Reverse Engineering is kind of a strange thing -- since it is quite difficult to prove... especially if you didn't decompile the program, and how can one prove that you accutually reverse engineered the protocol instead of actually completely clean boxing the code.
WINE is probally mostly designed clean boxed -- I don't see them use Microsoft code themselfs, it is written by the programmers themselfs, and is quite different for Microsoft code (that's why the widgits look different and why some programs refuse to work with WINE). AFAIK, none of the WINE developers have yet to get there hands on Microsoft Windows source code.
So you get a policy like this with WINE: It may be compatible with Microsoft Windows products, but it may not. This is an entirely new set of API's designed to make it easier for software developers to easily port there programs to Linux x86. No code was taken from Windows, nor was any code reversed engineered -- all was based on documented APIs (from various programing manuals or from populuar open source progams) and based on completely Free code.
Samba is kind of the same thing -- it was designed to work with other pieces of hardware, using a protocol similar and possibly comaptible with Windows NT SMB networking, but it contains no Microsoft code, and was not reversed engineered, therefore, none of the developers had access to Windows NT SMB source, it may not be compatible with your network, or it might be. Samaba is a totally new protocol, in which some cases may be compatible with Windows NT SMB, but it may not be.
Making sure that you don't claim your product is 100% compatible with a piece of software, and claiming that your software is totally new, and contains non code from a previous piece of software, makes proving reverse engineering almost impossible to prove.
As long as you can prove your code is 100% clean and contains no code from previous places, and was not reversed engineered, your okay. Since Open Source Software is completely open, and any person can look at it, it would be quite hard to find spefic lines that have been reversed.
Well, Reverse Engineering, I don't see to be much more enfourceble then in the past.
When do you call it reverse engineering? I don't see that clearly defined in UCITA, it basically states it's illegal. If you haven't looked at the source code or haven't tried to disemble the program is it reverse engineering?
Okay, I don't see Microsoft going after mom and pop small free software projects -- it would be just a pain in the ass for them (there are so many of them), it would be hard to prove, and most importantly it would be really bad press.
Think of the headlines in the bussiness section, "World's Largest Software Maker--Attempts to Crush all others via. Power of UCITA", "Microsoft Tries to make Populuar Free Product Illgal", etc.
Blunders like this don't look good -- they make the software company look like a large tyrant trying to hurt everybody out there. And if enought software developers just get fed up, they will choose to go with another existing software plaform, and build on top of that (such as Linux, FreeBSD, Mac OS X, Amiga NG etc.)
Reverse Engineering has been illegal for years -- but it's almost never enforcable for sevral reasons:
1) It's Hard to Create Enough Proof 2) Laws allow it partially -- as long as it's not an obivious copy. 3) It's too much time and money to waste on small companies. 4) At least with free software, who do you sue?
There is nobody to sue, as said in the license, except for maybe support companies, which already have rules in place to avoid that. Lawsuits against the FSF won't seem to go to far -- since FSF can claim that software is done by indivuals and not themselfs. And suing a non-profit organisation is quite hard to do legally, especially if they didn't blatently do something illegal.
Hell, you don't even need root for that. Just shut the machine off, turn it on while holding down command-option-pr, you should boot into the Mac OS (assuming that's installed). If BootX comes up, just put for kernel arguments 'linux single'. You'll boot in single user mode.
If it doesn't have BootX on the machine, try to find Bootvars, and use that to set OpenFirmware to boot into Linux.
You can also use the Apple disk tools disk and pop the bootX Extention on it, and use that to boot into Linux single mode.
I believe that crack.linuxppc.org is not on the connection as linuxppc.org, but on a smallish dsl connection of Jeff Carr. So that would help to explain the general hesitation of the site, but I am sure the old 604 132mhz doesn't help out... remember Microsoft2000test.com is on a fast T1 or T3, hooked up to a fast 350mhz PIII.
Give the PowerPC box a break. It's been up, but it's kind of lagged at times (especially with it doing a cat/proc/cpuinfo every time somebody loads the page), spawing 3 extra processes per hit.
