OSNews

And there is a new VM in the Kernel, as well as the CPU depending FPU optimizations which the kernel now understands.
Even newest APIC-Support and HT and EMT64 is on the way, folks.

My Dell Latitude D600 works with no problem.
even my Wlan card (intel 2100)works (only with 128 wep)

My intel 2200 are noy working but thats a bug (I don't get a ip.. and reported)

I hade to install the experimantal audio driver to get my sound to work.

and install the latest VLC 8.3 (night build from BeBits)for playing DVD

hmmm...if there were a livecd, I would try it out...

Notebook support _might_ be a problem. Zeta needs a 100 x86 System. A lot of notebook suppliers use to ignore standards of all kind. ^^ Well, for Windows it's OK. They just add their own drivers for their hardware. The best way to find a suiteble notebook is going to the shop with the Zeta-CD and asking if you may install it! The other way round, going to the shop with a Notebook and asking if you may install one of their copies of Zeta on it, could be difficult.

As a quote here stated out, there are significant kernel - related changes in the Zeta R1 (Ram limitation, bigger Tracker-Addon space, new VM model f.ex.)
Working on these portions without having the Kernel Sourcecode is simply impossible.
Working on that without Palm's Notice about that is impossible (if it occurs without their permission or whatsoever).

Meanwhile we still make some bla-bla about the sourcecode ownership of the Kernel of Dano etc., yT inofficially speaks out, that the next Update Version (R1.5) possibly will content a brand-new Kernel. But for sure in Version 2.0 including a 64-Bit Version of Zeta.

So go on living with rumours. Some even believe that OS X is a MediaOS with it's nearly 40year old legacy OS underlying in it. So it's ok.

>You say:
"Booting is a matter of, say, 15 seconds, completely blowing away any Linux, Windows or OS X install."

While it might be faster than the mentioned operating systems, I can boot my linux system into kde in 20 seconds.
And my windows xp takes about 20 to 25 seconds to boot.

So I would not have used the words "blowing away"

You got a point, however, who's to say that booting Zeta would take 15 seconds on your machine? Maybe it takes less? We can only know for sure when someone posts the boot times for the different OSes on the same hardware.

So I would not have used the words "blowing away"

Booting XP on my machine takes roughly 30 seconds. And yes, I find cutting boottime in half "blowing away" .

Can boot WinXP and Linux to KDE in 20s on P2 400mhz with 192mb of ram? I highly doubt it ;p

Put things in perspective,
1. What is your system specs
2. What tweaking have you done.

I ask this because I have the fastest and greatest laptop and still dont get 15 sec.

Your computer is not typical. Every version of BeOS and Zeta that I have installed on my various computers throughout the years has always booted far faster than any other OS I have ever installed (various Linuxes (slow!), Windows 98 and XP, some others). BeOS/Zeta boots consistently under about 15 seconds while all others boot in around a minute (sometimes longer (Linux)). BeOS/Zeta is by far the fastest, most responsive, and best multimedia capible OS I have ever used.

> While it might be faster than the mentioned operating
systems, I can boot my linux system into kde in 20 seconds.
> And my windows xp takes about 20 to 25 seconds to boot.

I'm curious which version of Linux and which HW you use to have such fast startup, do you use a normally configured distro or did you have to change the configuration to speed up the boot?
The good thing about BeOS is that the fast boot time is the default.

An average AthlonXP P4 System with a 3.0 GHz CPU or alike one won't need more than 5 Seconds for booting in Zeta R1.

Basically, it depends on the drivers which have to be loaded as an extra, especially if it's a IDE Card or a SCSI CARD with connected devices on it.
But honestly, this makes XP Pro also slower as well as Linux.

15 Seconds for a 400 MHz CPU with an average set of devices is not a practical Value on W2K, XP, OS X or Linux - it's utopic.

you all seems to forget that AmigaOs has the fastest boot time than any other OS mentioned here and this on an outdated 14Mhz 68k motorolla processor with 2MB of RAM.
Regards
Lio

What's your point? DOS can boot in seconds too.

