Should IBM's SOM/DSOM Be Open Sourced?

Esther Schindler sends a note about two journalists for very different publications (herself one of them) urging IBM to open-source, not all of OS/2 — they've consistently refused to do that — but instead one of its most powerful features: SOM, the System Object Model. Steven J. Vaughan-Nichols writes at desktoplinux.com, "IBM, I'm told by developers who should know, still has all of SOM's source code and it all belongs to IBM. It's because IBM doesn't have all the code for OS/2 and some of it belongs to Microsoft that IBM open-sourcing OS/2 has proven to be a futile hope." And Esther Schindler takes the developer angle in a blog post at CIO.com: "Could the open-source community use a library packaging technology that enables languages to share class libraries regardless of the language an application was written in? I dare say it could, especially since the code to accomplish that goal was written (and shelved) more than ten years ago. All it takes to make that code available is to ask IBM to release SOM and DSOM as open-source." What are the business issues that would convince IBM to assent?

Comments

To late?

[dryeo]

While SOM is very powerful and would of made a great addition to the Linux desktop I think it may be to late now as the common Linux desktop environments are quite entrenched. If only IBM had done this 10 years ago Linux could of had something to set itself apart.

Re:To late?

[Vectronic]

Set itself apart from what exactly?... Unix (variations) has used/had/does have SOM... and so has Apple, however Steve Jobs ended that "idea" when he returned to Apple... Microsoft has a very similar concept COM+

Seems more like it would bring Linux into this "group" rather than make it "special" for having SOM.

Re:To late?

[RobertM1968]

Set itself apart from what exactly?... Unix (variations) has used/had/does have SOM...

You must mean AIX - or you are predicting the future when IBM open sources SOM so your statement comes true...

and so has Apple, however Steve Jobs ended that "idea" when he returned to Apple...

You mean Taligent? Which was an IBM & Apple co-project?

Microsoft has a very similar concept COM+

You mean MS markets something they pretend is very similar, but any programmer who has used both will tell you they are light years apart, and that no matter how many ++++++++ MS adds to the end of COM, or how many times they change the name of OLE, it still will be light years behind SOM/DSOM.

You ARE on the right track with all of your comment - your just heading the wrong direction...

Re:To late?

[jockm]

> You mean Taligent? Which was an IBM & Apple co-project?

No he was referring to OpenDoc another IBM/Apple co-project which was all based on SOM/DSOM.

Re:To late?

[RobertM1968]

No he was referring to OpenDoc another IBM/Apple co-project which was all based on SOM/DSOM.

Oooops! Long night for me... yes I mean OpenDoc. :-) Thanks for correcting that for me.

-Rob

Re:To late?

[Vectronic]

I never said that they did it without IBM's help, however, if it was all (As in Linux, Microsoft, Apple, Sun?) based on IBM's SOM, then it could be benificial to everyone, and the entire computer industry (interoperability).

Nor did I say that I thought COM+ was good, infact its caused me many headaches both in developing my own software, aswell as other peoples software...

My point was basically that im not sure why the Linux Community as a whole hasnt put more effort into this, even if they had to go through corporate routes, like the more closed-source distro's (some might bitch and yell about that but it would also force/inspire the open source community to try their own)

Re:To late?

[SigmundFloyd]

I distinctly remember Mac OS 8 & 9 having an extension called "SOMobjects", which was used by/for a lot of different things, not just OpenDoc.

Re:To late?

[man_of_mr_e]

or how many times they change the name of OLE

This sort of response is typical of those that don't really understand the Microsoft stack.

COM is not OLE renamed. COM is the basic transport structure that OLE, and other MS technologies, are based on. OLE is the name of a specific set of COM interfaces involving such things as Compound Documents, Automation, In-place activation, etc.. OLE also describes a lot of things that don't technically involve COM.

Other examples include ActiveX, DirectX, COM+ (despite it's name, it's not a new version of COM, it's just a specific set of interfaces like the others I mention), OLEDB, etc..

So, in the sense that both COM and SOM are the transport layers of the technology, they are roughly equal, even if their implementations are very different. Other technologies get built on these layers.

Re:To late?

[RobertM1968]

or how many times they change the name of OLE

This sort of response is typical of those that don't really understand the Microsoft stack.

You mean, like you?

COM is not OLE renamed. COM is the basic transport structure that OLE, and other MS technologies, are based on. OLE is the name of a specific set of COM interfaces involving such things as Compound Documents, Automation, In-place activation, etc.. OLE also describes a lot of things that don't technically involve COM.

Actually, it kind of is... OLE was released in 1990, while COM was released in 1992... so COM started as a subset of OLE that OLE relies on. Revisionist history aside, or current dependencies aside, COM was based off the components of OLE that were "bundled" into what MS later called (and released as) COM.

Other examples include ActiveX, DirectX, COM+ (despite it's name, it's not a new version of COM, it's just a specific set of interfaces like the others I mention), OLEDB, etc..

Which are all based off OLE/COM - I fail to see your point.

"With Windows 2000, that significant extension to COM was incorporated into the operating system (as opposed to the series of external tools provided by MTS) and renamed COM+."

Which also invalidates your later statement at the bottom... since with SOM/DSOM, the extensions wouldn't be necessary... anyone can subclass or superclass a class/object to add whatever functionality they want without needing IBM to write "significant extension(s)" to SOM/DSOM... we've proven that already with OS/2 and eComStation by subclassing and superclassing numerous WPS "functions" to add tons of other functionality with very tiny code and not a single piece of the WPS needing to be re-written. Try that with COM+/++/+++/++++...

So, in the sense that both COM and SOM are the transport layers of the technology, they are roughly equal, even if their implementations are very different. Other technologies get built on these layers.

They are far from roughly equal - except in the limited sense you are discussing... it's like saying an apple and an orange are roughly equal because they are both fruit... SOM/DSOM is far more than "transport layers of the technology" - perhaps you just haven't programmed for both? See above for a very very basic example. Heck, playing with stuff like "IBM Works" or "Describe" or some of Stardock's Object Desktop components for OS/2 alone would help you understand that there are major differences in implementation and functionality - even with NO updates for SOM/DSOM in roughly a decade... it, just quite simply, is that good.

Re:To late?

[billcopc]

Aren't you just a charming troll ? With your critical tone yet lack of counter-points.

Yes, COM is the foundation layer upon which OLE, ActiveX, ADO and countless other interfaces are built. Is it as seamless and elegant as SOM/DSOM ? Who knows ? Is it in use today by millions of developers ? YES! Unlike SOM.

I'm not saying SOM is worse or better than COM; I don't know. On one hand, we're all suspicious of Microsoft, but what has IBM done for us lately ? They stopped being relevant the day they killed OS/2.

Re:To late?

[Richard Steiner]

Yes, COM is the foundation layer upon which OLE, ActiveX, ADO and countless other interfaces are built. Is it as seamless and elegant as SOM/DSOM ? Who knows ?

Well, the poster to which you are responding apparently knows, as do many other folks who have actually used both SOM and COM for development. Remember that OS/2 was a very popular operating system once (albeit a long time ago).

Is it in use today by millions of developers ? YES! Unlike SOM.

Yes, we all know that it's the number of people USING a technology which determines it's overall quality. That's why McDonald's and Burger King are the paragons of American cuisine. :-)

I'm not saying SOM is worse or better than COM; I don't know. On one hand, we're all suspicious of Microsoft, but what has IBM done for us lately ? They stopped being relevant the day they killed OS/2.

If you seriously don't know about the numerous things which IBM has done for the open source community over the past ten years or so, I'm afraid you weren't paying very close attention. And it hasn't been just technology (e.g., JFS) or patents from IBM that have made it into the open source ecosystem -- the very fact that IBM endorses the Linux software stack in an enterprise context in the first place is a huge validating factor for pointy-heads.

Re:To late?

[RobertM1968]

Oops... sorry, I misread the intent of your post... my apologies.

And I dont know why the Linux community hasnt put more effort into this either... though I can speculate (and my speculation would make the release of SOM/DSOM somewhat useless)... it could be because to use it in such things as the GUI - or applications - would require rewriting all or parts of said components.

That's probably the same reason why MS didnt do something of that nature as well... easier to simply continue with a few more add-ons as needed than re-invent the whole system to be more object oriented in a similar fashion to SOM/DSOM. I especially can't fault MS for that choice... there are too many things that require their current method that would be broken unless they figured out how to implement it (COM+, etc) into their new object based subsystem. There are probably more important things for them to worry about.

Robert

Re:To late?

[RobertM1968]

So, in the sense that both COM and SOM are the transport layers of the technology, they are roughly equal, even if their implementations are very different.