It's not been down yet (look at the live stats), although, until Jeff changed the page around a little bit the machine was putting on some serious lag (some people claim up to 2 minutes, for a bit I was getting maybe 1 minutes lag with telnet), mainly due to the 132 mhz 604 processor couldn't keep up with opening so many processes all at once. Also, remember people are hitting this machine with massive pings and trying to crack the root password by force (which is freaking retarded because by default LinuxPPC R5 does not let any remote root log-ins for any protocol.);-)
If you read http://crack.linuxppc.org/ you'll see the stats, the machine has been up for 3 days and like 5 hours. And yes they are live updated every minute, check the time on them, they are within the minute you loaded the page.
I would think the PowerPC box would have more people trying to hack it, since you hack that, you win a free computer, if you hack the NT box, you get nothing.
huh? And your same thoughs apply to Red Hat Linux 6.0 on x86?
LinuxPPC isn't really dumbed down, it's about as hard or as easy to work with as Linux x86. It has the standard RedHat 6.0 installer that we all know and love (and can use in your sleep), or a new X Linux installer which lets you use a graphical gtk-perl based installer.
Installation is much like RedHat Linux 6.0, the installer has virtually everything the same, including Xconfigurator, and all of the other standard tools. You can boot Linux via either Quik (sorta like LILO for PowerPC systems -- it uses OpenFirmware which is about the equvalant to x86 BIOS) or using the handy BootX utility that allows booting from the Mac OS, is easy to use, etc.
Yellow Dog Linux is much like LinuxPPC, since they are both RedHat-Linux based, so they share installers that look and feel the same and quite similar pakcages. I might mention that parts of Yellow Dog Linux Champion Server 1.1 are higher quality then LinuxPPC, and seem to work better.
Debian/PPC is still an unstable version of Debian, it doesn't yet have a PowerPC installer (you install RedHat-type Monolithic PowerPC Linux and then replace it with Debian).
TurboLinux/PPC is quite dated, the last time I checked it was still using glibc 1.99, instead of glibc 2.1, but that may have changed, since TurboLinux/PPC is more of an far east distro then other PowerPC ones. Again, RedHat-Linux based.
Lets, not forget MkLinux Release 1, which is another RedHat-based PowerPC distro, which is currently in developement. It uses MkLinux Genric 8 alpha something for a kernel, and well it should be released this fall if all goes well.
No, it obviously not a default install, since more services are enabled by this install (like telnet), then how it typically ships, with all inetd services disabled.
So they changed some things that would be typically changed on the server to make it 1) easier to use 2) more services 3) added services locked down better.
Yes, it would appear that way, since RedHat Linux 6.0 by default disables all inetd services, as do most populuar things based on RedHat 6.0, such as LinuxPPC R5.
Turning on Telnet on a server that you are trying to get secure seems a bit exterme... why have telnet enabled when you can have something like ssh enabled just when you need it. Of course if you are going to leave telnet enabled, you can at least use/etc/hosts.deny and/etc/hosts.allow to control what machine have access to that box.
(Not to say that populuar sites around the web do have telnet enabled for everybody on any machine as long as you know the login/pass, for example www.macnn.com).
This should be interesting to see what flaws (since we know they are not perfect) we learn about in RedHat Linux 6.0 and spefically LinuxPPC R5.
Of Course So. If you where on the Linux/PPC-developer's list (sponsered by LinuxPPC, Inc.) you would have recieved most of these patches as enclosers in your mail box.
Of course, that caused a problem -- lots of people would lose or accidently discard these patches, or were not on Linux/PPC-developer so now Motorola has a website (http://www.mcg.mot.com/linux/) that has most of them avalible to patch various Linux/PPC kernels you might be running.
Well, maybe not offically, offically, but they are certainly contributing several patches to the Linux/PPC Kernel (to support their hardware better--they are the main reason why Linux/PPC Kernel started to support RS/6000 / CHRP / PREP Machines), gcc patches to support the unreleased (yet) Altivec PowerPC processor extentions (think MMX on steriods), and patches to better support embeded devices. So Motorola (well, at least Motorola employees, have been 'unoffically' supporting Linux on the PowerPC for almost 3 1/2 years).
