Slashdot Mirror
If IBM Is Serious About Linux, What Do WE Want?
bfree asks: "Robert LeBlanc, Vice President, Software Strategy, Software Solutions Division says both that IBM would open source any part of AIX and that we would be better off taking bits and pieces and the expertise that IBM bring along with it. IBM's AIX Web site lauds Linux compatibility of AIX and the new AIXL only just slightly behind their statements such as 'A robust, scalable UNIX platform for critical applications.' It's clear IBM wants to be involved with Linux, and I feel that we should want that also. What should we ask them to do for us in return for their involvement? Networking scalability and redundancy, optimization and facilities for database systems (as the jfs has started) or systems management applications? It seems to me we have the offer on the plate from IBM to create a new joined project to bring Linux up another level if we can find a way from AIX. Surely we must take them up on this?"
1. IBM should introduce some of the design features and decisions of AIX, including their segemented memory model, down to their driver api. 2. Add support for token ring. AIX has a driver that does NOT have to reset the device if disconnected from the network for a short amount of time like the linux one does. 3. All their recent NUMA technology they got in the sequent deal. 4. Net dispatcher. 5. have IBM put their Performance Team in Raleigh up to the task of attacking linux kernel, and many of the freeware appliacations like Samba, and Apache, so that we have the best. 6. Push for a full port of Lotus smart suite, as well as a native notes client, not used under wine. 7. A full dfs client from transarc. I know ibm is pushing away from dfs, but in the mean time a full client on linux would be priceless. 8. Various AIX items such as: errpt srcmstr cfgmgr mksysb sysbak and almost forgot, ODM 9. the win32 subsystem developed by Deital, and Kogan for OS/2 that was never released, so that we may wrap up some of the missing tidbits in wine. 10. and can IBM PLEASE push the linux kittens in one direction so that we might possibly have some standard not already set by the lead selling distribs.... That is all... Ober@bsdconspiracy.net
IBM's ThinkPad moels A20, A21, T20, T21 or X20 will fail to boot every OS (including Windows) after installing FreeBSD until the hard disk is removed from the laptop and the FreeBSD partition is wiped clean. This is because IBM DID NOT FOLLOW OPEN STANDARDS.
The Slashdot story on this was here, and the BSDToday story (which Slashdot copied) is here.
Details:
IBM uses a Phoenix BIOS in their Thinkpads. Phoenix chose not to do 3 minutes of research on the internet, and they didn't find Andries E. Brouwer's List of partition identifiers for PCs. Because Phoenix didn't care about Open Standards, they used partition type 0xA5 for their hibernation partition type. FreeBSD has used this partition type for years Andries's Partition Identifier page has been around since 1995.
At boot time, when the FreeBSD partition is seen (and thought to be a hibernation partition), the BIOS panics, doesn't know what to do, and the entire laptop is dead in the water, after FreeBSD is installed. The only fix is to physically remove the hard drive from the broken laptop, install it into a working machine, and wipe clean the FreeBSD partition.
IBM claims to support Linux, and they claim to support Open Standards. Yet they chose Phoenix for their BIOS, despite the fact that the recent Phoenix BIOS is non- Open Standards compliant. Even if IBM had no knowledge of Phoenix's bad choice, they share some culpability for not testing and checking. (In the worst case, IBM is hugely culpable for purposely and wilfully violating Open Standards.) IBM used to have a support page in which they said that they would not support FreeBSD on their hardware; they have taken the page down.
If IBM will not do 3 minutes of research on the internet and check for Open Standards before implementing something, so that they don't violate Open Standards, how long will it take before Linux is affected? How long will it take before IBM violates some Open Standard related to Linux?
If IBM will purposely chose a company that produces software which is not Open Standards compliant, and will not do anything to remedy the problem, how long will it take before they chose software from another vendor that directly and negatively affects Linux?
If Linux support means anything, then it means adherance to Open Standards. IBM screwed the pooch on this one. Maybe the fact that they've taken their page down means that they are going to fix the problem. Maybe it means that they are going to try to hide the problem and pretend it doesn't exist. They have not yet apologized.
Recommendation:
For now, I am boycotting all IBM products, including their RS6000 line. I urge you to do the same. This screw-up (with no apology and no fix forthcoming) indicates that IBM really does not care about Open Standards compliance.
Kenneth J. Hendrickson
One Word! Documentation. Ever see AIX Docs? I have - There are about 15 volumes in my cubicle at work - fantastic sutff. In my opinion (I manage a couple-0-rs6000's) documentation is the only real difference between Linux and AIX... -- eric
IBM wouldn't have done what they've done already if we hadn't given them something they wanted --apache. Apache is now sold as the IBM http server (or something to that effect). They now recognize the benefits and are trying to learn to play the game correctly.
Remember that IBM has a huge services wing. If IBM becomes connected in the public conscience with Linux, who do you think the public is going to come to when they need help?
Now that IBM is listening and willing to help imporove Linux we have to really sit down with an IBM AIX box and figure out what is in those boxes that we don't have that we could use. Some of their development group websites might have technologies that we could use too.
Just off the top of my head, we need to investigate security models, printing systems (not that anything is wrong with cups) and getting recognized certifications for the firewall code.
I'm not sure if IBM has anything to offer as far as a graphical devlopment environment, but we could sure use a mature gui that uses the standard tools in the background. Hm... Cluster management tools would be nice and perhaps their graphical management tools are worth a look.
Thats a start for now. This is all just off the top of my head, so keep the flaming to a minimum (especially on the ide thing..okay?). Lets not waste this chance to pick through IBM's brain.
Peace out.
P.S. Thanks for jfs and those print drivers.
(This is diadian, I just couldn't remember my pw.)
ok.. I'm going to kill myself for being used to slam enter to next filed.. but that's another story..
What I'd love to see is crash dumping, and better debuging tools.. this is something enterprise linux needs in order to be accepted by big business.. come on.. even NT can dump ram to disk
Thanks for the info...I'd been meaning to do a search for a while...I just only seemed to remember when I was not around a computer.
:)
Now I just need CMS
I'd love to see XEDIT on Linux - free. Same with REXX...and while their at it, it'd be nice to have a full CMS system that would run under Linux.
Okay...the "Good Old Days" article got me real nostalgic!
The existing community does a good job of writing code. IBM could provide additional developers with unusual experience, but this is on more of an individual level than a corporate level. One thing I think the corporation could do usefully is provide testing platforms. Linux developers generally can't tell how their code will do on massive SMP machines with huge memory spaces, huge disk arrays, and so forth, because they don't have the machines to try stuff. IBM ought to do testing for people on some of the really big boxes, so that ultra-high-end performance can be tuned.
Niether of you seem to have read my post.
I spoke specificaly about making the desktop interface forever dependent on ViaVoice. Sure it's availeble. That's how a pore kid like me knows it works. However the level of integration I am talking about is not suiteble for a totaly closed software.
--= Isn't it surprising how badly I spell ?
What I would want IBM to toss into the open isn't a part of AIX at all. I want ViaVoice. The IBM Voice recognition/Sinthesizer software. Failing that I would like just the synchronization part.
No I don't think it would really benefit much from being open sourced or that the development would accelerate. However I do think the time has come for Linux to start embedding voice technology for the visually impaired right into the UI.
This means for me a choice of shells during install that includes "BLINUX". A yet to be designed shell optimized in every possible way for use with speech software. Also speech synthesizers embedded in the desktop. I have been told that this can be done on KDE without adding any bloat for those who do not use it.
A crucial missing piece is that once this is done the desktop will be forever dependent on "libViaVoice.o" and the KDE people don't want to go down that road again. So IBM help us out here. PCs now include sound cards by default so adding this to Linux would iliminate the current $1400 additional premium on computers for a blind user. ( $600 hardware synthesizer and $800 software )
"Equal Employment" is just a Politically Correct fantasy as long as it costs more money to employ a blind person.
--= Isn't it surprising how badly I spell ?
Actually, this definately does not do them justice. They are a genuinely old company, at least by American standards :) You can't even compare a company incorporated in 1911 to a .com. They just don't live in the same conceptual space.
Logi - I can do anything, but not everything.
You can't have your cake etc., etc.
If IBM wants to fork the codebase, isn't that their perogative under the GPL? As long as the source is released, then they are playing by the rules--you can't have a double standard.
Likewise, why should "WE" demand anything from IBM. They can play with Linux or not. If what they contribute is useful, then we can all use it. Just because they are IBM doesn't give anyone the right to demand some kind of compensation to "let" them contribute.
>K
And assuming we aren't just talking about kernel stuff, but user space as well, I'd like to see tools for the above fleshed out. Performance tools for hardware diagnostics and benchmarking would rock, too.
In short, there's lots of areas where IBM could contribute it's experience (if not code) to the free software community.
I used to be part of a team that did sysadmin for a set of about 500 RS6000's all over the globe for a major manufacturing company. When I was hired I was a Linux only kind of guy and had a sharp learning curve on the differences between AIX and Linux. For example, I once crashed a production sybase server by running killall -HUP inetd, which doesnt quite do what it would do on Linux. (For those that dont know, on Solaris and AIX killall does just that, kills all processes it can get its greedy hands on.) I cursed and swore daily about AIX. A coworker liked to say "AIX isnt an Operating System, it's a virus that's pretending to be an Operating System" But there's one piece of AIX that impresses me to no end, LVM, The Logical Volume Manager. The ability to intelligently allocate, size, and move filesystems was incredibly useful and lent itself to long uptimes. We once had a drive start to act flaky, so live, in production with no outtages, we moved a sybase database (The equivalent of a raw partition) from one drive to another, with Sybase smacking the database the whole time. I realize there's been several projects to move towards journaling file systems like IBM's JFS, but I havent seen anything to compare to LVM. THAT is what we need.
Their "Linux Certified" laptops have winmodems that don't work with open-source drivers.
... though low-cost PowerPC unix servers (particularly rackmount) installable as AIX and/or Linux would be nice..
;)
As would supporting the plex86 project or porting it to PPC
Hell, any LVM/JFS ports to Linux could be justified as R&D, allowing new eyes and new techniques to help with the speed, features and reliability.. Perhaps AIX would gain online filesystem reduction as well as enlargement?
Your Working Boy,
VA for Java is available for Linux.. IIRC it's even supported.
Not so sure about Smalltalk though. Websphere + VAJ is similarly useful for web applications, though the speed and reliability isn't there yet (though there are definitely some useful performance tuning redbooks for WS)..
Your Working Boy,
Why cannot IBM work on a Natural Language CLI where say one could type "Move file: "Quake" from desktop to games folder".
How about this?
(hey, if someone can map sysadm processes to DOOM, it could be a crazy enough idea to work...)
Your Working Boy,
...is for a heavy-hitter to come out vocally in support, not just of open source, but of freedom. Taking a stand against that hard drive from hell proposal next story over would be a great start.
That and all the other "end to end" ideas rely on deliberately crippled tech, where the user is not given free run of their own property. People need to see that this necessarily harms all freedom as a consequence, since this kind of tricks have no chance without coercive legal backup. IBM is big enough to let people see so - and to refuse to play ball, giving freedom a competitive chance.
Exactly. If IBM uses software patents mainly as a defence against being sued by other patent-holders, then they have little to lose from a situation where software patents are not enforceable. They wouldn't be able to sue others over software, but likewise others couldn't sue them; so no need to accumulate stockpiles of questionable patents as a deterrent. (Hiring fewer lawyers => less overhead => more profit.)
OTOH, maybe they like to have the 'nuclear option' available. Just think what would happen if IBM started enforcing its thousands of patents on things like cut and paste.
It's not so much that IBM accumulates software patents that is the problem - they have a need to defend themselves. It's the lobbying to extend the totally broken US-style swpat system to other places that I find objectionable.
