• Welcome to OS2World OLD-STATIC-BACKUP Forum.
 

News:

This is an old OS2World backup forum for reference only. IT IS READ ONLY!!!

If you need help with OS/2 - eComStation visit http://www.os2world.com/forum

Main Menu

OS/2 Vs. Windows... The Debate Rages On

Started by S.SubZero, 2007.11.02, 08:49:11

Previous topic - Next topic

El Vato

...Humm, there are no comments on the GUI zip utilities that, as far back as 1996 existed for OS/2.   Nor were there any comments regarding the alternative GUI utility resource that our moderator provided. 

And that brings me to a  necessary question ...Are you here for help or to provide any constructive criticism on the OS/2?  Or are you here merely to find out how many OS/2ers can you take to Zarathustra's cave ...er, that pond whose "wetness" property you can not seem to let go?

Additionally, I did not see any more comparisons to Solaris and/or Linux on such "important" applications as Flash and games –notwithstanding,  the current wedge against the OS/2 remains.

I thought that you finally had acknowledged the fact that if there are different operating systems, it is because some of us do not regard your operating system as the standard measure of computing.  Yet by your insistence on focusing on shallow metrics for arriving at your "data" you continue to miss that essential point: OS/2 is not (nor needs to be) like WinXX for us to use it.

For some of us, the Flash and movie stuff is irrelevant for we engage in real computing activities; notwithstanding for others, "clunky" or not, the OS/2 will get the job done for that multimedia stuff.

Quote from: S.SubZero on 2007.11.09, 00:13:24
Quote from: El Vato on 2007.11.08, 09:16:35
Let me summarize the argument:  If you are no "doctor," and in your own words "an average Joe," eCS is for you.  Otherwise the evaluation against current professionally developed/supported operating systems is, by logical necessity, walking on stilts.
First off, ECS would be nice to try, but the Live CD looks brutally limited and dated (Firefox 1.0.7?).

Secondly, Serenity's site lists nothing I either can't get free for OS/2 4.52 (from Hobbes) or otherwise simply don't need.  Xworkplace?  Check.  AE Editor?  Check.  InnoTek stuff?  Check.  Warpin?  Check.  ISDNPM?  Er..   ECS is OS/2 4.52 with fluff.  *My* 4.52 install is a newer service level all around than ECS 1.2R is.  I'd be curious to see what ECS 1.2R would be without that C:\ECS directory.  OMG THE THEMES!  :rolls eyes:

As an "average Joe" that is the only feature that you would be able to understand.   For users like you, coming from WinXX and placing so much emphasis on Flash and movies, eCS marketing arm assumes that a pretty GUI is what you want –why do not you let them know otherwise?   Needless to say, explaining the professionally developed ...humm "backend" --as you referred to it-- features would be out of your understanding.  Why would marketing engage engineering to dwell on those for someone who, evidently, up to this point in the thread has been merely regurgitating his/her prior babblings???

Quote from: S.SubZero on 2007.11.09, 00:13:24
  If I'm ignorant about this please correct me, as I am reading off Serenity's site right now and really, I see nothing of interest.  I especially see nothing warranting dedicating an entire PC to ECS, and ECS will not install in Virtualbox (resource.sys trap at CD boot, which ALSO happens on real hardware).

As I see the biggest impediment to an unbiased evaluation of the OS/2 are your MS induced notions.  First was the virtualization "instability" that for some reason I, or as a matter of fact others, are not experiencing in actual hardware.  Subsequently came the issue of the "expensiveness" of the eCS;  then it proceeded to issues of "average uses" of the OS/2 that –naturally-- do not measure up to your implicit "wetness" property of your pond (and what some have referred to as group mentality) of WinXX.

Accordingly, this is becoming sort of an concocted attack on OS/2, veiled under the guise of a purported "unbiased" evaluation by an purportedly former OS/2er that became indoctrinated into another environment. 

I can not help but be reminded of that certain enthusiast propagator of a certain religion who, not ready to convert others due to his engagement in games ...er, human pleasures, fell into the grasp of the Borg collective ...er, "savage" religion (whatever that means).   His psyche was re-educated to turn against his former brotherhood, and in the process the convert disparaged those former ways while exalting those of his master.
   
(Albert Camus, the atheist French Existentialist actually made that observation on that familiar human pattern of those who fall short on individuality.  Camus subsequently wrote an literary essay to illustrate that peculiar trait of collectively induced behaviour in The Renegade, one of six short works in his anthology, Exile in the Kingdom.)

Quote from: S.SubZero on 2007.11.09, 00:13:24
Quote from: El Vato on 2007.11.08, 09:16:35
In an experiment (especially if the intention is to arrive at an unbiased conclusion) the parameters are applied uniformly across the domain objects under consideration --else the conclusions that are arrived at simply reflect the subjectivity (narrow and familiar notions) of the experimenter/observer.
I draw most of my data from comparisons with XP, which came out around the same time as OS/2 4.52 (late 2001).  Really, XP came out in October and OS/2 4.52 in December.  Call 4.52 a maintenance release if you like, MS didn't really consider XP a big deal either... Win2K is NT5, XP is NT5.1.  A double irony, OS/2 went up over half a version number, while XP went up .1.

OS/2 W4's base is from around 1996-1997.  It was subsequently enhanced with the fixpaks and an stable and cutting edge Java component to serve the large enterprise deployments in existence under support contracts.  Needless to remark, the OS/2 aim never shifted to the consumer.
 
Obviously,  and even if written in your own words, you still do not grasp the concept that OS/2's aim has been different than WinXX.  The latter's code base, as judged from the behaviour of the innumerable members of that "family" has not remained consistent but actually half done hacks to focus on the marketing aims of each subsequent release.  Why do you think that I referred to them as in alpha and beta development releases???

Quote from: S.SubZero on 2007.11.09, 00:13:24
Quote from: El Vato on 2007.11.08, 09:16:35
That is not the main focus that you emphasized when you responded to our friend who was doing her/his OS/2 research.  On the other hand, I believe that some of the metrics under discussion were addressed in this thread.  "Primitive" or not, look around in the forum and you will find people doing all of those "average" tasks --and more.
I said they were tasks that were doable (to an extent) but they are clunky and complicated to set up.  Flash for Windows can be done in about 4 mouse clicks.  Flash for OS/2 (of a version that can play Youtube) involves consideraly more work.  Mind you, for that you get exactly a Flash that can play Youtube videos.  Any other video-related site (like cnn.com) would not work.

