It wasn't so much that Netscape 4 was a buggy piece of crap (which is true, but probably could be fixed) -- It was the fact that their entire DHTML engine was based on proprietary Netscapeisms such as document.layers and JSSS which were left over from the days when Netscape felt it could dictate web standards. The feeling was that it was impossible to make this things behave like W3C specs.
There were also a number of other issues. Linux users (the supposed worker bees in the open source project) didn't like the Motif GUI, Mail used some $ 3rd party components, etc. All-in-all, it's no shock that delta between Netscape 4.0 (1996) and Mozilla 1.0 (2002) is greater than the delta between Mosaic 1.0 (1993?) and Netscape 4.0.
CHRP is an open specification, and Moto/IBM will sell PowerPC chips to anyone who wants to buy one.
Apple sells proprietary hardware with the OS, true. For historical reasons, those systems happen to be PPC. But trying to claim that "Apple has a situation set up" where the market has NO demand for Open PowerPC systems is ridiculous.
The fact is that two huge companies (IBM, Motorola) tried to sell those systems and failed, and then pulled commercial OS support, and the motherboard market collapsed soon after. That has nothing to do with Apple.
I would hardly call an incomplete version with a super magic secret part number "shipping", and I would hardly call your evidence as proof of anything.
This was a Management decision that I am certain they don't regret.
This I agree with. Those machines would have sold about as well as the NT/Alpha and NT/MIPS machines did, or more accurately, didn't.
I remember when IBM sent some minions to my place of work talking these things up. By then the Pentium Pro was either shipping or about to, and if we needed more speed, there was always Alpha.
Of course, this goes back to why you can't buy a reasonably priced bare CHRP board in 2001 -- nobody wanted those systems, not from IBM, not from Motorola, not from the corner screwdriver guy. Apologies, just a rant against the Slashdotter who expressed a desire to buy hardware nobody wants (commodity PPC stuff) at the same price as hardware virtually everyone wants (commodity x86 stuff).
MacOS wasn't even supposed to run on CHRP until the vaporware Copeland release. If Moto built those boards for the MacOS market (and I content they did not), then they were stoopid. Motorola was making Mac clones at the time and they were seperate machines, and hopefully Motorola was aware of that.
The plan to commoditize PowerPC computers was never dependant on Apple, and Apple never played nice with the plan to commoditize PowerPC computers.
Out of the three of them in AIM, Apple is the only one who has ever moved a significant number of desktop PPC chips. You and I might disagree with their "5%" proprietary hardware strategy, but that was their call. The fact that they are still standing makes it easy to forget the titanic clusterfuck of Motorola and IBM's PPC PC efforts.
Building a personal computer from parts was no harder in 1995 that it is today. Small integrators (who purchase most of the loose x86 mainboards, not hobbyists) were just as prevelant back then.
Sure, there weren't as many boards on the market as x86 types, but what did you expect for a new platform. Unlike today, you could pick up the phone and order one at a price that didn't seem completely outragous. Not to mention that The PowerPC CPUs themselves were much cheaper (and faster) than Intel models back then.
And I would argue that the current "ubiquitous" x86 platforms make it more difficult to carve out a market for an alternative platform. Unlike back in the day, nobody's foolhardy nowdays enough to claim that Intel/AMD have a scalablilty wall.
Although, AMD is apparently planning to market the Sledghammer 64-bit chip purely to Linux/BSD users (no native commercial OS support planned), so maybe there's a market there. Of course, unlike PPC, that will run 32-bit x86 software damn fast.
Wrongo. That Moto ATX board wasn't designed to run MacOS, so Apple killing the clones had nothing to do with it.
IBM and Motorola figured that Intel couldn't scale the Pentium design and plotted to take a chunk out of the PC CPU market selling business machines running Windows NT and OS/2 on an open spec PowerPC system design.
MacOS was already long out of this market, and Apple was just along for the ride because they needed a new chip. What killed the PowerPC market was the fact that no businesses wanted to run PPC PCs. When IBM failed at shipping a workable OS/2-PPC and Moto refused to continue paying MS for NT-PPC, their only market channel left was Apple. So, of course they pointed figures at them, even though Apple's grandest market predictions only gave them 15% share for Macs + clones, and IBM/Moto was betting on a bunch more.