-- Ed Avis ed@membled.com
I thank IBM for there support of Linux on there high ticket hardware, but please start doing something about the consumer goods! IBM is starting to look very bad in my eyes when they return emails about *NOT* support linux on many of the products I own or can buy. It's time they start thinking of Linux for all of there products.
I am trying to create a "book-reader" for a friend who is losing his sight, & now cannot read print. He cannot claim against any insurance, so this has to be an "economic solution". Thus Linux is a very viable platform for us. I have found a suitable scanner controller, ( http://www.mostang.com/sane/, ), and a pretty good text to speech system ( http://www.cstr.ed.ac.uk/projects/festival/ ). We like the fact that it is possible to choose a voice with which we are comfortable. The missing link is the OCR component. There doesn't seem to be anything out there in ftp-land which works sufficiently well for us. As I only have a relatively limited amount of time and level of skill, and cannot catch up with the details of 20 years of Neural-Network technology overnight, this has to be a rather better goer than trying to get an "academic project" to work properly. If IBM could help with this one it would earn big Brownie-points I'm sure of that. I'd be more than happy to make available any glue files I create.
I was an OS/2 user and really miss the WorkPlace shell and the "e" ( Watson Works ) editor. IMHO "e" is a much better editor than anything currently available for Linux. If these things could be made available for Linux this /. reader would be a very happy chappie indeed. As OS/2 is really no longer a commercial earner, I'm sure that there wouldn't be any real financial loss in doing that, and there is the possibility that KDE and GNOME would benefit considerably.
Lastly is there a genuine implementation of REXX available for Linux?
First, I think they need to take the linux jfs support to equil ground with the AIX version (where file names dont have just one case, they have diffrent cases in their names).
Second, I'm Sure both samba projects could use all the Server Message Block docs they have, and maybe some code from the kernel.
Third, I think they could help on the user enviroment, maybe get XF to another level, or release their own X-Server.
Forth, I've only seen AIX running once, on a RS/6000 in a friends basement, but I'm sure they must have some useful apps they could work on, and port, maybe office tools, network admin tools, snmp monitors.
Fifth, I think we should give linux, and various apps that are nearly synonious with linux more features, and better support, so that ibm can better use it. IBM shoots hevely twards state contracts (schools) where they think win95 is the only os in town for workstations, and Netware 3.x/5 is the only thing that can share files.
Sixth, I think ibm should give us some wisdom, and lots of caffine. Perhaps the computing time and systems so that people could better develop some of this stuff.
:)
-LW
Well, I don't know what 'we' want, but I wouldn't mind a new BMW Z8. I also want a 10 room mansion, but built in swimming pool. Oh, and a house up in Lake Tahoe for the summer (and winter). And Alicia Witt as girlfriend. And a vacation to Australia.
And if IBM is unable to get me these stuff? Well, they're being parasitic anyways.
Je ne parle pas francais.
Someone mod this up.
It's ironic that the firm which inspired the phrase "nobody got fired for buying IBM" is now putting a high profile effort into supporting software over which it has no control.
Don't get me wrong - I think IBM have managed to reposition themselves really well. It's just not that long ago that they were the bad guys. With any luck, this pleasent trend will continue.
I second that - smit for linux would be interesting too.
What do we want IBM to give us?
Well it's easy enough to ask for code. There's lots of code to ask for too! We could ask them to open-source their OS/2 driver database, although there may be NDA's they have to honour that would prevent this from happening 100%. And I for one would like to see the OS/2 WPS ported to Linux as well. Although there are technical difficulties, I'm sure these could be overcome without too much stress. (Are there any legal problems with porting WPS?)
BUT. What I also hope for from a big, muscular company like IBM is their support in the NON-technical areas.
The more IBM talks Linux in the corporate arena, the greater the likelyhood of it being accepted by the pointy-hairs. And that would be of great benefit to the Linux community at large.
And I keep hearing about some smart-assed lawyer mounting a challenge to the (L)GLP. Frankly, I'd feel much happier if IBM footed the bill for the defense against such an attack, than if I was waiting for Linus to pay for it...
Sure, I want to see the code, but I also want the exposure and, if necessary, the protection!
OpenBSD is the most secure OS on the planet, FreeBSD outperforms linux and has a more stable filesystem, and NetBSD is portable beyond belief. So the question is, why does linux get the spotlight?
Yeah!! An' Betamax is better than VHS! But why does lin^H^H^H VHS get the spotlight? There's bound to be a sane and logical answer to this question too!
Disclamer: I am a Linux Zellot...
Former IBM loathing, then Apple Bashing, Now Microsoft bashing and occasionally Sun bashing... Zellot...
(IBM loathing only becouse I got in LATE.. I got into bashing the rest BEFORE it was populare.. and it's not populare to bash Sun.. YET)
Linux users rember OS/2 and quite often I hear "Will IBM do to Linux what they did to OS/2?" Or worse "IBM DID to Linux what they did to OS/2" the second being most insidous suggesting IBM has allready done "it"
But what is left out is... what did IBM do to OS/2? I don't think anyone knows or rembers...
Basicly IBM abandoned OS/2 when OS/2 needed defence... The PS/1 came with Windows preinstalled.. that sort of thing.. backing down when Microsoft started it's worst FUD attacks.
While OS/2 was IBM.. Linux isn't..
Then there is the other hand...
Linux has been shafted MANY TIMES and I suspect they want IBM to remeber we don't trust very quickly..
We have been burnt a few times... RedHat quickly "became" Linux.. and now RedHat is the worst example of Linux.. This isn't what we want to see from IBM.
RealMedia and other software venders jerking away Unix (and Linux) software in favor of (in some cases NEW) Windows Software.. (this also happend to Mac..)
Sun.. seemingly an allie.. we assumed WRONG... and Sun turnned and bit Linux on the butt... (Sun never did say they were allied so it's arguable the actions were consistent).
IBM is contributing advances.. they are taking the talk and walking the walk...
The sad part is others before IBM turnned into real snakes..
IBM has not been the good guy and is really on probation (as it were) by the whole computer industry...
Linux advocates are a really paranoid bunch..
I should know.. I am one...
But RedHat never did get the attention IBM gets..
Yeah we turn growl and bite...
We just want to remind you of where we stand on things...
It's going to take a long time before IBM gets accepted..
Yeah but think about this... IBM made more progress than anyone else...
IBM is in higher estime than Del, Compaq, Real Networks or RedHat...
IBM is taking Linux sereous....
Thats what we want...
We want long term commitment...
becouse we won't be foolled again...
I don't actually exist.
What "I" want or IBM is for IBM to move what makes AIX great into Linux.. direct ports rather than indirect ports...
This will degrade the marketability of AIX itself to nill.. So yes in a way I want Linux to suplant AIX. But I don't think thats what IBM wants.
IBM wants to sell hardware. AIX presents IBM hardware in the best light RIGHT NOW..
I want IBM to SELL and continue to sell other software solutions for Linux. I'm not to keen on open source ports of IBMs software.. I don't think that flys very well..
What I do want is a basic JVM that is open source.. like the old Blackdown...
What I want is an open source input "forum"..
Technical documentations for older IBM hardware...
I want my PS/2 portable fixed (but thats really not related to Linux.. it's a 386 anyway)
I want to see a "Standard" Linux type hardware.. like the PC.
But The PC has become property of Microsoft and Intel and really hard to get Transmeta or Linux into the picture.
So I want a NEW platform.. Something not depending on a CPU... The old case and powersuply is fine.. I'm not marryed to it but I'm not thinking anything specal for it eather... So go with ATX and bang the shit out of it for preformence...
Standard insert motherboard design looks fine to me.. The CPU is marryed to the motherboard so you gotta have a Transmeta MB for a Transmeta CPU... That looks fine...
What I want is a new BUSS...
if you gotta change anything to make it work.. then change it..
I want true plug-n-play.. I want my driver support in 4 styles...
1. Binary.. This is a driver dedicated to one CPU.. So the card would be lable say "Alpha" would have a driver for the Alpha..
The driver uses univeral hooks so all OSes can use it...
This driver is optional..
2. Source Code.. This would be a source code driver... again using universal hooks. But this time implemented in C or C++ with make files etc..
This is the alternitive to the binary.. it is located on the same ROM as the binary thus not requiring an additional CD for support..
3. portable driver.. This in a Java or Java like language.. takes a preformence hit but a valid alternitive to the binary. This also located on rom but in this case optionall.. This should be done when the card may be needed at boot time.. like a drive controller or a video card.
The need of the driver is to use on a computer that dosn't use the same CPU the cards binaru driver needs.. but is needed to compile the source code.. so this is a short term driver to provide just enough support to get things going...
4. Open source driver... Changes submitted to Linux in order to get the drivers into the Linux source code... This would provide long term.. stable.. and powerful support.
This should NOT replace the above... Relying on a kernel release can delay product release.. so the driver to rely on for a release should be the source code and binary drivers included on the card...
The drivers included into Linux should simply be "in addition"... This would allow improvments to be made to the drivers by other partys to improve and enhance preformence...
What I also want of IBM is probably something less of my benifit and more of IBMs benifit...
Dona rely on just Linux...
Simply put.. don't ride the wave...
Support BSD as well.... the drivers for Linux should also be provided to BSD...
Also give AtheOS a look...
If something (ghods forbid) happends to Linux.. Say Linus gose insain... or RedHat buys some larg valuable chunk of code and plays god with it (they do own Cygen. so they could conceavably screw with GCC) I'd like to have BSD as a strong alteritive.. and in many cases Linux isn't up to snuff so BSD really ends up being the only choice.. I'd like to keep that option open...
And of course as I said... I'd like maybe an open source committy.. just a group of people working on the hardware aspects.. IBM, VA, Peguin Computing, BSD (becouse they make servers), Cobalt and Rebel... could get together and work on some basic platform that everyone can agree on..
I'd like the preveous stated specs but.. you guys gotta marry the details... not I...
of course Slashdot would get some input vea VA..
Now to sell...
What dose IBM get out of the pacage if they do this?
Well with the BSD support they get BSD community and pritty much a more solidifyed commitment to the open source community.
With porting AIXes best features to Linux (and to BSD maybe?) yes AIX vanishes.. evetually BSD and Linux would suplant AIX anyway..
So this way you get early addopter status.. into the mainframe departmemt anyway...
You get credit for moving Linux from low to midrange into the high end..
You get free hacks to your software...
The downside is you get to deal with an outside OS again... You lose AIX and a revenue stream earlyer that you'd have lost it otherwise (It dies sooner or later.. and I know you'd like it to die later.. but it might be better to let it die quickly).
The hardware deal is very valuable.
Linux (and everyone else) is fighting it out on the PC.. at one time the PC had a standard setup.. as live moved on more standards were created but it still remainned standard.
But now companys won't work on a firmware standard for 3D cards.. or anything of the like...
The plug n play isn't the "Driver on rom" system TI put together (ironnicly for a Unix system) that Apple later addopted (and paid TI a fee for). In case anyone is wondering.. that system was a GUI Unix... it was one of the early failures in the GUI workstation market...
(hows that... GUI,, plug n play,... on UNIX... aimed at AI not users..)
Basicly the PC PnP system is burnned into Windows.. an ID on the card so Windows can pick out the correct Windows driver...
At one time many operating systems could live on the PC.. That hardware wasn't "Dos only".. Xenix and Minux had a fair chance...
I'd prefer a PnP "driver on rom" API standard rather than a firmware interface standard becouse firmware turns into legacy issues...
Gotta wonder about the ammount of additional cost in supporting the obsolete VGA firmware on todays 3D rendering cards. Thats a whole extra set of logic just to support legacy.. and thats only used just long enough to boot the system and get the drivers installed. What a waist... a cheaper Java like driver would save quite a bit in that area.