I, for one, do not have a need (or desire, as a matter of fact) to use that conservative resource.  I subscribe to the New York times and from time to time read articles from the Los Angeles Times.  My needs are not your needs; my choice of operating system is simply one more way to emphasize that --I thought that you had mastered that which was written in your own words in a certain prior post.

Quote from: S.SubZero on 2007.11.09, 00:13:24
Quote from: El Vato on 2007.11.08, 09:16:35
Perhaps if you had began your incursion into the forum with an unbiased and constructive approach, the thread would have taken a totally different twist.
I answered a question.  It was unbiased,

If this is your notion of "unbiasedness" when you pretend to compare the performance of an operating system  --in a virtualized setting-- against that of an equivalent "front end" marketing refocused equivalent, then you need to go back and reexamine the definition of "unbiasedness."  An appropriate and "unbiased" approach is required of anyone who is responsible for, say, the review of candidates' resumes for an competing position.

Seen from that perspective, it is no wonder then that the outsourcing of vital engineering processes continues to be sent to "developing" nations since those more serious local candidates "need" to engage in "gaming" to be considered "adequate."

Quote from: S.SubZero on 2007.11.09, 00:13:24
as I never said what OS I was using (for the record, I was in OS/2 in a VM under Vista x64).  I also never tried to directly compare OS/2 to any new Windows (I mentioned Windows 2000).

Quote from: El Vato on 2007.11.08, 09:16:35
Since you are at it, you may want to research more thoroughly (certainly, placing your baggage of assumptions stacked on the side --next to those "resumes") the conclusions that you arrived at on the domain of virtualization technology --and how the different approaches affect the virtualized operating system.  Those continue to escape your grasp.
Please list the conclusions I came to.

"Know thyself, first," attributed to Socrates, but actually in the general knowledge of many more ancient cultures than the Greek would better address that request. Review what you have posted and ...

Quote from: S.SubZero on 2007.11.09, 00:13:24
  I'm not approaching virtualization from the back-end, I'm approaching it from the front-end.  One of the biggest issues I think OS/2 faced was people on the back-end not understanding the people on the front-end.  Microsoft understood the people on the front-end, at a time when it was critical to do so.

"Front, back, right/left side, top/bottom side,"  whatever you want to call it in your "average Joe" jargon.  It does not relieve you of the responsibility for knowing its implications "before" babbling out something that simply fits your preconceived notions.  I could, out of ignorance and self-satisfaction like yourself, talk garbage about the performance of WinXX when virtualized, say under open source Xen, like in this screen shot (< http://www.chingonazo.com/xenFedora7_64-bitWinXppStart.png >) --but I know better.

For starters, the first thing that one would notice under a WinXX virtualized instance would be the the crappy driver issues –that has been addressed by XenSource and Novell in their subscription based offering, with Novell offering its Virtualization Pack under SuSE Linux Enterprise Desktop (SLED).

Indeed, I will tell you that which you are not perceiving:

You do not grasp the notion that unless the source code of the operating system kernel under consideration is open and modifiable to run in paravirtualized mode, it will "necessarily" trail substantially behind its natively hardware installed counterpart in performance and stability. 

You do not apprehend that the performance of an virtualized operating system is necessarily slower and unstable than if it were performing natively in actual hardware --especially when quirks like the WinXX drivers issue mentioned above are not taken care of.  When an operating system is resource rich, those problems will be addressed --as XenSource and Novell did.  On the other hand, if there is next to nothing support for that operating system --like is the case with OS/2-- well, ignorance or prudence will drive the conclusions drawn on the performance.

By what percentage is the performance affected?  Probably would be the natural question. 

That depends on the quality and maturity of the virtualization product –when MS Viridian comes out, it will not support live migrations, for instance.  Notwithstanding, (other) proprietary solutions from VmWare as well as those from open source alternatives like XenSource and Virtual Iron, as well as some based on KVM, do support live migrations of virtual machines now.

That also depends on whether the operating system has been fully virtualized with hardware (CPU) virtualization extensions.

That depends on whether the operating system has been virtualized in full software virtualization mode.

And that depends on whether the operating system has been paravirtualized (read closer to native installation performance) –which OS/2 is certainly not being done under Virtual Box and/or Virtual PC.  Accordingly, any conclusions of yours that the OS/2 performance under virtualization must reflect its performance on an native hardware installation is, by logical necessity, a shallow assumption and walks on stilts.

Against the backdrop of all of the above, could I ask you a favour???  Please, re-evaluate your reasons for visiting this "oasis" of the OS/2.  And if you can not be constructive (we need constructive input) nor seek advice on your "average" uses of OS/2 (we love to help out around here), then this is possibly not the right place for you.  OS/2 is not WinXX; OS/2 is not Solaris and/or Linux.  In short OS/2 does not have the capital nor human (or both) resources as those other alternatives and the comparison with an resource rich operating system is by default specious.

With all due respect, this discussion has become as redundant (and irritating) as answering the (reformatted at best) questions of a little 10 year old kid over and over who lacks the capacity  (or does not want) to acknowledge a basic concept; her/his voluntary instructor is in effect sending her/him back to kindergartner for her/him to learn the very basics as this last post in the thread is finished, capisci ??? 

(Door closes and El Vato is gone to México in his 1968 Plymouth Fury III)

lwriemen

Quote from: El Vato on 2007.11.09, 14:56:28
And that brings me to a  necessary question ...Are you here for help or to provide any constructive criticism on the OS/2?  Or are you here merely to find out how many OS/2ers can you take to Zarathustra's cave ...er, that pond whose "wetness" property you can not seem to let go?
The statement, "I do miss the advocacy group", may be a telling factor.