> Motorola had aboloutely no connection that i am aware of with the LinuxPPC project (www.linuxppc.org).
Well if you are referring to the Linux/PPC Kernel project see above. If you are refering to LinuxPPC, Inc., LinuxPPC, Inc. used to resell Motorola StarMaxes with LinuxPPC preinstalled. Of course that ended with Apple ending cloning.
> What is more, this is not the only linux distribution for PPC macs; there is also MkLinux.
You are again confusing Linux/PPC Kernel Developers and Linux Indepent Developers with LinuxPPC, Inc.
There are actually quite a few different distros that use the Linux/PPC Kernel -- Debian on PPC, Yellow Dog Linux, An unammed in development distro by the Linux am Mac people (for germans), TurboLinuxPPC and of course LinuxPPC Inc's LinuxPPC product.
So LinuxPPC isn't the only distro using the Linux/PPC Kernel.
Re:Memory Usage ?(or Memory more important than CP
on
Some KDE news
·
· Score: 1
I can attest to that. KDE and GNOME seem to work pretty good on a machine with only 32 megs of RAM.
If your machine doesn't have at least 32 megs of RAM, you can probaly still run KDE/GNOME but the preformance will be slightly slower (and likely unusable).
But the fact is, RAM is cheap. Virtually anybody can afford a machine with at least 64 megs of RAM, alot of new machines come with around 128 megs.
When people say Linux can run on a 386 with 4 megs of RAM, just remember they are referring to a min. equiped setup -- just the basics, with no X or anything.
Re:I'm looking forward to the day they ditch X
on
Some KDE news
·
· Score: 1
That may be a issue of how well the X Server supports your hardware (including things like Hardware Acceloration) -- most X Servers aren't good on all hardware -- for example XFree86 isn't the fastest or the best on several video cards.
What that means, is that Windows 95 has better video drivers then XFree86 (or what X Server you use) for your hardware.
Yes, some people with certain X Servers claim that they get far better preformance with X then with Windows 95.
When will we get better X Servers? When people convience there graphic card makers to support X development, by putting teams of people to improve things like XFree86. If enough people do this, your graphic's card vendor will feel required to improve it's XFree86 preformance.
Re:KImageShop as well...
on
Some KDE news
·
· Score: 1
They will try hard to ship KOffice in the same time frame as KDE 2.0, since they both need quite a bit of work (we are looking at around January '00 from what I have heard).
KOffice, according to what I have been reading is starting to have some stablity work done on it already, and doing some hookups in some of the apps. True KSpread and Kformula need some work, but KWord and KPresentor look as they will be well ready by the time of KDE 2.0.
KImageShop sounds nice, as does Killustrator. These also seem like important KDE-2.0 desktop apps.
Re:I designed this theme for testing
on
Some KDE news
·
· Score: 1
This "GUI Designer" theme building program, really excites me, and I hope it will live up to it's name by building powerful, quick and useful themes.
As a long time Kaliedoscope (a themeing program for the Mac OS) user, I often used a themeing program known as Kdesigner, which made building themes, relatively easy (well, not that easy, but it wasn't rocket science either).
Well, this should lead to dozens (literally) of both good and bad themes for KDE, but since everybody will have a choice this will be a good thing.
Adding a option to make the program support double clicking, probaly wouldn't be too hard--so feel free to implement it yourself.
Remember Kommander/Kfm uses a different idea to what files are -- it treats folders as if they were web pages (and you can modify the look and feel if you know how to), which is a totally radical deparature from anything Microsoft or Apple has ever done (which is invovation in my book). Since, most people agree the web is going to an important part of the future, we want to make the desktop easy to use and act just like the web (no matter what protocol you are using). What's nice about Kommander/Kfm is it treats a web page like files, you single click them to launch them.
Yes, this is different to what you may be used to in the old days, but I think this is a gui improvement.
With this new paradigm, you select files using the rectangle box, just like in the old days, but you can't do it with single clicks anymore (just like with web links).