Please read this... it'll explain a lot about how unequal they are both in capabilities and in design...

http://en.wikipedia.org/wiki/IBM_System_Object_Model

Unfortunately, many of the key points require an understanding of how they can be used... but I think some of the "Related Links" explain it if you are interested...

Re:To late?

[billcopc]

Then I must live in a distant universe devoid of IBM. Pointy heads around here don't give a crap about Linux, largely because it's a government area and they won't touch Linux with a ten foot pole. It's Windows all the way, not that I agree with it.

Most people in my city think of IBM as that formerly ginormous company that now makes crappy little PCs and PC-based point-of-sale systems. Their support is absolute shite (30+ days turnaround for RMAs to million-dollar regional accounts). The hardware itself is a disaster, they build those PCs worse than the shadiest Chinese OEMs... I guess I should be grateful, IBM gear creates a ton of lucrative support calls for me. The fact that I could build a more reliable machine for less than the cost of one onsite call, well that's just not good no matter how you spin in.

As much as I believe what you're saying, and want to agree with your points, there's just not enough good in IBM's presence to convince the masses. A handful of mainframe techies in California simply isn't enough to drive any PR.

Re:To late?

[XO]

erm, dude, IBM hasn't turned out a PC in over a decade...

Re:To late?

[man_of_mr_e]

OLE was released in 1990, while COM was released in 1992

That's correct, but it's irrelevant. COM existed before that, they just didn't commercialize it outside of OLE yet.

The point is that while OLE includes COM, COM does not include OLE, therefore COM is not OLE renamed, though you could say it's a subset of OLE renamed.

And no, ActiveX, DirectX, and COM+, nor OLEDB (despite it's name) have nothing to do with OLE, other than they use COM. OLE refers specifically to those bits surrounding Automation, Linking and Embedding, Compound Documents, etc.., none of that exists in other technologies based on COM.

Re:To late?

[man_of_mr_e]

As I said, their implementations are very different, but they accomplish mostly the same goals. Sure, SOM has inheritance and runtime dispatch, but COM can simulate inheritance with aggregation. The only limitaiton is that you can't superclass in COM (as I said, subclassing can be simulated with aggregation).

You're confusing features, with purpose. They have the same purpose, they just go about it with different feature sets.

Re:To late?

[RobertM1968]

As I said, their implementations are very different, but they accomplish mostly the same goals. Sure, SOM has inheritance and runtime dispatch, but COM can simulate inheritance with aggregation. The only limitaiton is that you can't superclass in COM (as I said, subclassing can be simulated with aggregation).

You're confusing features, with purpose. They have the same purpose, they just go about it with different feature sets.

Ah... I see what you are trying to say. I guess I am just looking at it differently.

(As an example) If I wanted to replace any part of the (OS/2 or eComStation) GUI, I could... if I wanted to "replace" the whole thing, while still leaving every other method in place for programs that relied on it, I could, by simply superclassing the classes I needed - while everything lower in the class tree would inherit the features, and my classes could pass anything "standard" to the original classes to handle. To me, those are features - as they were considered in the design as features of the implementation... I guess that's why I'm looking at the two's features differently.

:-)

Re:To late?

[Richard Steiner]

Most people in my city think of IBM as that formerly ginormous company that now makes crappy little PCs and PC-based point-of-sale systems. Their support is absolute shite (30+ days turnaround for RMAs to million-dollar regional accounts). The hardware itself is a disaster, they build those PCs worse than the shadiest Chinese OEMs... I guess I should be grateful, IBM gear creates a ton of lucrative support calls for me. The fact that I could build a more reliable machine for less than the cost of one onsite call, well that's just not good no matter how you spin in.

It's interesting how people can see the same company in different (sometimes even opposite) ways.

When I think IBM, I think of refrigerator-sized boxes running OS/390 and zOS, stacks of IBM blade servers running Linux contentedly humming in our corporate data center, and sleek jet-black IntelliStations (which were very well designed and constructed x86 desktop machines, not the consumer PC crap that you're thinking about which was sold to Lenovo quite some time ago). Oh, and Thinkpads. :-)

Re:To late?

[billcopc]

Ugh... Thinkpads. Sure, they were sturdy, but they were perpetually nine months behind the curve for features and performance. It was like owning an Acer with plate armor; a well guarded piece of shit.

I have yet to see an IBM server with my own eyes, but the datacenters are full of commodity PC hardware. That's probably because there is a glut of commodity PC developers in my town (yay government). I barely trust these drones with a Celeron, there isn't enough Ibuprofen in the world for me to let them loose on big iron. That's just one big-ass infinite loop just dying to happen with these asshats. The only saving grace is the inevitable memory leak that ultimately crashes the process, not so much garbage collection but garbage regurgitation.

I've seen quite a few large SPARC installations. Those have a bit of traction because they're both cheap and effective for what little Unix filth gets thrown around in these parts.

Who needs the code?

[DoofusOfDeath]

If the great innovation offered by SOM is basically a design pattern or interface technique, do we really need IBM's source code? It seems to me that the great thing about SOM is the idea of how something is done, and that we could pretty quickly write our own implementation of that idea. No?

Re:Who needs the code?

[RobertM1968]

If the great innovation offered by SOM is basically a design pattern or interface technique, do we really need IBM's source code? It seems to me that the great thing about SOM is the idea of how something is done, and that we could pretty quickly write our own implementation of that idea. No?

Sadly no - on all counts. In over a decade and a half, no one (but maybe Apple) came close. DSOM/SOM hasn't been worked on in many years, and still, with kludge after kludge, MS cant come close. (some of) The Linux community wanted the WPS open sourced just because of how powerful it was - even though I dont think they even realized that it meant also open sourcing SOM/DSOM. With many attempts at numerous windowing environments, though the Linux community has made both some pretty and some pretty useful windowing environments, they still haven't come close...

And of course, SOM/DSOM is far more than the WPS... (just a requirement for the WPS to work).

Also, saying SOM/DSOM is just "the idea of how something is done" is like looking at cars, bicycles, sneakers and skateboards and calling the car engine "an idea of how something moves" - it is far more than that. It is a technology that allows anyone on almost any language, to interact with and integrate with any other device, network resource, app, GUI or OS that is SOM/DSOM enabled. Almost 8? 10? years of little to no development on SOM/DSOM and there is still nothing half as powerful for any PC based operating system. Yeah, MS can keep writing inelegant, bloated (which is a massive understatement when compared to SOMObjects of better capabilities) kludges to achieve some of the functionality on a limited scale...

SOM/DSOM is truly what most OO programmers truly want - even if they dont know it (which would simply be because they dont understand it, or love VB that much).

Re:Who needs the code?

[LLKrisJ]

Sadly no - on all counts. In over a decade and a half, no one (but maybe Apple) came close. DSOM/SOM hasn't been worked on in many years, and still, with kludge after kludge, MS cant come close. (some of) The Linux community wanted the WPS open sourced just because of how powerful it was - even though I dont think they even realized that it meant also open sourcing SOM/DSOM. With many attempts at numerous windowing environments, though the Linux community has made both some pretty and some pretty useful windowing environments, they still haven't come close... I use this quote but to me, one thing stands out in reading nearly all posts in this thread;

SOM/DSOM was apparently a very powerful system with some advantages which supposedly make it a desirable tool to have within the Linux environment.

Apparently _nobody_ (this is important) has been able to recreate anything even remotely like it independently from IBM. So _nobody_ succeeded in doing this and yet everybody seems to be very happy to just take some cheap shots at MS for creating their 'kludge' COM+ attempts.

I can't help but feel that that's just stupid. At least they have tried to do something, even if their solution has shortcomings.

People should stop with this whole MS = evil, Linux = good crap. Software platforms should be rated on merit and unless the Linux community comes up with a good solution in this particular case the should probably turn it down a notch or two with their criticisms already..

Re:Who needs the code?

[cyber-vandal]

The trouble with being a habitual criminal is that no-one trusts you ever again. That's where MS = evil comes from and is not likely to stop until they show a consistent period of obeying the law and playing within the rules. Since that is antithetical to the MS corporate culture MS = evil will continue to be held as an opinion irrespective of whether WinXP isn't as god-awful as Win98.

Re:Who needs the code?

[Simulacrus]

It is a technology that allows anyone on almost any language, to interact with and integrate with any other device, network resource, app, GUI or OS that is SOM/DSOM enabled. Almost 8? 10? years of little to no development on SOM/DSOM and there is still nothing half as powerful for any PC based operating system.

Smalltalk / Squeak has been doing something similar for years. Anything in the system can message anything else. System components / GUI elements can be freely inspected, subclassed, modified - and all on the fly while the system is running. Of course, this means that all source is very literally open (other than virtual machine primitives).

Re:Who needs the code?

[Marcos Eliziario]

MODERATOR, MOD PARENT INSIGHTFUL.