I like in kernel drivers becouse it just seems to me more talored. But I don't want hardware venders to rely on that. So you should have the vender driver (in card) and the Linux driver (in kernel).. The in kernel driver would be the Linux equal to Microsoft certifyed drivers.. and with incard drivers to keep Linux honnest...
I don't actually exist.
Allow me to insert something into this debate..
IBM is getting credability...
LinuxOne got it.. RedHat got it... what did we get? A scam IPO and a buggy distro...
IBM needs credability even more than LinuxOne and RedHat...
IBM needs it BADLY...
Thats all IBM is asking for... and thats what we WILL NOT give them...
Thanks for the code.. we need proof of commitment...
Anyone can deliver code.. We need something only IBM can deliver...
We need big machines... We need IBMs big machines... and we need Linux running on them.
Now I think we can ask a few things of IBM but... IBM will only do it if they think they'll benifit...
So it's not so much what will "WE" do for IBM.. but how will IBM benifit from what we want from them.
We shouldn't expect IBM to do something they can't benifit from and we shouldn't offer "in trade" anything for the moment...
So.. what I want is Linux on IBMs machines.. and some other things IBM could benifit from...
I don't actually exist.
I'm currently responsible for maintaining an IBM SP for a scientific computing facility. In this scenario, it's actually fairly close in overall design and function to a beowulf cluster.
The two things that have blown me away about this system are AIX's LVM and their high-performance shared filesystem, GPFS. Linux's logical volume manager comes close to the AIX LVM, but doesn't implement all of the functionality. I'm sure there's plenty to be learned.
BTW, a logical volume manager allows you to treat physical disks as a raw pool of resources that can be drawn upon for filesystms. You can create new filesystems, grow them in size, add more disks to the pool, all on the fly, without even having to unmount the filesystem.
GPFS is a distributed shared filesystem -- it lets you combine disks from many machines in the cluster into a shared filesystem that's visible to all machines in the cluster. There are other projects that accomplish this under linux, but IBM has been doing it for years, and the performance is unbelievable, bordering on magical.
These packages are the killer app for any type of clustering application of linux machines, be it web servers or beowulf-style computing.
-- Scott
No, Linux does not need anything from IBM. We don't need their journaling filesystem, their scalabiltiy, clustering software or clout in the marketplace.
But they would be damn nice things to have, and I think we could do better than to sneer with NIH syndrome when somebody offers something.
I'd really love to see a press release from IBM there it states that it is commited to use its (HUGE) patent portfolio to defend Linux and its developers. This kind of support would be priceless, IPR is the sector there Open Source moment is going to have the biggest problems...
Ville Oksanen
My DeCSS archive:
IBM is a business, they exist to make money. From that, it's only a simple leap to knowing what they want.
They want to sell stuff.
From my perspective, that's an easy one. Thinkpads are already the best notebook I know of and if our company's Netfinity M20 server is representative, they make damn fine server hardware, too. And the service... *sigh* A month after we got the Netfinity, one of the 35GB drives in the primary RAID array died, late on a Friday afternoon. The service engineer was in our building by 7:00 PM (we're in a small town; she had to drive here from another town,) another service engineer drove the replacement part from Chicago while the first one was on the way (she'd ordered one of everything she thought could be the problem!) and we were back up that night. And I didn't have to listen to elevator music once during the entire process. Yeah, if we were a big company we'd have had spare drives on hand, etc. but we're not big: our company's annual turnover would be laughably small to most of you. We've got the one netfinity server and aren't likely to need another one for years, etc. We still get great service and I'm awfully impressed.
mjs
Sigmentation fault - core dumped
I work for IBM also and while I have no idea what management and/or marketing are planning for Linux, I do know that there is a lot of internal enthusiasm for Linux and Open Source/Free Software in general. I would say that most of the developers I work with use Linux either at work or at home and are very familiar with it. Also even when we are working with other OS's, we regularly use tools such as emacs, apache, samba, gnumake, gcc, bash, perl, gdb, etc. Even when I have to develop for WinNt I write the code in Xemacs on RedHat and compile it using gnumake! You seem to see quite a few penguin or /. tee shirts or people lately too!
Okay, sure, when it's released in the *NEW* kernel it is GPL'd. What about before it is released? Huh? It isn't GPL'd before this. What is stopping IBM from releasing it and claiming it is their own creation?
this would happen once. at that point it would be on the front of slashdot, and ibm would loose all credability witht he linux community. so, ibm will have to decide if having control of that driver for that printer that makes 3 diminsional plots in linux is worth it.... i dont think it would be.
use LaTeX? want an online reference manager that
-- john
Anyway, the gist of it is that massive success of the internet requires a portable language (Java), portable data format (XML) and portable system services (Linux). Essentially, if you an put Linux everywhere, you can guarantee a set of portable system services that allow you disconnect the software from the platform.
What this does imply is that IBM might not be interested in making Linux vastly more efficient for certain classes of machines if it means that ease of portability is lost.
Agreed. Disclaimer -- I work for Intel Online Services. We sell a lot of Sun to our Unix customer base. They'd like to use Linux, and we'd like to sell it to them, but the only place we use it is relatively small web servers.
I've worked through two big jobs lately with open source companies who began designing with Linux on Dell and switched to Solaris on E420's at the last minute. They didn't switch because they don't trust Linux -- they make their money from Linux. They switched because of superior I/O and scalability beyond 2 CPUs and 2 GB RAM.
On the other hand, we use a _lot_ of Apache. If you look at the history of Apache, you see heavy IBM involvement in making it the stable and scalable solution it is now. I don't see anything wrong with IBM's involvement in Linux -- rather I see Linux's lifeline in the enterprise server market. No matter how small the company, there's a intention to get big. People don't want to limit themselves to systems that they think they'll need to leave when the company gets bigger, so the "Enterprise" market is vital to large-scale acceptance.
Desktop dominance would be nice too. I used Linux on a laptop and a server exclusively at my last job and quite enjoyed the stability. Make no mistake about it, though -- the only thing stopping me from doing my job on Linux is Microsoft Exchange Server. GNOME and KDE are good enough for now. StarOffice (OpenOffice I suppose) needs work, but will do. Netscape needs a lot of work, but it and Mozilla combined are enough to hold out until a good browser is available. But inability to "natively" (i.e. without POP3/IMAP) interoperate with Exchange is a showstopper for any corporate installation. A lot of sysadmins would love to get rid of the Exchange/Outlook combo, but it isn't going to happen overnight and interoperability is required.
"Nothing was broken, and it's been fixed." -- Jon Carroll
IBM is definately getting better with their hardware support, their people have been most helpful with getting me technical specs information for the drivers I maintain. Check out developerWorks for the stuff they've done.
:)
Now if they want to make me an offer to head up an IBM/Linux hardware support and development team
This isn't "advanced technology" they are contributing for the good of the world, it's messy legacy functionality they want to get into a publically maintained software base so that they don't have to pay for software maintenance for their legacy customers.
Hey now, that's not completely true. There may be some attempts to add legacy crap to the kernel, but the Linux community isn't going to accept crap into the source tree. A lot of the work we're doing is fixing bugs and adding features. Some features are accepted, others aren't. Some features have to be maintained in our own kernel tree, precisely because they support legacy crap, like you said. But those features never make it into the main kernel. I wish this attitude of 'IBM can't do anything good' would go away, it's not very helpful.
I do agree that some parts of IBM are exactly as you describe. IBM's a big company, and has a lot of different people with a lot of different perspectives and motivations.
But hey, judge the patch, not the person.
I've got to agree with others - IBM is in danger of throwing away most of its open source credentials if it supports the closed, proprietary hard disc locking system.
I support copy protection; I don't support patent-protected, proprietary mechanisms for doing it.
For the IBM open source software guys; thanks a lot and have a good xmas and new year - but lean on the other guys a bit please!
How about a bsdlicensed yacc?
Yeah, it's selfish, but you asked. ;)
If IBM where to opensource OS/2, then maybe we can start to get rid of the whole windows platform. That would freak MS a little I think.
In addition to sharing their expertise in development...
Testing, documentation and applications would help "us" and IBM the most. Granted, what they really want is better support for "big" machines, but that is going to take some time (and debating with the folks who want linux optimised for embedded uses).
I suppose I'm not too threatening, presently, but wait till I start Nautilus
As a person who has had personal (and quite recent) experience working w/ IBM on linux software, I think I can answer your concern regarding the possibility that IBM will " possibly taking credit for major chunks of the kernel and networking code for themselves": it's GPL'd, man. Although there will always be the pessimist who will argue that the GNU General Public License hasn't withstood a real legal battle (which allows for the possibility of some MegaCorp. lawyer finding a way to exploit it), it's obviously the most widely accepted OSS license around today. I can see absolvtely no motivation for IBM to want to try and find a loophole in the GPL, anyway. It seems to me that they sincerely want to help make a fscking awesome OS even better.
-Will the Chill
Legalize It.
Creator of RPerl, Scouter, Juggler, Mormon, Perl Monger, Serial Entrepreneur, Aspiring Astrophysicist, Community Organiz
I've worked with AIX in the past and let me tell you that we don't want SMIT and ODB. SMIT is the System Management Interface Tool or some such. Linuxconf is much easier to navigate.
We definately do not want to use the IBM ODB (Object DataBase) that they use for storing system configuration information. This is worse than the Windows Registry because there is no way to troubleshoot problems. I've had systems take 45 minutes to start up because the IP information in the ODB was incorrect. Not fun. I think that text files are the best way to handle system configuration issues. If you mess up the file, you can look in it to see the problem. You can still build tools to do the configuration and populate the text files. Databases for system configuration is unnecessary obfuscation.
Just my 2 cents...
Topic says it, thou, i dont think the source is available but the program is there for you to exploit^H^Hre it ;)
--
yush
Agree - device drivers or their specifications
are always welcome. Getting device specifications
has always been a struggle. Lately this situation
has eased.
Funny, I though that "shotgun" approach was what made Linux, and the whole Open-Source movement, so reliable and powerful.
http://www.tuxedo.org/, and read "The Cathedral and the Bazaar", if you haven't already.
I'm not saying its impossible.
It's surely dirty.
Take a look at db2:
1. You can use rpm directly (I don't like this)
2. Or alien
3. Or do a by hand instalation (alien to tar)
All the nice things that db2setup does (like creating instances) are unavailable. It's all handwork. Compare this with installing oracle 8.1.6i...
It is a sad day today, I didn't think of this as a joke, more of a good idea.
Matt Newell
Running KDE CVS with AA text, Alpha Blended icons, and lots of other cool stuff(some of it is even productive:)
What's to stop them doing it now?
.oO0Oo.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
I wasn't joking
.oO0Oo.
And I actually think it woul dbe a good idea for them. I'm not sure that they have no interest in the movie market. A move towards open source software means they have less invested in the software copy protection angle.
One of IBM'm main benefits of OSS is that they don't have to invest so much in OS development for their new hardware.
From only reading what was posted earlier about ATA OSS will suffer from this idea. Certainly I have yet to see any benefits for drive consumers, or producers for that matter.
If I have it right then it's like a printer cartidge manufactuer persuading paper people to make sure their paper can't be photocopied so that people have to buy ink cartridges instead of photocopier cartridges.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
how about getting them to promise to continue to build quality open use hard drives.
.oO0Oo.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
If Linus gets hit by a bus, the progression of Linux is fairly clear.
If Linus gets hit by a bus, I think there's a >99% chance Alan Cox will take his place. He's already Linus' right-hand man...
--
For "a beowulf cluster of these". Come on trolls, you're letting the rest of us down! Happy/merry all.