RobertM

#62
SubZero,

Here are some points contrary to your "argument" against OS/2 and eCS:

  • My Netfinity M10 7000 server has been running since the day we purchased it to now, with NO crashes and only the occasional reboot to install/activate a new service. The machine is used both for develpoment work AND for running continuous use HTTPD, FTPD, NetBIOS Domain Control, MySQLD, NTPD and other services. It proves the stability of OS/2 (Warp Server for e-Business CP2 PF, NO fixpacks, NO updates) - even under high use (it currently serves almost a dozen of our and our clients' sites, including much dynamically generated content requiring the seamless use of various running services including the REXX subsystem, various add-on DLLs, various image manipulation tools, DLLs, scripts and programs, MySQLD, the FTPD server and Domino GoWebserver to complete a single request.
  • All of our EMS Database Software (which is web driven to the clients' WinXP workstations) is running on eCS v1.2MR workstations (with no fixes or updates applied). To date, one machine has crashed (due to a faulty memory module), yet all get near constant daily usage that includes the eCS machines running and seamlessly using the REXX subsystem, various add-on DLLs, various image manipulation tools, DLLs, scripts and programs, MySQLD, the FTPD server and Web2. Some of these systems have been running for over 2 years straight. Most are running for a year straight and the remaining two hit their one year anniversary in December and March respectively.
  • OS/2 (and eCS) really shine in numerous areas when an OS/2 native program (or even a Linux port) is used. For instance, many programs and certain functionality under Windows (XP/Vista/Take Your Pick) will either when loading, or during use; utilize a large amount of system resources, degrade either the responsiveness or usability of the Windows system (or a combination of those). Prime examples are any version of Acrobat Reader in comparison to Lucide (which is orders of magnitude faster); media handling (XP or Vista require loading a resource intensive, very large, monolithic player to provide the same functionality to do simple tasks such as playing MP3 files or video - in comparison to integrated "Playlist Folders" on eCS or various ported media players such as KMP, mPlayer and ffMPEG). While some of those dont have the "fancy GUI" that Media Player has, I can assure you, from having programmed for years, the bloated aspect of Media Player and the greater CPU and memory resources it consumes have nothing to do with the fancy GUI.
  • OS/2's and eCS's threading model and multitasking capabilities still surpasses that of Windows (when running native apps or ported Linux apps), and does a far better job of thread and process scheduling, making the system feel more responsive overall - even in instances where a high amount of swap file usage is required (such as when I do audio editting with Ceres Sound Studio and it needs to swap out half a gigabyte worth of temp files on a machine that has only 300MB of RAM). As a matter of fact, its thread scheduling and handling model is so good, that even the Linux community wanted access to those parts of the OS/2 kernel to try to duplicate it, and then later decided (thanks to an inability to ever see that code) to try to re-implement their threading model to achieve such performance - with the help of IBM I might add.
  • OS/2's firewall is REAL, and actually does what you tell it to. The "firewall" in XP is a joke (doesnt do proper traffic filtering, and doesnt allow you to filter certain types of traffic - especially if it is destined to go to Microsoft), and the "firewall" in Vista is only slightly better.
  • Virtually any OS/2 service (media, folders, the WPS, kernel, file system parameters, the firewall, networking features and components, hardware and driver control) can be controlled or modified in a plethora of methods, including the new methods and settings boxes included with eCS (or added to OS/2 via xWorkplace). Virtually all of these can also be controlled from the commandline, or via custom apps, via REXX (directlly through REXX calls or through a REXX script that modifies the parameters in an ini file). It thus allows an open ended method of controlling or dealing with virtually any aspect of the operating system... (ie: "Do it your way, or our way, or any way you want" - instead of the MS mentality of "do it our way or not at all - or hope someone writes a hack so you can do it their way instead of ours")
  • OS/2 scales FAR better than Windows (and even Linux) on multiple CPUs. Today with virtually every CPU coming with 2 or more cores, that (to me) is a major issue. If I have mutliple CPUs (or cores) in my system, I expect my OS to use them. On our Netfinity (Quad Xeon system), I've done load testing that requires the concurrent use of the REXX subsystem, MySQLD and Domino GoWebserver, where the system, both before it reaches its processing power limit (at over 30 connections served a second) and after it reaches that limit, will successfully utilize OVER 90% of each CPU's processing power - while you will be lucky if you manage 70-80% utilization on an XP or Vista machine, thus decreasing potential or expected performance when under load.
  • Because of the way OS/2 is written, as well as because of the way OS/2 handles processing system calls from other apps, even single threaded apps can utilize multiple CPUs for a performance increase on an OS/2 machine (since many calls done by any app will trigger a call to an SMP capable subsystem in OS/2, that single thread app suddenly is (unbeknownst to it) using more than one CPU - all with OS/2 transparently handling such functions in the background). On the other hand, you need an SMP capable, SMP aware app to benefit from SMP on XP and Vista - or need to be running multiple apps at the same time (and then still lose expected/theoretical performance since Windows does not handle SMP (ie: scale) nearly as well as OS/2 does).
  • Still on the SMP front, OS/2 SMP (Warp Server Advanced, WSeB, eCS v2, eCS v1 with SMP) has ALWAYS supported up to 64 CPUs, while MS struggled (and was VERY late) to get their 2 then 4 then 8 CPU support out the door, and needed IBM's help to manage 32 and the (I've never seen, but MS has claimed) 64 CPU support. As a matter of fact, many of the bugs that were fixed in XP were done by IBM - you can search online and you will find a few articles that point out that fact... and it cost MS a pretty penny to get IBM to help fix bugs that apparently MS themselves couldnt fix. That trend extends to even beyond more complex things like the OS... all the way down to extremely simple things such as many years worth of mouse drivers, for MS's own mouse hardware, where if you check the driver info itself, you will find that IBM was needed to finish and/or fix the drivers, with as recent or more recent copyright notices than MS's own copyright claims on the driver, that are in IBM's name.

Windows isnt superior... it's just more prevalent. And I have yet to find it more stable than a properly configured WSeB or eCS system on good hardware.


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


Terry

#63
Even without the in-depth "under the hood" experience of some in the the OS/2 community, it is a privilege to share the long-standing operating system stability track record that is OS/2.

Some of us eCS'ers are end-users who did no more than actually take the time to understand what OS/2 could do over ten Micro$oft kernels in the past => meaning the days of OS/2 v2.1 and DOS/Windows v3.1 dual-boot partitioned on a state-of-the-art 2 gig hard drive, and then Windows 95 replacing Windows 3.1 while hogging more of the hard drive's real estate.

There is an extremely easy test to determine the capabilities of the latest Windows Vista Home Premium vs. the latest eComStation by using the same program on the same hardware setup without running either operating system inside virtualization but a dual-boot scenario:

Load Lotus Word Pro and print just a one-page document, immediately Load Lotus Word Pro again and print that same one-page document, while both documents are printing load Lotus Word Pro a third time on top of the first two Lotus Word Pro loads and print that same one-page document once more, and keep doing just that again and again and again...

...now, which operating system will be left standing before one of them has a system crash?

We all know what the answer to that simple question will be without ever performing that simple test.  It isn't "rocket science" to recognize the how an operating system will perform.