So the whole "Apple killed the Open PowerPC market" story is complete and total bullshit. Windows users killed that market by not buying the product. (The Mac clones weren't standard PReP/CHRP machines anyway - they were all Apple-designed custom motherboards with an Apple ROM on them.)
OK, why isn't Netscape keeping up with the Mozilla release schedule? Instead of leaving people hanging with the piss-poor 6.01 release, if you were tracking Mozilla milestones, you'd have a fairly usable 6.05 (Moz.9) version out shortly.
You're probably right that Mozilla will never have a bug-free feature-complete 1.0 done in the next year. But unless you guys at Netscape take what they've done, kill the obvious problems, and slap your logo on it, what has been accomplished will never see wide usage.
Hypothetically, if RedHat ran 100% of games 100% as well as Windows did, I'd think you'd find that a vast majority of those KICK-ASS game boxes would still be running Windows for a variety of other minor reasons.
People have to see Unix as value-add for their primary computing tasks in order to switch to it. For the most part, the good things about Unix have nothing to do with game playing, so it's a wash if gaming is what's important to you.
I guess what I'm saying is that you don't use your Linux box because you don't really need a Unix, and that's fine.
the major complaint I hear amongst people wanting to move over to Linux isn't that there aren't enough games but that their existing game collection would go out the window if they were to switch
Maybe you are trying to convert the wrong people over to Linux? From a pragmatic standpoint, the operating system has to provide applications that the user wishes to run. It's really a simple checklist to determine who is most likely a potential Unix user and who isn't.
+ AOL (stay with Windows)
+ All the latest games (stay with Windows)
+ Powerful scripting and text manipulation functions (Unix)
+ Free programming tools (Unix)
+ Powerful and logical system management functions (Unix)
+ Basic websurfing and e-mail (Both, either, anything).
For the most part, games don't bring in the users -- users bring in the games. A vast majority of games are sold to the casual gamer who is doing real work on his/her machine. The bleeding edge 'Wintendo' gamer crowd is essentally subsidized by this broader market.
Really, Loki has the right idea by trying to sell proven hits to a userbase that wants to relax after a hard day of using 'grep'. Even then, it's a pretty much marginal market, and certainly not large enough to attract someone from the dark side.
I agree, go build the technical infrastructure for portability if that's what turns your screws. But, even if it's technically possible to port games, that's not going to make it economically possible until there's a larger desktop userbase.
Re:There is _no_ reason to stick with Netscape
on
Mozilla 0.9 Out
·
· Score: 1
From a featurepoint level, the Mozilla DOM is excellent. However, recently I tried to use it (with Moz.8), and I ran into all sorts of problems.
There's a number of basic incompatiblities I ran into with Netscape 3-style form objects. (These aren't W3C, but they are essentially de facto standards based on Netscape's excellent documentation. 98% of existing scripts would die without them).
Then, I tried some of the DOM methods. Apparently trying to change certain runtime attributes of tables causes big nasty crashes. This is pure W3C code, BTW.
Anyway, the whole experience was "looks good, isn't good yet". (And yes, Bugzilla knows about all of the issues.) Looking forward to 1.0.
Re:People only use Mozilla to spite MS...
on
Mozilla 0.9 Out
·
· Score: 1
Note that he is apparenlty talking about Netscape 4, which has a renderer so admittedly broken that Netscape/Mozilla rewrote it from the ground up. No need to defend it.
Specifically, the W3C has made a clear distinction between text-level and block-level elements. Netscape 4 doesn't seem to groak this, which leads to all sorts of bizarre mismatches and even crashes. Not to mention his points about the table-renderer being so unpredictable to be essentially broken.
IE does have all sorts of "best guess" features which generally make a page look better. However, for the most part you can override them and it will render it the way you told it to render (although there's a whole bunch stupid little CSS issues wrt forms).