[Connection closed by foreign host]
Well, I'm assuming this is a sell the desktop request. I hate to say this, but IBM already attempted this with a superior product, OS/2.
And if you have not used the Warp 4 Workplace Shell and compared it to GNOME or KDE, try it before flaming. It has a beautfully consistent object model based on SOM. Of course, Warp 4 does not support the new hardware....you have to go back to a 1996 machine or so. One of the main reasons Warp could not win was MS Office did not run on it. Also, IBM is not good at selling consumer items like that. They will not charge up *that* hill again!
IBM will sell the server side vigorously (though that will not be 100 million any time soon) and perhaps form partnerships to sell embedded/pervasive Linux (that could get to 1 billion in short order).
With the exception of providing a good desktop development environment and reasonable productivity tools for developers, it's a waste of time to go after enterprises or home users without the key apps. That was the last war; it's over, even with the cheaper desktops. The enterprises are locked into Office; the home users into the games, productivity tools, education SW and other multi-media on Windows. The only other opportunity are vertical application desktops (e.g. terminals, network computers, kiosks, etc... dedicated to a single app. And that app needs to be ported to Linux.).
BTW, I don't work for IBM, but used to and was on the core team that LeBlanc mentions here.
Aide: Grant drinks too much to command an army. Lincoln: Find out what he drinks and give it to my other generals!
Linux on the desktop: Linux is almost there. All it needs is a bit of a push. I think your HCI people should download Linux-Mandrake 7.2, if they haven't already. This is probably the most cutting edge of the newbie distributions. This is extremely significant and adding a little bit of help here will significantly destabilize rivals like Microsoft.
Linux on the Thinkpad/(insert other IBM hardware here). It'd be nice if IBM stopped shipping Aptiva's with winmodems. Also I have recently been trying to get ahold of a Thinkpad. I don't like that I can't get either a linux distribution pre-installed or a blank HD without M$ 9x/NT. There are a few models at this page but the links lead to intellistations not thinkpads.
Linux & IBM PR: Ibm should keep helping the community and promoting itself as a "smarty pants/deep thinking tech company", and it will reap the rewards of this strategy. Shenanigans like CPRM on hard drives will negate these positive effects.
Wishlist: I need something to locally cache web pages, and keep them orderly on linux. I remember IBM had a product like this for Windows. I wouldn't mind this being OSS if it isn't already.
Please contact me if you'd like some further feedback, I wouldn't mind working for IBM at some future stage.
So true. I work with many OSS developers. Most of them have no clue what "software engineering" really is - or strives to be. Just because you get million monkeys with million typewriters does not mean you get Shakespeare. The only oss project that has been really well thought out is the GCC compiler. It has been ported easily to many more platforms than any other software. But maybe that says something about the people contributing instead of the movement as a whole.
The shotgun approach served well for those areas that are more attractive, have lot of fun, not very tedious, and results can be observed almsot inmediately: i.e. kernel development.
But the above is not true for applications that need a lot of research, tedious works, comprehensive testing and documentation for allowing further development: programming frameworks, software engineering tools (CASE and the likes), user interface, hardware development toolds (drivers), etc.
This is being demonstrated every simple day:
- There is/was no development tools comparable to VS, Delphi, CASE tools, etc. For example, give an example of CASE and DB development tools for Linux comparable to Oracle for MS or Sun.
- User interactive applications are much more tedious and need a lots of research compared to the kernel. Why Linux doesn't have yet a unique or compatible user interface system? Why the open source community couldn't develop a web browser comparable to IE5?
I can find a common denominator for those issues: it needs research, tedious work, it takes a long time to finish a single product, it requires lot of documentation and standarisation procedures. In few words, it needs formal software engineering approaches.--ricardo
sgis ddo ekil t'nod i
You're right of course... they can't open up the entire code base for AIX right off the bat, but I think that within 3-5 years AIX and Linux will be basically one and the same animal, and quite possibly with a lot of code-sharing. Plus right now most of the AIX code is optimized for the RISC 6000 processors, so it's not like you could download it and run it on a PC - you'd be buying IBM hardware to run it anyways. Background: I'm a sysadmin who works with OS/390 and AIX, and I think IBM is looking forward to community involvement to improve some of their drivers etc. The AIX ones have been known to have some problems in the past *cough understatement cough*.
This strategy would allow Big Blue to concentrate on hardware - I'm no IBM bean-counter but my experience leads me to believe that they make more bucks on hardware and services than anything else, plus the licencing fees for other software (especially middleware) such as Websphere, MQSeries, DB2, and even VM to run virtualized Linux on S/390 (aka Z900).
The OS is something they're pretty much obliged to provide for customers - they have enterprise-scale hardware and have to have an enterprise-capable OS with enterprise-level OS utilities to sell that hardware. It's not an option to sell the customer the dandy hardware and then tell them "Piss off and find an OS to run on it", they wouldn't be in business long (makes me wonder how MS got so big - they essentially did the reverse. Here's a * passable server OS *, now piss off and find some enterprise-level hardware to run it on). So if Linux slowly took that responsibility off of their hands for the most part, or helped them to maintain a 'feature edge' over other commercial *nixes it's worth the ROI for them. Plus I'm sure that they figure they can compete pricewise on hardware with Sun, EMC etc., so if the OS is commoditized and can run on anybody's hardware maybe it puts them into a more competitive position.
IMHO this is probably one of the smartest moves that IBM has made recently. Here's a link to a very interesting interview that indicates IBM is basically willing to open-source ANY code that the community feels is useful (it's with Irving Wladawsky-Berger, who is a heavy at IBM, I think this inteview was mentioned in a Slashdot article previously). I haven't lost my mind, I don't expect them to open up the entire source code to DB2 or similar profitable products, but I don't see AIX or many of AIX's utilities as being such.
IBM is not shorting the community on their commitment from what I can see. Just my $0.02
----
----
Slán leat agus go n'eirí an bóthar leat
ibm has been contributing to apache for quite some time.
a ch e.html
http://www.linuxworld.com/linuxworld/expo/lw-ap
the open source movement should be wary of ibm's "the enemy of my friend is my friend" approach. everyone should remember that at one time ibm was just like microsoft, except in hardware... changing interfaces to force out competition and create vendor lock-in, forcing many companies out of the biz. they were the preinnovators of computing monopoly.
jim
I don't defend redhat 7 for releasing something that wasn't perfect, but I don't believe for a minute that it looks inviting to businesses to become a part of the community when we roast the companies that are a part of the community.
Redhat was very professional, fixed the problems, or explained why they made the decisions they made, and moved on. Slashdot continued to rail against it longer than necessary, and made it into a brief holy war.
Why would any other company want to expose themselves to that kind of nonsense?
A host is a host from coast to coast, but no one uses a host that's close
Im sorry - your post is mindless shit. Do you have an actual fucking point?
Id like to ask IBM to move the LSB forward. Move development and target features to a Board of Uber-Super Geeks we can respect - who will not turn into Markatroid whores at the earliest opportunity. Not that I dont think Linus isnt/hasnt done a great job - but a little more co-hesion between distros will give something to shoot for by the Software Industry. People can still use the 'roll your own' philosophy all they like... and still meet an accepted Standard. Afterall - standards are a real power in computing...
Perhaps IBM will reconsider their strongly pro-software-patent policy.
Not a chance, IMO. I'm an IBMer (though *not* any sort of official spokesman) and from where I sit it appears that patents of any and every sort are of great importance to IBM. They're both a status symbol and a very practical lever in legal disputes.
OTOH, at least IBM rarely (I've never seen it) chooses to enforce software patents for direct financial gain, and is generally quite willing to offer very inexpensive licenses, as long as it doesn't appear that the legal leverage provided by the patents will be needed. It's been explained to me (informally) that the main reason IBM accumulates so many patents is so then when other companies sue IBM over claimed patent infringements IBM can just point out the IBM patents being infringed upon and settle quietly out of court, usually just with a "we won't complain if you don't". At the same time, IBM is extremely careful (sometimes irritatingly so, from my viewpoint) to avoid stepping on anyone else's patents.
To summarize, I think that, currently, IBM's collection of patents is fairly benign. However, future management may look on that patent database differently.
In the spirit of full disclosure I should point out that I'm moderately opposed to the notion of software patents but enough of an intellectual whore to be perfectly happy to submit software patents of my own and collect the associated bonuses.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Employing people on large Open Source (FreeBSD,Linux,GNU) projects, open up sources.
How about demanding that IBM withdraw their Content Protection of Recordable Media (CPRM) proposal to the ATA specifications body? As Alan Cox and Stallman have both said, this would mean the death of Linux and all Free Software/Open Source.
IBM has a lot of gall to be buddying up to the Linux community, while proposing industry-wide hardware standards that advance the intellectual-property fascist agenda that would obliterate accepted consumer free use practice and basic computer science principles such as file system abstraction. You couldnÕt even backup your hard drive if these standards are implemented!
Sure IBM wants to work with Linux, and Microsoft wants to work with Java. Wake up and demand that IBM back-off this proposal, or the party is over.
Check out the coverage here http://www.theregister.co.uk/content/2/15684.html
As Cox said, ÒWelcome to the United Police State Of America."
Instead of providing solutions branded just 'linux'
they should help promote the projects/distribution that made the solution possible. So we dont have another redhat, or valinux give the impression of been 'the linux company'.
We want to be part of a independant community with commercial companies support. Instead of a company with community support.
IBM can contribute anything they like. They cannot, however, hinder our Linux kernel, as any contributions they offer become ours. No IBM control.
However, it is entirely possible that linux developers and IBM devolopers will be writing new code for the new linux kernel. IBM may be able to take this code and then put it into their OS before the Linux version can be released under the GPL thereby circumventing the GPL and getting free linux programming to boot!
--
303infinity Rocks, buy their CD's.
No, they're not wrong because Microsoft did NOT buy the original Windows code. It was written by Microsoft people at the Microsoft building at 10700 Northup Way in Bellevue, Washington. Perhaps you're thinking of MS-DOS 1.0 which was derived from Seattle Computing's QDOS (and didn't support hard drives, pipes, directories, etc. which Microsoft added)?
Examples:
This post is right on the mark. Problem is that IBM has an extortive software licensing charge policy on which they are addicted for revenue, and they have not demonstrated much wisdom in general in the 390 group on this subject. For example, if you get an older lower end machine, a Multiprise 2000, and you get an equivalent model of the Multiprise 3000 with the same machine capacity (MIPS), you can pay $5000/month on the 3000 whereas you would pay $11000/month on the 2000. IBM's 390 group is insulated from the real world of most linux people. They need to get a clue and take Mr. Maynard's suggestion seriously, with no commercial restrictions on such a license under a certain machine capacity level- say 150 MIPS.
I don't care what they do, as long as it makes linux better. Let's not be picky here it's a major company if they spend one red cent on improving the linux experience I'm happy.
All my co-workers that have an interest in Linux, all ask the same question, Is there an office package for Linux? My department usually creates all techical documentation for our account, and we all must use Microsoft Office. We all admin over 6000 *NIX systems, and most of us would rather stick with a *NIX OS! A native Linux office package is what I want to see. One that will allow you to work with the most commom document types, spreadsheets etc. Let it be able to interact with the database of your choice (postgres, mysql etc). One that portrays the true blue IBM technical ability. If IBM could capture that market on Linux, I believe that could be the "sky is the limit" moment, and help get other major software companies producing more everyday useful stuff.
You really didn't answer anything... What I asked is WHY linux got the limelight in the first place. I understand companies go more to the popular platform rather than the 'superior' one, but how and why did linux get to be the press darling it is? There's always some technology that's caused the adoptation in the first place (i.e. beta had better picture quality, it only supported 1 hour of video on a tape, this technology gap was how VHS got the edge).
