A lot of information about OS/2's USB Support can be found on the web, and as others have said a lot of video support is being provided by Scitech. Compare the numbers between the OS/2 and Linux versions.:-)
Don't know about wireless networking, but some info can be found here.
but it wasn't SOM-based at all. It looked like (and acted like) like the WPS to a certain extent when it came to basic file/folder manipulation, but that's about the extent of it.
When I have a few thousand source modules to search through for various things, I'd rather use a tool like @CULL to create a pre-processed search database and a tool like IACULL or FINDREF to search for keywords. Much faster. Think cscope on steroids.:-)
If you're running online transactions rather than batch and have an application which can spike at various times due to unforeseen events (e.g., bad weather), you probably want to keep your peak CPU at a lower level than a batch-only system.
Odin's Win32 support works well enough for me to get my old Palm m105 data (and PalmPix data) transferred back and forth under OS/2 via the Win32 version of Palm Desktop 3.0, and it also runs Adobe Acrobat 4.0 and the latest IrfanView fairly well (a few glitches in things like video and thumbnail creation, but it mostly works).
In the normal GeoManager, one would use the left mouse button to rubber-band-select the files to copy or move, and then use the right button to perform the actual copy/move operation.
It seems like Windows was the odd man out when it came to the assigned use of mouse buttons, at least in the x86 world in the early 1990's.:-)
OS/2 Warp was the third major 32-bit release from IBM.
OS/2 2.0 was the first 32-bit release of OS/2, and the one that started gaining popularity in the PC hobbyist community. It had the WorkPlace Shell from IBM, the MVDM (Multiple Virtual DOS Machine) subsystem that made it so good at running DOS programs, and WinOS2 (the Windows 3.x subsystem that could poke holes in the PM desktop and run OS/2 and Windows programs together "seamlessly".
OS/2 2.1 (1993) introduced a Windows 3.1-compatible WinOS2 (the WinOS2 in OS/2 2.0 only did Windows 3.0), and OS/2 Warp 3.0 (1994) introduced dial-up TCP/IP networking (SLIP and PPP) in the box as well as a graphical web browser (the original IBM Web Explorer).
The OS/2 Warp 3.0 "Connect" variant introduced NIC support and peer-to-peer network support in the box, and IBM released OS/2 Warp 4.0 in 1996 with all of that networking stuff (and two browsers) as part fo the standard package.
There's a huge distinction between copyright law as it applies to source code and patent law as it applies to software, and I suspect most of us do not group the two subjects together arbitrarily as you seem to be doing above.
Most of the open source programmers I know are quite respectful of copyright but quite skeptical about patent law relating to software.
One of the ways to handle such situations is to try to prevent the opponent from having the time and resources to porc up in the first place, but I guess that's pretty easy to say from the sidelines.:-)
Once he's porced up, though, he isn't expanding, and that gives you the upper hand in the long run.
I know I'd try to porc up if I thought I'd get away with it, but I'm not good enough to actually *win* that way. Sometimes it's better to go out with a bang, anyway...
Windows 95 was an attempt to shoehorn the Win32 API onto the old Windows 3.x architecture and dressing it up with a very poor functional workalike of the OS/2 2.0 WorkPlace Shell as the desktop (sans most of the powerful OO features).
It wasn't really an attempt to push tech forward as much as it was an attempt to create a stop-gap 32-bit solution that would hold their existing desktop markiet in place until their NT platform could mature enough to be generally useful.
Many of the games features were designed to reduce micromanagement, freeing the player's brain (and mouse) for more important things.
For example, you could direct a given airfield to assign a team number to all units produced, and you could use the Shift key and mouse to define a multi-vector patrol route for the aircraft to navigate once built. Assuming you also had a few repair platforms around and activated, the aircraft would automatically assume a patrol along the assigned route, attack anything that came in range, and automagicaly fly off to repair themselves if they were wounded too badly.