Saijin_Naib

#64
No we dont, and I feel that you are heavily baised in favor of OS/2. Windows, whether you feel the need to admit it or not, has come a long way since the days of Win 95/98. That however, is simply not something that is to be appreciated in this forum seeing as its an OS/2 eCS community forum. However, if we are to tout the virtue of being unbiased, I do believe it should go the other way as well. When I had a w4 machine, I definately prefered it to the windows machines my friends had, though at the time, I did not understand what the difference was. After many years I have found that Windows 2k and XP (2k more so) are incredibly stable, provided that the user KNOWS what they are doing and how to handle the operating system. If you have read any of my previous posts, you will see that I have admitted to a very poor track run with eCS owing mostly to my unfamiliarity with it. However, when I say that one is more stable in certain applications or faster than another, I do so by providing RESULTS from tests that I perform. You sir, however, do NOT, indicating a high degree of bias and fan-boydom which frankly, I find a bit disconcerting considering you are not some impetuous child, like I could rightly be considered. But in any case, you are free to have your opinion, but I would prefer if you could perform the test you have outlined and please, post the results here so that we may see them. I am a very emphaitc supporter of the value of knowledge gained through scientific pursuits.
/rant

RobertM

Saijin_Naib,

I presume you are commenting on Terry's post. On a technical standpoint, I can discuss it (and if someone finds me copies of both the Windows and OS/2 version of Lotus Word Pro, I'll run the test).

WinVista has the advantage of allowing a similar (and in some cases greater) memory addressing space - but it also has the major deficiencies of wasting a LOT of resources (still, to this day) to do so, and a very poor threading model, and very poor SMP support. While Vista will support SMP on a few more machines than eCS (which is changing with the new ACPI tools, which should keep it in FAVOR of eCS - which will also run on numerous specialized SMP servers that Vista WONT run on), Vista does not make use of the additional cores or CPUs for any single app running. Hence, lists of SMP enabled apps are all over the Internet (here's one: http://hardforum.com/archive/index.php/t-854892.html). OS/2, as I mentioned, doesnt have that problem. That alone means that any such test like Terry outlined can benefit, because the thread scheduler can move threads from CPU to CPU, which in many cases can reduce system bottlenecks (tough to explain how in non-technical terms, but I can try if you would like).

Also, back to the resources issue... you can duplicate the resource issue by loading lots of copies of anything on Windows, track your memory usage (in something better and more accurate than the tools that come with it), and see what it frees up when done, and how much it is using for each copy. Native OS/2 apps, and well written ports consume less resources and (usually) free their resources properly when exiting, so a similar test with OS/2 will show a far different resource usage.

As a matter of fact, another wonderful example is doing something that incurs a big swap penalty on both OS's... for instance load a __________ (or bunch of ___________) that grows your swap file to a few HUNDRED megabytes, or even a gigabyte... (1) OS/2 will usually be more responsive, and (2) when the programs (or whatever) are closed, OS/2 will free the physical ram footprint they were using, then swap back anything else *as needed*, AND then shrink the swap file to it's normal size - all in idle cycles. That increases overall performance when freeing allocated RAM, over the Windows method which generally results in wasted swap space, and incomplete clearing of the no longer used memory footprint (meaning you also get more of those resources back - both RAM and disk).

So, there is a test you can run for yourself... as a matter of fact, you can download the trial of PMView for both OS's and try it... open tons of them... open tons of large images in them each... close them, see what happens before, during and after the test...

After reading your post, I do think it best if you try the test yourself. Try it on both a single CPU and on an SMP machine... either way, OS/2 will win... hands down.



|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


os2monkey

In the words of my man Rodney King:
"People, I just want to say, you know, can we all get along? Can we get along?"

Thank you.

S.SubZero

Quote from: El Vato on 2007.11.09, 14:56:28
...Humm, there are no comments on the GUI zip utilities that, as far back as 1996 existed for OS/2.   Nor were there any comments regarding the alternative GUI utility resource that our moderator provided. 
If you want a comment I'll give you one.  I couldn't find ezip in any format except a zipped archive.  That's the one in my original post.  While the interface is considerably nicer than the command line, it still has moments of "I should be able to do this, why can't I?" to it.  It just feels like it takes more to do things than it should.

Quote from: El Vato on 2007.11.09, 14:56:28
Additionally, I did not see any more comparisons to Solaris and/or Linux on such "important" applications as Flash and games –notwithstanding,  the current wedge against the OS/2 remains.
I saw no further comment as necessary.  I'm not going to beat a dead horse.  Games?  Neither Solaris nor OS/2 is going to win this one, tho Solaris comes with more ways to spend idle time.  Flash is more important than you may realize, given that many websites are beginning to assume the user has it.  One only needs to spend time in a browser that can't do Flash to realize how much the web has adopted it, even sites that don't necessarily need it.

Quote from: El Vato on 2007.11.09, 14:56:28
I, for one, do not have a need (or desire, as a matter of fact) to use that conservative resource.  I subscribe to the New York times and from time to time read articles from the Los Angeles Times.  My needs are not your needs; my choice of operating system is simply one more way to emphasize that --I thought that you had mastered that which was written in your own words in a certain prior post.
Did you know the front page of nytimes.com has flash embedded in it?
Did you know the front page of latimes.com has flash embedded in it?

Quote from: El Vato on 2007.11.09, 14:56:28
If this is your notion of "unbiasedness" when you pretend to compare the performance of an operating system  --in a virtualized setting-- against that of an equivalent "front end" marketing refocused equivalent, then you need to go back and reexamine the definition of "unbiasedness."  An appropriate and "unbiased" approach is required of anyone who is responsible for, say, the review of candidates' resumes for an competing position.
Nothing I talked about was affected by running the OS in a virtual machine.  A dated desktop is still a dated desktop.  The SIQ problem is still the SIQ problem, one that has plagued OS/2 since long before anyone ran it in a virtual machine.  The app problem was around since long before Virtualbox was a gleam in a programmer's eye.  Again you speak of "performance".. In what context?

Quote from: El Vato on 2007.11.09, 14:56:28
You do not grasp the notion that unless the source code of the operating system kernel under consideration is open and modifiable to run in paravirtualized mode, it will "necessarily" trail substantially behind its natively hardware installed counterpart in performance and stability.
You throw around this "performance" word, and claim that I have some performance comment that is negative.  I actually said that OS/2 runs *fast* under Virtualbox.  It runs fast!  OS/2 runs fast under Virtualbox.  Under the Virtualbox, OS/2 runs fast!  Fast, OS/2 runs, Virtualbox using.  OS/2 R TEH FAST IN VB!  How many more times do I need to say it? ; ;

I did not make any attempt to blame any virtualization shortcoming on OS/2.  Stop trying to make it sound like I did.