Lately, there it seems to be a bunch of people here on /. that like to portray any microsoft criticism as childish manifestations of fan-boyism.
It's good to have someone remembering from time to time that our problem with microsoft has to do with their illegal and anti-ethical behaviors, behaviors who had the intent to monopolize the market, at the cost of destroying innovation.
Linux is good not because it's a l33t thing, but because Linux was instrumental on keeping innovation alive on the computing field.

Re:Who needs the code?

[RobertM1968]

I can't help but feel that that's just stupid. At least they have tried to do something, even if their solution has shortcomings.

People should stop with this whole MS = evil, Linux = good crap. Software platforms should be rated on merit and unless the Linux community comes up with a good solution in this particular case the should probably turn it down a notch or two with their criticisms already..

I never implied any such thing. SOM/DSOM is still superior to COM/COM+ or any related technology that MS has planned (that isnt vaporware). The merits (or lack thereof, depending on your point of view) of MS's business practices has nothing to do with that comparison.

COM and all it's later incarnations, quite simply, are kludges in comparison - that has nothing to do with whether MS is evil or not. And, if you actually bothered to read my post instead of taking snippets of it and misconstruing them, you would have seen I also indicated that as of yet, the Linux distros dont have anything as powerful yet either.

As a simple example, when the WPS came out (till recently) it could only display icons using the OS/2 .ico format, at up to 40x40 pixels, had three (standard) folder views; Icon, Tree, Details (and some specialized ones like "LightTable"). Since then, with well under a meg of code, the WPS has been superclassed to enable use of PNG icons (and Cairo) with transparencies at virtually any size, including individual sizes per object (icon, for those of you who don't understand that an icon in OS/2 is really an object reference) (no set size - pick what you want for whichever object), active widgets with transparencies (not pretend to be active widgets), various new folder views and additions to older ones (from toolbars, search panes, status/info windows), tool-tip help pop-ups, media folders (that dont use more than a few bytes of resources to show or play media), object preview windows/panes, extensible (in and out of folder) toolbars... and the list goes on - and on and on... but that's enough to start.

How many updates to SOM/DSOM were required? None.

Why your post was not modded "Troll", I dont know... especially since you are the one trying to turn a legitimate discussion into a "MS is evil" - "MS isnt evil" - "you only say that because you think MS is evil" conversation, when that wasn't the point of it at all.

Re:Who needs the code?

[RobertM1968]

Ooops... sorry for the all bold... not intentional...

Re:Who needs the code?

[RobertM1968]

Smalltalk / Squeak has been doing something similar for years. Anything in the system can message anything else. System components / GUI elements can be freely inspected, subclassed, modified - and all on the fly while the system is running. Of course, this means that all source is very literally open (other than virtual machine primitives).

That's very interesting... can you superclass the stuff as well? And does it work with a GUI, thus making everything on the GUI an object (that can be both subclassed and superclassed)? And are objects created using it interoperable with other objects? And are you limited to using SmallTalk, or can you use any almost language with the relevant code to do all the above? Does it support full inheritance? Are the resulting DLLs as small - and as fast?

Not picking on your statement... I have a copy of SmallTalk (someplace) - but never installed it, much less looked into it's capabilities... so your answer, whatever it is, will be a (welcomed) learning experience for me. And perhaps I need to dig out my copy of SmallTalk and play with it...

Thanks,
Rob

Re:Who needs the code?

[LLKrisJ]

And, if you actually bothered to read my post instead of taking snippets of it and misconstruing them, you would have seen I also indicated that as of yet, the Linux distros dont have anything as powerful yet either. I actually did. I actually repeated your 'Linux doesn't have anything either' statement... So I could use the good old 'read what I wrote' on you as well.

What bothers me about your post or those of others for that matter is the simple fact that if MS tries to do something which turns out a little bit flawed it immediately gets pegged 'flawed', 'utter crap' or any other diminutive term. Even if the competition is equally crap and/or lacking in comparison. I would hardly call that flame bait or trolling. You're being overly sensitive.

If I read some other posts on this topic SOM/DSOM was an equal 'kludge' in many respects :)

But hell... does it really matter? Probably not...

Cheers mate

Re:Who needs the code?

[RobertM1968]

But hell... does it really matter? Probably not...

You are probably very correct!!! :-)

Robert

Re:Who needs the code?

[demon driver]

It's good to have someone remembering from time to time that our problem with microsoft has to do with their illegal and anti-ethical behaviors, behaviors who had the intent to monopolize the market, at the cost of destroying innovation. You may praise him for that, but still both of you don't see that such behaviour is in fact quite mandatory; nothing else can be expected from a company which has become so successful that there is a chance to build a monopoly with one or more of their products. Everything else would be strictly against shareholders' interests and thereby out of question. That's how capitalism works, and this is the more true the more globalized it gets.

Cheers,
d. d.

Re:Who needs the code?

[Marcos Eliziario]

No, disobeying the law is not how capitalism works, it's the exception and not the rule, and that's why we pay taxes to live in a civilized world. Don't forget that several times microsoft has been found guilty of anti-competitive behavior and had sanctions applied not only in the US, but also in the EU.

Where's the 'notgonnahappen' tag?

[Caspian]

Srsly, folks. This is IBM we're talking about. They aren't foolish or desperate enough to give up complete control of one of the more useful features of an OS that they've already declined to open up. And MS is involved here... puh-lease. Not gonna happen. Give it up.

Re:Where's the 'notgonnahappen' tag?

[FooAtWFU]

I dunno. IBM isn't really in the OS market anymore, at least, not the regular mass-produced OS market - the big-machine Z/OS UNIX mainframe stuff and the like is an entirely different matter altogether (and probably wouldn't be substantially affected by such an offering anyway). Giving something like this to Linux+friends, if it helped, could boost Linux and be seriously annoying to Microsoft. And IBM likes Linux - certainly a lot more than they like Microsoft. It could happen. Maybe.

IBM Open-sourcing Experience

[fsckr]

IBM has been pretty good at taking sucessful closed source technologies and open sourcing them (think eclipse, webspere community ed, jikes and all the patents they've made available to the os community). I think IBM's genius has been in fostering communities to ensure that the technologies are well supported.

That said, SOM & DSOM are old tech from the dinosaur mainframe days. With so many distributed apps using more flexible interoperating technologies (SOAP, XMP-RPC etc) I don't really think open sourcing D/SOM will make that big of a difference to most new application developers.

Re:IBM Open-sourcing Experience

[ciggieposeur]

think eclipse

I used to work within IBM, and here is the story I got from a senior developer...

Eclipse was a rogue project within IBM whose purpose was to take down MS Visual Studio 97. When it became obvious that there was absolutely no way that Windows developers who were used to the speed, ease of use, and rather decent integration with Windows of VS97 would shell out money for a Java-based IDE with its own oddball help system, poor integration with the Win32 toolchain, and most importantly was slow as molasses even on decent hardware, IBM decided to in essence use the $30+ million in sunk costs to buy goodwill in the F/OSS community. (And Eclipse really needed that OSS shot in the arm: NetBeans was killing it in the Java IDE landscape.)

Re:IBM Open-sourcing Experience

[Locutus]

Eclipse was based on VisualAge for Java which was, IIRC, based on VisualObjects or something like that for Smalltalk. IBM bought that company/product and added C++ capabilities to it. There might have even been a Cobol version. So it was first Smalltalk and then C/C++ but written in Smalltalk. When Java came along, IBM added the Java modules to it. IBM was also doing killer work in the JIT/jvm area and since they effectively had 2 vm's in VisualAge , why not create a universal vm so that VisualAge used it for its runtime and also allowed Java in VisualAge to use it. We are now talking about the 1995/97 or so timeframe. VisualAge for C++ also had some great SOM capabilities built into the framework for OS/2 and AIX applications. So by this time, IBM has given up on OS/2 on the desktop, server, and workstation while AIX is still going strong. Linux is picking up and so is Java on the server since Microsoft bastardized it on the desktop. It looks to me that Eclipse is built from VisualAge for Java with the licensed portions removed. Basically the really cool VisualBuilder was yanked and it/Eclipse started as a Java IDE. C++ was added back in with the CDT a few years later and now other languages are being added.

Remember, in the early 90s there was Borland and Watcom along with IBM who had compilers and SDKs for OS/2. OOP/C++ frameworks were/are a threat to Microsoft in that they abstract the underlying APIs and untie developers from their platform. So, Microsoft VisualC++ was pretty much sold at a lost for a good number of years to tie developers to Windows. Borland senior software designers were purchased away from Borland and helped in their demise(a court case won Borland some bucks from Microsoft for the brain-drain ). Watcom was not allowed to ship anyone elses class library framework along with Microsofts MFC so Watcom eventually was obsoleted by only shipping with MFC. The great era of cross platform C++ class libraries in the early to mid 90s entered the late 90s with C++ frameworks all but extinct. IBM on the other hand had spent a good decade attempting to find a cross platform development system for their mainframes, workstations, and PC developers. First with Smalltalk, then C++, SOM, and then Java. I'm sure the attack by Microsoft on cross platform frameworks also upset an number of IBMers. Once you go MFC, you're hooked and a massive port is required. And IBM spent alot of time and money attempting to build cross platform tools to ease developer reuse across its business divisions.

