It can be argued that there's not much point in trimming working sets when an application is minimized, but it definitely doesn't have such a huge impact on performance as you describe.
It sure looks like something causes this huge imact on performance from where I'm sitting...
I also thought it was paging on minimize and that it took so long on my XP laptop because laptop HD's are so slow.
I don't want a command-line based sysadmin OS. I want an easy GUI with all the latest features targeted to average home users.
I've heard the "90% of users" idea. I assume it to be true and take it into consideration. That's why I didn't say it could be done in less than 64MB.
Windows95 can be less than 64MB with almost no loss in functionality if any, but I'm shooting for a lot more than that. I'm trying to match what users are doing now and what Longhorn is trying in this build.
(And I don't pine for VMS. I only said that old mainframe OS's implemented modern advanced core features in a small fraction on the size.)
VMS along with BeOS, NeXT, Plan9, Multics, etc. each demonstrate their respective modern "bloat" features (such as modern kernel features (SMP, virtual memory, hot swapping), vector-based device independent gui, almost overuse of exception handling and error checking/recovery, advanced database filesystems, rich set of media and codec libraries, numurous sets of useful API's, fully-featured networking, advanced security, etc.) all while staying quite small and fast.
If you put all the fattest and most useful parts of these OS's together, you end up with something still relatively small--and you'll be asking yourself, "What has XP/Longhorn got that I haven't got?"
Give it all the help and tools you can think of, some cool eye candy in the windowing system, a scripting language, add some encryption or online transaction support, compression built in... you're pushing 200MB or 300MB on disk, depending on how you go about it. (Maybe 400MB if you're reeeaallly lax about it: a lot of duplication, too much wrapping, unused legacy code, suboptimal implementation, go with your first write or little code cleanup / refactoring, etc.)
It's hard to imagine it getting any bigger just for the OS. What is this missing? What has Microsoft added that this doesn't have? What more could Longhorn possibly have that must stay in memory?
Yeah, in the distant future we won't even need to worry about releasing memory that we allocate. We'll just lose the pointer and cackle with glee as we drive around in our flying cars.
No, no, no. You assume no paradigm shift in what we expect to run on them.
Text will be represented by 1MB/char (on average) as every app will have its text incoded in every language and as pre-recorded sound all the time.
The OS will take 9/10 of the RAM (magnetic, BTW) to provide necessary services, such as dynamically modifying all the apps so they can run at a decent speed on the latest downloaded processor map and also providing a frustrating and inefficient 3D interface.
Future PDAs with 1TB of RAM will be able to edit one plaintext document, no larger than 64 chars long. Anything more requires swapping onto the new IntarWeb(TM) network--and at 5GB/sec (theoretical) only takes several seconds for every swapped local memory operation.
"Why is it obvious that an OS in two years won't consume 400mb of ram?"
Because there's no good reason for it to. It's hard to think of a good reason for it to take 400MB on disk!
This is just the same kind of bloat that took word processors / desktop publishing from floppies to multi-CD bloat-fests. Yes, things should naturally get a little larger with these windowing systems etc., but not so many orders of magnitude larger (or slower).
What does XP give me with that > 1GB install that I couldn't otherwise have in less than 100 Meg? I excpect any disassembly of most modern software will find 100MB chunks of NULLs to increase hardware sales. And perhaps there is one copy of each file for each day of the month.
RAM/HD space is for the data your applications work on, not your OS!
Even Multics, VMS, and MVS back in the day had kernel features comperable to modern, BeOS had the database FS and it was small and fast, Plan9 is just 64MB and the whole windowing system is TINY, NeXT had DisplayPS way back when and was smaller and faster than modern systems...
Let's design an OS.
We'll give it preemptive multitasking w/ hard realtime support, memory protection, ErOS type capability security, sync & async I/O, rethink the API like BeOS, make it multiuser, full-featured shell (before Berkeley [cat should not have flags]), a database filesystem (hell, several lesser FSs too), a DisplayPDF or PS vectored windowing system, give it TCP/IP OpenGL etc., support for filesharing of various types, some initial apps (basic UTF-8 editor, HTML4+CSS browser, image preview)...
How much space? 'Prolly well under 100 or 200 Meg.
It should be noted that on top of parts generally being attributed to certain authors (book of Amos, etc.), analysis shows many of these parts actually had multiple authors. For example, parts believed to have been written by Moses were apparently written by several different people. There are also sections that look like edits.
I was thinking the exact same thing. I spend a lot of time trying to get rid of the crap that comes with XP (ex: http://www.msfn.org/board/index.php?showtopic=1049 7).