Mozilla doesn't have as many of these "best guess" features mainly because it's goal is to emulate NS4 behavior when something is undefined. As far as I've seen, this leads to some minorly annoying situations where you need to 'over-specify' the style of the element, but I haven't seen any point where the renderer does something incorrectly.
To your point - Pages are "meant to appear" the way the page author specified it. Netscape 4 doesn't do that in many cases where IE generally does. In fact, you often have to create 'incorrect' HTML to get NS4 to work 'correctly'. I think the folks like yourself who believe that it's only the "bad M$ HTML" that doesn't render in Netscape are either living in fantasyland and/or you've never created any page which uses HTML 3.2+ features.
--Posted from Mozilla 0.9
Re:Backwards compatibility vs extensibility
on
Evolution Of RSS
·
· Score: 1
There's a plug-in called "CleverContent" or something (used at the commercial version of terraserver) that does some voodoo to disable the Windows printscreen function. I'm sure it's defeatable if you are willing to hack around the video driver or something.
What, then, is to stop a company from working with a small distributor and dictating the configuration of the "OEM install" such that it mirrors the company's default install.
We used to do this some years ago with a large Compaq/IBM reseller. Basically, we would build a disk image (including Windows, Office, hardware-specific doohickies, networking config, etc) and give it to them. They would dupe it onto the quite large number of PCs we were purchasing, and we would pay Microsoft according to the site licence agreement.
To get away with this, we couldn't use the vendor's preinstalled Windows however (which usually had a bunch of bundled crap anyway) -- we would rebuild it from scratch using stuff from the vendor's download site (quite annoying for things like ThinkPads).
My understanding is that Microsoft ended this program because, naturally, it was costing them useless OEM sales. Not to mention the middle-men of the PC industry pretty much dried up, and now you order Dell-style direct-from-vendor, who Microsoft has more control over.
You've got it all wrong -- here's how those small businesses operate:
A) They limp along with a bunch home computers they bought at "The Good Guys" and their AOL accounts until that doesn't scale any more.
B) They call the local computer store on the corner. For about $5000, the guy sells them a box running "Microsoft Small Business Server" which is a bundle of Exchange, Proxy Server, and SQL Server. He also sets them up a domain and maybe gets DSL installed. He installs a nightly reboot script and sits back waits for them to get hacked or otherwise have him rebuild the whole thing.
Now imagine if the computer guys specs out and sells them a Linux server for $3000. It runs the usual firewalling, MTA, IMAP server, HTTP server and so on. For $100/month, he'll telnet in and remote administration and upgrades. This kills Microsoft's 'entry market' solution
But of course, just about the only thing worse than a bundle of Microsoft crap that you could drop on a small business would be Lotus Domino running on some form of Unix. Not only is a product purely designed for large businesses with in house developers, it costs $100/seat. Leave it to IBM to find a way to do that and screw a great market opportunity.
There's all sorts of operations on MacOS that prevent multitasking. Using "Toast" to burn CDs for example. It only works well when you are using apps that are designed to play nice (such as Photoshop), and in co-op multitasking, it is the app's responsibility to do so. Even so, app switching is noticibly slower on a MacOS machine than say on Windows.
In some respect, the institution of preemptive multitasking is going to hurt Apple. They're a whole bunch of professional Mac video and sound gear that essentially depend on the software being able to monopolize the hardware. By in large, this market dislikes NT because the feeling is that the hardware's less reliable, but this might have more to do with the fact that NT apps are forced to play nice with everything else going on. Of course, there's almost no interest in this market to move to OS X either.
Slashdot Mirror
It wasn't so much that Netscape 4 was a buggy piece of crap (which is true, but probably could be fixed) -- It was the fact that their entire DHTML engine was based on proprietary Netscapeisms such as document.layers and JSSS which were left over from the days when Netscape felt it could dictate web standards. The feeling was that it was impossible to make this things behave like W3C specs.
There were also a number of other issues. Linux users (the supposed worker bees in the open source project) didn't like the Motif GUI, Mail used some $ 3rd party components, etc. All-in-all, it's no shock that delta between Netscape 4.0 (1996) and Mozilla 1.0 (2002) is greater than the delta between Mosaic 1.0 (1993?) and Netscape 4.0.