But, just like AmigaOS, its old, antiquated, and defunct. Sure, I loved my Amiga back in the day and would take one over a DOS machine anyday, but both of these systems are very simple compared to all the capabilities of a modern OS. BeOS just happens to offer almost every modern feature known plus it can boot up really fast.

Users can still run Pixel image editor from http://www.kanzelsberger.com

you can boot XP and linux on a Pentium2 400Mhz in 20seconds approx...

Not in my experience..

While many people like the ability to skin their windows, it detracts from one of the best aspects of BeOS's little yellow tabs (and yellowTab's namesake): that you can drag the title tab of a window, and layer them on top of one another, just like tabbed folders.

But you couldn't move the tab in relation to it's window, could you? Because don't you end up with a full window on the left, and a very thin window on the right when doing this?
----
When OpenTracker added single window file browsing, that pretty much wrecked the whole of the tracker.

How is opening new windows for everything you click a good thing compared to using back and forward? I hated the clutter it gave on Windows and same on BeOS.

But you couldn't move the tab in relation to it's window, could you? Because don't you end up with a full window on the left, and a very thin window on the right when doing this?

Yes, you could. Moving the tab doesn't change the shapeo f the window, it moves the tab. Instead of having it on the far left, you could put it in the middle, or on the right, or whatever.

How is opening new windows for everything you click a good thing compared to using back and forward? I hated the clutter it gave on Windows and same on BeOS.

Having one window per folder strengthens the "folder == window" metaphor, as well is making it possible to actually use drag & drop to manage files. Most Windows users I know don't use drag & drop, they use copy & paste (because going to another folder means you can't see your original folder anymore). The problem is, there is no way to copy and paste files without either knowing how to use the right-click menu or knowing how to use the keyboard shortcuts (which is a UI design no-no).

Having used spatial Nautilus for a good six months in GNOME, I have to say its kind of hard to go back to anything else. BeOS's Tracker wasn't really spatial, but it was better than the "browser" paradigm.

"Having used spatial Nautilus for a good six months in GNOME, I have to say its kind of hard to go back to anything else"

amen! that is one thing I thought was horrible.... UNTIL i used it and got used to it now it is like tabbed browsing - nothing you miss until you use it then you wonder how you ever lived without it....

Couldn't have said it better myself.

Any MacOS user from the "good old days" will know this feeling. (=

Spacial is neat, relationalt database is better

Cheers,
. Knut

Yes, you could.<snip>

Ah, that elucidates a lot. Now I've only got to figure out how to shift the tabs..

If I recall correctly, you press Shift and drag the tab with your mouse.

Long time between drinks though.

Having used spatial Nautilus for a good six months in GNOME, I have to say its kind of hard to go back to anything else. BeOS's Tracker wasn't really spatial

What's not spatial about it? Unless you have "single window browse" turned on, the only non-spatial feature I've come across is that Tracker (irritatingly) lets you open additional windows for the same folder *if* they're in different workspaces. Originally it didn't do that, but some misguided soul "fixed" it when Tracker was made open source.

Yes, you could slide the tab in relation to the window. If I remember correctly you just hold the ctrl key down, click and drag to move them.

I never moved them much, just used lots of workspaces.

How is opening new windows for everything you click a good thing compared to using back and forward?

Because then you don't have to constantly resize windows to fit their contents.

I hated the clutter it gave on Windows and same on BeOS.

Clutter is not intrinsic to a spatial file manager, especially when you know how to use it.

I agree. The nice thing about BeOS, was that it was a very Mac-ish UI, except fast and on a PC. Zeta seems very different, from the bloated preference app to the gaudy bling-bling tabs to the WinXP rip-off scrollbars.