You could also queue up a few dozen build orders for a selected number of construction vehicles (nice for putting together some fairly organized defenses), committing those units to that action for a half-hour or more, and then you could set a few more construction units (level 1 construction aircraft were good for this) on patrol in the general area. Not only would the patrolling const a/c automaticaly perform repairs on wounded units as they passed (or harvest enegry or metal if needed), but they would also stop and assist the building units with their tasks, adding their own construction power to the power of the original unit until the given target was completed.
The Shift key could be used to queue almost any type of order, be it builds, attacks, movements, or whatever, and I wish more RTS games would impelement that simple concept...
The EMX libraries (a DLL that maps the POSIX API to native OS/2 API calls) have existed for years, allowing many text-mode Linux programs like slrn, pine, lynx, links, Midnight Commander, etc get ported to OS/2.
In addition, XFree86 has been ported to OS/2 for years, and OS/2 users have had access to a native port of GIMP 1.x for some time.
Apache is ported to OS/2 already, and OpenOffice has been ported by Innotek. Note that the original StarOffice (on which OpenOffice is based) was written for OS/2, so OS/2 users have a long history with that code base.
One thing that would help OS/2 immensely, IMO, is support from the QT and GNOME folks.
It's only an 8MB palm, and it only runs PalmOS 3.5, but it's an inexpensive model (usually US$40 on eBay), people are dumping accessories for them (folding keyboards, modems, etc) on eBay for less than US$5 in many cases, and a set of AA batteries can last a month even with moderate usage.
Yep, and for my Proliant 2500 with 64MB a similar pattern applies:
mem=exactmap mem=640k@0M mem=63M@1M
However, once I disabled the on-board video and put a Compaq-branded Matrox Millenium card in a PCI slot it got a little more complicated.
Mandrake 8.2 just wants a mem=64M directive, but a lot of newer live CDs (DSL, INSERT) won't work with either the previous solution or the simpler one. Must be a 2.6 kernel thing...
...they come with hot swappable hard drives and power supplies and lots of other fancy stuff. Once you figure out how to use MEM kernel directives to get around the memory hole(s) you're all set!:-)
Most of the comments I've seen w.r.t. Solaris, BSD, and Linux performance or features have been anecdotal in nature, but I've seen very few hard numbers to back them up.
I'd love to see how the latest mainstream Linux, *BSD, and Solaris/x86 kernels perform in various situations, how smoothly they can adjust to load differences, and how well their filesystems work under various scenarios, but I've not found much data along those lines.
It should also be kept in mind that the features or performance tweaking which might make a given kernel close to optimal for one type of task might well cause problems for other types of tasks, so the whole concept of "the best kernel" might not even exist as a general concept except in the minds of platform advocates.
Here in Smyrna (a NW Atlanta suburb), we can get excellent Chinese food delivered (from Orient Express in Vinings) and fairly good Italian food (from Bella's in Smyrna), and it's replaced pizza for us. Good stuff. Atlanta doesn't have good pizza places, anyway.
We even had some good options when I lived in a SW suburb of Minneapolis -- Green Mill in Eden Prairie delivered their entire menu, including various pasta dishes and sandwiches as well as pizza, and I seem to recall at least one other place in the area that did Italian food delivery.
A shell running in a fullscreen console would have no idea which window manager you were running, or in some cases which X server occurrence has priority for things like "themes".
Also, if X isn't running, using the X clipboard would not be very useful.:-)
There are some features that would be nice to see when a shell program is running in an xterm, but keep in mind that shell programs also have to run in environments which are totally detached from the X server...
More and more it seems like "a sane economic policy" means "making decisions to obtain short-term savings and line my own packets at the expense of long-term viability".
I back up to an 8MB Northstar Mobile MemorySafe module somewhat frequently (I actually alternate between a pair of them), and when I change the batteries I simply transfer the MemorySafe software from one of the other palms we have and then restore the old data from the module.
It hasn't failed me yet! I also keep an emergency backup on the PC, but I've not had to use it for quite some time...
Slashdot Mirror
A lot of information about OS/2's USB Support can be found on the web, and as others have said a lot of video support is being provided by Scitech. Compare the numbers between the OS/2 and Linux versions. :-)
Don't know about wireless networking, but some info can be found here.
but it wasn't SOM-based at all. It looked like (and acted like) like the WPS to a certain extent when it came to basic file/folder manipulation, but that's about the extent of it.