CHRP is an open specification, and Moto/IBM will sell PowerPC chips to anyone who wants to buy one.
Apple sells proprietary hardware with the OS, true. For historical reasons, those systems happen to be PPC. But trying to claim that "Apple has a situation set up" where the market has NO demand for Open PowerPC systems is ridiculous.
The fact is that two huge companies (IBM, Motorola) tried to sell those systems and failed, and then pulled commercial OS support, and the motherboard market collapsed soon after. That has nothing to do with Apple.
I was responding to
"Five years ago, it was almost unheard of for someone without a technical degree to build their own computer, while it's become very common today."
Read my other posts - I'm aware of the OS situation. (And if the only OS that runs on your product is Windows, that's not the worst situation!)
I would hardly call an incomplete version with a super magic secret part number "shipping", and I would hardly call your evidence as proof of anything.
I've been wondering that for years. Netscape 4.x loads pages at the exact same speed no matter what you set your disk cache to.
This was a Management decision that I am certain they don't regret.
This I agree with. Those machines would have sold about as well as the NT/Alpha and NT/MIPS machines did, or more accurately, didn't.
I remember when IBM sent some minions to my place of work talking these things up. By then the Pentium Pro was either shipping or about to, and if we needed more speed, there was always Alpha.
Of course, this goes back to why you can't buy a reasonably priced bare CHRP board in 2001 -- nobody wanted those systems, not from IBM, not from Motorola, not from the corner screwdriver guy. Apologies, just a rant against the Slashdotter who expressed a desire to buy hardware nobody wants (commodity PPC stuff) at the same price as hardware virtually everyone wants (commodity x86 stuff).
MacOS wasn't even supposed to run on CHRP until the vaporware Copeland release. If Moto built those boards for the MacOS market (and I content they did not), then they were stoopid. Motorola was making Mac clones at the time and they were seperate machines, and hopefully Motorola was aware of that.
The plan to commoditize PowerPC computers was never dependant on Apple, and Apple never played nice with the plan to commoditize PowerPC computers.
Out of the three of them in AIM, Apple is the only one who has ever moved a significant number of desktop PPC chips. You and I might disagree with their "5%" proprietary hardware strategy, but that was their call. The fact that they are still standing makes it easy to forget the titanic clusterfuck of Motorola and IBM's PPC PC efforts.
Building a personal computer from parts was no harder in 1995 that it is today. Small integrators (who purchase most of the loose x86 mainboards, not hobbyists) were just as prevelant back then.
Sure, there weren't as many boards on the market as x86 types, but what did you expect for a new platform. Unlike today, you could pick up the phone and order one at a price that didn't seem completely outragous. Not to mention that The PowerPC CPUs themselves were much cheaper (and faster) than Intel models back then.
And I would argue that the current "ubiquitous" x86 platforms make it more difficult to carve out a market for an alternative platform. Unlike back in the day, nobody's foolhardy nowdays enough to claim that Intel/AMD have a scalablilty wall.
Although, AMD is apparently planning to market the Sledghammer 64-bit chip purely to Linux/BSD users (no native commercial OS support planned), so maybe there's a market there. Of course, unlike PPC, that will run 32-bit x86 software damn fast.
Motorola and a number of other companies shipped NT/PowerPC systems. There was a product.
Wrongo. That Moto ATX board wasn't designed to run MacOS, so Apple killing the clones had nothing to do with it.
IBM and Motorola figured that Intel couldn't scale the Pentium design and plotted to take a chunk out of the PC CPU market selling business machines running Windows NT and OS/2 on an open spec PowerPC system design.
MacOS was already long out of this market, and Apple was just along for the ride because they needed a new chip. What killed the PowerPC market was the fact that no businesses wanted to run PPC PCs. When IBM failed at shipping a workable OS/2-PPC and Moto refused to continue paying MS for NT-PPC, their only market channel left was Apple. So, of course they pointed figures at them, even though Apple's grandest market predictions only gave them 15% share for Macs + clones, and IBM/Moto was betting on a bunch more.