On my r5 machine OpenTraker has the option of single or multiple window browsing...so the choice is up to the user. I use single window on my laptop because it only supports 800x600 resolution so opening a lot of windows can make me feel crowded.

the trick was to use CTRL + doubleclick. This would open a new windows and close the old one. I really like the multiwindow approach combined with all the shortcuts.

>all the look and feel comes directly from the post R5 appserver codebase.
This look and feel 'post R5' was purely for testing/beta purposes, never released to public by Be,inc

AFAIK the app_server also runs some major changes since Dano - just look at for in the version on R1 ;-)
In fact there are realy just a few portions of 'Dano' which weren't touched by yT, or rewritten!

On my dual P4 Xeon, it supports both CPUs, works with my SCSI disks, and prints perfectly. I don't have a scanner, so I can't say anything about that.

The test was done on a 400MHz machine, so in comparison, its not bad at all!

It wasn't. The PII 400 is just a side-machine I have. The test itself was done on a AMD Athlon XP 1600+, Ati Radeon 9000 128MB, 512 MB RAM. If you looked at the screenshots, you would've known .

At work I need to have my computer started at a certain time or else my boss will think I'm late. It takes about 10 minutes to start up Windows with all the applications that's needed for my work (Windows itself takes about 5-6 minutes). If I could cut that boot time in half I could miss a subway train and still have time over. So don't diss boot times. Even though my situation depends a lot on how the autostart apps behave I would've saved time if Windows could behave better.

I care about boot times and maintenance-free computers because computers are made for the purpose of serving us humans, not vice versa, belive it or not. I view my *personal* computer similary to how I view my scientific calculator; which means that old fashioned mainframe-like behaviors (like maintenance, loud fans, heat, or silly boot times) disturbs me.

Best regards
Henrik K.

I fully agree. I prefer not to buy a ,,cat in a sack'' as Polish proverb says. To see it's to belive. :-)

"I prefer not to buy a ,,cat in a sack'' as Polish proverb says."

Nice. We say almost the same in Italian: "the cat in the sack"

hmm, im starting to suspect that its a kinda universial expression as it exist here in norway to

In Sweden we say "pig in the sack"

>In Sweden we say "pig in the sack"
Which is obviously ( ) the most correct. I mean, come on, who the fuck _buys_ a cat?

Granted, not many people buys pigs these days either supposedly the sayings are fairly old.

Round here you "buy a pig in a poke" .....it might be a moggy but you won't know untill you "let the cat out of the bag!"

Windows XP doesn't feel like Windows 95 either, but still, it still has the same feel."

Correction: "Windows XP isn't the same as Windows 95 either, but still, it still has the same feel."

I wouldn't say that was comparable, as windows doesn't have a single feel. It randomly uses both mdi and sdi (or both, in office's case), toolbars move around, dialogs are used for different sorts of things, and so on.

BeOS on the other hand was very consistent, with only some odd apps really needing to be learned, most you would just start using and feel comfortable with.

Moving to a garish look without a spatial Tracker isn't "moving on", its "moving backwards". Having enormous, space-wasting icons in Tracker is moving backwards. Having a media player that looks like a children's toy instead of a proper BeOS application (like SoundPlay does) is moving backwards. Having an "all-in-one" configuration app that violates several UI principles (eg: egregious amounts of blank space) is moving backwards.

for you it isnt but dont you remember that Yt is aiming for the NEW users in stead of the old user base? they want to sell...and many people already do...

as for the icons, you can make them smaller
as for the "garnish look" you can choose for R5 theme.
as for the Media Player..People want eyecandy..

i dont think its bad..look at Win XP everybody was saing the Teletubby Interface...and what hapened?? everybody uses it now...

just my 2 cents

Every menu-link to the separate preference apps still remain, they only link directly to a page in the new preferences app.
So you can still invoke a sub-part of the preferences the old way.

@ emil:

> Hibertnation. Suspend mode.

No. I need full startup to detect network settings and so on. You know, I have a laptop to take it to different networks.

> a instant-Linux based on initrd or a microLiveCD/USB

way slow compared to DOS on a harddisk.