RobertM

#68
Quote from: S.SubZero on 2007.11.10, 09:31:49
Quote from: El Vato on 2007.11.09, 14:56:28
...Humm, there are no comments on the GUI zip utilities that, as far back as 1996 existed for OS/2.   Nor were there any comments regarding the alternative GUI utility resource that our moderator provided.
If you want a comment I'll give you one.  I couldn't find ezip in any format except a zipped archive.  That's the one in my original post.  While the interface is considerably nicer than the command line, it still has moments of "I should be able to do this, why can't I?" to it.  It just feels like it takes more to do things than it should.

eZip is a Zip Front End - since you thus need ZIP/UNZIP to make it work (which already come on eCS), installing the package is easy. Download it to some folder, right click on the folder, select "Open Command Line", type "unzip ezip.zip" (or whatever the name is), follow the installation instructions. And if you are running OS/2 Warp (which doesnt come with it), simply saying "Hey gang, I downloaded eZip (and forgot to read the website where it says "You need ZIP/UNZIP... get it here"), how do I unzip it?" - we would have responded "Hi S.SubZero, go to Hobbes, and type in UNZIP, and grab the latest one. It's a self extracting archive... put it in your OS2\DLL directory and run it, or make a directory for it and run it there - then add the directory to your path and dpath statement" (not mentioning that you missed that section on their website - because we all make mistakes - I've done ones just like that).

As for moments of "I should be able to do..." I have no clue what you are talking about. You can very easily do whatever with it, and it has a plethora more capabilities and options than the built in Windows counterpart. The ONLY drawback is in trying to drag and drop large amounts of files from a zip container - which can simply be done through the "Unzip to..." option, without the delay. Modifying the behavior (to avoid that delay) is very easy, but I dont have VisPro REXX (or I would do it).


Quote from: S.SubZero on 2007.11.10, 09:31:49
Quote from: El Vato on 2007.11.09, 14:56:28
Additionally, I did not see any more comparisons to Solaris and/or Linux on such "important" applications as Flash and games –notwithstanding,  the current wedge against the OS/2 remains.
I saw no further comment as necessary.  I'm not going to beat a dead horse.  Games?  Neither Solaris nor OS/2 is going to win this one, tho Solaris comes with more ways to spend idle time.  Flash is more important than you may realize, given that many websites are beginning to assume the user has it.  One only needs to spend time in a browser that can't do Flash to realize how much the web has adopted it, even sites that don't necessarily need it.

And is being taken care of by Paul Smedley with the Gnash port. That will give us a highly reliable, quick, not as bloated alternative - and like all of Paul's work, should (a) work well from the get-go, or (b) will be tweaked or fixed by him as people point out issues with it. More likely given Paul's track record it will be (a)... but in either event, Paul is always very quick to respond to both personal emails and forum posts about the software he ports - and has always been very helpful.

So, you are right, eCS does need an updated Flash... but that is soon to be a moot point.


Quote from: S.SubZero on 2007.11.10, 09:31:49
Quote from: El Vato on 2007.11.09, 14:56:28
I, for one, do not have a need (or desire, as a matter of fact) to use that conservative resource.  I subscribe to the New York times and from time to time read articles from the Los Angeles Times.  My needs are not your needs; my choice of operating system is simply one more way to emphasize that --I thought that you had mastered that which was written in your own words in a certain prior post.
Did you know the front page of nytimes.com has flash embedded in it?
Did you know the front page of latimes.com has flash embedded in it?

Many Flash sites work properly under eCS... some that check for Flash 9 or 8 dont, and some that require it but dont check will either work or be quirky... but again, read my comment above.


Quote from: S.SubZero on 2007.11.10, 09:31:49
Quote from: El Vato on 2007.11.09, 14:56:28
If this is your notion of "unbiasedness" when you pretend to compare the performance of an operating system  --in a virtualized setting-- against that of an equivalent "front end" marketing refocused equivalent, then you need to go back and reexamine the definition of "unbiasedness."  An appropriate and "unbiased" approach is required of anyone who is responsible for, say, the review of candidates' resumes for an competing position.
Nothing I talked about was affected by running the OS in a virtual machine.  A dated desktop is still a dated desktop.  The SIQ problem is still the SIQ problem, one that has plagued OS/2 since long before anyone ran it in a virtual machine.  The app problem was around since long before Virtualbox was a gleam in a programmer's eye.  Again you speak of "performance".. In what context?

EVERYTHING you talked about is affected by running OS/2 or eCS in a virtualized environment. If you read some of the forums elsewhere, and posts/announcements by companies who have worked on virtualization software for OS/2, you will see that making a virtualized environment that even RUNS OS/2 is difficult (check out Bosche for one, which goes into technical detail about the issue).

Heck, some companies, citing how difficult it was, gave up entirely on getting their virtualization software to run OS/2.

Nonetheless, eCS does NOT perform as well, or as stable in a virtualized environment, especially because its thread scheduler and the method it uses memory is vastly different than Windows... that is definitely a cause and certainty for PERFORMANCE and STABILITY problems in a virtualized environment.

Now, as for the app problem (I am presuming you mean a lack of the games you want), I dont see how lack of apps (games or otherwise) is a PERFORMANCE problem - so I dont know why you bring that up in your rebuttal to El Vato. Makes no sense.


Quote from: S.SubZero on 2007.11.10, 09:31:49
Quote from: El Vato on 2007.11.09, 14:56:28
You do not grasp the notion that unless the source code of the operating system kernel under consideration is open and modifiable to run in paravirtualized mode, it will "necessarily" trail substantially behind its natively hardware installed counterpart in performance and stability.
You throw around this "performance" word, and claim that I have some performance comment that is negative.  I actually said that OS/2 runs *fast* under Virtualbox.  It runs fast!  OS/2 runs fast under Virtualbox.  Under the Virtualbox, OS/2 runs fast!  Fast, OS/2 runs, Virtualbox using.  OS/2 R TEH FAST IN VB!  How many more times do I need to say it? ; ;

I did not make any attempt to blame any virtualization shortcoming on OS/2.  Stop trying to make it sound like I did.