So the whole "Apple killed the Open PowerPC market" story is complete and total bullshit. Windows users killed that market by not buying the product. (The Mac clones weren't standard PReP/CHRP machines anyway - they were all Apple-designed custom motherboards with an Apple ROM on them.)
CP/M-86 shipped with the IBM PC. It cost $300, while IBM (MS) DOS was $50.
Bare PowerPC boards were on the market back when there was commercial operating system support for them (94-96).
Short answer is nobody bought them. Do you expect it to be any different this time?
OK, why isn't Netscape keeping up with the Mozilla release schedule? Instead of leaving people hanging with the piss-poor 6.01 release, if you were tracking Mozilla milestones, you'd have a fairly usable 6.05 (Moz .9) version out shortly.
You're probably right that Mozilla will never have a bug-free feature-complete 1.0 done in the next year. But unless you guys at Netscape take what they've done, kill the obvious problems, and slap your logo on it, what has been accomplished will never see wide usage.
Running IMAP/SSL is not the same thing as using SMIME (which allows you to sign and encrypt mail). Support for this should be included in Mozilla 1.0.
Believe it or not, 4 of the remaining 6 Netscape users use the mail client.
(Sorry about that, there's actually 600 Netscape users left. But dropping the mail component wouldn't help that number...)
Hypothetically, if RedHat ran 100% of games 100% as well as Windows did, I'd think you'd find that a vast majority of those KICK-ASS game boxes would still be running Windows for a variety of other minor reasons.
People have to see Unix as value-add for their primary computing tasks in order to switch to it. For the most part, the good things about Unix have nothing to do with game playing, so it's a wash if gaming is what's important to you.
I guess what I'm saying is that you don't use your Linux box because you don't really need a Unix, and that's fine.
the major complaint I hear amongst people wanting to move over to Linux isn't that there aren't enough games but that their existing game collection would go out the window if they were to switch
Maybe you are trying to convert the wrong people over to Linux? From a pragmatic standpoint, the operating system has to provide applications that the user wishes to run. It's really a simple checklist to determine who is most likely a potential Unix user and who isn't.
+ AOL (stay with Windows)
+ All the latest games (stay with Windows)
+ Powerful scripting and text manipulation functions (Unix)
+ Free programming tools (Unix)
+ Powerful and logical system management functions (Unix)
+ Basic websurfing and e-mail (Both, either, anything).
For the most part, games don't bring in the users -- users bring in the games. A vast majority of games are sold to the casual gamer who is doing real work on his/her machine. The bleeding edge 'Wintendo' gamer crowd is essentally subsidized by this broader market.
Really, Loki has the right idea by trying to sell proven hits to a userbase that wants to relax after a hard day of using 'grep'. Even then, it's a pretty much marginal market, and certainly not large enough to attract someone from the dark side.
I agree, go build the technical infrastructure for portability if that's what turns your screws. But, even if it's technically possible to port games, that's not going to make it economically possible until there's a larger desktop userbase.
From a featurepoint level, the Mozilla DOM is excellent. However, recently I tried to use it (with Moz .8), and I ran into all sorts of problems.
There's a number of basic incompatiblities I ran into with Netscape 3-style form objects. (These aren't W3C, but they are essentially de facto standards based on Netscape's excellent documentation. 98% of existing scripts would die without them).
Then, I tried some of the DOM methods. Apparently trying to change certain runtime attributes of tables causes big nasty crashes. This is pure W3C code, BTW.
Anyway, the whole experience was "looks good, isn't good yet". (And yes, Bugzilla knows about all of the issues.) Looking forward to 1.0.
Note that he is apparenlty talking about Netscape 4, which has a renderer so admittedly broken that Netscape/Mozilla rewrote it from the ground up. No need to defend it.
Specifically, the W3C has made a clear distinction between text-level and block-level elements. Netscape 4 doesn't seem to groak this, which leads to all sorts of bizarre mismatches and even crashes. Not to mention his points about the table-renderer being so unpredictable to be essentially broken.