> init 1

hmm, yes but forgot about that one. thanks for reminding :-)

Well, on Linux all you need is hibernation and a icon pointing to "service network restart". That's going tp take all of 2 seconds+ whatever your DHCP takes.

the boot time is only true if you are using a non-Gentoo linux OS. With gentoo, my laptop boots in a manor of seconds, without a bloated kernel, and runs fine. But installation time must be factored in I guess...

if you use this, it gives you a simple shell really fast:

lilo: linux rw init=/bin/bash

*Don't*.
If your distro uses an initrd image or your kernel doesn't have all the required driver compiled into the binary image, at best you'll have an empty OS under your hands.
sysinit is there for a reason.
Even init 1 may cause grief on some distros.

Gilboa

OK, thanks, I just stick to {
lilo: linux rw 1 /or/ lilo: linux rw single } then

you're an idiot, you've taken his comment out of context, its clearly sarcasm.

Oops, it was sarcastic.
Oh and you're a mean poopy-face.

that's the funniest thing I've heard in the past day...

OpenBeos = Haiku

They just changed the project's name.

I tried to change it to cocaine, but "sipping" didn't work.

The rest of the world uses ID3 but BFS is blind to it.

This is the kind of thing that's built into more fleshed out tools like Spotlight. Oh, and heaven forbid you ever decide to change an ID3 tag, because you'll need to run the tool again to re-sync up the BFS attributes. This is also true of other standard metadata information, such as EXIF tags, Word document metadata, movie sizes, you name it. BeOS fans have tried to sell the world on BFS for years without ever acknowledging the massive amount of work it takes (speaking as someone who ran BeOS for four years) the user to keep the attributes correct and up-to-date.

You make valid points about content-search, but I disagree about it being a BeOS failure that it doesn't sync ID3tags with the attributes.

Theoretically, it should be easy enough to implement into the programs you use to change tags to also change the attributes, or to make some nifty OpenTracker plugins for it (anyone? . But attributes are awesome, and the world should've adapted to attributes, not the other way around (and it's only happening slowly now). Having different formats for song-information on ogg, mp3, midi and eac isn't making things better (does Spotlight support all those tag-formats btw?). Having a singular attribute naming the song's artist is a whole lot better. It would've been really nice if those attributes were downloadable across the internet unzipped by now.

Attributes are a nice idea in theory, but we live in a world of commodity operating systems, and it seems attributes just didn't make the cut. At this point, I'm far more intrigued by Hans Reiser's ideas about exposing embedded file information (eg: id3 tags, EXIF tags) through the filesystem itself.

Theoretically, it should be easy enough to implement into the programs you use to change tags to also change the attributes

Up to this point in the BeOS's history, it has never happened. And because stale data is usually worse than no data at all, it's a boil the ocean problem: every program touching any kind of metadata would have to be updated, because any chance whatsoever of updating a file without updating the metadata means the chance for stale (meta)data.


or to make some nifty OpenTracker plugins for it (anyone?

It is probably not the place of OpenTracker to do this sort of thing, although I guess it could be hacked in, unless Tracker wants to handle notifications for every file modification in the system. (I could be wrong on this point, but I believe that currently Tracker only registers for file modification notifications for files/directories that are currently visible.)


But attributes are awesome, and the world should've adapted to attributes, not the other way around (and it's only happening slowly now).

The world didn't. You can make whatever theoretical argument you want, but in 2005, if I download a (presumably legal) MP3 from the internet, it will have ID3 tags set but not extended attributes. If I receive an e-mail containing photos from someone, they will have EXIF tags set, but not extended attributes. The BeOS implementation remains theoretically cool but practically not very useful. Saying "the entire internet and most programs should be re-written, and then it will be useful!" is not really a good answer. Metadata-embedded-in-data-files is here to stay, at least for a while. Implementations like Spotlight have adapted to the realities of the world and managed to make things useful. It's not like the concept of an Importer is rocket science. Be just never had the time / resources to really turn BFS into a end-user solution, and it seems that Yellowtab aren't doing so, at least for now, either. Selling BFS as a convenient end-user tool is just dishonest.


does Spotlight support all those tag-formats btw?