No... there is the problem... you run OS/2 *in* a virtualized environment, complain about it's speed and it's stability, but fail to realize that it could entirely be caused by the virtualized environment - and not the OS itself. You keep ignoring the points at hand. We have been trying to tell you that many of the performance and stability issues you claim OS/2 has is DUE TO the virtualized environment you run it under. You keep ignoring and sidestepping each comment. Now you are claiming it runs FAST in VirtualBox... which is counter to your claim it runs slow in your tests... which is it? Regardless, many of the claims and complaints you have had can be traced right back to the fact that it is running IN VirtualBox instead of on real hardware. My Netfinity and the 6 eCS servers I have installed for clients prove that. Others here have posted their experiences as well - which are also contrary to yours.

Yes, OS/2 isnt perfect, but other than the Flash issue (being resolved), mentioning the SIQ issue (which since Warp 4 hasnt been too big of an issue, and since a later fixpack and the inclusion of CAD with eCS has barely been an issue) - other than that, you have yet to claim anything broken or worse about OS/2 over Windows - other than it doesnt have a plethora of games or big name software. And each of your claims still ends up indirectly or directly claiming that those "problems" are performance and stability related (because you seem to like coming full circle back to that).

I really dont understand your point here... any of them. I dont know why you are even here, or still using eCS.

There was a guy in the newsgroups years back, who sounded just like you. He claimed to be (in his own words) an avid OS/2 fan and user, and used that as the cornerstone for wild, erroneous complaints about OS/2. You keep ignoring almost every part of any post that brings up a counterpoint... either by totally skipping it in your replies, or by answering it with something totally unrelated or totally wrong. There are times I truly wonder if that person is you...



Regardless, here's the oddest thing about this whole thing. You seem to have a ton of complaints about eCS/Warp... yet continue to use it - FINE. No one is stopping you, and I dont in ANY way begrudge you that choice. The part that makes no sense is this... this forum, outside of IBM and some of the gang at Serenity, probably has the biggest collection of OS/2 experts on the planet.

Instead of continuous rants about how "OS/2 does this poorly", "OS/2 crashes when I do this", "OS/2 needs this", "I cant figure out how to do this (because OS/2 isnt intuitive)"... why dont you simply find the right forum section and start a topic saying "Hey, how do I do this on OS/2?" or "OS/2 is crashing when I do this? Can anyone help me fix that?"


Look around on the forums and you will see what I mean in those last two paragraphs. Heck, dig through my post history for that matter... I for one, after using OS/2 *CONTINUOUSLY* since v2.0, am still learning about new things it can do (or new ways of doing the same things). Often, that means I need help with a specific thing... so... I post a question here... and you know what happens? A half dozen people jump in with solutions, suggestions or more questions to narrow down my exact needs to then follow up with an answer. Keep digging through my posts and you will also find I (and many others) are also on the giving end of that help and support when someone else asks a question.

Heck, on the OS/2World front page alone there is:
- Creating Java Icons on the Desktop? (answered)
- How ican join OS/2 on active direct... (answered and ongoing for more questions)
- dbExpert(?)  (answered and ongoing for more questions)
- GUI sftp client?  (answered)

Maybe many of your issues are simply because, unlike the people who started those questions (and then received answers to them), you choose to complain about the issues as opposed to asking one of us how to solve the issue. Beauty is we wont candy coat a problem with nonsense... we'll either find you an answer or say "That cant be fixed, it's an issue with this base component of OS/2 which IBM wont release the code to" (and there are posts on the forum to prove that).

Interestingly enough, we are slowly finding, through the resourcefulness of others, that numerous things we thought werent changeable without the source code actually ARE! For instance, not too long ago, numerous of us posted that getting larger icons and adding PNG support wasnt possible without the source to certain components of the WPS/PM - we were proven wrong.

But either way, maybe it's time to stop this ridiculous argument, and for you to actually choose to utilize this forum in a similar fashion to how the rest of us do. Having a problem? Have a complaint about OS/2 that you want resolved? Ask us. Unless it's something you KNOW is out of our powers to deal with like "Gee, there arent enough games... can someone fix that?" you will find that with our combined knowledge, someone here will be sure to help.

Think about that for a while... here's a great starting point:

  • Kim and I have a decent amount of experience with virtualized environments.
  • El Vato has a lot of experience with virtualized environments, and a much better understanding of them
  • Paul Smedley not just has that experience but is porting VirtualBox to OS/2
  • Numerous others here have a lot of experience with virtualization
  • All of us, individually and combined have a tremendous amount of experience with OS/2, fixing issues, making it stable, and properly installing it.

With that in mind, think how your OS/2 experience could be right now if you came to this forum and made a post like this:

Quote
Hi,

I'm running OS/2 Version ___ under VirtualBox Version ___ on an XP system with 4GB RAM (1GB assigned to the virtual session). I'm having problems with the OS/2 session locking up or becoming unresponsive. Does anyone have any suggestions, or is there any more info you may need to help me resolve this?

Thanks,
S.SubZero


I guarantee you that you would get replies and offers of help.

Think on it before you keep complaining about the issues you have (that others say is NOT the norm), without stopping to ask someone for help.

-Robert


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


S.SubZero

#69
Quote from: RobertM on 2007.11.09, 23:28:52Virtually any OS/2 service (media, folders, the WPS, kernel, file system parameters, the firewall, networking features and components, hardware and driver control) can be controlled or modified in a plethora of methods, including the new methods and settings boxes included with eCS (or added to OS/2 via xWorkplace). Virtually all of these can also be controlled from the commandline, or via custom apps, via REXX (directlly through REXX calls or through a REXX script that modifies the parameters in an ini file).
One of the tricks here is that if the solution you want doesn't already exist, the chance that it will is slim.  Creating REXX scripts for stuff is fun I'm sure, but it's not something many users really want to be troubled with, especially when other OSs don't require programming to do things.  As far as "plethora of methods" of doing things, that's currently more than the "zero" methods I have in OS/2 of controlling the volume control with UNIAud.

Quote from: RobertM on 2007.11.09, 23:28:52OS/2 scales FAR better than Windows (and even Linux) on multiple CPUs. Today with virtually every CPU coming with 2 or more cores, that (to me) is a major issue. If I have mutliple CPUs (or cores) in my system, I expect my OS to use them.
One of the clarifications I have been looking for is OS/2's ability to use multiple cores.  ECS very carefully words the ability for OS/2 to *install* on multi-core systems, but there is no mention of utilization of those cores.  Nobody I have asked knows, and even Google has proven useless.  I also think your overall statement of utilization is a bit hopeful, as I am sure it's possible to code an OS/2 app that scales to multiple CPUs or cores poorly.  There is also the fact that OS/2 is an older, less resource-intense OS that has more headroom for additional CPUs/cores to begin with.
Quote from: RobertM on 2007.11.09, 23:28:52Still on the SMP front, OS/2 SMP (Warp Server Advanced, WSeB, eCS v2, eCS v1 with SMP) has ALWAYS supported up to 64 CPUs, while MS struggled (and was VERY late) to get their 2 then 4 then 8 CPU support out the door, and needed IBM's help to manage 32 and the (I've never seen, but MS has claimed) 64 CPU support.
Well IBM is a hardware company.  It makes perfect sense for software makers to ask hardware makers how to support their hardware.  Also, while OS/2 SMP may be more scalable, Windows is more available.  Windows has also supported 2 CPUs even in the home-friendly incarnations for several years now.   