IE does have all sorts of "best guess" features which generally make a page look better. However, for the most part you can override them and it will render it the way you told it to render (although there's a whole bunch stupid little CSS issues wrt forms).
Mozilla doesn't have as many of these "best guess" features mainly because it's goal is to emulate NS4 behavior when something is undefined. As far as I've seen, this leads to some minorly annoying situations where you need to 'over-specify' the style of the element, but I haven't seen any point where the renderer does something incorrectly.
To your point - Pages are "meant to appear" the way the page author specified it. Netscape 4 doesn't do that in many cases where IE generally does. In fact, you often have to create 'incorrect' HTML to get NS4 to work 'correctly'. I think the folks like yourself who believe that it's only the "bad M$ HTML" that doesn't render in Netscape are either living in fantasyland and/or you've never created any page which uses HTML 3.2+ features.
--Posted from Mozilla 0.9
From http://slashdot.org/slashdot.rdf :
x -ns#" xmlns="http://my.netscape.com/rdf/simple/0.9/">
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-synta
Maybe not legal, but still in use as "RDF". I hate to pick nits too, but once it wos published it was out there.
There's a plug-in called "CleverContent" or something (used at the commercial version of terraserver) that does some voodoo to disable the Windows printscreen function. I'm sure it's defeatable if you are willing to hack around the video driver or something.
What, then, is to stop a company from working with a small distributor and dictating the configuration of the "OEM install" such that it mirrors the company's default install.
We used to do this some years ago with a large Compaq/IBM reseller. Basically, we would build a disk image (including Windows, Office, hardware-specific doohickies, networking config, etc) and give it to them. They would dupe it onto the quite large number of PCs we were purchasing, and we would pay Microsoft according to the site licence agreement.
To get away with this, we couldn't use the vendor's preinstalled Windows however (which usually had a bunch of bundled crap anyway) -- we would rebuild it from scratch using stuff from the vendor's download site (quite annoying for things like ThinkPads).
My understanding is that Microsoft ended this program because, naturally, it was costing them useless OEM sales. Not to mention the middle-men of the PC industry pretty much dried up, and now you order Dell-style direct-from-vendor, who Microsoft has more control over.
You've got it all wrong -- here's how those small businesses operate:
A) They limp along with a bunch home computers they bought at "The Good Guys" and their AOL accounts until that doesn't scale any more.
B) They call the local computer store on the corner. For about $5000, the guy sells them a box running "Microsoft Small Business Server" which is a bundle of Exchange, Proxy Server, and SQL Server. He also sets them up a domain and maybe gets DSL installed. He installs a nightly reboot script and sits back waits for them to get hacked or otherwise have him rebuild the whole thing.
Now imagine if the computer guys specs out and sells them a Linux server for $3000. It runs the usual firewalling, MTA, IMAP server, HTTP server and so on. For $100/month, he'll telnet in and remote administration and upgrades. This kills Microsoft's 'entry market' solution
But of course, just about the only thing worse than a bundle of Microsoft crap that you could drop on a small business would be Lotus Domino running on some form of Unix. Not only is a product purely designed for large businesses with in house developers, it costs $100/seat. Leave it to IBM to find a way to do that and screw a great market opportunity.
There's all sorts of operations on MacOS that prevent multitasking. Using "Toast" to burn CDs for example. It only works well when you are using apps that are designed to play nice (such as Photoshop), and in co-op multitasking, it is the app's responsibility to do so. Even so, app switching is noticibly slower on a MacOS machine than say on Windows.
In some respect, the institution of preemptive multitasking is going to hurt Apple. They're a whole bunch of professional Mac video and sound gear that essentially depend on the software being able to monopolize the hardware. By in large, this market dislikes NT because the feeling is that the hardware's less reliable, but this might have more to do with the fact that NT apps are forced to play nice with everything else going on. Of course, there's almost no interest in this market to move to OS X either.
First of all I commend your user agent for supporting obscure and unwanted corners of the CSS2 specification, such as 'text-decoration:blink'.
But, then, since it's misinterpreting < as a < and > as >, I have to wonder. You say you wrote it yourself?