So, saying that Eclipse was created to take down MS VisualStudio seems off mark. I would not doubt that there are still a few in IBM who find that saying the word "Microsoft" is like a psychedelic breakfast and leaves a bad taste. And Eclipse, being for Java, came way after MS Visual J anyways so how was it to take down VisualStudio anyways? Giving the OSS community something with the power of VisualAge probably had more to do with some IBMer wanting to continue using the IDE instead of letting it collect dust on the shelf like SOM/DSOM has. People are not forced to use Eclipse but look at the marketshare it has.

BTW, IBM also has a bunch of tools to generate SOM objects to make it much much easier to deal with. Still, SOM is not for PHBs in the sense of MS Visual Basic. It's for real developers and many did use it and used it well. Also, where the heck are the StarDock developers? They'd know a bit about SOM and the WorkplaceShell in that WorkplaceShell extension package they sold. It was incredible and mostly because of SOM and their use of it.

LoB

Re:IBM Open-sourcing Experience

[toriver]

Eclipse was based on VisualAge for Java

There's not a smidgen of VisualAge in Eclipse; the former was a Smalltalk-based IDE family as you point out, the latter is written from scratch in Java with a custom native GUI toolkit (SWT) to mess things up.

I have used both. If Eclipse had any of the horrid mess from VisualAge in it it would have been dead, and Borland's IDE codebase (used for JBuilder and Oracle's JDeveloper) would have won even though it was closed-source. VisualAge is an IDE from the first generation of Java IDEs, when it was obvious they were trying their way along as the methods of development matured.

Re:IBM Open-sourcing Experience

[ciggieposeur]

I'm pretty sure VAJava and Eclipse had radically different lineages, especially since I could get VAJava in 1999 but not Eclipse.

I can't remember the name of the organization that did Eclipse anymore, but it started with O, like "Orion" or something. I remember the dev talking about how they weren't "really" IBM: they had funding and access to everyone else's code, but they didn't allowed anyone else to see theirs. They were apparently the brain child of a DE or two (Distinguished Engineers). The dev thought that their re-engineering of SWT was stupid when they had Swing already available and that it delayed the release by at least two full years.

I do remember hearing or seeing some kind of presentation about it. People were asking what was going to replace VAJava for WebSphere development and some bigwig said that this project was going to do it with its "plug in architecture" and would replace the IDE for C and C++ along with Java. It was supposed to be better than SlickEdit, better than "anything else out there". AND it was supposed to be sold to make money. Somewhere along the line it got split into the F/OSS version and WSAD which is sold for money.

Re:IBM Open-sourcing Experience

[Locutus]

Before there was VisualAge for Java, there was VisualAge for C++. VisualAge for Java was based on the same foundation VA C++ was based on and that was a Smalltalk VM. IBM put alot of work into VAJ to merg the Smalltalk VM with Java and you might even find references to a universal virtual machine referencing this fact.

It's been a while but Eclipse reminds me of VA of old but with a different skin and some of the things removed. Even VA C++ had method based compiling instead of full project compilation.

Maybe there's someone who worked on the code will chime in, otherwise we'll just have to disagree.

LoB

Re:IBM Open-sourcing Experience

[Locutus]

I'm pretty sure VAJava and Eclipse had radically different lineages, especially since I could get VAJava in 1999 but not Eclipse.
why would you then not consider Eclipse was based on VAJ lineage since it came before it?

It sounds like maybe VAJ/VAC++/VASmalltalk might not have been too modular and the Eclipse people may have built it from the guts of the VA platform. From what I've seen, there's a common feel to it having used VAC++ and VAJ in the 90s.

LoB

Re:IBM Open-sourcing Experience

[ciggieposeur]

why would you then not consider Eclipse was based on VAJ lineage since it came before it?

Yeah, I wasn't very clear there. I meant that if Eclipse really was the future of VAJava I would have expected to see or hear something like "the next release of VAJava will be called Eclipse" or "Eclipse is the new product from the VAJava / VAC++ line". Instead Eclipse seemed to spring suddenly out of the blue after years of semi-secret development, only shortly after the last official VAJava release, and with a pretty different interface.

Re:IBM Open-sourcing Experience

[Locutus]

ok, that makes more sense. As far as that goes, it might have been that announcing open sourcing of the product would harm the current sales channel so they didn't announce it. Or maybe they didn't know if it was going to be released until someone approved it. Of all the things that the anti-trust case against IBM did, one very evident result was that they would not and did not pre-announce products like they used to. Microsoft will announce stuff years before they see the light of day but with IBM, you're lucky if you get 6 months notice unless you are an insider.

BTW, get the latest Eclipse tarball and run the following command against it( you might need the CDT ):
strings eclipse/configuration/org.eclipse.core.runtime/.extraData.2 | grep VisualAge

How many times do you see the name "VisualAge"? I count 5 in the runtime module reverencing VAC++ versions. Not conclusive evidence but it shows 'some' connection.

LoB

Re:IBM Open-sourcing Experience

[Locutus]

There's been some question as to Eclipse being derived from IBM's VisualAge products. FWIW, Wikipedia does a decent job at spelling it out but isn't too linear in its description. Object Technology International( OTI ) made a product in the early 90s which was based on Smalltalk and had a click-n-drag GUI application builder which included building apps with both GUI elements and non-visual/GUI elements. VisualAge was based on this product and IBM bought this company. After IBM built VisualAge for C++ ontop of the Smalltalk product, they then built VisualAge for Java and added Java capabilities to the Smalltalk runtime of VisualAge. Then, trimmed down and pure Java version of VisualAge was created and called VisualAge for Embedded and targeted Java on embedded devices. This, Java-based product was what eventually became Eclipse.

Here are some links to follow up on the synopsis above:

http://en.wikipedia.org/wiki/Object_Technology_International
BTW, I saw this product demo'ed on OS/2 in the early 90s and was pleased to then see the same kind of visual builder in the VisualAge product. Unfortunately, if you build it, they don't always come when they only see through Windows.

http://en.wikipedia.org/wiki/VisualAge

Getting back to SOM -
This is a review of VisualAge for C++(Windows) v3.5 with SOM v2.x and COM linkage:
http://www.ddj.com/windows/184415551

LoB

Re:IBM Open-sourcing Experience

[JAlexoi]

Ahem... Webspere Community Edition is the other way around. Webspere Community Edition = Apache Geronimo with pretty interface and IBM's name on it :)

Licensing and open source

[dryeo]

One of the advantages of SOM is that it allows a closed source environment to be extended. Don't like the file dialog, subclass it with a better one. Or a recent example, need transparent png bolted on your 10 year old OS, well create a few new classes and use Cairo to display them. Suddenly you have modern transparent icons, transparent widgets on the desktop etc.
Unluckily with GPL you can get into issues of whether closed source or just incompatible licensed libraries can be added. One of the ideas behind SOM/DSOM was that anyone could write a DLL and extend the WPS. Now it seems that in free software land you often have to worry about incompatible licenses.
If IBM ever does open source SOM/DSOM I hope it is with something liberal like the LGPL. Don't have to think about issues with linking and the important source stays open.

Re:Licensing and open source

[IamTheRealMike]

I would note that you can do all those things on Windows with COM. In fact that's how transparent PNG support was added to Internet Explorer, and is why you have to invoke them in such a funny way (MS could have done a better job of the IE/COM integration in the images department).

Re:Licensing and open source

[Hal_Porter]

Don't like the file dialog, subclass it with a better one

You can do that in Win32.

http://www.codeproject.com/KB/dialog/DavidKotchanFileDialog.aspx

Actually, you can subclass any Windows window even in plain C

http://furix.net/subclassing-child-controls

Re:Licensing and open source

[XO]

Not exactly - how is it that you could, in Windows, change -all- applications that use the standard File->Open box, to use a new one tha you've written?

I don't know a lot about windows coding, but I'd expect that if it were possible, there'd be a lot of replacement File->Open boxes out there.

IE 6 supported PNGs, but not transparent PNGs, and DirectX supported display of transparency, so you had to make up some horrendous CSS hack to get windows to pipe the PNG through DirectX before displaying on screen. it was almost accidental that it worked, not a clever side effect of a functional object model.

