 |
|  |
 |
|  |
 |
|  |
 |
|  |
 |
|
Tolls
|
 |
Hereford, UK
Jan 1970 time: 05:19
|
|
Asher:
"No, not in computers.
1 single chip architecture for PCs is DEFINITELY a good thing. And we have that. It's called x86. Athlons, Pentiums, Cyrixes, etc all run x86 as an architecture.
If we had competing ones (which we WILL in a couple years, mind you -- AMD's x86-64 vs Intel's IA-64), you'll have to download different patches for each architecture, you'll need to buy different software for each architecture, and it'll be a complete hassle to developers to actually make more than one version.
1 chip architecture is a VERY good thing in PCs."
Well, not for all computers, but let's stick to PCs anyway.
This only comes about because the instruction set is open, allowing other companies (ie AMD) to get in and produce their own versions. This allows competition.
Conversely, then, for a healthy OS market a standard set of APIs would need to be defined which would allow applications to be easily ported between OSs. Without something like this (that everyone would stick to) you wouldn't have a healthy industry.
These APIs would need to be pretty well carved in stone (like x86) so that you wouldn't be trying to hit a moving target. Since I can't see this ever happening, the only way to maintain some level of competition to help drive the market is to have multiple OSs available...
Is there another industry out there that has the same structure as the OS/application industry? I can't think of one offhand that could be used for comparison...Imran?
|
|
|  |
 |
|
Asher
|
 |
Calgary, Alberta
Nov 1999 time: 22:19
|
|
quote: Originally posted by Tolls
This only comes about because the instruction set is open, allowing other companies (ie AMD) to get in and produce their own versions. This allows competition.
Conversely, then, for a healthy OS market a standard set of APIs would need to be defined which would allow applications to be easily ported between OSs. Without something like this (that everyone would stick to) you wouldn't have a healthy industry.
These APIs would need to be pretty well carved in stone (like x86) so that you wouldn't be trying to hit a moving target. Since I can't see this ever happening, the only way to maintain some level of competition to help drive the market is to have multiple OSs available... |
Unfortunately this would definitely not work with software.
Unlike CPU instruction sets, OS APIs change rapidly with time. Look how often a new DirectX build comes out, for example.
There are general "multiplatform APIs" like OpenGL for graphics, and SDL(?) for audio, but for OS interaction they all function differently with the OS and programs, and all OSes frequently change their API calls as the product evolves, changes, and improves.
One possible solution to this is Microsoft's .NET. It's sort of like Java. The idea is if you have a .NET runtime on the computer (there are Mac, Unix, and Windows runtimes in development) it'll compile and run the code on multiple environments, on multiple instruction sets. Unlike Java, .NET doesn't restrict you to one specific language either. They can develop in tons of programming languages. The main .NET programming language is called C#. Microsoft designed it and submitted it to a 3rd party body to make it an open standard.
This is somewhat like a universal API for operating systems, and can be updated with a bit of ease too.
|
|
|  |
 |
|
Tolls
|
 |
Hereford, UK
Jan 1970 time: 05:19
|
|
"One possible solution to this is Microsoft's .NET."
But would the structure of .NET be open to allow, for example OpenNET to be made, which would do the same thing?
"There are general "multiplatform APIs" like OpenGL for graphics, and SDL(?) for audio, but for OS interaction they all function differently with the OS and programs, and all OSes frequently change their API calls as the product evolves, changes, and improves. "
These are the sort of things I was thinking of...this limits the amount of rework for an OS to updating one (albeit big) product (ie OpenGL, or OpenAL). The problem with these is that they are both fairly low level. SDL is higher level, but not quite as flexible...it's largely a graphics engine, but there is an audio plug in of sorts.
My point, essentially, is that I can't see the OS market as a monopoly being a healthy market..at least for the tech side of things.
...
Oh...and GO TO BED!