This would be of a greater value than regular XP if it were just a naken windowing shell, say (only?) a bloated 150-200MB.
"Oh, I absolutely agree that most people wouldn't go to the trouble to put meta-data on every single file. However, they also won't go to the trouble of renaming every file either."
You can just lump badly named files into a well-named folder. And you can batch-rename.
"So, since putting meta-data on is easier (even if only slightly), they are at least more LIKELY to tag relevant information to a photo than they might have been otherwise. It certainly doesn't worsen the situation."
Why do you assume adding meta-data is easier? I don't think adding any meta to a photo that "Bob is in it" is any easier or descriptive than name/dir. A complicated symlink system to break away from a tree structure, maybe... but since we're talking about a search function, that might be unnecessary.
And there is a cost. This meta-data system is going to add so much bloat... oh, I feel the weight right now.
A system built on file metadata is doomed to be incompatible with anything but the latest datatypes designed for it.
That whole post was quite illucid, but I think this line was about the problem the Mac had before? Files moved over the internet (UNIX's fault?) or to filesystems that don't support all the metadata lose their information.
You have to ask, how will WinFS support metadata on files downloaded off the internet? Will it warn me as I move files onto a FAT32 or EXT3 disk?
For instance, how about if I wanted to search for "my wedding photos?" Neither X1, nor Windows XP's built in search would find your wedding photos. Why? Because they have useless names like DSC0001.jpg and there's no metadata that says they are wedding photos."
That's why you can change filenames and organize things into directories.
"when someone says they have written something that breaks conventional wisdom, and the only response is from people saying it's rubbish"
That's just the thing. He wants us in a state of wonder over what most scientists already know is true. Very little in the book was not already published by others when he started writing.
And no, that's not what makes him a crank. He's a crank for his "this looks like a snowflake, therefore this explains snowflakes" attitude. No it doesn't.
I've done my best to read it. It's derivitive in its best parts, and just wrong or so unspecific and unclear to be almost meaningless in its worst. All he succeeded in doing is transforming pre-existing real science into new pseudoscience.
Some good reviews by real scientists & such: http://arxiv.org/PS_cache/quant-ph/pdf/0206 /020608 9.pdf http://www.ams.org/bull/2003-40-01/S0273-09 79-02-0 0970-9/S0273-0979-02-00970-9.pdf http://groups.go ogle.com/groups?selm=b9401f8a.0302 062325.194d0c8d%40posting.google.com
I saw a lecture of his streamed over the internet. He started out saying that it would be more of a quick introduction or summary and wouldn't be as in-depth as the book. Well, I've seen the lecture and read the book. While there's a whole lot more text in the book, it didn't seem any more in-depth. The very things I wanted more detail on were brushed off (sometimes with an almost identical phrase from the lecture) in the book.
Surprisingly, despite his continuous repetition that this is a great revolution, he doesn't do a great job defending that position. Take his writing on fluid flow and the inadequacy of equations. He rehashes the traditional problem, offers the CA take on it, and fails to give us anything of any practical use. If he could compute a solution faster, with more accuracy, or give the solution to an unsolved problem with this technique, that would be great, but he can't.
Or take his views on biology. He says there could be a small set of genes that are a "leaf generator" and with a few small tweaks, generate all known leaf types. No duh. It isn't the only possible view on it, but its many people's naive theory. Ditto for shells.
This happens all throughout the book. He finds something surprising that I think most wouldn't. He than shows that it's not surprising from the CA perspective, and he basically stops there. Lameness, I say.
(If I'm wrong, do your/. duty: call me an idiot and correct me with something intelligent.)
Like in XP when I click on "Always use this program to open these files" and then... can you guess what? It doesn't always use this program to open these files. (Shock)
It just keeps on giving me that box, even if I play with the file types in explorer. I have to edit the registry to fix it.
The best keyboards use mechanical switches. I'm using a Focus 2001 right now and I could keep typing for hours with few mistakes and less pain than with some squishy membrane.
Re:Two things you can't say
on
What You Can't Say
·
· Score: 2, Informative
"We don't let women do things like crew submarines or fly combat jets in battle."
For submarines, the reason I hear is that women would want/need seperate showers etc. and there is not room on a sub.
As for flying jets, I think it's the male desire to protect the female; keep her from falling into the enemies hands where who knows what would happen to her.
A little drink for a young teen at a family meal == teaching good drinking habits.
Keeping all alchahol away until 21 == making it more desireable than it should be, with habits formed at underage unsupervised parties.
No matter how obvious this is, it still gets people upset.