Anyone who doesn't understand how incredibly awesome the SOM/DSOM (and WPS) were, seriously needs to install OS/2 and Stardock's Object Desktop. (yes, they made an "Object Desktop" for Windows that had some of the functionality, but it's nowhere near as awesome)

All this talk about things that have recently been done, makes me want to go and get eComStation ... unfortunatly, I don't even have the hard drive space to tinker with Linux :(

Re:Licensing and open source

[XO]

Interesting article you've pointed out, however it seems to loosely indicate that this set of file dialog mods is primarily useful to developers, as it does not extend the existing box for everything else.

Re:Licensing and open source

[Hal_Porter]

Actually you can Global Subclass

http://msdn2.microsoft.com/en-us/library/ms997565.aspx

However, there are ways you can add subclassing functionality to every process. Once you get a function inside the address space of a process, you can subclass anything in that process. There are a few ways to do this. The easiest (and most brutal) approach is to add a dynamic-link library (DLL) name to the following key in the registry:
HKEY_LOCAL_MACHINE\Software\Microsoft\Windows NT\CurrentVersion\Windows\APPINIT_DLLS

This key causes Windows to add your DLL to every process in the system. Your DLL would need some way to wake up after every event that the DLL would want to subclass after. A WH_CBT hook usually does the trick. The DLL can watch for the HCBT_CREATEWND event, then subclass the desired windows. The CTL3D sample application uses the WH_CBT hook to do its subclassing, although it does not contain the registry entry that makes subclassing a part of every process. Applications that want CTL3D can link it into their process. Applications loaded CTL3d to get new style 3D look controls on Windows 3.1. By Windows 95 all controls were 3d by default and it wasn't needed. Oddly enough one of the features of Office 2000 was new "flat effect" controls.

But making it opt in was a (very wise) design decision. You could write something like CTL3D and add it to HKEY_LOCAL_MACHINE\Software\Microsoft\Windows NT\CurrentVersion\Windows\APPINIT_DLLS and it would affect (infect?) all applications. Spyware, adware and toolbars use this technique to inject code into all processes. I've actually seen these sorts of DLLs crash third party applications. In fact it's almost inevitable since there is no way the company making the injected code can test it on every possible Windows process. In the case of Spyware they probably don't even try to.

Re:Licensing and open source

[XO]

Wise? The whole thing looks kludgy and a horrible security issue.

Re:Licensing and open source

[Hal_Porter]

Using APPINIT_DLLS is bad because you can never test it properly and that's why the CTL3D's author's decision not to use it was wise. But it's not a security issue anymore. Now you need local admin rights to change it, and if you have admin rights you could destroy the system in lots of other ways. I think the same is true for WH_CBT global hooks.

But if you're writing an application, want 3D controls and choose to load CTL3D.DLL it is up to you to test. There's nothing really wrong with this as far as I can see - it allows people to use old applications which would be broken by the new controls since they won't load CTL3D and people who had time to test the new controls and didn't find a problem could add one line of code to load CTL3D to their application.

Whether it's a kludge is subjective I guess. Kludge is just a way of something complicated that you don't understand in an OS you don't like. And this stuff is more complicated than Posix.

A non kludge way would be to change the controls and tell people who depend on applications that made questionable assumptions about the old controls and thus crashed to upgrade. But for a lot of people that wasn't possible since the app vendor had long since gone out of business. If they had to choose between upgrading and running some vital application, they'd pick the app. So there are valid business reasons for the complexity.

Mind you, I actually find this sort of thing fascinating, and I think it's good that Microsoft does it. From my point of view it's better that they do stuff like this that will cause lesser programmers to give up because that minimizes the competition I face and probably increases my salary due to supply and demand effects. Mind you I wouldn't use something like APPINIT_DLLs because it will fail some small percentage of the time and cause people to call tech support and that will give me a reputation for producing unreliable code which could decrease my salary. But kludginess of the level of CTL3D is fine with me, it just sorts the wheat from the chaff.

Incidentally, CTL3D isn't actually used anymore. But there are other transitional things that are equally kludgy. And long may that continue.

Its already there

[blackpaw]

Could the open-source community use a library packaging technology that enables languages to share class libraries regardless of the language an application was written in?

Its called .NET or Mono

Re:Its already there

[sigzero]

You have no clue. They are not in the same class at all.

Re:Its already there

[sl0ppy]

.net and mono are functionally equivalent to a SWIG wrapper - a bit easier to deal with, but similar nonetheless.

this is a bit different. for one, it allows you to redefine how an API is implemented. imagine a generic database API. instead of programming to a client library, all software can be written to a single standard API that then has its guts replaced with a simple call to be database specific. need to move from mysql to postgres and have 30-40 different maintenance programs and scripts that are programmed to deal with the API? simple, replace the guts of the API with the postgres API, and you're back in business. no need to write wrappers or even recompile.

Re:Its already there

[cnettel]

And nothing stops you from implementing that in COM (where the "interface" aspect is very much a part of the concept) or .NET. In fact, that's the way ADO data access worked in Windows, to continue your example. That doesn't magically solve the issue of incompatible DB syntax. I think there were some plans on implementing a version of Gecko that exposed the MSHTML interfaces, but I don't think they ever got close to actually finishing it.

Re:Its already there

[man_of_mr_e]

Actually, .NET is far closer to SOM than COM is. One of the advantages of SOM was that it was, in essence, a runtime object system, so you could "inject" new methods into objects if you wanted (kind of like how smalltalk works). .NET allows just that sort of thing as well (though it's not widely known) through a process called Reflection for runtime discovory of objects, and you're allowed to modify objects on the fly.

No, SOM and .NET aren't exactly the same (and .NET is a lot more than that, including sandboxing, a virtual machine, etc..).

Re:Its already there

[demon driver]

So what? Just superclass them with the same class!

Still not clear how it will help?

[William Robinson]

I could be wrong...but I am not too sure I understand how SOM, another kind of Object Model, simplifies building an Operating System.

Linux Desktop environment use bonobo implementing some of its services to achieve reusability and that too is not universal. (Gnome uses it). And, though not often, messed up interfaces could prove to be counterproductive.

IMHO, a good disciplined approach could produce solid architecture, providing robust environment. Though unrelated here Apache mesmerizes me, how it could achieve many concepts of OO.

Re:Still not clear how it will help?

[Knuckles]

Linux Desktop environment use bonobo implementing some of its services to achieve reusability and that too is not universal. (Gnome uses it).

No, Bonobo failed and is being phased out: http://live.gnome.org/Bonobo

It's the best outcome

[jsse]

The good things about OS/2 I described before is blessed by the wonderful SOM. I really can't tell you all the good things about SOM here, but during the time I wrote apps at IBM in C Set, SOM really save us a lot of time and efforts.

Open Source should be the default "archive" choice

[TakeyMcTaker]

Ignore IBM and OS/2 and everything, just for a second, and review this hypothetical situation on its own:

A Big Computer Giant (BCG) purports to be very Open Source friendly. They defend OSS products and licenses, even using their own lawyers, and make a lot of money using/supporting OSS, in their own hardware, and in huge consulting contracts. It turns out they have this collection of source code that they aren't really using anymore, and they have complete rights to at least part of it. Let's just say they only have 2 real options when archiving the source code they own the rights to:

1. Keep it locked in some internal media or shelf, never to see the light of day, unless an internal developer finds it interesting and digs it up for an internal-only project. The internal project may never see the light of day either.

2. Put it on the Internet, and Open Source License it, preferably with an existing OSI license. Not only could the online repository be considered the source "archive", but it also leaves the possibility of growing a redundant, living archive. The source could then be provided by various OSS repositories and mirror hosts.

I know I'm preaching to the choir here, but shouldn't #2 always be the default, or at least the first option considered? Even if you're not an OSS nut (like me), you have to admit the hypothetical company looks pretty darn hypocritical if they don't choose #2, when given the choice, early and often.

So am I using a hypothetical situation to say that IBM (BCG) is a big hypocrite if they DON'T release and apply an OSI License to SOM/DSOM source, ASAP? Why yes I am! How could you tell?...

Of Course!

[Tragek]

All should be open source. Source should never be hidden. It violates all ethos, it violates all truth, it violates my first GPL right to see the code!

Re:Of Course!

[Anonymous Coward]

What colour is the sky in your world?

Re:Of Course!

[Idiot with a gun]

What colour is the sky in your world? Why, I can easily recompile the sky to be whatever color I want it to be, now that you mention it.

Re:Of Course!

[pyrrhonist]

Why, I can easily recompile the sky to be whatever color I want it to be, now that you mention it.

Not unless you're using Firmament Beta 12. The latest stable release only supports Skies with 4 bits per pixel.

Re:Of Course!

[Arimus]

Ah, but he's got the source so he can modify to run 16 bits per pixel - just needs one hell of a CPU and graphics card, and if you do a poor job your sky ends up panicking.

