't fix the problem while staying in it) or telling their bosses "Sorry, you'll just have to write off a third of the Internet as incompatible with our site.". 3 guesses how the boss will react to the second option. IMHO this would be a Good Thing.
I had something better. The only Yahoo e-mail address I have was one from a Pepsi promo campaign. I don't need it anymore. I took the attitude of "I'm sorry, if I didn't change those settings then any changes are invalid and on your heads be the consequences.", and added a rule to my .procmailrc to EXITCODE = 77 anything coming in to that address. All that address is good for now is clogging up the sender's mail system with messages that can't be delivered.
But I can run X-Windows without a window manager. In fact there's very good reasons for doing so. For example, a cash register needs to run the register application and only the register application, the cashier shouldn't be able to close or move windows or get at the desktop except as the register application allows them. Ie., they need a system without a window manager or desktop, just the application running in a full-screen window.
If you replace the IBrowser implementation completely, then what IE's implementation used underneath the hood doesn't matter anymore. Why should a Mozilla-based browser call IE's rendering and manipulation routines, after all? It's got it's own.
So Microsoft isn't using their own recommended method for displaying HTML: use COM to create an IBrowser object and pass what you need displayed to it? That's what they want everyone else to do, and if they follow their own recommendations then you can drop any browser or other app that implements the IBrowser interface into the system and it'll get used just fine. At least as long as what's being displayed is standard HTML and not MS-proprietary non-HTML stuff.
One major difference, though. If I go to Fiori's every day, the staff there will get to know me and my habits. If I go to Giovanni's, the staff at Fiori's don't know what I ordered there, and it's none of their business. If I want them to know, I'll tell them. My objection to the way businesses want the Internet to work is that it's the equivalent of Fiori's and Giovanni's not just sharing my habits without my permission but the local Restaurant Association snagging the information from both of them to target advertising at me without my permission to do so. If I had wanted Fiori's or the RA to know what I ordered at Giovanni's, I'd've told them. I didn't, so I probably don't and I don't appreciate their spying on me behind my back.
And if I'm going to the Kelly Blue Book site, I probably don't want ads for particular cars. I want information on the cars I ask for. This is where most businesses get tripped up, trying to guess what the customer wants instead of just asking him and then giving him what he asked for. I don't appreciate stores where the staff are constantly hovering behind me interrupting while I'm looking, and I don't appreciate it on-line either.
It's not that software engineers don't want to remove errors from their code. It's that their business-type managers don't feel there's a business case for removing the errors, since that would delay the product and cost money and won't have any noticeable effect on sales.
The Brilliant client gets executable code downloaded from the Brilliant servers and download of the code is under the control of the servers, not the client. If someone got control of the Brilliant servers they could download code to your machine that either used your access or exploited a security hole to gain admin access and completely compromise your machine. It could then set up a server like Back Orifice and wait for orders.
Scenarios like that are one reason I refuse to install software that does things under the control of someone else's servers. I can control my machine and what I do, I can't control their servers and what they do, and if I don't have control I have no way of insuring that nothing happens that breaks security.
Actually it would be significant. It's an indicator of people's behavior. A relatively constant session time would indicate people are most likely browsing, following links and wandering around in the on-line equivalent of window-shopping at the mall. A steadily decreasing session time would indicate that people are instead using search engines, bookmarks or some other method of locating what they need/want, going there, doing what they need done and then leaving. The first would be indicative of entertainment or casual use, the second is more indicative of purposeful, goal-directed use, with all the implications for things like advertising the different entails.
I think you miss the point: Microsoft won't negotiate on that point. You either take a contract to ship Windows on every PC or you won't have a contract. With the retail price of Windows being greater than their profit margin on a box, this means they've got to increase the price of their systems noticeably (10% or so) above the prices of their competitors or lose money on each sale. You can't do either of those and stay in business.
For cache checking, yes you'll need to check URI in the request, Host header if present and other things along with the original destination IP address. Any cache should tag the cached data with the IP address of the server as well, since it's not guaranteed that two servers with the same name will return the same data if they've different IP addresses (in fact under round-robin DNS it's possible for two servers to be required to return different information).
I was speaking, though, of how the cache determines which server to talk to when it needs to talk to the destination server. At that point a transparent proxy should only connect to the actual IP address used by the client. Anything else leads to legal configurations which the proxy will break.