Internet Intergration is a good thing in my book.;-)
Linus has expressed some opinions on replacing the old fashion VGA text modes on the x86 with Frame Buffered text console modes, like almost every non-x86 platform out there. The facts, are that frame buffered text mode is almost as fast as the old fashion VGA text modes, and how often do you really test the speed of hardware VGA text modes versus frame buffered modes in day to day work? Not much.
There are obvious benfits to using a frame buffered console, for one you can include graphics in it, such as when the system boots. Also if you are using a frame buffered console you can actually possibly run graphical programs without X11 (like Mac-on-linux or XF68_FBDev, a frame buffer based X server based on XFree86). This saves some of the bloat of X when you don't need it.
And yes, since kernel 2.2, x86 machine also support frame buffered consoles, but you have to spefically enable them when compiling your kernel.
I agree, the key to Linux for the masses is advertising. Linux hardware vendors seem to the first ones to get into the TV advertising bussiness. Some day, you will turn on your TV to some channel and be watching commericals and see:
"They're hot, inexpensive and stable"
"Includes the lastest Linux technologies, and life- time upgrades for most populuar Linux software*"
* depends on the package, see offical licenses for details.
"True Internet Intergration"
"We provide hardware and software otimized for our hardware--an @*&% exlusive!"
"Linux - Combines Ease of Use, Low Prices and Flexablity"
People will see those phrases on TV, and will start running for stores to get Linux machines. They will start to see the benfit to Linux, and see that they will be enjoying maxium flexablity.
The only problem with XML is that not all XML documents are combatible. XML is a very open and flexable standard, each program writes and reads XML slighty different then the other. (That's why AbiWord can not open and decently read Gnumeric or KWord documents).
XML does make the document fairly readable, so building convertors from one XML format to another should be as hard as to Microsoft Word format to AbiWord format, but still there are some challenges.
The good thing about XML is that at least for now all programs that write XML, their XML is at least some what humanly readable (then again the same is true with Microsoft Word documents.
You might try the shareware open-source program boches (www.boches.com), which costs $25 bucks, compiles on virtually every Linux (including PowerPC).
Obviously, since this program emulates an entire x86 you can run your favorite OS's on it, including Windows NT, DOS, BeOS x86, Linux, etc.
At this point the preformance kind of stinks for any kind of long term usage, but it certainly okay for a once and a while usage of Microsoft Excel for Windows 95 or some great DOS program.
Sheepshaver, I kind of doubt because it's closed source, although I guess somebody could send the SheepShaver team a RS/6000, although that's kind of unlikely since RS/6000's aren't cheap. At any rate I want to see a final version of SheepShaver running on LinuxPPC before we jump to conclusions.
Mac-on-Linux according to the ReadMe file (see ftp server at: ftp://ftp.nada.kth.se/home/f95-sry/Public/mac-on-l inux/) says a port to non-Power Macintosh, PowerPC's (like the RS/6000 or BeBox, etc.) wouldn't be too difficult, since they do run the same exact kernel. A few Macintosh-spefic things would need to be patched, but besides that it could be done with relative ease.
Microsoft is using AOL's properity OSC protocol, which is copyrighted and may not be reversed engineered by anybody including Microsoft.
The Free clients (GAIM, FAIM, etc.) are using TOC, which is fully documented, and you are encouraged to create clients that are compatible with it.
Microsoft chose to reverse engineer OSC because they felt that OSC users get better service (since they are using AOL technology and are viewing the AOL ads that are paying for it) and better features.
It's too bad they couldn't have gone the TOC protocol method... it would have saved them alot of bad publicity.
Slashdot Mirror
Reverse Engineering is kind of a strange thing -- since it is quite difficult to prove... especially if you didn't decompile the program, and how can one prove that you accutually reverse engineered the protocol instead of actually completely clean boxing the code.
WINE is probally mostly designed clean boxed -- I don't see them use Microsoft code themselfs, it is written by the programmers themselfs, and is quite different for Microsoft code (that's why the widgits look different and why some programs refuse to work with WINE). AFAIK, none of the WINE developers have yet to get there hands on Microsoft Windows source code.