The only solid point you got was the GPL... BSDs run on S/390s as well, and FreeBSD has the SMP support anyone would want. You're perfectly welcome to use FreeBSDs kernel with SMP support and stick that kernel in an otherwise completely OpenBSD system without major hurdles. I admit a distro is easier, but hey, it wasn't long ago that there was no linux/GNU distros period.
You've been spouting Rhetoric again buddy... The SINGLE OpenBSD exploit you mention is insanely rare. Besides, it's a root exploit only in the most perfect of conditions. You can't take the one OpenBSD exploit and compare that to the tons of Linux exploits comming out every day. Besides, FTPD is not enabled by default, you have to know what you are doing and go manually enable it. You say Linux smokes FreeBSD but it's simply not true. If you've got SEVERAL RECENT benchmarks to prove it you go right ahead and link them... As far a portablility, NetBSD is currently beign ported to the iPAQ handheld. Besides, what good is a huge, bloated linux OS on a handheld anyhow? Not that linux can't be small, just that it isn't and never has been thus far.
I was just pointing out where the BSDs beat linux. They do just as Damn well in a linux side by side comparison. The single thing linux has that OpenBSD comes up short on is SMP. But! SMP is rarely needed (only on extremely high end machines) and if you want SMP, swap the OpenBSD kernel with the FreeBSD kernel and you've got the best of both worlds (or ABCD XYZ if you prefer)
Okay, you said there are reason to go with linux, but other than the GPL you haven't given me one...
You really can't say FreeBSD isn't as featureful because it can run linux binaries unmodified, and can compile source written with linux in mind without problems (assuming the right libaries are in place) You also mention firewalls/proxies but the fact is, FreeBSD does do a better, fast job as a firewall/proxy (in fact, that's the main use of freebsd) eventhough there aren't any versions that fit on a floppy that I know about.
Please remain calm, I do not intend to start a flamewar. Why do soooo many people-as well as the press, companies, and so on-flock to Linux and not another Open Source OS?
OpenBSD is the most secure OS on the planet, FreeBSD outperforms linux and has a more stable filesystem, and NetBSD is portable beyond belief. So the question is, why does linux get the spotlight? I really would appreciate a sane and logical answer from anyone who thinks they know.
Many who have grown up using a Graphical User Interface (GUI) look back on Command Line Interfaces (CLI) as old and primitive. Well they are half right because a CLI is also very fast and powerful. Unfortunately most CLI's have VERY arcane phrases and words to do the most simple of tasks. Why cannot IBM work on a Natural Language CLI where say one could type "Move file: "Quake" from desktop to games folder". This makes a lot more sense then what one has to type now. I know in most Linux systems you can already type "Shutdown now" but more progress needs to be made. Also those who are used to present CLI's wouldn't have to use this they could stick to bash or tsch if they wanted to.
Mac OS X and Windows XP working side by side to fight back the night.
While X is great for networked workstation clients, it's a bit overkill for single-user desktop machines. Would it be possible to GPL some of the stuff from OS/2 toward such a single-user GUI that talks more directly to the hardware (like Windows or Mac) as opposed to a networked GUI like X? (My explanation is probably weak - Somebody please help me out here)
Also, does IBM hold any TT fonts which could be GPL'd with this too?
close enough
Its obvious by now that Linux is very good for "big iron" and embedded projects. It was IBM who has put Linux on its smallest platform ( a wristwatch) and the world's second largest super computer (largest commercial).
~Either IBM will actually commit significant coding resources/assets to linux [necessarily under the GPL], or they won't. I of course hope for the latter.~ I think one would hope for the *former* in this case..IBM committing resources towards linux and not just reaping the rewards.. at least that's MY hopes in all this.. ;
.sig under construction
Kerberos is still open - widely open...
Just because Microsoft has filled few places with their proprietory extentions - that doesn't make an open protocol proprietary...
Ask Redhat - they install kerberos as default if memory serves correct..
Hetz (Heunique)
The most important thing for Linux is always market penetration, especially market penetrations in large companies. The best and simplest thing IBM can do for Linux is to sell it, especially sell it to large companies.
Then there are a lot of interesting things that are part of AIX. These ideas are usually ahead of their time and they are badly implemented.
Take the AIX system administration tool SMIT for example. It is using a binary registry for much configuration data which is a pain in the ass to maintain - text files and a binary cache regenerated from the text files dynamically, as KDE does it, is much better.
But SMIT is the only system administration tool which I have ever used which is actually helpful and easily extensible. SMIT is based on a number of command line programs to edit system configuration, things like adduser and deluser, but for anything on the system. SMIT generates command lines to execute system administration functions and you can preview these command lines before they are executed. So while SMIT is a GUI tool for the inexperienced user, at the same time it educates this user and brings the user onto the path to automated system administration.
SMIT has menus, and these menus are nowhere hardcoded. If you know the SMIT extension API, you can add menus and functions to SMIT and integrate your own software into the tool. That makes SMIT a truly universal administration tool, which can be taylored to a sites needs.
And finally SMIT keeps a log of all changed done to the system in the form of an endless shell script. So if you actually use SMIT to do your everyday system administration, you get a complete protocol of all changes to the system since the last install.
SMIT comes in two variants, SMIT with an X interface and SMITTY, with a tty interface. Both have a command line interface which allows you to shortcut into any menu screen from the command line so that you do not need to navigate through a menu hierarchy for the 1000st time, but can jump directly into medias res.
SMIT is broken in current AIX, but it can be fixed - get rid of the binary object database aka registry and enter a KDE style binary cache for textual configuration file to speed access to large parseable text files.
Also, SMIT was lacking cryptographically authenticated and encrypted remote administration capabilities the last time I checked (early AIX 4.x releases), so you still had to rlogin in order to administrate a remote station. A mass administration utility that can remotely control each single remote workstation as well as groups of them would be a great extension to SMIT.
There are other parts of AIX which are worth a very close look and evaluation. Much of it does not really fit the current Unix model as it is, but is extremely interesting and just needs a good ripoff session and a heavy dose of realworldly interaction in order to become a great addition to Linux.
© Copyright 2000 Kristian Köhntopp
All rights reserved.
Simulatneous releases of all software (server & desktop) for linux as well as NT/AIX (Note the 'desktop' software should include a Notes client and Lotus SmartSuite). At work we are a websphere shop ( :-( ), and I can't get Enterprise versions of WebSphere for Linux, so I can't test all of our apps on equal platforms to show the phbs that Linux can do a good job (vs. AIX/Solaris running an Enterprise version of WebSphere). I can show results vs. Advanced & stantdard edition, but then the phbs snidely says, 'Yeah, but we use the Enterprise version!". Anyway...Just my little gripe...Even though I totally dig what IBM is doing.
Also, they need to comit to porting a lot of their software to S/390 Linux. I think it's great that Linux runs on the S/390, but other than opens source software (and this is one of the powers of OSS software, I'm not knocking it), what can I run on it (OSS would be fine for me, but for most corps, it ain't enough).
Also....A Notes client & Lotus Smart Suite...If IBM did a NATIVE (i.e. Not WINE) port of Lotus Smart Suite (taking advantage of KParts or Bonobo) and integrasting with one or both of the desktops, they would sweep the Linux Office market. Add to that a resonable price for the Suite and I could get a LOT of people to switch to Linux....
Palin...
Only two things:
1) Open Source the WorkPlace Shell! Every other GUI I've ever used, including Gnome & KDE, feels clumsy by comparison. We could take that, prettyfy it up for the modern world, and watch Steve Jobs turn green with envy.
2) DRIVERS. DRIVERSDRIVERSDRIVERS. Did I mention DRIVERS?
Autodetecting hardware and installing the right drivers. The fantastic WPS GUI. Oh yeah, I'd be in heaven if that was in Linux!
As far as things they could do other than what they already own:
Add more people to the job of auditing the Linux code. There's no reason OpenBSD should be more secure than Linux with all the resources the Linux community has on hand. That's just absurd.
Oh, I just remembered - the OS/2 BootManager was pretty nice. Not sure how it compares to GRUB (probably not well these days), but it'd be nice to have it available...
I think one of the best things that RMS can do is to make sure that the GPL and the IBM public license work together. I don't think RMS should give as far as ideals go, but they should probably meet and work out the license differences, so that IBM Public License contributions are fully integrateable with everything else.
Engineering and the Ultimate
by "supporting linux", we're not talking about porting proprietary programs, we're talking about extending linux itself - which they have done quite well.
Postfix - the greatest mailer ever - checkout postfix.org
Port to S390 - again a great accomplishment
Jikes compiler - great stuff
JFS - this will be killer when its done
Releasing open hardware boards for LinuxPPC
Modifying the kernel so it can run heavily multithreaded apps better
There's probably a million things I haven't thought of, too.
These things are great for _every_ distribution. Everyone has benefitted from IBM's help, even if their proprietary stuff only runs on RPM-based distros.
Engineering and the Ultimate
Actually, they can only use their own contributions for that. They cannot use the contributions of other Linux people for their own projects, even if they do that before a "release"
Engineering and the Ultimate
Perhaps IBM will reconsider their strongly pro-software-patent policy. They were the ones who started this nonsense in the US and have lobbied for its extension to other parts of the world.
OTOH, pro-swpat attitudes often come from a company's legal department, which has a life of its own separate from what the techies think, so I wouldn't be too hopeful.
-- Ed Avis ed@membled.com
As great as it would be to incorporate great new techniques into linux and thereby spread the use of linux among people, that should not be the end-all of the interaction. Instead, that should be the result (or byproduct) of an increasing dialogue between linux developers and IBMers.
:) Ds.
Linus et.al. has many times indicated an interest in expanding the playground for linux into 'big iron' as well as embedded systems. Why not listen to IBM, and aske them what their customers would find lacking in linux as it stands? By becoming more informed about a market of which most of us really know little about, better choices can be made, old mistakes can be avoided, reinventing of wheels will not take place, etc.
That is not to say that linux developers should just lie down and think of Finland when IBM comes knocking; the goals for linux are at least partially different than the goals for a commercial system, and the development process tends to be very different. That would be lessons IBM could do well in learning from a fruitful dialogue (and they seem to have picked it up fairly well already).
New filesystems, failover capabilities etc. would then fall out fairly naturally.
Ps. open sourcing the OS/2 drivers would make for a nice midwinter gift
Trust the Computer. The Computer is your friend.
IBM can and should fork Linux. It'll be a cold day in hell before Linus accepts a patch that isn't emailed, inline, not as an attachment, under a few K, and can be understood personally by Linus in under an hour of reading.
If Linus gets hit by a bus, the progression of Linux is fairly clear. If he decides to reject patches because he's having a bad day, the answer seems a little less clear.
--
I've finally had it: until slashdot gets article moderation, I am not coming back.
IBM has some great software development tools like VisualAge for Java & VisualAge for Smalltalk (I'm specially interested in a port of the latter.)
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
They just need to understand that they oughta keepa their handsa offa da source , and they'll get wonders of cooperation, especially if it is aims squarely against Redmond.
Besides, nothing could prevent them from being generous towards Open Source developpers (k00l hardware, seminars, whatnot) as a genuine, legitimate R&D expense.
Ever wondered why Minnesota Mining and Manufacturing offers so many ingenious products? Well, it's their R&D policy that does the trick. Every researcher is forced to set aside some 20% of his research budget to investigate side effects, and they are forbidden to use it on their main research activities.
They could very well do the same for Linux.
--
Game over, 2000!
--
Game over, 2000!
What the interview really says is that IBM is not planning to open source AIX, however it is willing to open source any piece of it that the Linux community considers useful. It's claimed that dumping all those lines of source out there at once "wouldn't be prudent".