Re:Things labelled heresy on slashdot
on
What You Can't Say
·
· Score: 2, Insightful
#1 & #3 are not heresy. They are just not true. Saying the Earth is flat, 2+2=5, and the sky is a tea-cup are also not heresy. They're just close your eyes, cover your ears, deny all facts style contrarianism.
#2 wasn't true when Feminism meant "give us an equal chance, let's see what we can do." Now Feminism, after being hijacked several times by fringe groups, has more to do with lesbianism, anti-maleness, and moronic post-modern philosophies.
But they had power almost only when they had good ideas and lost power when they didn't.
Some said 3d interfaces were the future. We'd all be using them in the year 2000.
Like flying cars, it was wrong on so many levels.
Are capability systems a blanket solution?
on
20 Years of Virii
·
· Score: 1
(Warning, ignorance ahead.)
I think a capability system (ex. EROS) is theoretically invulnerable because a virus would never have the rights it needs to infect. I barely understand how a capability system works, but I think it goes like this:
Your e-mail client (for example) can't see anything but itself and e-mail (not even the file system), and it doesn't have authority to write onto itself.
This is possible because every process or program has it's own set of "keys" that grant it rights to see/read/write/modify/execute a file, resource, or process. This is in comparison to giving universal rights/limits to users which apply equally to every process and program run under that username.
I've been looking at FAQs and asking questions here and there, but never got a grasp at how it secures without disabling the ability to administer a computer or what it means for viruses.
Yeah yeah. Saudi. I often spell Iran with a P. (And you missed my missing "the".)
Except the media in the rest of the world, which was reporting that they were Saudi within a couple of hours of the event.
I don't know exactly how long it took the US news, but it wasn't days and they did report it many many times. Like I said, I tried to remind people they were Saudi--not Iraqi--even before it bacame "the thing to do."
Of course, the White House wasn't helping at all with that...
Slashdot Mirror
It can be argued that there's not much point in trimming working sets when an application is minimized, but it definitely doesn't have such a huge impact on performance as you describe.
It sure looks like something causes this huge imact on performance from where I'm sitting...
I also thought it was paging on minimize and that it took so long on my XP laptop because laptop HD's are so slow.
I don't want a command-line based sysadmin OS. I want an easy GUI with all the latest features targeted to average home users.
I've heard the "90% of users" idea. I assume it to be true and take it into consideration. That's why I didn't say it could be done in less than 64MB.
Windows95 can be less than 64MB with almost no loss in functionality if any, but I'm shooting for a lot more than that. I'm trying to match what users are doing now and what Longhorn is trying in this build.
(And I don't pine for VMS. I only said that old mainframe OS's implemented modern advanced core features in a small fraction on the size.)
VMS along with BeOS, NeXT, Plan9, Multics, etc. each demonstrate their respective modern "bloat" features (such as modern kernel features (SMP, virtual memory, hot swapping), vector-based device independent gui, almost overuse of exception handling and error checking/recovery, advanced database filesystems, rich set of media and codec libraries, numurous sets of useful API's, fully-featured networking, advanced security, etc.) all while staying quite small and fast.
If you put all the fattest and most useful parts of these OS's together, you end up with something still relatively small--and you'll be asking yourself, "What has XP/Longhorn got that I haven't got?"
Give it all the help and tools you can think of, some cool eye candy in the windowing system, a scripting language, add some encryption or online transaction support, compression built in... you're pushing 200MB or 300MB on disk, depending on how you go about it. (Maybe 400MB if you're reeeaallly lax about it: a lot of duplication, too much wrapping, unused legacy code, suboptimal implementation, go with your first write or little code cleanup / refactoring, etc.)
It's hard to imagine it getting any bigger just for the OS. What is this missing? What has Microsoft added that this doesn't have? What more could Longhorn possibly have that must stay in memory?
Yeah, in the distant future we won't even need to worry about releasing memory that we allocate. We'll just lose the pointer and cackle with glee as we drive around in our flying cars.
No, no, no. You assume no paradigm shift in what we expect to run on them.
Text will be represented by 1MB/char (on average) as every app will have its text incoded in every language and as pre-recorded sound all the time.
The OS will take 9/10 of the RAM (magnetic, BTW) to provide necessary services, such as dynamically modifying all the apps so they can run at a decent speed on the latest downloaded processor map and also providing a frustrating and inefficient 3D interface.
Future PDAs with 1TB of RAM will be able to edit one plaintext document, no larger than 64 chars long. Anything more requires swapping onto the new IntarWeb(TM) network--and at 5GB/sec (theoretical) only takes several seconds for every swapped local memory operation.