So you get a policy like this with WINE: It may be compatible with Microsoft Windows products, but it may not. This is an entirely new set of API's designed to make it easier for software developers to easily port there programs to Linux x86. No code was taken from Windows, nor was any code reversed engineered -- all was based on documented APIs (from various programing manuals or from populuar open source progams) and based on completely Free code.
Samba is kind of the same thing -- it was designed to work with other pieces of hardware, using a protocol similar and possibly comaptible with Windows NT SMB networking, but it contains no Microsoft code, and was not reversed engineered, therefore, none of the developers had access to Windows NT SMB source, it may not be compatible with your network, or it might be. Samaba is a totally new protocol, in which some cases may be compatible with Windows NT SMB, but it may not be.
Making sure that you don't claim your product is 100% compatible with a piece of software, and claiming that your software is totally new, and contains non code from a previous piece of software, makes proving reverse engineering almost impossible to prove.
As long as you can prove your code is 100% clean and contains no code from previous places, and was not reversed engineered, your okay. Since Open Source Software is completely open, and any person can look at it, it would be quite hard to find spefic lines that have been reversed.
Well, Reverse Engineering, I don't see to be much more enfourceble then in the past.
When do you call it reverse engineering? I don't see that clearly defined in UCITA, it basically states it's illegal. If you haven't looked at the source code or haven't tried to disemble the program is it reverse engineering?
Okay, I don't see Microsoft going after mom and pop small free software projects -- it would be just a pain in the ass for them (there are so many of them), it would be hard to prove, and most importantly it would be really bad press.
Think of the headlines in the bussiness section, "World's Largest Software Maker--Attempts to Crush all others via. Power of UCITA", "Microsoft Tries to make Populuar Free Product Illgal", etc.
Blunders like this don't look good -- they make the software company look like a large tyrant trying to hurt everybody out there. And if enought software developers just get fed up, they will choose to go with another existing software plaform, and build on top of that (such as Linux, FreeBSD, Mac OS X, Amiga NG etc.)
Reverse Engineering has been illegal for years -- but it's almost never enforcable for sevral reasons:
1) It's Hard to Create Enough Proof
2) Laws allow it partially -- as long as it's not an obivious copy.
3) It's too much time and money to waste on small companies.
4) At least with free software, who do you sue?
There is nobody to sue, as said in the license, except for maybe support companies, which already have rules in place to avoid that. Lawsuits against the FSF won't seem to go to far -- since FSF can claim that software is done by indivuals and not themselfs. And suing a non-profit organisation is quite hard to do legally, especially if they didn't blatently do something illegal.
Hell, you don't even need root for that. Just shut the machine off, turn it on while holding down command-option-pr, you should boot into the Mac OS (assuming that's installed). If BootX comes up, just put for kernel arguments 'linux single'. You'll boot in single user mode.
;-)
If it doesn't have BootX on the machine, try to find Bootvars, and use that to set OpenFirmware to boot into Linux.
You can also use the Apple disk tools disk and pop the bootX Extention on it, and use that to boot into Linux single mode.
Security, the Macintosh way.
Just to point this out: finger is disabled by default on LinuxPPC R5, and is usually not even installed by default.
;-)
Next?
There are no user accounts on the machine, since this is an default install of user accounts in RedHat Linux Installer / LinuxPPC X Linux Installer.
So there would be no reason for Jeff to add a non-root account, since that just wouldn't make sense.
And yes, the other system generated accounts are locked down, you can not log in to them normally either.
AArthur On a LinuxPPC R5 Machine.
I believe that crack.linuxppc.org is not on the connection as linuxppc.org, but on a smallish dsl connection of Jeff Carr. So that would help to explain the general hesitation of the site, but I am sure the old 604 132mhz doesn't help out... remember Microsoft2000test.com is on a fast T1 or T3, hooked up to a fast 350mhz PIII.
Give the PowerPC box a break. It's been up, but it's kind of lagged at times (especially with it doing a cat /proc/cpuinfo every time somebody loads the page), spawing 3 extra processes per hit.