Re:Of Course!

[CAIMLAS]

Dude, you remind me of a Dr. Bronner's Magic Soap label:

Absolute cleanliness is Godliness! Balanced food for mind-body-soul-spirit is our medicine! Full-truth our God, half-truth our enemy, hard work our salvation, unity our goal, free speech our weapon!

Re:Of Course!

[Tragek]

Pretty much what I was aiming for :P

Re:Of Course!

[Locutus]

What colour is the sky in your world? if your sky were a SOM object, I'd subclass your sky, add a sun and some puffy clouds.

LoB

Re:IBM not so open

[the_womble]

In what way is Red Hat not FOSS? All the cod in the Red Hat distro - that is why Centos can exist. I think Suse is all FOSS as well.

Business issues for open-sourcing SOM/DSOM

[Duke]

What are the business issues that would convince IBM to assent?

I can think of three:

1. That it would hurt Microsoft
2. That it would hurt Microsoft
3. That it would hurt Microsoft

Re: Business issues for open-sourcing SOM/DSOM

[jzhos]

mod parent up. that is right on the point.

OK, that's cool, but...

[niteice]

Why is everyone raving about SOM? Admittedly I never used OS/2, especially never coded for it, but from what I've read it seems like it does lots of the same things as COM. Why is SOM such a hot topic? People are saying that it could make Linux an attractive option for programmers if open-sourced and ported, doesn't DBus fill that niche? (on that note, is comparing SOM and DBus an apples-oranges comparison?)

So why do programmers care about SOM?

Re:Open Source should be the default "archive" cho

[Anonymous Coward]

It's not hypocritical because they didn't tell other people that they must open source everything. Also, you are somewhat familiar with the logistics involved by projects, archives, etc. You might consider that open sourcing is far from a 'default' choice, there are plenty of considerations:

1. it takes time to look over something to make it ready to see the light of day. You have a reputation to uphold.
2. you might want to make money off the software, now or in the future. as much as i love and support and contribute to open source, there's nothing wrong with that.
3. thanks to certain lawsuits, there is some perception in the industry that open source is risky. someone might sue you because you use linux. So, it takes work (lawyer time) to make sure code is clean
4. open source needs a community to really thrive. interested contributors, maintainers, etc. you would really like to see it 'picked up' by someone if it is going to be thrown over the wall
5. i don't think the safety of archiving is a major concern. probably more true for small companies that are not as likely to be around for the long term

demanding (not asking) that something on someone else's shelf be released is not really going to give co's a warm feeling about putting anything out there. you might reconsider your statement in terms of damage it could cause to open source...

just some things to consider.

Re:SOM sucks ! Fragile OS/2 Desktop proves it

[jzhos]

For those who do not know much about SOM, this wikipedia link compares it to MS COM http://en.wikipedia.org/wiki/IBM_System_Object_Model/.

The following quote supports the parent's point

The most notable difference between SOM and COM is support for inheritance -- COM does not have any. It might seem odd that Microsoft produced an object library system that could not support one of the most fundamental concepts of OO programming, but they did have their reasons. The main issue is that it is difficult to know where a base class exists in a system where libraries are loaded in potentially random order. COM demands that the programmer specify the exact base class at compile time, making it impossible to insert other derived classes in the middle (at least in other COM libraries).

SOM instead uses a simple algorithm, looking for potential base classes by following the inheritance tree and stopping at the first one that matches; this is the basic idea behind inheritance in most cases. The downside to this approach is that it is possible that new versions of this base class may no longer work even if the API remains the same. This possibility exists in any program, not only those using a shared library, but problem can become very difficult to track down if it exists in someone else's code. In SOM only solution is extensive testing of new versions of libraries, which is not always easy.

CORBA?

[jockm]

I would ask the question of what SOM/DSOM would bring to the party, when much of that technology went into making CORBA. CORBA is here now, has multiple implementations, both commercial and open (source and beer).

The only thing I can point to is GUI components, but those were either tied to a specific implementation (OS/2) or to an additional frameword (OpenDoc) and I am not sure they would be of much value.

Re:Who needs the SUN?

[Nazlfrag]

I guess you meant CORBA, and no, even if OSS was an entity (perhaps you meant OSI) it is far from abandoned.

SOM = CORBA 1.0

[micromuncher]

SOM implements a subset of the CORBA specification. Other technologies implement CORBA. Some already are open source (MICO). So... consider porting to another ORB.

And for the person who mentioned Apple... Apple implemented a subset of SOM specifically for OpenDoc. Though highly cool at the time, it was too castrated to be useful and has been surpased by other technologies for robustness (like J2SE/J2EE). Don't forget cool stuff like Spring... Lots has changed in 10 years.

Re:SOM = CORBA 1.0

[micromuncher]

SOM/DSOM was an implementation of the CORBA spec, and yes they had meta-classes, but the meta-class (though cool) isn't much different than a java Class and was often used for object caching / singletons.

Linux doesn't need it. It has D-BUS

[Cyberax]

Linux already has more more powerful D-BUS system (http://en.wikipedia.org/wiki/DBUS). It's already a base for PolicyKit, HAL daemons, soon it will be used in Upstart and so on.

It's MUCH MUCH easier to use than COM or SOM. And I still remember working with OpenDoc, so I don't really share good feelings toward SOM.

Re:Linux doesn't need it. It has D-BUS

[DrXym]

D-Bus is an interprocess communication layer. It doesn't do you any good if you want to embed one visual component inside another, or save a compound document etc. So while it might be great and all, it's only part of the puzzle.

Anyway, COM isn't hard to learn - it's actually incredibly simple. It's all the OLE2 related stuff that makes a mess of programming visual components. Steer clear of automation, OLE and it's pretty easy to knock up stuff in COM.

Re:Linux doesn't need it. It has D-BUS

[sjvn]

D-BUS is great. D-BUS is wonderful. But, DBUS and SOM do entirely different things. DBUS is meant to be a universal IPC (interprocess communication) mechanism for the Linux desktop. See http://www.desktoplinux.com/news/NS4449390454.html for more details. SOM's a set of object libraries. You would use them together. In fact, D-BUS, since both the GNOME and KDE communities have embraced its use, would make an ideal interface for SOM. Or, in other words, DBUS would enable applications to more easily access the power of SOM objects.

Steven

Re:Linux doesn't need it. It has D-BUS

[Cyberax]

What use is SOM then? You can write D-BUS services directly in your native language (most D-BUS language bindings have their own IDL and proxy compilers).

SOM is more comparable with KParts and Bonobo. KParts have already evolved quite a bit ahead of SOM and Bonobo is slowly evolving too.

Objective-C bridging?

[Jimithing DMB]

Not ever having used SOM I can't compare it directly but it seems to me that the basic idea is to have an object model with a hard to break ABI (i.e. you can add methods without breaking it) and interoperability with other languages.

Apple has distilled various Objective-C bridging efforts into a common bridge library that is now used by PyObjC (which predates the Apple/NeXT merger) and RubyCocoa and is slated for use with .NET. By virtue of being able to mix Objective-C into a C or C++ file (more or less simply by changing the extension from c to m or from cc/cpp to mm) you instantly get interoperability with C and C++. If you absolutely need to call Objective-C from plain C or C++ you can use the underlying objc_msgSend call.

Both the PyObjC and RubyCocoa bridges allow you to implement Objective-C classes in Python or Ruby so as long as you can stomach writing your interface declarations in Objective-C you can get most of the way there to the holy grail of cross-language object interoperability.

If you think about it, it's not much different from any other object model it's just that Objective-C happens to have a direct implementation in Objective-C whereas COM and SOM don't. The biggest downside of course is that you do need an Objective-C compiler and the only one is GNU. Then again, GNU is a decent enough compiler and runs on almost all platforms so it's not a huge deal. If you have really performance-critical code paths you can always write them in plain C with your vendor compiler and link to them.

Is it useful?

[Z00L00K]

Will that source code really be useful? A component like this may be completely useless in any other environment or maybe it can be seen as a curiosity and nothing more.

Some of the ideas may survive in new solutions, but it can be a really bad idea to take this out of concept and trying to graft it upon a different platform. The amount of work involved for that operation may exceed the amount of work to re-create the functionality without source.

object models...

[sentientbrendan]

there are already a lot of object models out there.

There's various incarnations of CORBA on linux, COM on windows, XPCOM which is used for firefox components, DBUS on Gnome and now KDE.

Component software is a Good Thing(tm), but all the various implementations essentially do the same thing, which is to allow a cross language interface to be specified in a way that doesn't require wrapper libraries to be written. Also, it is common to include some kind of remote procedure call specification.

>>In over a decade and a half, no one (but maybe Apple) came close.
Funny story, Apple uses COM.