I've been working with IBM software on Linux and AIX (and even NT) for about a year now. Mainly WebSphere + Visual Age, DB2 and MQSeries. I think there are a few places IBM could improve their product line.
.ser file and it essentially generates and compiles the code needed to make a BMP Entity bean on the fly. I've looked at the code for JBoss (a GPL'd EJB container) and this is definately something only WebSphere does. Also, I'm sure this will come with EJB 1.1 support, but it needs ejb-ref.xml support and application.xml support in the deployer.
WebSphere: First, EJB 1.1 support, we need this now! Better integration with Unix admin utils. WebSphere is either admined through an undocumented XML config file or through a GUI that works sometimes. We need command line tools that work with websphere. Sometimes administration must be done remotely through a firewall, and there's no port available for the GUI to pass through. Secondly, what's with that CMP Entity bean hack? On WebSphere, you deploy a jar with your EJB and
Visual Age: no integration with CVS, I want to use vim as a replacable text editor rather than the one in VA (this might be a stretch, but hey, it would be an improvement). Try using less than 80+ MB of memory when it runs. More support for standard packages and less compiling on the fly. It takes me way too long to write code in Visual Age, especially EJBs. I wouldn't even have touched VA if the EJBs I wrote in vim would work in WebSphere to begin with. How about better docs on how to do this. The whole reason I write J2EE apps is portability. It seems that IBM is trying to pull an M$ and provide something that looks like it conforms, but doesn't really.
DB2: Actually, I have very few complaints about DB2, I kinda like it. One place for improvement though is CPU usage. DB2 appears to push everything through the kernel I/O buffers, since it doesn't support raw devices. I've watched x86 boxes jump to 100% CPU utilization during even basic queries. The simpler the query the less time it stays at 99%, but with a lot of queries this gets very worrisome for management who is making the final call. Man pages for the utilities. I understand that there are docs available in PDF, HTML and on paper, but there are always times when you need that man page, simply because of the convienance. The db2 shell should support bash rather than only ksh, or there should be docs on how to use bash in the db2 shell if it does currently support this. Most Linux guys are used to bash and are pretty quick with it. Solaris 8 also uses bash, so now more than ever Unix guys are using bash.
One other database related suggestion would be tools for Linux/AIX like ERWin from CA and more analysis and DBA tools. IBM may have some, but I haven't found them yet. Maybe support for the GNOME database GUI (Gnome Transcript?).
MQSeries: Could we get a release date please? Also in general better docs on how to use JMS would be nice. It would also seem like a good idea if IBM went ahead and built EJB 2.0 support into WebSphere so we could use Message Driven Beans with MQSeries. Many of the same things with DB2, I'd like to see in MQ as well (man pages, bash utils).
Also some nice information about how to integrate all of these together would be nice. I've never had just one of these to work with, it's always the whole sheebang. I've developed on these platforms both internally at IBM and with other clients. Quite frankly you need better information about these tools, b/c both inside and outside of IBM there seems to be a lack of knowledge about how to use all of them. I've been on my own a lot, and the tools have typically been non-intuitive about how to get things to work. I've read the RedBooks, the docs at IBM website and the newsgroups. It was still difficult.
Another nice thing to see would be Linux native on the Intel NUMA boxes. That would just be cool.
I'd like to see IBM succeed in the Linux space. I think they make damn good hardware and some of the software is pretty sweet (DB2), but others are going to get eaten for lunch (WebSphere, VA) in the Linux space unless changes are made. IBM needs to reach out and make it easier for the Linux hacker to learn IBM software and get it working as well as some of the systems we have put together using Open Source stuff (PHP+Apache+MySQL or Perl+Apache+PostgreSQL, or Java+Jboss+Tomcat+Apache, or some combination thereof), IBM needs to integrate better into the Unix world, (I haven't spent much time in the AIX world, so the admins may look more like NT guys, HHOS).
I have a few other suggestions, but those are slipping my mind at the moment, this should be enough for now...
Discuss amongest yourselves. I'm feeling vaklempt.
Arrogance is Confidence which lacks integrity. -- me
I know device drivers aren't as sexy as filesystems or memory architechtures. But when IBM got their Thinkpad "Red hat certified" and then had to admit the modem was incompatible, they looked pretty dumb.
IBM makes a lot of hardware. Certainly and important way they could help out would be to make sure it all works with Linux
OK, so what do we need from them? Open Source software is built on open standards, depends on open standards. We've seen software vendors 'embrace and extend' standards to the point that they've ceased to be open (can anyone say 'Kerberos'?)
The biggest christmas present any of the big comupting companies could give the Open Source community is a cast-iron, board-level committment to support and comply with open standards wherever appropriate, and most significantly of all, never to 'extend' an open standard without making those extensions freely available under the same terms as thr original standard.
I'm old enough to remember when discussions on Slashdot were well informed.
So IBM is opensourcing AIX. Just because it is open source doesn't make it Linux. Its still AIX.
Sure, once some of the neat drivers, modules, and applications are open on AIX, there will be a cross pollinisation between Linux and AIX. But that isn't going to make AIX into Linux. They will remain separate systems, both will be better for IBM's moves.
The opensourcing of AIX should be hailed as a great triumph of the free software model. Now instead of Linux as the only project to point to, we can point to a completely separate system. Next, I'd love to see Digit^H^H^H^HCompaq open source VMS, just for nostalgia's sake.
IBM makes most of its money in the services arena. AIX sales were a tiny fraction of their revenue, so opening the source is not going to hurt their bottom line in any way. What would be really good is for their services group to start supporting Linux installations for large organisations, that would bring Linux into the corporate mainstream.
the AC
Hemos is like...sci-fi fans;he thinks technology is cool, but he hasn't bothered to understand the science it's based on
Disclaimer: I'm an IBMer.
IBM's knocking on Linux's door. Don't let the opportunity get shot down. IBM's a business, while Linux was NOT developed under a business model. Find a bridge, and Linux will have the oportunity to join up with one of the biggest names in the computer industry.
No need to find a bridge, I can tell you, from what I've seen and heard, IBM is building a 10 lane highway. The connection is there and is not going away anytime soon.
IBM is trying real hard to gain mindshare for the S/390. Putting Linux on it is a major step in that direction: How many of you thought of the S/390 as a practical Linux platform?
There's one thing that IBM can do that would help both the Linux community and itself: Hobbyist licensing for its mainframe OSes. Very few folks would ever run a single Linux image on a 390. By far and away, they'd use the VM/ESA operating system, which allows multiple virtual machines on one physical one, or VIF, a cut-down VM intended for hosting multiple Linuxes. Unfortunately, a VM/ESA license costs about $25K on the smallest machine, going up from there. (I don't know how much VIF will cost.)
Before you say, "But no hobbyist will run their own S/390!", I'll point out that the Hercules S/390 emulator will run VM/ESA just fine. Now that running an S/390 is within reach of the hobbyist, IBM would gain a lot of mindshare by letting hobbyists run real configurations on their Linux boxes and get familiar with how it all goes together.
--
Disinfect the GNU General Public Virus!
the only thing stopping me from doing my job on Linux is Microsoft Exchange Server...inability to "natively" (i.e. without POP3/IMAP) interoperate with Exchange is a showstopper for any corporate installation.
HP Openmail is also available for Linux right now and supports Exchanage based mail delivery to Outlook and Entourage, with automated migration options.
This is just a personal thing. I'm no purchaser, and I have no power anywhere. So, by helping me, you'll get relatively little return on your investment.
:)
:)
;) But in all honesty, QT/e looks to be a good way to go. Develop some games, some productivity apps, use QT/Windows and QT/X for the synchronization software, and you're all set.
:)
I am, however, an Average Joe(tm), so if what I want is what everyone wants, you could make a big return on your investment
First, I'd like to congratulate you on the work you've done so far. I don't know a lot about the in-depth stuff, but I like what I've been seeing. Good job.
What I'd like is to run Linux on my Palm Pilot. The Palm is remarkable. An example in successful adaptation. The UI is perfectly suited to the device, the speed is great(considering it's running on what, 16MHz?), and the hardware is cool.
I think this is something that Linux is perfectly suited for. Look at TrollTech's QT/Embedded. Amazing piece of software. What I'd like IBM to do is build the hardware, make some basic software, and sell a few million units. Use a standard UI tookit(like QT/E, or GTK+ nothing specifically "designed" for a palmtop; I'd like to, one day, be able to run the same apps on my palmtop as I run on my desktop today).
Now, this is just the end-goal that I'm talking about. But all the infrastructure work that would have to be done, if done right, would be great.
So, to sum up:
I'd like a Linux handheld, about the size of a PalmPilot, with similar cost($~200).
Software to run on it. I'd suggest QT/E, because I can read the code
The reason I'm suggesting this, as opposed to the other things I'd really like to see come out of IBM, is that I think it's feasable(sp?). I honestly think that within two years, IBM could release a handheld computer that I'd buy.
Thank you for your time, and sorry if I'm a bit incoherent. Working off a hangover
Dave
Barclay family motto:
Aut agere aut mori.
(Either action or death.)
Barclay family motto:
Aut agere aut mori.
(Either action or death.)
Get real! We should be asking what we can do to help them. They are the ones helping us here and you think we should take that as a sign to hold out our hand and ask what else they can do for us? How ungrateful are you?
Prevent email address forgery. Publish SPF records for y
It seems to me that IBM "Gets" the open source movement and how to make money with it better than any other company in the industry, possibly even Redhat and the other Linux companies. And in many ways their goals align quite nicely with ours.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
But you still can't get a Linux version of Lotus Notes. And running it in wine works, but is very unpleasant -- in managed mode you get menus that pop up and never go away and in unmanaged mode it forces itself to the top in all of your virtual desktops.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
If I could be selfish and ask for three things from IBM doing Linux work, they would be:
1. Compiler technology to improve code generation of the GCC tools (including gdb).
2. TrueType Fonts. Surely IBM has some fonts lying around from years ago that they would not mind making public.
3. Video CODECs, (integration into the OGG/Vorbis project).
IMHO, IBM Should focus on what they need in Linux to migrate their AIX userbase over to IBM Linux first though (attempting to not NIH existing resources). Supporting both AIX and Linux is a long-term support and resource nightmare.
There is actually already a windowing system called Y windows.
find it here
X isn't going to be replaced anytime soon. It doesn't really need to be either... Xprotocol was designed to be added to and updated, it's just that people haven't had the motivation to do it until recently.
I trust IBM as far as I can throw them -- which is to say, about as far as they're willing to be thrown given their own financial self interest. They (like many others in the Open Source movement) see value in supporting the movement. As such, I think that their support is about as sincere as anybody else who sees supporting the Open Source model as valuable.
------
In a lot of ways, it seems to me that the open source model is going back to the original copyright law intent -- by giving people an incentive to create a work that would ultimately go into the public domain. Now we have laws which, for all practical purposes, never let any creative work go into the public domain. I think that the Free and Open Source movements are now proving that the original idea of putting intellectual creations into the public domain is socially -- and even financially -- a sound concept. IBM is simply riding this horse forward.
`ø,,ø!
Free Software: Like love, it grows best when given away.
Yeah, they are older than time_t.
Well, if you read the interview you will see that it is clear the IBM are talking about Linux and not the other *nixs. I think they understand better than most what they should and shouldn't do in this new arena, and it seems that what they want if to develop a Free software system which allows them to have a common OS BUT they do not see the BSD alternatives as worthwhile becuase they would risk everything they had done been taken and reused in a closed way without reaping the benefits. I think it is clear that IBM understands the deal with Linux (and essentially the GPL) which says "use and/or help, but tell your users what you did". If you think about what IBM is known for and good at, this suits them down to the ground. In 20 years the MSCE's of this world could be the ILCE's (IBM Linux Certified Engineers), all the big boxes could be IBM machines running Linux (with Sun et al free to come in their to compete, but if they want the benefits they must join the code sharing club).