;-)
It's not been down yet (look at the live stats), although, until Jeff changed the page around a little bit the machine was putting on some serious lag (some people claim up to 2 minutes, for a bit I was getting maybe 1 minutes lag with telnet), mainly due to the 132 mhz 604 processor couldn't keep up with opening so many processes all at once. Also, remember people are hitting this machine with massive pings and trying to crack the root password by force (which is freaking retarded because by default LinuxPPC R5 does not let any remote root log-ins for any protocol.)
If you read http://crack.linuxppc.org/ you'll see the stats, the machine has been up for 3 days and like 5 hours. And yes they are live updated every minute, check the time on them, they are within the minute you loaded the page.
I would think the PowerPC box would have more people trying to hack it, since you hack that, you win a free computer, if you hack the NT box, you get nothing.
huh? And your same thoughs apply to Red Hat Linux 6.0 on x86?
LinuxPPC isn't really dumbed down, it's about as hard or as easy to work with as Linux x86. It has the standard RedHat 6.0 installer that we all know and love (and can use in your sleep), or a new X Linux installer which lets you use a graphical gtk-perl based installer.
Installation is much like RedHat Linux 6.0, the installer has virtually everything the same, including Xconfigurator, and all of the other standard tools. You can boot Linux via either Quik (sorta like LILO for PowerPC systems -- it uses OpenFirmware which is about the equvalant to x86 BIOS) or using the handy BootX utility that allows booting from the Mac OS, is easy to use, etc.
Yellow Dog Linux is much like LinuxPPC, since they are both RedHat-Linux based, so they share installers that look and feel the same and quite similar pakcages. I might mention that parts of Yellow Dog Linux Champion Server 1.1 are higher quality then LinuxPPC, and seem to work better.
Debian/PPC is still an unstable version of Debian, it doesn't yet have a PowerPC installer (you install RedHat-type Monolithic PowerPC Linux and then replace it with Debian).
TurboLinux/PPC is quite dated, the last time I checked it was still using glibc 1.99, instead of glibc 2.1, but that may have changed, since TurboLinux/PPC is more of an far east distro then other PowerPC ones. Again, RedHat-Linux based.
Lets, not forget MkLinux Release 1, which is another RedHat-based PowerPC distro, which is currently in developement. It uses MkLinux Genric 8 alpha something for a kernel, and well it should be released this fall if all goes well.
bah... It's probaly just the old LinuxPPC, Inc. web/mailing list/ftp server, that they finally retired last month.
That machine was a great, state of the art, 90mhz 601 machine, I think a PowerMac 7200.
In the past, that machine seemed to lag quite often with all of the stress it had on it, and was partcally due to all of the load.
RedHat 6.0 has no internet services turned on by default, you must manually enable FTP, telnet, etc.
w orking.shtml
Obviously, LinuxPPC has made a few changes to the server, for example they disabled some scripts and enabled telnet (by far a fair compremise).
Trust me, enought people have asked why telnet and FTP are disabled with the default install, if you don't believe me, see:
http://www.linuxppc.com/updates/telnet-ftp-not-
That's also true with RedHat 6.0.
No, it obviously not a default install, since more services are enabled by this install (like telnet), then how it typically ships, with all inetd services disabled.
So they changed some things that would be typically changed on the server to make it 1) easier to use 2) more services 3) added services locked down better.
Yes, it would appear that way, since RedHat Linux 6.0 by default disables all inetd services, as do most populuar things based on RedHat 6.0, such as LinuxPPC R5.
/etc/hosts.deny and /etc/hosts.allow to control what machine have access to that box.
Turning on Telnet on a server that you are trying to get secure seems a bit exterme... why have telnet enabled when you can have something like ssh enabled just when you need it. Of course if you are going to leave telnet enabled, you can at least use
(Not to say that populuar sites around the web do have telnet enabled for everybody on any machine as long as you know the login/pass, for example www.macnn.com).
This should be interesting to see what flaws (since we know they are not perfect) we learn about in RedHat Linux 6.0 and spefically LinuxPPC R5.