Even so, it really wasn't a bad Windows shell.
When I have a few thousand source modules to search through for various things, I'd rather use a tool like @CULL to create a pre-processed search database and a tool like IACULL or FINDREF to search for keywords. Much faster. Think cscope on steroids. :-)
If you're running online transactions rather than batch and have an application which can spike at various times due to unforeseen events (e.g., bad weather), you probably want to keep your peak CPU at a lower level than a batch-only system.
Odin's Win32 support works well enough for me to get my old Palm m105 data (and PalmPix data) transferred back and forth under OS/2 via the Win32 version of Palm Desktop 3.0, and it also runs Adobe Acrobat 4.0 and the latest IrfanView fairly well (a few glitches in things like video and thumbnail creation, but it mostly works).
In the normal GeoManager, one would use the left mouse button to rubber-band-select the files to copy or move, and then use the right button to perform the actual copy/move operation.
:-)
It seems like Windows was the odd man out when it came to the assigned use of mouse buttons, at least in the x86 world in the early 1990's.
OS/2 Warp was the third major 32-bit release from IBM.
OS/2 2.0 was the first 32-bit release of OS/2, and the one that started gaining popularity in the PC hobbyist community. It had the WorkPlace Shell from IBM, the MVDM (Multiple Virtual DOS Machine) subsystem that made it so good at running DOS programs, and WinOS2 (the Windows 3.x subsystem that could poke holes in the PM desktop and run OS/2 and Windows programs together "seamlessly".
OS/2 2.1 (1993) introduced a Windows 3.1-compatible WinOS2 (the WinOS2 in OS/2 2.0 only did Windows 3.0), and OS/2 Warp 3.0 (1994) introduced dial-up TCP/IP networking (SLIP and PPP) in the box as well as a graphical web browser (the original IBM Web Explorer).
The OS/2 Warp 3.0 "Connect" variant introduced NIC support and peer-to-peer network support in the box, and IBM released OS/2 Warp 4.0 in 1996 with all of that networking stuff (and two browsers) as part fo the standard package.
Heh heh. :-)
Or was that Tholenbot? I don't remember anymore...
There's a huge distinction between copyright law as it applies to source code and patent law as it applies to software, and I suspect most of us do not group the two subjects together arbitrarily as you seem to be doing above.
Most of the open source programmers I know are quite respectful of copyright but quite skeptical about patent law relating to software.
One of the ways to handle such situations is to try to prevent the opponent from having the time and resources to porc up in the first place, but I guess that's pretty easy to say from the sidelines. :-)
Once he's porced up, though, he isn't expanding, and that gives you the upper hand in the long run.
I know I'd try to porc up if I thought I'd get away with it, but I'm not good enough to actually *win* that way. Sometimes it's better to go out with a bang, anyway...
Windows 95 was an attempt to shoehorn the Win32 API onto the old Windows 3.x architecture and dressing it up with a very poor functional workalike of the OS/2 2.0 WorkPlace Shell as the desktop (sans most of the powerful OO features).
It wasn't really an attempt to push tech forward as much as it was an attempt to create a stop-gap 32-bit solution that would hold their existing desktop markiet in place until their NT platform could mature enough to be generally useful.
Many of the games features were designed to reduce micromanagement, freeing the player's brain (and mouse) for more important things.
For example, you could direct a given airfield to assign a team number to all units produced, and you could use the Shift key and mouse to define a multi-vector patrol route for the aircraft to navigate once built. Assuming you also had a few repair platforms around and activated, the aircraft would automatically assume a patrol along the assigned route, attack anything that came in range, and automagicaly fly off to repair themselves if they were wounded too badly.
You could also queue up a few dozen build orders for a selected number of construction vehicles (nice for putting together some fairly organized defenses), committing those units to that action for a half-hour or more, and then you could set a few more construction units (level 1 construction aircraft were good for this) on patrol in the general area. Not only would the patrolling const a/c automaticaly perform repairs on wounded units as they passed (or harvest enegry or metal if needed), but they would also stop and assist the building units with their tasks, adding their own construction power to the power of the original unit until the given target was completed.