Actually, it would be more accurate to say that apple doesn't have a component model (you have to use wrapper libraries or objective-c++). In a few instances where this is a pain (plug-ins) they just use com.
http://www.macdevcenter.com/pub/a/mac/2004/04/16/com_osx.html

It should be noted that Apple's com implementation is *very* basic and not really meant for developing cross platform COM libraries like the article suggests.

Re:object models...

[maestroX]

There's various incarnations of CORBA on linux, COM on windows, XPCOM which is used for firefox components, DBUS on Gnome and now KDE.

.. and OpenOffice, which also has his own version.

The beauty of SOM/DSOM is/was that it's fast (unlike CORBA), stable, elegant (no dll-hell), well thought-out and very extensible.

I've seen various incarnations of object models in KDE/Gnome alone, why start over everytime and set yourself years back? It took the Gnustep several years to implement an existing standard.

Re:object models...

[sentientbrendan]

>I've seen various incarnations of object models in KDE/Gnome alone,
>why start over everytime and set yourself years back?

The Linux community has a strong case of Not Invented Here

http://en.wikipedia.org/wiki/Not_Invented_Here

which is why you see every library being redesigned over and over again, instead of just fixing the existing standards. In my opinion, the inability to use existing standards developed for other operating systems, and the insistence on reinventing the wheel is probably the biggest thing holding back Linux.

We're finally standardizing on DBUS, but the equivalent technology (COM) has been in windows since 16 bit; however, the Linux community always views any technology from outside, especially from Microsoft, as somehow ineffably inferior to their home-brewed concoctions. Instead we've sat around for decades with the C ABI as the dominant "object model" (if you can call it that), on Linux for the inability to standardize around any kind of component technology developed outside of the Linux world, including CORBA, COM, SOM, etc.

Even now, DBUS uses the GObject model, which is C ABI based.

Why resurect old stuff + What about Thrift library

[ivec]

Is SOM/DSOM really that unique?
Rather than trying to resurrect a 10 year old source base, and to re-create a community around it, would it not be preferable to join an existing, related, project, and expand/improve it in order to cover SOM's unique features?

SOM/DSOM being 10 years old, I suspect that its support for Java, Python, PHP, and Ruby isn't very mature (ahem).

For instance, I have heard good things about this multi-language interface library (open-source project lead by facebook):
      http://developers.facebook.com/thrift/
Would someone know how it compares to SOM/DSOM ?

Re:Open Source should be the default "archive" cho

[TakeyMcTaker]

It's not hypocritical because they didn't tell other people that they must open source everything. The hypocrisy I intend to imply does not stem from what IBM (or BCG) is telling others (though it would be interesting if IBM was really requesting other companies Open their source, yet not doing the same). My BCG hypothetical company didn't do that at all. What I was speaking about was a company that directly benefits from Open Source, which is provided freely by others, yet doesn't participate in it whole-heartedly with their own source, when given the opportunity. Unused internal source is ripe ground for Open Sourcing -- you never know what will take off, until you have the courage to put it out there, and let others have a go at it. I don't mean to imply Open Sourcing old, unused code is an easy task; but I would say it is "low hanging fruit", compared to starting an OSS project from scratch.

demanding (not asking) that something on someone else's shelf be released is not really going to give co's a warm feeling about putting anything out there. you might reconsider your statement in terms of damage it could cause to open source... Bravo! Spoken like a true Anonymous Coward! ;P

Kidding aside, you make plenty of good points. Those points are not necessarily related to my statements, though.

I didn't say that the two forms of archiving (private vs. public) don't pose any difficulty, or that they are even of equal difficulty. I was just bringing up the fact that both forms of "archiving" are *available*, in this case and many others. Each method choice results in a different potential set of public opinions.

Calling someone a "hypocrite", and "demanding" something from them, are two very separate acts. One is a matter of stating opinion, and the other of calling to action. If you choose to equate them, that just shows a personal bias against the word "hypocrite", more than anything. ;) Perceived "hypocracy" can be debated in the public sphere. "Demands" are a one-way street.

Re:Open Source should be the default "archive" cho

[sw155kn1f3]

Because in real world companies exist to make money, not please OSS crowd.

Re:Open Source should be the default "archive" cho

[greg1104]

Releasing internal code as open-source has many costs associated with it, from software packaging issues to legal ones. Unless it's clearly a case where that code has some value to a larger community, paying those "free the code" costs may have no payback for the company. It's completely reasonable to say that the default position is "do nothing", which also costs nothing, and only consider the costs of releasing as open-source when there's a demonstrated need or desire to use the code outside of the company.

To think of it another way: if you were IBM, would you rather spend a hypothetical $1M cleaning up crufty old code and having your lawyers review it so you can release it, or spend the same amount of resources improving/extended existing open-source projects? Choosing to ignore your old code just because you're friendly to the open source community isn't necessarily hypocrisy, sometimes it's the only rational business decision.

Re:Open Source should be the default "archive" cho

[TakeyMcTaker]

Because in real world companies exist to make money, not please OSS crowd. Yes, but when a big chunk of that money is made based on products *provided by the OSS crowd*, doesn't the opinion of that "crowd" become important? Marketing is a huge factor in making money, and public perception is a big part of that.

Re:Open Source should be the default "archive" cho

[TakeyMcTaker]

Choosing to ignore your old code just because you're friendly to the open source community isn't necessarily hypocrisy, sometimes it's the only rational business decision. I wasn't speaking about the relative costs of each archive type (public vs. private) at all. I was just pointing out that it seems hypocritical to benefit so much from a system that you are not willing to contribute to, whenever the opportunity arises. IBM benefits financially from OSS, so any expenditure in contributing back to OSS could be seen as returning the favor. It could also be seen as a marketing cost, since it obviously has more to do with long-term industry good will than "the short-term bottom line".

I look forward to hearing more about IBM's "rational business decision" in any case. I find it funny that so many people are defending the decision before it's even made public. The term "rational" requires factual substantiation, not hyperbole. The term "hypocrite" has no such requirement. ;)

Re:Open Source should be the default "archive" cho

[sw155kn1f3]

And how many %% of OSS crowd really care about OS/2? I don't care, for instance. It's just small %% who remembers what OSS is and wants its source. You can't please everybody with finite resources.
As another poster in this sub-thread said, open sourcing any code isn't free at all.

D-BUS != COM/SOM

[jopsen]

As far as I'm aware D-Bus is a message passing bus for interprocess communication... not a object binding library.
I think the main difference, in terms of approach, is that D-Bus can't communicate with libraries, but with running applications or daemons. Am I right?

Re:D-BUS != COM/SOM

[Cyberax]

It can be used for intraprocess communication, but is not really the best tool for it.

It's much easier to use GObjects or KParts or your favorite component tool. And then use D-BUS to connect all of them together (for drag&drop, for example).

Re:D-BUS != COM/SOM

[fast penguin]

GObject is a library that adds some semantic sugar and utilities to make developing in C object oriented like. I don't know what GTK+ developers can use that is equivalent to KParts. I think Bonobo or maybe GtkPlug?

Re:D-BUS != COM/SOM

[Cyberax]

GObjects take the same niche as non-OLE2 COM objects in Windows. GNOME Bonobo is roughly the same as OLE2 and automation in Windows. Bonobo is now deprecated and GNOME team plans to move away from CORBA to D-BUS (http://live.gnome.org/DoYouKnow) and other solutions.

COM/SOM/CORBA are just plain too hard to use, so they are becoming obsolete very fast.

Re:Open Source should be the default "archive" cho

[IntlHarvester]

Remember when Netscape put Communicator source on the internet, and everyone just whined how it didn't compile and how useless it was?

And that's when Netscape was the most popular browser, not some 12 year old abandonware desktop middleware that nobody used in the first place.

I don't think anyone active in the OSS desktop community wants a bunch of old unpopular OS2 crap. If anything, they have too many duplicate component models and they need to start standardizing, not add a new one.

just leave them alone

[Bizzeh]

did any of you write os/2? no... did any of you pay for os/2's development? no... so who do any of you think you are to demand the release of its source?
rather than demanding source code for closed applications, go outside, and realise there is a real world where people couldnt care less about things like this. there are more important things than wether something that has been long since dead is forced to be open source or not. just let it go.

Maybe because it would cost them money?

[Britz]

They would have to check the code. Maybe the responsible developers have left the company, or maybe they are on an active project. It would take time to assemble the original team (or what is left of it) and ask them: Do we have the exclusive rights to this code? And if they were unsure they would loose even more time trying to find out. Time is money. If those people get decent wages it could be a lot of money. If they were working on important projects IBM was to earn a lot of money for they would loose that too. If it would be free for them to release code I suppose a lot of companies would do that. Simply because they don't use it anymore.

Why is SOM / DSOM needed?