"Why is it obvious that an OS in two years won't consume 400mb of ram?"
Because there's no good reason for it to. It's hard to think of a good reason for it to take 400MB on disk!
This is just the same kind of bloat that took word processors / desktop publishing from floppies to multi-CD bloat-fests. Yes, things should naturally get a little larger with these windowing systems etc., but not so many orders of magnitude larger (or slower).
What does XP give me with that > 1GB install that I couldn't otherwise have in less than 100 Meg? I excpect any disassembly of most modern software will find 100MB chunks of NULLs to increase hardware sales. And perhaps there is one copy of each file for each day of the month.
RAM/HD space is for the data your applications work on, not your OS!
Even Multics, VMS, and MVS back in the day had kernel features comperable to modern, BeOS had the database FS and it was small and fast, Plan9 is just 64MB and the whole windowing system is TINY, NeXT had DisplayPS way back when and was smaller and faster than modern systems...
Let's design an OS.
We'll give it preemptive multitasking w/ hard realtime support, memory protection, ErOS type capability security, sync & async I/O, rethink the API like BeOS, make it multiuser, full-featured shell (before Berkeley [cat should not have flags]), a database filesystem (hell, several lesser FSs too), a DisplayPDF or PS vectored windowing system, give it TCP/IP OpenGL etc., support for filesharing of various types, some initial apps (basic UTF-8 editor, HTML4+CSS browser, image preview)...
How much space? 'Prolly well under 100 or 200 Meg.
//end pathetic, illucid rant
I hate verbose languages. My VB code would be much more readable if MS had decided to use curly brackets instead of words to end blocks.
Or even just indentation.
It should be noted that on top of parts generally being attributed to certain authors (book of Amos, etc.), analysis shows many of these parts actually had multiple authors. For example, parts believed to have been written by Moses were apparently written by several different people. There are also sections that look like edits.
I was thinking the exact same thing. I spend a lot of time trying to get rid of the crap that comes with XP (ex: http://www.msfn.org/board/index.php?showtopic=1049 7).
This would be of a greater value than regular XP if it were just a naken windowing shell, say (only?) a bloated 150-200MB.
"Oh, I absolutely agree that most people wouldn't go to the trouble to put meta-data on every single file. However, they also won't go to the trouble of renaming every file either."
You can just lump badly named files into a well-named folder. And you can batch-rename.
"So, since putting meta-data on is easier (even if only slightly), they are at least more LIKELY to tag relevant information to a photo than they might have been otherwise. It certainly doesn't worsen the situation."
Why do you assume adding meta-data is easier? I don't think adding any meta to a photo that "Bob is in it" is any easier or descriptive than name/dir. A complicated symlink system to break away from a tree structure, maybe... but since we're talking about a search function, that might be unnecessary.
And there is a cost. This meta-data system is going to add so much bloat... oh, I feel the weight right now.
Ach, I shold just stop bitching about it.
Silly me. It looks like Anon was talking about requiring metadata to be built into every file format, thus leading to incompatibility.
It's been pointed out: that's not how it works.
A system built on file metadata is doomed to be incompatible with anything but the latest datatypes designed for it.
That whole post was quite illucid, but I think this line was about the problem the Mac had before? Files moved over the internet (UNIX's fault?) or to filesystems that don't support all the metadata lose their information.
You have to ask, how will WinFS support metadata on files downloaded off the internet? Will it warn me as I move files onto a FAT32 or EXT3 disk?
For instance, how about if I wanted to search for "my wedding photos?" Neither X1, nor Windows XP's built in search would find your wedding photos. Why? Because they have useless names like DSC0001.jpg and there's no metadata that says they are wedding photos."
That's why you can change filenames and organize things into directories.
I'm thinking I should post these in HTML.
8 9.pdf 0 0970-9/S0273-0979-02-00970-9.pdf 2 062325.194d0c8d%40posting.google.com
http://arxiv.org/PS_cache/quant-ph/pdf/0206/02060
http://www.ams.org/bull/2003-40-01/S0273-0979-02-
http://groups.google.com/groups?selm=b9401f8a.030
"when someone says they have written something that breaks conventional wisdom, and the only response is from people saying it's rubbish"
6 /020608 9.pdf9 79-02-0 0970-9/S0273-0979-02-00970-9.pdfo ogle.com/groups?selm=b9401f8a.0302 062325.194d0c8d%40posting.google.com
That's just the thing. He wants us in a state of wonder over what most scientists already know is true. Very little in the book was not already published by others when he started writing.