Bottom line I would NOT want IBM to release under a BSD style licence, they would hold back on technologies that MS et al would like because they could take them and give nothing back. We all need to work together and forget about using open code to make money, if you want to make money from binaries, start fresh and write something new. Do you really want MS to embed ViaVoice into windows and pump lots of money into research....only to use any benefits of this research to strengthen windows giving nothing but windows binaries to their users? Or would you rather that maybe IBM will release something so special that MS will be faced with the choice of working with GPL style software, staying behind or spending massivly on research to replicate it? GPL code is usable by everyone, you just gotta play the reciprocating game...in a season like this have you got a problem with giving?
Never underestimate the dark side of the Source
Lots of interesting and valid points....ask any of my friends and they will tell you I am a devil's avocate so take everything I ever say in that frame of mind. Before I redress your points may I just say the reason I like /. is because these argumetns can take place freely.
Firstly, if you follow the story link to the interview and scroll down to the second question you will see IBM say Linux and NOT BSD. That's their call but I agree with it....
All of your ViaVoice points are valid (I assume you are correct on L&H). I was taking this as an example. If we asked for code and worked on it, would we want a company taking all that work and releasing a closed version, possibly with some useful feature we want? I think this is the BSD/Linux fissure, but really it is liek software patents and IP to me... I just can't see how it makes sense to allow people use your work without benefiting you if you can even consider an alternative (release GPL). I have to say though the ViaVoice is not GPL, but would it be if we asked? If IBM GPL'd ViaVoice, how much attention and use would it receive. Would Linux strive ahead of Windows thanks to this (come on now, voice control and recognition is a BIG thing to mere mortals)? Is it the thing we should ask for (it ain't from AIX, but if they saw we would go for it...)? If it was GPL, what would MS do? Any great hacks would have to be reverse engineered or accepted as missing features. It would be fun to watch money versus Freedom! If ViaVoice was GPL....well then we can watch it go into windows....what's the point?
GPL code is usable by everyone who is willing to use it! If you want to open the code further (i.e. allow it to be closed) then you can't. If you are willing to allow the code to be owned for eternity by it's users then you are sorted. If you can't enter into the GPL spirit you can stay outside and leave our code alone (tone provactive so sue me). Again, why should anyone be alllowed to close given away code (if it uses the code to provide function then it is not a unique invention, if it is a unique invention make it independant of any other code other than usable libraries).
Finally, I hate "Open Source". You can fuck off with it afaic. If it is Free (i.e. I can do what I want with it for myself, and give it to other people, and no-one can ever deny anyone any part of anything based on it) than I am happy, other wise get stuffed. I want software to be as good as everyone who is interested in it can make it. I don't want software that is as good as the current market environment allows.
Never underestimate the dark side of the Source
Does IBM really support Linux? Or just RPM based GNU/Linux Distributions?
Have you ever tried to install DB2 on a Debian box? It's possible, but only after a lot of tweaking, and, you are on your own... Probably the same is true for other non-RPM based distros.
On a web event done by IBM I asked for DB2 support for other distributions. They were kind enough to answer (that's good), but the answer was most disappointing.
I think this problem exists for evey IBM product for Linux.
So, my first whish is: Support for non RPM distributions, like oracle BTW.
I'd also like to see DB2 client libraries freed, so that DB2 could be easily (out of the box) supported be free software. Like gnome-db for instance.
First, more device drivers for non-IBM hardware. IBM has made the fantastic decision to make its own hardware work with Linux. Could IBM use its superb business skills to persuade other hardware manufacturers to do the same?
Second, better documentation. IBM documentation tends to be very thorough and accurate. Linux documentation tends to be ... do I even have to say?
Third, an excellent GUI. Linux will never dominate the world if its presence on the desktop is insignificant. Developing an excellent GUI requires closeness with end users and top ergonomics skills--both of which IBM has in abundance.
Whatever IBM does though, let's be clear that we are extremely happy for it.
At present, Linux relies too much on a Monte Carlo shotgun approach to software engineering (thousands of folks randomly hammering on software to shake the bugs out). What is lacking is the industrial strength support for formal engineering methods of design, testing, and QA. IBM has plenty of expertise in this area. My hope is that IBM will step in and plug the the hole.
Of course IBM's help in educating developers in the benefits of formal industrial strength development methods wouldn't hurt either. SGI has done some work in this area, offering a suite of regressions tests for the kernel. However, how many people know about them, how many actually use them?
In the case you mention, IBM acquired the company. They bought it, they owned it, it became a part of IBM; although claiming that the original innovations of that company had come from IBM is a bit of a semantic stretch by implying that the research resources of IBM had been capable of achieving those results, when in fact IBM had simply bought it, it is not precisely technically inacurate.
As linux is GPL'd, IBM could never do such a thing with it. I hope that IBM's motives are sincere, and I have no reason to think they neccesarily aren't... The comoditization of OS's may well be in their strateig favor, if they want to focus on being a hardware, and software, and services, provider.
The real results will speak for themselves. Either IBM will actually commit significant coding resources/assets to linux [necessarily under the GPL], or they won't. I of course hope for the latter.
---
man sig
---
the pen is mightier then the sword. the sword is mightier then the court. the court is mightier then the pen.
The linux group (software) is separate from the PC company group (hardware).
/. beat up RH on the 7.0 release, talk about eating one's own family!)
Remember, in a great giant company, these two separate worlds may not see eye to eye.
I had the opportunity to speak to one of the higher-ups in the pc company, and he said to me that he's looking to linux, because they hate Microsoft, but that the community still appears small and splintered. (Look at how
The point is, Software Group is embracing Linux. PC company (which has been bleeding money for years) is scratching their heads wondering how to make a buck at it, like any good business should.
(I don't speak for IBM. my opinions are my own.)
A host is a host from coast to coast, but no one uses a host that's close
How about they head a project to design a replacment for XWindows. They could form a group and set about creating a new modern standard...
And just for kicks call it YWindows as in: Why Windows?
90 out of 100 SAs I talk to equate CLI with the 1980s
;)
Why dont you correct them...? If that dosnt work: stop talking to them
You've got it all wrong. We don't need more enterprise stuff; we need IBM to open up (or at least port) their Lotus SmartSuite so we can have another office suite to sell the suits on. We need desktop and workstation tools, people, and IBM should help support that.
However, I'm a long time user of both Linux and AIX. Linux from 386's to Beowulf clusters. AIX from tiny F40's up to 30-50 node range SP/2 complexes. I have an overall good view of IBM and their products. Here are the exceptions to that good view, and they are strong:
- AIX is horrible. Top to bottom. For some quickie examples: The TCP stack is horribly inefficient, often the bottleneck as opposed to the network interface speed; The text tools (sed, grep, awk) have ridiculous limitations like "no lines longer than 2048 characters"; Most of the crucial system configuration information is stored in a proprietary binary object database (as opposed to textfiles in
/etc). The commands for interacting directly with this database are largely hidden and undocumented.
- The SP/2 clustering software sucks. For examples: Their redundant failover for control workstations (HACWS) is vastly overcomplicated, and has had a continual chain fatal problems in every release I've ever used; Their monitoring tools often reporting utter lies (spmon and freinds); Their version control and upgrading proceedures for the OS and the clustering software are beyond abysmal, you're better of just re-installing everything for each minor change.
I could go on for days.. but you get my point... don't let that AIX or SP/2 code into the Linux codebase before you check it out _really_ well. I bet a solid 65% of their stuff wouldn't pass muster in the Linux community.11*43+456^2
"the only thing stopping me from doing my job on Linux is Microsoft Exchange Server...inability to "natively" (i.e. without POP3/IMAP) interoperate with Exchange is a showstopper for any corporate installation."
Well, I know a product that may solve some of your interoperability problems. Look into bynari products here.
The have client software for linux called 'TradeXCH' that will communicate with exchange and outlook. They also have their 'Trade Server' product which is a standards based suite of mail, directory and collaboration services.
As most of you are aware, IBM is one of the companies looking into the new ATA spec. Keep in mind, this is ATA spec and thus SCSI is not affected. Even if SCSI is an unencumbered alternative, any of the new ATA drives that creep into the organization will cause trouble.
We as a community, and by that I mean all free software advocates - the BSD's, Linux, LPI, GNU, FSF, etc.. you name it, should unite and form a coalition against this new technology.
More specifically, Linux companies and organizations and their their respective leaders should pressure IBM to keep their hard drives open if they truly wish to be part of the open community that Linux represents. IBM may be a big company, but they would be stupid to alienate themselves from the very developers and support of the Linux community should they persist in making this new ATA spec a reality.
I have been following the Linux/390 situation for a bit of time now, and I find myself left with some questions and concerns. I write this to ask thoughts of others. It's all fine and good that IBM is actually supporting the notion of Linux running on the S/390 and their new 64 bit zSeries mainframes. What are the real benefits to the Linux community as a whole, and is there any real commercial relevance to this whole thing? For example- as far as I know, Linux/390 still can't operate with most IBM I/O devices, or take full advantage of the rich error and recovery approaches built into the S/390 hardware. And most astonishingly of all, it still can't deal with a tape drive directly. On top of all this, like the worst things which the linux community has seen in recalcitrant video hardware manufacturers, IBM refuses to fully document their network interface hardware and provide open source drivers for it. In addition, much of the IBM hardware itself in terms of its direct programming is not documented. Along with this, the effort for development and refinement of Linux/390 is concentrated in Boblingen, Germany and consists of code drops from a team there. The fact that most Linux developers have no access to a System 390 means that the development of that port lags radically behind others and is slow to respond to bug situations and potential development enhancements. So, from the Linux community point of view, what is the real contribution of Linux/390 to the mix? As for the commercial side, we are hearing little to nothing of how Linux/390 is making any real inroads in the mainframe arena. If you talk to IBM Business Partners, a group who sells mainframe computers to businesses, they will tell you that Linux is not even on the radar of the people they deal with. And a lovely singular sale like Telia AB, who bought a high end S/390 Parallel Sysplex machine, does not speak to how IBM is potentially succeeding in the small and medium size business arena in selling machines like the Multiprise 3000 with VM and Linux. It seems that for the most part Linux/390 is the province of large conglomerates with either excess capacity on existing equipment, or a very limited cost-sensitive server consolidation situation. Beyond that, there's nothing. There are no benefits to the open source community's involvement with Linux because they can't get to a 390 if they wanted to. Other IBM efforts are probably more useful, and hopefully will be encouraged.
Unfortunately, until the commitment to Linux, it didn't bother to even document these extras. And while docs have trickled out of North Carolina since, they've left it to the Linux community to fill in the holes rather than provide any support themselves.
All I want from IBM, and any other hardware company, are technically excellent pieces of hardware, and decent, well documented, drivers to go with them. An IBM MWave driver, and something to drive the TV capture hardware in my Thinkpad, would be all I could wish for from them.
--
You are not alone. This is not normal. None of this is normal.