Spotlight allows anyone to write Importers for any file type they want, much like Be's concept of media translators. As for whether Apple wrote Importers for all those filetypes and shipped them as part of Tiger, I do not know. Certainly they did for MP3 and AAC, although I would be shocked if they bothered with OGG. Doesn't stop anyone else from writing one, though.


Having a singular attribute naming the song's artist is a whole lot better

...which is exactly what Spotlight does. The Importers Apple ships match the Artist tag from both MP3 files and AAC files to a Spotlight tag called "Artist". Say a new format comes out and uses "Performer" instead of "Artist". A Spotlight Importer could still map that tag to the "Artist" tag in the Spotlight database.


It would've been really nice if those attributes were downloadable across the internet unzipped by now

But they're not and they won't be any time soon, so devising a "solution" that relies upon that being implemented is foolhardy. I think a good analogy is Be's (excellent) file type handling. If the file has a MIME type specified, great, use it! But if not, rely on the .extension to help identify it. In practice, it worked great. It's a very similar approach to what Apple used in Spotlight. "No metadata for the file? Well, look to see if there's any metadata embedded in the file that we can expose!" I imagine that Be's engineers would have eventually implemented this, but the fact is, it's not there, and Thom's discussing it like it's a complete, ready-for-the-user solution like Spotlight is simply dishonest.

If you're willing to do a lot of work (i.e. manually syncing everything yourself, effectively meaning "use third-party sync-it programs like Spotlight Importers, but instead of having the system take care of the re-syncing for you, do it yourself"), BFS is extremely cool. But I can testify that as a multi-year user, it's too much work to be worth the effort. And its coolness has been lessened with the years, as tools like Spotlight not only bring that kind of syncing to the table, but also allow for full-text searching and other features.

"If you're willing to do a lot of work (i.e. manually syncing everything yourself, effectively meaning "use third-party sync-it programs like Spotlight Importers, but instead of having the system take care of the re-syncing for you, do it yourself"), BFS is extremely cool. But I can testify that as a multi-year user, it's too much work to be worth the effort."

Hrrrmm. I agree with many of your points, and disagree with some, but I don't think Thom was selling it as a Spotlight-buster. And changing and resynching attributes are not really that hard... and how often do they change if they were well thought-out in the first place?

I still have R5.0.3. I am also bucking for Haiku (and have Zeta RC3 as well). BUT! My primary computers are all Macs running OS X 10.4.x. So I understand what you are saying... but to me, BeOS was great and BFS was fantastic and quite useful. I wrote code to use file attributes to map information to my wife's artwork which could be pulled up and used via python on her website... it was fast and worked well and easy to add new images and map new attributes to it (since the data mapping could be done programmatically OR THROUGH the tracker).

What you are missing about BFS is that all attributes are immediately available via the user interface and modifiable as well. This is something Spotlight lacks. It is also faster and was programmatically very [easily] accessible since it is an integral part of the file system - granted having a large index file for spotlight is no slouch and works quickly as well.

Again, I understand what you are saying, since I am a BeOS user as well as a Mac OS X user (and a linux user and a windows user (on a regular basis)), but I think you are unfairly tearing BFS and file attributes down and punditing on your soapbox.

I always found it a li dumb not to be able to search other FS eg NTFS partitions in BeOS.

What about Be_indexed? It should do that Apple 'importer' does, tried it? I'm curious.

Also I'd like to search in specific folders only but the search function in BeoS can't even filter that out : (

Nutela

What about Be_indexed? It should do that Apple 'importer' does, tried it? I'm curious.