[DrXym]

SOM, COM, CORBA are all much of a muchness. They let you define APIs to objects and interfaces in language neutral IDL, generate stubs that you implement and let you call those objects from other libraries, processes or even machines.

If Linux really wants language neutral interface bindings, it already has it. There are numerous CORBA implementations, idl compilers and so on. For example GNOME offers ORBit (a Corba implementation) for embedding UI components via Bonobo. Bonbo appears deprecated and is being replaced by another language neutral offering called D-Bus. KDE also appears to be moving from DCOP which is yet another IPC model to D-Bus. I don't know what either platform's plans are for components since D-Bus is for IPC, not embedding components, maybe Bonobo and KParts will live on for a bit yet. Firefox is also built on top of a language neutral object model - XPCOM is similar to MS COM but cross-platform and has bindings for several languages.

I do think OS/2 should get something like WINE - an emulator / runtime that lets OS/2 apps run on Linux. Part of that could of course be SOM & WPS support. I'm less sure why Linux would need native SOM since there are already so many choices. Another one ported from 90s codebase isn't going to offer much.

Re:Open vs. Close

[ta bu shi da yu]

If only someone on slashdot could filter out the above stupidity. That's the third article I've read this one. It contributes absolutely nothing to the topic being discussed.

Anonymous coward: SHUT UP!

Not much worth without the WPS

[krischik]

As other pointed out: there are enough CORBA implementations out there. The only advantage of SOM was that ist offered igh performance in an non distributed environment while beeing compatible with it's distributed peer DSOM. But even that is not so cool any more.

What was realy cool was the only real application build with the SOM: The Workplace Shell. Neither KDE nor GNOME can to what the WPS could do.

nobody has really tried

[Trepidity]

Yes, SOM/DSOM had some good points. However, it was also pretty horribly complex. Especially to implement, and even in many ways to use. As a result, almost all other approaches have gone for something simpler rather than trying to recreate it. One of the better ones IMO was NeXT's Portable Distributed Objects (PDO), which was so dirt-simple to use that NeXT engineers developed an infamous reputation in the 90s for writing letters to the editor in response to CORBA articles showing how to do the same thing in PDO in some ludicrously small fraction of the code. It was also incredibly fast. GNUstep has a reimplementation of PDO, though I don't think it's broken out into a nicely reusable library (I could be wrong).

Basically: weighing all the pros and cons, nobody else reached the conclusion that writing their own version of SOM/DSOM was the best option available, so they all did different things. I don't know if this was necessarily the right conclusion, but it's hardly that SOM/DSOM is some magical bit of code that nobody else could've reimplemented had they wanted to.

If they combine this with some other bits...

[downix]

Sitting here thinking on it, perhaps what IBM should do is analyze all bits of OS/2 to see what they do own. This is not just an exercise. One of the things I know IBM co-owns with Microsoft is the kernel, for example. Once they know what they ownint he free and clear, remove all bits that they do not own, and begin replacing them with pieces of Linux, writing layers to give them a new, IBM distribution. Call it Linux/2, or Linux OS/3. Compatibility with legacy OS/2 apps, features like SOM, a window-manager based on PM, and Linux brings with it Windows compatibility from WINE. Would be a nightmare scenario for Microsoft.

Frankly...

[CAIMLAS]

Frankly, I don't care how it is achieved, but I am really getting tired of having different UIs to perform the simplest of tasks, particularly when newer applications seem to be defaulting to the GTK2 file dialog more often than not. That dialog is irritating enough on its own, but not having it be identical, from one application to another, really starts to get under one's skin.

If SOM is the best method to do this, then bring it on. It seems to me that, in 2008, there could at least be something which developers are attempting to use.

You Don't Want It

[scherrey]

I did two contracts with IBM in the early 90's, one on the OS/2 2.1 change team. In both I "got" to deal with SOM and its implementation source code. It's a giant nasty C macro & function pointer hack. OS/2's Workplace Shell was very cool but the underlying implementation was pretty nasty stuff. One of my fixes was dealing with how slow it was populating icons in folders. SOM is a good example of the "Prototry" anti-pattern where one does an initial implementation of a cool trick then ends up shipping & extending it rather than ever bothering to architect it right in the first place. I can understand why IBM doesn't want this source out in public. FWIW - I also had to deal with some of the Microsoft source code, especially device driver stuff. Was the worse C code I've ever seen in production...

If you like SOM & Workplace Shell features you'll find it far easier to implement on top of Qt/KDE or wxWidgets or a smart functional integration of some Boost library features & a GUI than you'd ever have hopes of getting that code to work with anything modern or useful today. I loved OS/2. Borland had a Beautiful C++ compiler for it and CSet/2 was one of the better standards compliant compilers at the time as well. They're all bit rot by now though. Appreciate the memories but let this one die.

Re:You Don't Want It

[Esther Schindler]

That was definitely true of SOM in the 2.1 days. SOM 2 took it from a cool, technically interesting hack to real code.

Re:You Don't Want It

[scherrey]

I find this difficult to imagine as IBM was definitely not interested in investing in re-engineering any of the source code. A friend of mine and I put together a proposal to re-write Presentation Manager in 32bit in a few short months and it was simply not considered. A re-work of SOM would have required a rework of WPS and I'm fairly certain I would have heard about that effort. Once they moved from Boca to Austin that was it for real OS/2 dev that I'm aware of.

Re:You Don't Want It

[Esther Schindler]

I covered OS/2 actively for several years past that point. I knew several of the developers, and was one of the lead perpetrators of the OS/2 WarpTech conference in 2000. So yeah, I have some idea of how much work continued to go into development. And how little went into marketing. ::sigh::

Re:You Don't Want It

[scherrey]

These (and yourself) must have been some dedicated guys & gals. :) Well - would be cool for you all to make a BSD licensed version of the same concepts and maybe use a language that better naturally enforces these ideas. Might get me all nostalgic again!

Pure API Linking

[Doc Ruby]

Yeah, why does it matter anymore which language some binary was written in, in order to link it to another binary library? Everything uses a stack of pointers or values to pass arguments, and a symbol table to resolve references to instructions and data external to the binary package stating the reference. That seems to offer a language-neutral resolution of those references across binary packages, regardless of which language was used to generate them.

Sure, there could be some differences, like argument ordering on a stack, and even datatype conversion between them (eg. Pascal "count+data" vs C "NULL terminated" strings), even byte endianness. But those are deterministic differences that could be directly mapped to one another in a wrapper (even at the expense of some performance). And having the sourcecode for both could allow analysis that reverse engineers any incompatibility at link time to wrap or rewrite it.

Is there really no hope for reconciling the binaries for inter-language linking without extra programmer intervention?

Re:Wah? Esther Schindler long since dead ?

[Hal_Porter]

Probably she asked Cutler and Ballmer to open source old versions of Windows NT. Then she woke up six months later in hospital and decided that maybe suggesting open sourcing OS/2 was a better idea.

petition

[elkosmas]

Online petition http://www.petitiononline.com/sombsom/

Re:Open Source should be the default "archive" cho

[rastoboy29]

How could they be a hypocrite for not releasing that code if they ARE helping the open source movement with their lawyers and stuff?  They're only a  hypocrite if they take and take and never give anything back, which they do.  They don't have to give up EVERYTHING they could just to not be hypocritical.

human nature

[grikdog]

18-1/2 missing minutes? What 18-1/2 missing minutes? I love the private sector.

Re:IBM loves Windows, flip flops on Linux

[demon driver]

Right. BTW, it originally was MS's threat to stop selling Windows (probably Windows 95, then) to IBM's PC division which made them stop actively marketing OS/2 for the home and small business market in the mid-nineties. At a time when the battle was not yet won for MS, IBM gave in, not having enough confidence in their own, much better product.

Re:Author

[XO]

I'll make a case for PHP. It makes a rather fine language to write things that are best output in HTML. Like dynamic HTML pages. Kinda like what it was originally built for.

It's easy to learn, if you're familiar with the C syntax ...

Adding all the other garbage to it is an Epic Fial.

Re:Author

[rbanffy]

"It's easy to learn, if you're familiar with the C syntax ..."

There is a trade-off. This ease of assimilation drags PHP's expressive power and level of abstraction closer to what you see in C. This is a very low standard.

Re:Petty issues for open-sourcing trees.

[hey!]

So would closing down the chair industry.

If I had that kind of control over the chair industry, I could hurt Microsoft a lot more than that.

Re:Author

[XO]

Well, I didn't have a lot to say for it. It seems to me like it's very easy to write some seriously godawful stuff in PHP, and although most of the people that have used the stuff I've written consider me brilliant, if they were programmers, they'd probably be all like "WTF?" looking at it.

But then, I look at most everyone else's PHP code, and it makes absolutely no sense whatsoever, so I guess we'er all in the same boat there.