And no, that's not what makes him a crank. He's a crank for his "this looks like a snowflake, therefore this explains snowflakes" attitude. No it doesn't.
I've done my best to read it. It's derivitive in its best parts, and just wrong or so unspecific and unclear to be almost meaningless in its worst. All he succeeded in doing is transforming pre-existing real science into new pseudoscience.
Some good reviews by real scientists & such:
http://arxiv.org/PS_cache/quant-ph/pdf/020
http://www.ams.org/bull/2003-40-01/S0273-0
http://groups.g
I saw a lecture of his streamed over the internet. He started out saying that it would be more of a quick introduction or summary and wouldn't be as in-depth as the book. Well, I've seen the lecture and read the book. While there's a whole lot more text in the book, it didn't seem any more in-depth. The very things I wanted more detail on were brushed off (sometimes with an almost identical phrase from the lecture) in the book.
/. duty: call me an idiot and correct me with something intelligent.)
Surprisingly, despite his continuous repetition that this is a great revolution, he doesn't do a great job defending that position. Take his writing on fluid flow and the inadequacy of equations. He rehashes the traditional problem, offers the CA take on it, and fails to give us anything of any practical use. If he could compute a solution faster, with more accuracy, or give the solution to an unsolved problem with this technique, that would be great, but he can't.
Or take his views on biology. He says there could be a small set of genes that are a "leaf generator" and with a few small tweaks, generate all known leaf types. No duh. It isn't the only possible view on it, but its many people's naive theory. Ditto for shells.
This happens all throughout the book. He finds something surprising that I think most wouldn't. He than shows that it's not surprising from the CA perspective, and he basically stops there. Lameness, I say.
(If I'm wrong, do your
"computers do EXACTLY what you tell it to do"
Like in XP when I click on "Always use this program to open these files" and then... can you guess what? It doesn't always use this program to open these files. (Shock)
It just keeps on giving me that box, even if I play with the file types in explorer. I have to edit the registry to fix it.
'Does exactly what I ask... Bah!
Yes, yes, yes, yes, yes, yes... what?
What the hell is wrong with hollow-point bullets?
I'm still using 95. /I'm important
or is Lobbyist some kind of profession in USA?
That's what you'd put on your tax form. It's all legal.
The best keyboards use mechanical switches. I'm using a Focus 2001 right now and I could keep typing for hours with few mistakes and less pain than with some squishy membrane.
"We don't let women do things like crew submarines or fly combat jets in battle."
For submarines, the reason I hear is that women would want/need seperate showers etc. and there is not room on a sub.
As for flying jets, I think it's the male desire to protect the female; keep her from falling into the enemies hands where who knows what would happen to her.
I have to agree.
A little drink for a young teen at a family meal == teaching good drinking habits.
Keeping all alchahol away until 21 == making it more desireable than it should be, with habits formed at underage unsupervised parties.
No matter how obvious this is, it still gets people upset.
#1 & #3 are not heresy. They are just not true. Saying the Earth is flat, 2+2=5, and the sky is a tea-cup are also not heresy. They're just close your eyes, cover your ears, deny all facts style contrarianism.
#2 wasn't true when Feminism meant "give us an equal chance, let's see what we can do." Now Feminism, after being hijacked several times by fringe groups, has more to do with lesbianism, anti-maleness, and moronic post-modern philosophies.
But they had power almost only when they had good ideas and lost power when they didn't.
And I'll second your #4.
Some said 3d interfaces were the future. We'd all be using them in the year 2000.
Like flying cars, it was wrong on so many levels.
(Warning, ignorance ahead.)
I think a capability system (ex. EROS) is theoretically invulnerable because a virus would never have the rights it needs to infect. I barely understand how a capability system works, but I think it goes like this:
Your e-mail client (for example) can't see anything but itself and e-mail (not even the file system), and it doesn't have authority to write onto itself.
This is possible because every process or program has it's own set of "keys" that grant it rights to see/read/write/modify/execute a file, resource, or process. This is in comparison to giving universal rights/limits to users which apply equally to every process and program run under that username.
I've been looking at FAQs and asking questions here and there, but never got a grasp at how it secures without disabling the ability to administer a computer or what it means for viruses.
Anyone?
Yeah yeah. Saudi. I often spell Iran with a P. (And you missed my missing "the".)
Except the media in the rest of the world, which was reporting that they were Saudi within a couple of hours of the event.
I don't know exactly how long it took the US news, but it wasn't days and they did report it many many times. Like I said, I tried to remind people they were Saudi--not Iraqi--even before it bacame "the thing to do."
Of course, the White House wasn't helping at all with that...