Partially, but not exactly. I haven't personally used BeIndexed, but from looking at its description, it looks like it's simply a full-content searching engine. To the extent that Importers also provide the content to Spotlight, they are similar in that way. But really, BeIndexed is more akin to Panther's search engine (i.e. "periodically index the drive") than Spotlight.

Spotlight's Importers also provide the metadata to the Spotlight database. The BeOS equivalent would be an importing engine that set the BFS attributes, so that you could do all the typical fun BFS things, like live queries, etc., which you can do with Spotlight. BeIndexed does not do that; it's completely separate from BFS. And, of course, one of the most important things is that Spotlight is hooked into the kernel, so that the database information is updated when the file is -- it's always up-to-date. BeIndexed provides for no such thing.

So really, BeIndexed is more akin to Sherlock in Mac OS 9 or Panther's full-text searching. It's not really at all to the level of Spotlight, or even really BFS.

Searching by content is so much more useful that I'll gladly take a speed hit for it when I want that functionality.

Eh, so will I - that's why they have this neat little program called "grep".

Sure, ummm, but the author "conveniently" forgot to add the single most important thing! How does the BFS MP3 attributes database get populated? Unlike Spotlight or similar technologies, it doesn't have the concept of "Importers" or "Syncers" or whatever you want to call them. Even if you've spent hours meticulously setting up your ID3 tags, Zeta won't be able to find your music. You have to either manually enter the same information for each song again, or you must download a program that will read the ID3 tags and then set BFS attributes for them.

Yes, Spotlight is a much more automagic system. The trade off is that it's significantly less flexible than queries.

In other words, unless you rip every CD using a BeOS program that can set BFS attributes and don't have any MP3s you want to use from your other operating systems, *you do need specialized applications.*

Or you can just play them in SoundPlay, which will write the ID3 tags to attributes the first time you play a file.

The rest of the world uses ID3 but BFS is blind to it.

Um, of course it is. Why would the filesystem have awareness of MP3's internal metadata format?

"At least you get your money back if you can't get it to work on your system with help of the telephone support "

Are you serious? That'd be good news for me! Where did you read that, please?

That German sale manager said so.

"Whoever buys Zeta and can't get it to run on his machine with the help of our telephone support can send back his CD and will definately get a refund."

Not sure whether this is only valid for German customers though.

use id3attr to convert id3 to attributes

Dude, did you even read my comments before responding? In both I specifically mentioned that you can convert from ID3 tags (and other common types) to attributes with the help of third-party tools. Here's the Cliffs Notes version:

or you must download a program that will read the ID3 tags and then set BFS attributes for them

...*you do need specialized applications.*

Oh, and heaven forbid you ever decide to change an ID3 tag, because you'll need to run the tool again to re-sync up the BFS attributes

If you're willing to do a lot of work (i.e. manually syncing everything yourself, effectively meaning "use third-party sync-it programs like Spotlight Importers, but instead of having the system take care of the re-syncing for you, do it yourself")


Of course, there are still several problems:

1. The user has to know what the difference is between embedded metadata, such as ID3 tags, and BFS attributes, and that you can only search based on BFS attributes. (Indexed BFS attributes, at that, but that's a different issue.)

2. The user has to understand that the two, being distinct, can get out of sync, and understand when and why he/she must manually sync them (i.e. run the converter tool again).

3. The user must realize that every new file not originating from his/her system (i.e. if a CD is ripped in BeOS the program will probably set the BFS attributes, but files acquired otherwise won't) will have to be manually converted, and must be vigilant in doing so.

Compare that to a system like Spotlight, where all that metadata is automatically populated and automatically kept up-to-date.

> Compare that to a system like Spotlight, where all that metadata is automatically populated and automatically kept up-to-date.
And if a user don't want that? Or if a user wants automatic updates for a select number of different file formats?
And if the separate metadata used by Spotlight crashes or gets corrupt?

Apparently, Spotlight has the advantage of being userfriendly with eyecandy whereas BeOS/ZETA has the advantage of the journaled filesystem and the architectual solution that has by far the most flexible approach, that paves the way open for anything you can do with a attributed filesystem.