Of Course So. If you where on the Linux/PPC-developer's list (sponsered by LinuxPPC, Inc.) you would have recieved most of these patches as enclosers in your mail box.
Of course, that caused a problem -- lots of people would lose or accidently discard these patches, or were not on Linux/PPC-developer so now Motorola has a website (http://www.mcg.mot.com/linux/) that has most of them avalible to patch various Linux/PPC kernels you might be running.
> motorola isn't "officially supporting" anything.
Well, maybe not offically, offically, but they are certainly contributing several patches to the Linux/PPC Kernel (to support their hardware better--they are the main reason why Linux/PPC Kernel started to support RS/6000 / CHRP / PREP Machines), gcc patches to support the unreleased (yet) Altivec PowerPC processor extentions (think MMX on steriods), and patches to better support embeded devices. So Motorola (well, at least Motorola employees, have been 'unoffically' supporting Linux on the PowerPC for almost 3 1/2 years).
> Motorola had aboloutely no connection that i am aware of with the LinuxPPC project (www.linuxppc.org).
Well if you are referring to the Linux/PPC Kernel project see above. If you are refering to LinuxPPC, Inc., LinuxPPC, Inc. used to resell Motorola StarMaxes with LinuxPPC preinstalled. Of course that ended with Apple ending cloning.
> What is more, this is not the only linux distribution for PPC macs; there is also MkLinux.
You are again confusing Linux/PPC Kernel Developers and Linux Indepent Developers with LinuxPPC, Inc.
There are actually quite a few different distros that use the Linux/PPC Kernel -- Debian on PPC, Yellow Dog Linux, An unammed in development distro by the Linux am Mac people (for germans), TurboLinuxPPC and of course LinuxPPC Inc's LinuxPPC product.
So LinuxPPC isn't the only distro using the Linux/PPC Kernel.
I can attest to that. KDE and GNOME seem to work pretty good on a machine with only 32 megs of RAM.
If your machine doesn't have at least 32 megs of RAM, you can probaly still run KDE/GNOME but the preformance will be slightly slower (and likely unusable).
But the fact is, RAM is cheap. Virtually anybody can afford a machine with at least 64 megs of RAM, alot of new machines come with around 128 megs.
When people say Linux can run on a 386 with 4 megs of RAM, just remember they are referring to a min. equiped setup -- just the basics, with no X or anything.
That may be a issue of how well the X Server supports your hardware (including things like Hardware Acceloration) -- most X Servers aren't good on all hardware -- for example XFree86 isn't the fastest or the best on several video cards.
What that means, is that Windows 95 has better video drivers then XFree86 (or what X Server you use) for your hardware.
Yes, some people with certain X Servers claim that they get far better preformance with X then with Windows 95.
When will we get better X Servers? When people convience there graphic card makers to support X development, by putting teams of people to improve things like XFree86. If enough people do this, your graphic's card vendor will feel required to improve it's XFree86 preformance.
They will try hard to ship KOffice in the same time frame as KDE 2.0, since they both need quite a bit of work (we are looking at around January '00 from what I have heard).
KOffice, according to what I have been reading is starting to have some stablity work done on it already, and doing some hookups in some of the apps. True KSpread and Kformula need some work, but KWord and KPresentor look as they will be well ready by the time of KDE 2.0.
KImageShop sounds nice, as does Killustrator. These also seem like important KDE-2.0 desktop apps.
This "GUI Designer" theme building program, really excites me, and I hope it will live up to it's name by building powerful, quick and useful themes.
As a long time Kaliedoscope (a themeing program for the Mac OS) user, I often used a themeing program known as Kdesigner, which made building themes, relatively easy (well, not that easy, but it wasn't rocket science either).
Well, this should lead to dozens (literally) of both good and bad themes for KDE, but since everybody will have a choice this will be a good thing.
Adding a option to make the program support double clicking, probaly wouldn't be too hard--so feel free to implement it yourself.