Team, I had an opportunity to spend 3 years with IBM. In that time I attended courses on AIX internals, and had my hands on some fairly hot hardware, the SP2. One thing that IBM could bring to the table that would be fairly significant, is their modifications to the Unix filesystem; namely their JFS filesystem support and their Virtual Disk management suite. Don't get me wrong, I've seen better file system code produced by the now defunct company Data General (those boys KNEW how to build filesystems!), but IBM has built a fairly nice mechanism for logical disk management, being able to expand filesystems, etc. They've integrated that support into their systems-administration interface (smit/smitty), and it's pretty darned easy to grow a file system on-the-fly without requiring a umount. Of course you can't shrink filesystems like you could with DGUX, but nonetheless, its a start. If you're not aware of it, IBM has been for years, using the Open Source community software. They used a modified version of Tcl/Tk in their SP2 maintenance software, and they're using a tcl interpreter (God forbid, it was re-written as a java-class jar file) bundled with their WebSphere Application Server 3.5. They're also bundling the Apache Web Server with their WebSphere distribution, as well as InstantDB. I'd like to see IBM put up their virtual filesystem code, along with their SMIT utilities, and really start supporting the Linux community with their cash-cow products. I'm refering to their storage management software, ADSM. Yes, there is a Linux client port available, and I used it while I worked there. But their is no support for running an ADSM server on Linux. I'm not real fond of their ODM (for you IBM'rs out there...it really stands for Odious Data Mangler), but it may provide the open-source community to some insight on how AIX works. The real point here...I'll show you mine if you show me yours. Unfortunately IBM, has in the past, made their code available, but without sources. Take for example their modifications to the Apache Web Server. They added a caching module, along with an SSL module...but no source. IBM is in business to make money...of course aren't we all? But I would be cautious about how much of my work I'd be willing to give IBM for them profit on. One final point. When was the last time you saw AIX running on an Intel platform? IBM was involved with project Montery, but who knows what state that is in with the downfall of SCO. I think that IBM is interested in Linux as an Intel platform solution that they can market as "total-solutions". Today, you don't see AIX running on Intel, like you do with Sun.
I work at IBM and I see that, while still a Multi-National Corporation (capitalists) , they're actually willing to move into a position of reciprocation with the Linux community. They make our OS better, we like them more, we buy their hardware. Hey, IBM makes some cool hardware anyway, so where's the problem?
/is/ the future. They're certainly not losing any time in gaining market dominace in hardware to support our OS.
Like I said, I work(ed) and will work at IBM on developing software for Linux. I got a chance to meet and talk w/ Jeff Smith, whose official position is VP of eBOSS (Electronic Business Operating Systems Solutions, I think) last summer during my first internship, and we discussed IBM's plans for Linux. I am active in two LUG's, as well as the Texas Tech ACM, and Jeff was able to come speak at Tech last semester. What did he talk about? IBM's plans for the future, wich seem to contain a heavy helping of Linux development and support. Why is IBM doing this? Because they're smarter than M$ insomuch as they realize that OSS
And they pay me to write GPL'd software. IBM kicks ass.
-Will the Chill
Legalize It.
Creator of RPerl, Scouter, Juggler, Mormon, Perl Monger, Serial Entrepreneur, Aspiring Astrophysicist, Community Organiz
A whole bunch of people think Linux? Are people so un-educated that they don't know about BSD? Open Sourced OSes are *MORE* than GNU/Linux.
From the person who asked the question, showing his lack of knowledge:
"Robert LeBlanc, Vice President, Software Strategy, Software Solutions Division says both that IBM would open source any part of AIX and that we would be better off taking bits and pieces and the expertise that IBM bring along with it.
IBM is committed to Open Source where they think it makes a good business choice.
Yet, a whole group of people see Open Source and jump to the conclusion that means Linux. IBM has contributed to the BSD project (via the whistlejet)
It seems to me we have the offer on the plate from IBM to create a new joined project to bring Linux up another level if we can find a way from AIX. Surely we must take them up on this?"
What if IBM writes code that EVERYONE can use? Be they BSD, Linux, SCO, etc? Does this make them non "Open Source"? If IBM is truly an Open Source company they should consider making their code work with EVERYONE.
What is better to encourage IBM to do?
Ask them to release code for Linux, and pretend that is Open Source, *OR* ask them to make sure what they release is useable by EVERYONE who believes in Open Source code?
If it was said on slashdot, it MUST be true!
...and find someplace to land with Linux before it's too late.
.com's. For you to hand them an OS and say, "We want this and this and this and this and this and this and this..., oh, and while you're at it, throw in a nice screen saver or two as well!", you'll get a simple reply of "Why?"
..."
Networking scalability and redundancy, optimisation and facilities for database systems (as the jfs has started) or systems management applications?
Try and understand this from a business model. IBM has been around for an EXTREMELY long time (in computer years, I mean). They've survived longer than many computer companies (Packard Bell, for instance), and outlasted many
IBM doesn't want Linux just so it can become part of the "basement programmer Linux movement." The last thing they want is an anarchal OS that's being developed for EVERYTHING. Anything that's aimed for everything is looked on by a business as 1) Egotistical, and 2) Doomed to fail. (Yes, MS Windows is an exception, but there are always exceptions to the rule.)
IBM is going to want a target to shoot for. They want to get Linux and have someone tell them, "Market it towards the office. Here's why: It's free. Offices love free. All we need from you is to make it into a scalable office OS and you'll make money." Or..."Market it towards the Enterprise Server. It has remarkable scalability, is a Network OS, and all we need from you is
IBM's knocking on Linux's door. Don't let the opportunity get shot down. IBM's a business, while Linux was NOT developed under a business model. Find a bridge, and Linux will have the oportunity to join up with one of the biggest names in the computer industry.
On a more serious note, I'm not sure that IBM's contribution has to necessarily be directed at the OS itself. They could probably make good contributions to MySQL, maybe Apache, possibly a web broswer. These are things that, if this were Microsoft, would be thought of as components of the OS but are certainly necessary pieces if Linux is to make inroads on the desktop.
Given that they have loads of Unix experience I imagine they could help improve security as well. Linux is lacking in good GUI based security management. No doubt many Slashdotters are perfectly happy with CLI (I am) but 90 out of 100 SAs I talk to equate CLI with the 1980s.
Anyway, I'm not sure that the community needs to ask IBM for anything. Linux is given away free as in speech. I don't think it you can give software away for several years and then start asking for favors once companies have embraced it. This isn't to say that they don't have an ethical responsibility to give a little back, I'm just not sure that we should be the ones who decide what. If their past history is any indication (didn't they just open an FS?) they will no doubt make a contribution.
Icebox
We can't keep IBM from porting this stuff to Linux, but I hope major Linux distributions will have the good sense not to put it into the standard installations. I switched from AIX to Linux when I had the choice. There is virtue in being simpler: Linux was faster than AIX on slower hardware, it was more reliable, and it was simpler to use for server applications. If there is one thing I wouldn't mind seeing IBM contribute, it's their Fortran 90 compiler. Linux lacks a good, free, native Fortran 90 compiler right now. Beyond that, I can't think of anything in AIX that I would like in Linux.
And don't get confused about why companies like IBM and SGI are "donating" this stuff: they have a big legacy software problem, and they want to rid themselves of their albatrosses. This isn't "advanced technology" they are contributing for the good of the world, it's messy legacy functionality they want to get into a publically maintained software base so that they don't have to pay for software maintenance for their legacy customers. Most of the customers who want that stuff come from a different computing era and environment than Linux, and I suspect many of them will want to continue using proprietary software (meaning NT or Solaris).
Making software open in order to reduce development and systems integration costs is as good a motivation as any for contributing to open source software. But while it may be very attractive for a vendor to have such stuff added to the Linux kernel and OS, ultimately, the open source community may pay a steep cost in delayed releases and reduced usability as Linux gets ever more complex. Open source needs to be selective about what it accepts. Maybe Linus should start charging money for putting formerly commercial software into the kernel.
Disclaimer: I'm an IBMer.
IBM has big plans for Linux, as I'm sure everyone knows. There is a Linux Technology Center in Austin, Texas. Many good people down there doing nothing but working on Linux. IBM also has many open source projects, most under the IBM Public License, at the developerworks open source site. And I'm sure there are many other things IBM is doing that I don't know about (I'm just an employee).
However, IBM is not a complete friend of Open Source, just yet. It took me over 4 months of constant effort to get approval to attempt to contribute to the Linux kernel, and even then I only have approval to modify 4 files!!! Let's not forget that IBM is, by *far*, the world leader in patents. There are many parts of IBM which use Linux and open source, but are very hesitant to contribute to open source.
That said, there are many good people in IBM who really do want to do good things. It's just not that easy to get paid by IBM to do them. But it is possible.
So, back to the question of what Linux can do for IBM. I know exactly what the community can do for IBM. Accept us. I have received a lot of resistance to any help I offer on mailing lists. Patches and comments are ignored. My initial contact with a certain kernel developement mailing list was met with...quite a bit of hesitation, to say the least. I think a lot of kernel developers are very skeptical of IBM employees. Maybe they think we're only sending patches or commenting/discussing so we can sell some product, and in some cases, they may be right. But please, don't assume someone should not be trusted just beacuse they have @us.ibm.com in their sig. Just because I get paid by IBM doesn't mean I don't want to help.
First, I think the Free Software community needs to do some serious thinking about what technologies IBM can bring to the table. However, I'm not going to try and answer that question, because I think there's another question that needs to be asked in almost the same breath.
If IBM is willing to cater to our needs, then how can we reciprocate?
The foundation of free software is sharing resources, ideas and solutions--helping the other guy get his job done--building a community out of relationships, not between programs but between people.
If IBM is serious about joining the community (and if any IBMers are reading this, let me say that I really hope you are), then we need to reciprocate their seriousness with seriousness of our own.
We need to seriously ask, what can we give IBM?
I think the more IBM sees a willingness from us to work with them, the more willing IBM will be to work with us.
In this light, I would only ask this of IBM: Sell a hundred million copies of Linux, especially to the big businesses. Show the commercial world that there are good alternatives to Windows. Do what ever it takes do this. Adding software will be a small part of the project, which you may not even have to write yourself, if you discuss your needs with us. Build your own distribution, sell commercial software on the side, build a world-wide support network, print manuals, what ever it takes. Make sure that when Linux reaches total world domination, a good part of it is IBM Linux.
In Murphy We Turst
Linux is absolutely amazing for typical workstation class systems, but the one place that it seriously lacks in comparision to the other serious [and propriatary] unix'es [unices?] is in support for really big systems.
...Now, if only this could all somehow be a compile time option, that would be spectacular. The problem, though, is that we are talking about significant differences in basic architecture and of what the kernel has to worry about, these are significant design decisions. I don't think it's as easy as putting a few #ifdefs in the source.
What, you say? Isn't linux used on many supercomputer class systems, and heck, IBM even ported it to the S390! Well, yes... and no.
On clusters, linux is an excellent choice, due to it's efficiency and perhaps especially due to it's lack of licensing costs/hassels. But the key thing with clusters is that they are just that, clusters of workstation class machines. The `big'ness of them comes from the parallelism of many small machines working together, which it turns out is quite usefull for certain classes of a problems.
The S390 linux port is the same sort of thing. A massive big iron system, running hundreds of different simultanious copies of linux, each in it's own virtualized address space. Linux couldn't manage the resources and power of the entire machine, but as a 'process' in it, it does just fine; and this too is usefull, for say, consolidating what would otherwise have required rooms of rackmounts, with their own upkeep hassle, into one box [plus, cluster type applications can take advantage of the much faster [then ethernet] internal system bus for message passing, thus achieving comparatively much better performance then `real' clusters.
So, to make a long story short, what does linux really need? Big iron stuff: quality NUMA integrated into the memory management code [as a compile-time option, of course]. True support for high order SMP [say, 64 or 256 processors] and all that that requires [such as distributed kernel threads across multiple CPU's].
One of the problems, of course, is that adding this extra architecture into the kernel would slow it down on good old fashioned workstations, which is where linux is [currently] almost exclusively used [although that S390 thing rocks!].
---
man sig
---
the pen is mightier then the sword. the sword is mightier then the court. the court is mightier then the pen.