YellowTAB should not make the mistake of providing solutions that only affect the outcome (Ten's of years of non-improving filesystems). They should not look to other OS-es but build on the immense strength of the BeOS architecture, but when providing interfaces to the user they should ask themselves what a user wants and how a user wants to invoke a function intuitively and speedily.
Strength comes from the inside, i.e. core design of an OS.

And if a user don't want that? Or if a user wants automatic updates for a select number of different file formats?

Why would a user not want this? Are you telling me that users will say "yes, please let my indexed metadata becomes out of date?" Would then, for example, recommend that users can say "hey filesystem, don't ever update the 'modified' date for .doc files!"


And if the separate metadata used by Spotlight crashes or gets corrupt?

...and what if the data in the BFS filesystem gets corrupt? In both systems, you're storing the data somewhere and using it in the same fashion. Both therefore have a similar chance for corruption. The nice thing about a system like Spotlight is that the database could be rebuilt. If you had corruption within BFS itself, all your data would most likely be hosed.


Apparently, Spotlight has the advantage of being userfriendly with eyecandy whereas BeOS/ZETA has the advantage of the journaled filesystem and the architectual solution that has by far the most flexible approach, that paves the way open for anything you can do with a attributed filesystem.

I don't think you really know what you're talking about. For one, Mac OS X's filesystem, HFS+, is (now) both Journaled and supports extended attributes. So BFS has no "architectural" advantages, and can do nothing as an attributed filesystem that Spotlight cannot. (Yes, Spotlight can query the extended attributes, just as BFS can).

Now, as for "most flexible", this is where your point starts to get really absurd. With either Mac OS X or with BeOS, files can have extended attributes. How is BeOS more flexible? With Mac OS X, any (non-network) filesystem can be searched. With BeOS, you can only search BFS partitions. How is BeOS more flexible? With Mac OS X, any attribute -- or content -- can be searched. With BeOS, you can only search based on user-indexed attributes and cannot search based on content. How is BeOS more flexible? With Mac OS X, metadata is automatically populated and automatically synced. With BeOS, it's left up to the user. How is BeOS more flexible?


Strength comes from the inside, i.e. core design of an OS.

Agreed, and you have made it abundantly clear that you are rather ignorant about the architecture of completing products.

And even if it took WinXP 2minutes to boot it won't change anything. Now that most OSs are so stable you're only supposed to boot one time... It doesn't change anything if it takes 2minutes or 15seconds.

Thing is, I have a life, which means that I don't *want* to let my computers running all the time. It eats power, which costs money and is bad for the environment. At night, my computers go off. When I'm away from home for too long, they go off.

100% of the people I know do the same.

Thing is, I have a life, which means that I don't *want* to let my computers running all the time. It eats power, which costs money and is bad for the environment. At night, my computers go off. When I'm away from home for too long, they go off.

100% of the people I know do the same.

Funny, I have a life too, which is precisely why I leave the machine on. All my applications and documents are open so that I can return to what I was doing instantaneously, without waiting for the machine to boot up and to relaunch applications. A properly designed machine with a modern OS in "sleep" mode eats as much power as your DVD player plugged in and not doing anything else. And as for anecdotal evidence, so what? 100% of the people I know leave their machines on and compete for longest uptime.

ACH! I'm sorry, my comment (above) should have been a reply, not a new comment.

Absolutely

I also add that a faster boot OS indicates an OS much closer to version 1. That could also mean no features yet or no bloat.

I see W2K boots on a pair of different xp2400s takes what seems several minutes (web pc) vs maybe 1m for disconnected. The slow machine uses a crappy old 4G drive which probaly slows things down a bit.

I still hope one day OS writers will get to a 1 sec boot from the HD spin up time.

These are a bit older, but nicely outline some of BeOS' special features:
http://www.birdhouse.org/beos/byte/