;-)
Remember Kommander/Kfm uses a different idea to what files are -- it treats folders as if they were web pages (and you can modify the look and feel if you know how to), which is a totally radical deparature from anything Microsoft or Apple has ever done (which is invovation in my book). Since, most people agree the web is going to an important part of the future, we want to make the desktop easy to use and act just like the web (no matter what protocol you are using). What's nice about Kommander/Kfm is it treats a web page like files, you single click them to launch them.
Yes, this is different to what you may be used to in the old days, but I think this is a gui improvement.
With this new paradigm, you select files using the rectangle box, just like in the old days, but you can't do it with single clicks anymore (just like with web links).
Internet Intergration is a good thing in my book.
Linus has expressed some opinions on replacing the old fashion VGA text modes on the x86 with Frame Buffered text console modes, like almost every non-x86 platform out there. The facts, are that frame buffered text mode is almost as fast as the old fashion VGA text modes, and how often do you really test the speed of hardware VGA text modes versus frame buffered modes in day to day work? Not much.
There are obvious benfits to using a frame buffered console, for one you can include graphics in it, such as when the system boots. Also if you are using a frame buffered console you can actually possibly run graphical programs without X11 (like Mac-on-linux or XF68_FBDev, a frame buffer based X server based on XFree86). This saves some of the bloat of X when you don't need it.
And yes, since kernel 2.2, x86 machine also support frame buffered consoles, but you have to spefically enable them when compiling your kernel.
I agree, the key to Linux for the masses is advertising. Linux hardware vendors seem to the first ones to get into the TV advertising bussiness. Some day, you will turn on your TV to some channel and be watching commericals and see:
"They're hot, inexpensive and stable"
"Includes the lastest Linux technologies, and life- time upgrades for most populuar Linux software*"
* depends on the package, see offical licenses for details.
"True Internet Intergration"
"We provide hardware and software otimized for our hardware--an @*&% exlusive!"
"Linux - Combines Ease of Use, Low Prices and Flexablity"
People will see those phrases on TV, and will start running for stores to get Linux machines. They will start to see the benfit to Linux, and see that they will be enjoying maxium flexablity.
The only problem with XML is that not all XML documents are combatible. XML is a very open and flexable standard, each program writes and reads XML slighty different then the other. (That's why AbiWord can not open and decently read Gnumeric or KWord documents).
XML does make the document fairly readable, so building convertors from one XML format to another should be as hard as to Microsoft Word format to AbiWord format, but still there are some challenges.
The good thing about XML is that at least for now all programs that write XML, their XML is at least some what humanly readable (then again the same is true with Microsoft Word documents.
You might try the shareware open-source program boches (www.boches.com), which costs $25 bucks, compiles on virtually every Linux (including PowerPC).
Obviously, since this program emulates an entire x86 you can run your favorite OS's on it, including Windows NT, DOS, BeOS x86, Linux, etc.
At this point the preformance kind of stinks for any kind of long term usage, but it certainly okay for a once and a while usage of Microsoft Excel for Windows 95 or some great DOS program.
Sheepshaver, I kind of doubt because it's closed source, although I guess somebody could send the SheepShaver team a RS/6000, although that's kind of unlikely since RS/6000's aren't cheap. At any rate I want to see a final version of SheepShaver running on LinuxPPC before we jump to conclusions.
l inux/) says a port to non-Power Macintosh, PowerPC's (like the RS/6000 or BeBox, etc.) wouldn't be too difficult, since they do run the same exact kernel. A few Macintosh-spefic things would need to be patched, but besides that it could be done with relative ease.
Mac-on-Linux according to the ReadMe file (see ftp server at: ftp://ftp.nada.kth.se/home/f95-sry/Public/mac-on-
Microsoft is using AOL's properity OSC protocol, which is copyrighted and may not be reversed engineered by anybody including Microsoft.
The Free clients (GAIM, FAIM, etc.) are using TOC, which is fully documented, and you are encouraged to create clients that are compatible with it.
Microsoft chose to reverse engineer OSC because they felt that OSC users get better service (since they are using AOL technology and are viewing the AOL ads that are paying for it) and better features.
It's too bad they couldn't have gone the TOC protocol method... it would have saved them alot of bad publicity.