Quote from: RobertM on 2007.11.09, 23:28:52
Windows isnt superior... it's just more prevalent. And I have yet to find it more stable than a properly configured WSeB or eCS system on good hardware.
You're not looking very hard then.



S.SubZero

Quote from: RobertM on 2007.11.10, 10:26:04
EVERYTHING you talked about is affected by running OS/2 or eCS in a virtualized environment. If you read some of the forums elsewhere, and posts/announcements by companies who have worked on virtualization software for OS/2, you will see that making a virtualized environment that even RUNS OS/2 is difficult (check out Bosche for one, which goes into technical detail about the issue).
VMWare says they can't get it to work.  Nobody else appears to be having a problem.  I tried VMWare, and while they support 64-bit guests which is cool, the software is big and bulky, and doesn't like Vista x64 at all.  Virtualbox runs OS/2 4.52 fine, without the bulkiness.  Then again, all of the major players in virtualization have kind of downplayed the importance of OS/2 virtualization.  The VB guys even said the vast majority of their customer base was Windows users running Windows virtualized.  Note that Parallels exists primarily because of a need to virtualize OS/2.

Bochs can certainly run OS/2, but Bochs does things a bit different than virtualization software, it's more of an emulator. 

Quote from: RobertM on 2007.11.10, 10:26:04
Nonetheless, eCS does NOT perform as well, or as stable in a virtualized environment, especially because its thread scheduler and the method it uses memory is vastly different than Windows... that is definitely a cause and certainty for PERFORMANCE and STABILITY problems in a virtualized environment.
I think ECS runs poorly in VPC2007 because VPC2007 handles OS/2 begrudgingly and with low priority to improvement.  I think ECS fails to install on VB for the same reasons that could affect real hardware.
http://news.ecomstation.com/article.php?id=1396&group=ecomstation.support.virtualpc#1396
QuoteI have a core 2 duo PC and tried to install eCS 1.1 and eCS 2 beta 4,
but both failed with the same resource.sys trap I get when I try with
real hardware (on the same machine)

Quote from: RobertM on 2007.11.10, 10:26:04No... there is the problem... you run OS/2 *in* a virtualized environment, complain about it's speed and it's stability, but fail to realize that it could entirely be caused by the virtualized environment

I am officially stepping out of this thread because all you guys are doing now is putting words in my mouth.  I do encourage *anyone* with an unbiased, fair mind to read over every single one of my posts and quote where I said OS/2 was slow without *specifically* saying it was because of virtualization.  Mind you, once again, I will say VB RUNS OS/2 FAST.

I will not continue to be insulted like this by the users *and* moderators.

RobertM

#71
Quote from: S.SubZero on 2007.11.10, 11:41:04

I will not continue to be insulted like this by the users *and* moderators.

I have YET to insult you. Dont put words in my mouth. On the contrary, I offered to HELP you with what issues you are having. My offer still stands. Nor did I say *you* said OS/2 was slow because of virtualiztion. *I* said it was slow because of that. I definitely think you are misinterpreting my words - so at least ask me to clarify what I am saying next time. Why would I insult you and then offer my help?

So again, you are still missing my point. You say you are having certain issues... I have offered to help, and have advised you that others here would as well.

For instance, the dual core issue... have you tried the newest ACPI builds? there have been some major revisions since beta 4(?). Try RC3 and let us know. If you are still having problems, some people here have already posted tips.

At this point, I will only respond to your posts if they are a request for help or information since you seem to be misreading my posts. Perhaps english isnt your first language and I just am not making myself clear enough to make up for that? (and DONT take that as an insult - it's not - we have numerous non-native-english speaking members). I'm just trying to understand why you think *I* have insulted you. I've repeatedly asked you to respond to the same points - and you have not. You keep misreading what I say (namely most recently the virtualization issue - which I went on for paragraphs about). I want to understand why you dont understand what I am saying so I can explain it more clearly, and I want you to ask for help with your problems so we can resolve them instead of just hearing over and over again that you have them - when you shouldnt (and that isnt an attack on you either... it's just a fact... I've run into my share of problems with OS/2 on certain hardware or in certain environments... an it just wont work right - period - until the cause is found and resolved. Thus, with help, you shouldnt have the problems you are having).

So, how about we drop this thread, and you post what you are having problems with - that way we can tackle this from a more constructive position, and help you resolve them. With some of the issues you have touched on, your continued tenacity at running OS/2 is to be applauded, so let us help you focus that tenacity into getting it running stably and properly.

Robert


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


RobertM

#72
As I said, I'm ignoring the "debate" and am going to try to offer you my help... so...

Quote from: S.SubZero on 2007.11.10, 11:12:48
Quote from: RobertM on 2007.11.09, 23:28:52Virtually any OS/2 service (media, folders, the WPS, kernel, file system parameters, the firewall, networking features and components, hardware and driver control) can be controlled or modified in a plethora of methods, including the new methods and settings boxes included with eCS (or added to OS/2 via xWorkplace). Virtually all of these can also be controlled from the commandline, or via custom apps, via REXX (directlly through REXX calls or through a REXX script that modifies the parameters in an ini file).
One of the tricks here is that if the solution you want doesn't already exist, the chance that it will is slim.  Creating REXX scripts for stuff is fun I'm sure, but it's not something many users really want to be troubled with, especially when other OSs don't require programming to do things.  As far as "plethora of methods" of doing things, that's currently more than the "zero" methods I have in OS/2 of controlling the volume control with UNIAud.

You dont need to write a REXX script for the functionality I described... you COULD if you want. Or better yet, if you want certain functionality (via a program, via a pop-up menu, whatever), simply ask, and I am sure I can whip something up in my sleep. I dont expect everyone to be REXX programmers... I was just mentioning it could be done... and again, if you have a certain need, I can almost assuredly and quite willingly do it.

So, want some of that added functionality for your zip/unzip files? Let me know...