The Shift key could be used to queue almost any type of order, be it builds, attacks, movements, or whatever, and I wish more RTS games would impelement that simple concept...
The EMX libraries (a DLL that maps the POSIX API to native OS/2 API calls) have existed for years, allowing many text-mode Linux programs like slrn, pine, lynx, links, Midnight Commander, etc get ported to OS/2.
In addition, XFree86 has been ported to OS/2 for years, and OS/2 users have had access to a native port of GIMP 1.x for some time.
Apache is ported to OS/2 already, and OpenOffice has been ported by Innotek. Note that the original StarOffice (on which OpenOffice is based) was written for OS/2, so OS/2 users have a long history with that code base.
One thing that would help OS/2 immensely, IMO, is support from the QT and GNOME folks.
It's only an 8MB palm, and it only runs PalmOS 3.5, but it's an inexpensive model (usually US$40 on eBay), people are dumping accessories for them (folding keyboards, modems, etc) on eBay for less than US$5 in many cases, and a set of AA batteries can last a month even with moderate usage.
Yep, and for my Proliant 2500 with 64MB a similar pattern applies:
mem=exactmap mem=640k@0M mem=63M@1M
However, once I disabled the on-board video and put a Compaq-branded Matrox Millenium card in a PCI slot it got a little more complicated.
Mandrake 8.2 just wants a mem=64M directive, but a lot of newer live CDs (DSL, INSERT) won't work with either the previous solution or the simpler one. Must be a 2.6 kernel thing...
...they come with hot swappable hard drives and power supplies and lots of other fancy stuff. Once you figure out how to use MEM kernel directives to get around the memory hole(s) you're all set! :-)
That way, platform compatibility is a nonissue.
I use a dedicated PPro box running Coyote Linux myself, but there are far more robust solutions out there...
That's a very important qualifier. :-)
Most of the comments I've seen w.r.t. Solaris, BSD, and Linux performance or features have been anecdotal in nature, but I've seen very few hard numbers to back them up.
I'd love to see how the latest mainstream Linux, *BSD, and Solaris/x86 kernels perform in various situations, how smoothly they can adjust to load differences, and how well their filesystems work under various scenarios, but I've not found much data along those lines.
It should also be kept in mind that the features or performance tweaking which might make a given kernel close to optimal for one type of task might well cause problems for other types of tasks, so the whole concept of "the best kernel" might not even exist as a general concept except in the minds of platform advocates.
Here in Smyrna (a NW Atlanta suburb), we can get excellent Chinese food delivered (from Orient Express in Vinings) and fairly good Italian food (from Bella's in Smyrna), and it's replaced pizza for us. Good stuff. Atlanta doesn't have good pizza places, anyway.
We even had some good options when I lived in a SW suburb of Minneapolis -- Green Mill in Eden Prairie delivered their entire menu, including various pasta dishes and sandwiches as well as pizza, and I seem to recall at least one other place in the area that did Italian food delivery.
If there isn't an "owner" tag on it or some record that identifies a previous owner, you're basically screwed.
A shell running in a fullscreen console would have no idea which window manager you were running, or in some cases which X server occurrence has priority for things like "themes".
:-)
Also, if X isn't running, using the X clipboard would not be very useful.
There are some features that would be nice to see when a shell program is running in an xterm, but keep in mind that shell programs also have to run in environments which are totally detached from the X server...
More and more it seems like "a sane economic policy" means "making decisions to obtain short-term savings and line my own packets at the expense of long-term viability".
Is that really sanity? Or is it just greed?
Not that it matters, me being a single data point and all, but the number of OS/2 users these days is certainly non-zero.
I back up to an 8MB Northstar Mobile MemorySafe module somewhat frequently (I actually alternate between a pair of them), and when I change the batteries I simply transfer the MemorySafe software from one of the other palms we have and then restore the old data from the module.
It hasn't failed me yet! I also keep an emergency backup on the PC, but I've not had to use it for quite some time...