Actually the ISP is broken. The customer opens a connection to IP address A.B.C.D, but the ISP is delivering a connection to IP address M.N.O.P instead. It shouldn't matter which DNS roots the user is using, or whether they're using DNS for that matter (I access some of my company's servers using /etc/hosts entries because their specific names aren't in DNS, for example). The ISP should be looking at the destination address in the IP packets and nothing else when doing a transparent proxy. If it does that then it behaves correctly for all customers no matter what they're using for name resolution. If it doesn't, then it breaks completely legal setups. The ISP is broken.
In California at least the employment contract does not trump the law on this matter. California law places limits on what work the company may lay claim to, and explicitly limits contract terms to no more than what the law allows. I can dig the exact wording up out of my files, but it amounts to saying the company can claim work that:
- is done on company time.
- is assigned to you by the company to do, whether on company time or not.
- falls within the normal type of work you would do for the company (which is read fairly narrowly) and which the company, when notified of it's existence, elects to use.
If you work in California, at least, your employer is not permitted to claim anything and everything you create.Simple: notarization and all that stuff aren't required when buying or selling a house. A simple oral agreement is all that's required. All the rest of the notarization and escrow and such is there not as a requirement of the sale, but as insurance by each party that the other will have a really hard time denying the terms later on. It's hard for either party to deny they signed a contract, for example, when the other guy has a notarized contract with the denying party's signature on it. Same for escrow, it's just a way of insuring that when you hand over the deed or the check for a couple of hundred grand the other party won't take it and then refuse to hand over their part of the transaction. None of it's legally required, technically, it's a framework that allows you to trust that what's been agreed to will happen even when you don't trust the other party completely.
Jon, sanity check here. Electric lights aren't high-tech, they're near a century old now. Pick a few dozen of your friends and ask them how lights work. I'll bet that at least half of them give an answer along the lines of "You flip a switch and the magic electricity flows and the bulb lights.". IOW they know how to work a light but not anything about how and why it actually lights up. They know the electric grid and generating plants work, but for all they know about how and why they work the whole process of getting electricity to a light bulb and making it glow might as well be a magic spell. It might not be obvious on the surface, since probably you understand it, but if you go probing you'll find that most people don't.
Now, if after a century we have a large percentage of people who don't understand the stuff that makes something as simple and ubiquitous as a light bulb work, why should we expect any significant percentage to understand what's behind something as complex and relatively new as a computer? That's what seperates the tech-savvy from the rest: we're the relatively small percentage who're actually interested enough to learn the whys and wherefores and whats behind using the technology. We know how the wires are connected behind the walls, how generating plants and electric grids work and what the switch actually does to the electrons, not just that flipping the switch makes the magic bulb glow. We've actually ferreted out why things work, not just how to make them work.
That's also why the gap won't go away: you can't teach curiousity. You can encourage those that have it to use it, but if they don't have it you can't force-feed it to them. Trying just annoys them, and they start asking why they need to know all this technical stuff just to use X.
This is also, IMHO, why the average savvy person's choice of reading is so different from the mainstream: we like things that make us think, the mainstream doesn't.
Counter-point: if AOL announces the change-over, no web site out there can afford to not work with Mozilla when the change occurs. You simply can't afford to tell 30% of your user-base "Sorry, we don't want your business.". If they're under the gun, I suspect the web designers will scramble to accomodate this large chunk of their user-base. It's not like they don't have the tools available, after all.
It's not the preferred form for compiling, though. What they compile is irrelevant, the GPL specifies the preferred form for making modifications to the code. Even if they built for distribution from the obfuscated code, unless that's actually the code their programmers work on it's not in preferred form per the GPL. The GPL also says preferred form, not preferred format, and form implies things in addition to just, say, text format.
Read the section of the GPL that mentions "preferred form of the work for making modifications to it". Deliberate obfuscation of the code with the intent of making it useless for the purposes of modification could easily be construed as an attempt to violate this part of the license, and I'm sure the FSF, after recent court decisions, would be happy to discuss the point before a judge.
In the eyes of the recording companies and some musicians, yes you are ripping them off by buying used CDs. However, as Garth Brooks found out, the law doesn't see it that way. When he tried suing to prevent resale of his CDs by used CD stores, the judge told him flat-out that copyright law ended his rights to royalties at first sale. After that, he had no rights to royalties from additional sales of that copy and no legal right to prevent those sales.