Now, as for UniAud and your volume problem, download the appropriate UniMix and/or UniMixPM. I'd grab both. One issue I have with UniAud (and certain software) is that one particular program I use, when first initialized, resets the volume, which requires me opening the mixer and changing the level from the near 0 the program sets it to, back to it's normal setting. If THAT is the problem you are having, that too is no big deal. Let me know. I already wrote a REXX script to take care of it, and would gladly send it to you. You can create a program icon for it so it resets the volume by clicking on it, or you can add it to your startup folder if your volume issues are simply that the volume is too low or high when the machine starts. So, again, let me know. I'll be glad to help.

Quote from: S.SubZero on 2007.11.10, 11:12:48
Quote from: RobertM on 2007.11.09, 23:28:52OS/2 scales FAR better than Windows (and even Linux) on multiple CPUs. Today with virtually every CPU coming with 2 or more cores, that (to me) is a major issue. If I have mutliple CPUs (or cores) in my system, I expect my OS to use them.
One of the clarifications I have been looking for is OS/2's ability to use multiple cores.  ECS very carefully words the ability for OS/2 to *install* on multi-core systems, but there is no mention of utilization of those cores.  Nobody I have asked knows, and even Google has proven useless.  I also think your overall statement of utilization is a bit hopeful, as I am sure it's possible to code an OS/2 app that scales to multiple CPUs or cores poorly.  There is also the fact that OS/2 is an older, less resource-intense OS that has more headroom for additional CPUs/cores to begin with.


Don't worry, it doesnt just install on multiple CPUs, it uses them as well. Somewhere in the eCS announcements, that fact is listed (you can check Serenity's site if you want to verify that yourself).

The list of SMP aware versions of OS/2 are as follows (all in the list will support up to a max of 64 CPUs/cores):
- Warp Server 4 *Advanced* (Standard does NOT support SMP)
- Warp Server for e-Business (ANY release - gives you the SMP option during install on SMP machines)
- eCS (any 1.x version) WITH ADD-ON SMP UPGRADE PACK ONLY
- eCS 2.x (NO add-on needed)

Now the differences in support vary as follows Warp Server 4 Advanced, WSeB and eCS 1.x with SMP Upgrade will support multi-CPU setups that meet the MP1.1 and MP1.4 specs (usually a few year old single core multi-CPU setups, like my Netfinity, and newer soon to arrive xSeries). They will also support SOME dual core CPUs (older ones that followed those specs).

eCS v2.x will/should support the latest (and presumably all) multi-core CPUs since the ACPI support is based off the same base code and specsas everything else that supports multi-core. The betas were not as advanced in that support compared to RC3 (RC3 if I recall correctly is at ACPI level 3.? while I think the betas were at v2.?)

So, there's your SMP answer... if you have any other questions about it, feel free to ask.


Quote from: S.SubZero on 2007.11.10, 11:12:48
Quote from: RobertM on 2007.11.09, 23:28:52Still on the SMP front, OS/2 SMP (Warp Server Advanced, WSeB, eCS v2, eCS v1 with SMP) has ALWAYS supported up to 64 CPUs, while MS struggled (and was VERY late) to get their 2 then 4 then 8 CPU support out the door, and needed IBM's help to manage 32 and the (I've never seen, but MS has claimed) 64 CPU support.
Well IBM is a hardware company.  It makes perfect sense for software makers to ask hardware makers how to support their hardware.  Also, while OS/2 SMP may be more scalable, Windows is more available.  Windows has also supported 2 CPUs even in the home-friendly incarnations for several years now.   

Quote from: RobertM on 2007.11.09, 23:28:52
Windows isnt superior... it's just more prevalent. And I have yet to find it more stable than a properly configured WSeB or eCS system on good hardware.
You're not looking very hard then.

21 years of being a tech... for DOS and Windows based machines, yes I am... please dont let a couple other people's negative experiences prevent you from learning how to make OS/2 as stable as it should be. My Netfinity IS used all the time... from the console/keyboard. Daily. For hours a day. As is this Thinkpad.

Yes, it's not all rosy... as I said in one of my earlier posts, my older (mostly HP, with various added crap) machine crashes a lot (which was since I mixed and matched different memory modules in it - which the OS/2 docs specifically state isnt a good plan if you want a stable machine). I live with it because that machine is slated to be replaced with the Systemax I have. My newer Systemax Intel based machine doesnt. OS/2 is pickier about hardware, and my HP compared to the Intel or my 2 IBMs proves it.

So, instead of worrying about why a bunch of people are telling you it is more stable than you claim, while a bunch of people are telling you (or posting elsewhere) that it isnt stable... lets instead work on helping make YOUR setup more stable. I think that's all that really matters in the end.

-Robert


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


melf

#73
As the question came up in the beta newsgroup at http://news.ecomstation.com/article.php?id=2058&group=ecomstation.beta#2058, I just want to show these data regarding eCS on my dualcore (LG F1 22PTV), regarding respond and bootup times on winXP vs eCS. I define bootuptime as the time until harddrive acticity is near zero, not until desktop shows. Both cores are working in both OS. OO preloads in both systems. !:st and 2:nd refers to opening program first and second time. eCS installed on JFS, winXP on NTFS.

winXPeCS RC3
bootuptime   3 min 0 sec   1 min 15 sec   
FF 1:st   7 sec   9 sec   
FF 2:nd   3 sec   2 sec   
OO 1:st   10 sec   5 sec   
OO 2:nd   2 sec   2 sec   
/Mikael

Saijin_Naib

#74
Here are my results. I measured boot time as the time beginning immediately following my enter keystroke on the IBM boot-manager until the point where the desktop is displayed and the HDD activity is negligable (read: none). Note, for XP, I do not have auto-login enabled so I must press enter when the login is displayed, adding a bit of time to the boot. I timed program times as the time until the program is loaded, fullscreen, page is displayed (browser http://www.blackle.com) or ready for text input (office suites). I chose OpenOffice writer for eCS and XP, and Office XP Word for XP. I do NOT have pre-load for OpenOffice, firefox, or OfficeXP on my computer.
Firefox on both platforms is the 2.0.0.9 release with the Phoenity Modern Skin.

eCS-HPFS
boot: 1:33
FireFox1: 22
FireFox2: 14
OpenOffice1: 39
OpenOffice2: 37

XP-NTFS
boot:36
FireFox1: 6
FireFox2: 2
OpenOffice: 30
OpenOffice: 4
Internet Explorer 7-1: 1.48
Internet Explorer 7-2: .97
Office XP: 5
Office XP: 2.36