|
|
|  |
 |
|
Thue
|
|
Copenhagen, Denmark
Jan 1970 time: 06:19
|
|
There is a story on slashdot that isn't too kind to microsoft; it did seem pretty outrageous. Some points:
|
|
|  |
 |
|
Asher
|
 |
Calgary, Alberta
Nov 1999 time: 22:19
|
|
quote: Originally posted by Tolls
But would the structure of .NET be open to allow, for example OpenNET to be made, which would do the same thing? |
Yes, actually.
MS just gave away the source to a .NET runtime engine called Rotor, and I believe that's what Ximian is using to make "Mono", their .NET runtime (which is open source). Ximian's the same company who makes Gnome, etc.
quote: A bold assertion, for sure. |
It WOULD work, but it wouldn't benefit anyone.
A perfect example of why it wouldn't work is OpenGL.
OpenGL is so very, very far behind from DirectX in terms of functionality that all the new features are added by the board makers in extensions, which means they need to give you new software every time to run those extensions anyway, which means they need to develop that code for each specific platform anyway. A large committee-style API group is simply slow to move. We'd probably still be stuck using Win3.1's API because "it worked", and graphics rendering would be stuck at the stock OpenGL 1.2 (blech).
It's certainly possible, but I don't see it actually benefitting the consumer.
I also don't see very many people who will be ready to run away from Windows, which they've used for years, and run on a different platform because it's $25 cheaper or free. Of course, the Linux zealots would be extremely happy and would throw a party, but I don't see many people outside of them actually utilizing that.
And what does everyone else get out of it? Stagnated development, frozen OS APIs, slow-moving committees changing things and requiring each developer to update their OS and how it handles APIs accordingly.
And then there's the support problem.
All software has bugs. It'd be a support nightmare if the same executable ran on many OSes with many settings and many developers and many quirks. 
|
|
|  |
 |
|
moomin
|
 |
Moo Like In Moomin
Jul 2000 time: 05:19
|
|
quote: Originally posted by Asher
I don't understand your Kerberos comment, but it seems to be the typical troll these days against MS: MS extended Kerberos. MS' version is a superset, it should work fine with existing implementations and should cause you no headache.
|
Look, Asher, it might not always have been so in this thread, but I'm kinda trying to have a serious conversation here. If you feel better throwing terms like "troll" around, perhaps I should leave it to you and your friend Cloud 9 to duke it out?
As for Kerberos, MS has hardcoded the resolution of the allowed time quanta a ticket is valid. Not only have they hard-coded it, they have also made their bracket a wee bit different from what most other manufacturers provide - enough so that any auth system relying on good tight clocking is going to "sync out" once every day or so. A lovely little feature to track down and correct. Correcting, btw, in this instance, meant resetting the other systems so they followed the MS "standard". I shudder to think of what would have happened hade some part of that setup not been under our direct control. And MS software is full of little nasty surprises like that.
quote:
Where did you pull that figure out of?
|
Gut feeling, based on my own experience. I've got a bunch of MS projects Gant charts to back it up, but sinve there's no way I'm gonna post 'em here you can choose to believe it, or not.
quote:
Splitting up MS would result in having an OS maker doing its own thing, and an Application maker doing its own thing. It would have zero effect on things like Kerberos and LDAP that you've mentioned.
|
Perhaps it would at least allow me to purchase a better Kerberos for Windows, no?
quote:
The reason why the splitup was thought of was because people thought that if the OS and its Applications were split up, this would allow companies like Corel and Netscape to compete better.
|
And do you have any reason to doubt this?
quote:
Here's a newsflash: People don't use Office because of Windows. They use Office because everyone else uses Office.
|
Actually, I use office because I think it's the best Office package available. But I wouldn't mind using it on a Sun Workstation, for instance, a market the MS App group will never have any incentive to target as long as they're busy by the, admittedly very profitable, word sells windows sells words sells windows loop.
quote:
If you're doing multiplatform work, did it ever occur to you to use a multiplatform environment like Java? |
The day they port the JWM to MRI scanners and Gene Splicers I'll take your advice under consideration.
|
|
|  |
 |
|  |
All times are GMT. The time now is 05:19. Apolyton Time is 00:19. |
top of page
|
| archivepost |
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|