System iNetwork

I think the laments regarding the death of the standalone AS400 are way overblown. Good riddance to geared down, over priced hardware. It is the IBM i OS that counts. That is what runs our applications. What we have now is great, market priced hardware to run our apps.

Looking forward, I would like IBM to focus its efforts on providing the basic building blocks of the OS which 3rd party software vendors can build upon. An example is the ILE language precompiler and integrated debugger. IBM does not have to provide new features to CL, RPG and even SQL procedures. What it needs to do instead is provide the base features in the pre compiler APIs that customers and vendors can use to enhance those languages as they would like.

The precompiler and debugger view APIs enables customization of the RPG compiler. Would you like RPG to support a new BIF that makes it easier to call a web service? You can use the precompiler to translate the RPG code that contains your new BIF to RPG code which implements that BIF ( similar to how a Macro is expanded to actual executable code ).

To complete my example, I would like IBM to improve on the pre compiler and integrated debugger view APIs. Currently, you can't map variable names between debugger views. This is why ( guessing ) sql procedure debugging is such a hassle. Also, the mapping of source code lines from one view to another is not as clean as it could be ( use F10 in the debugger to step thru a running sql procedure to see what I mean )

-Steve

Hey Chris -
Good thoughts, but not complete. IBM does market an all-in-one machine- the mainframe. Like the '400, the mainframe has suffered through some serious naming issues; S/390 -> zSeries, etc.

The difference is that the '400 is aimed for small to medium business, though it can grow to support fairly large business sites. The mainframe was initially aimed at the government, a few universities, and only the very largest business entities of the time.

The mainframe also has a much longer history. It is what made IBM sucessful of course, and the IBm 360 mainframe was wildly sucessful when introduced in 1962. (1964?)

But here's the key, the IBM Model 360 computer was a "bet the company" deal back in 1960. IBM spent close to a billion *1960* dollars on it.

IBM officially expected to sell a couple dozen of them - but wihtin six months of its introduction, they had more than a 100 orders for 'em. And those orders kept pouring in.

Companies like RCA, then competing in the computer market, took one look at the 360 and exited the computer business.

Incountable billions have been spent on the mainframe since then, counting both R&D, Design, and of course, customers buying the things!

IBM certainly hoped for as similar buying frenzy with the AS/400. They wanted a wildly sucessful platform that would generate an IBM controlled marketing space in the SMB market.

The 400 and the S/36 S/38 machines before it, were affordable to even very modest businesses. Far less expensive than say a Vax, and much more capable than a PDP-11.

And compared to the UNIX machines of the day, such as VAX's - they were fantastically well supported.

But, they were also critically crippled in comparison to a mainframe. IBM did not exepct any compeition between the platforms.

And as expected, the AS/400 platform shone in its day, and as you mentioned, it generated not only a fanatically loyal following, it also generated an entire ecosystem of products around it, very similar to the mainframe.

But it never made profits like the mainframe systems - how could it? It was affordably priced, but not so affordable that a SMB wanted to replace or upgrade it. And they could get along pretty well with one or two programmers too!

Once people bought one, they stopped spending on 'em! Horrors of horrors! What had IBM done?

They created a micro sized mainframe and an ecology to support it, and did it so well that they could not make outrageous profits off it that everyone had been led to expect.

Add one other factor, IBM has always portrayed the '400 as the machine you don't need to know anything about to use. You don't need to know how the processor operates! Or operating system internals. You could not study the assembler code and make the program more efficient, or handle bottlenecks, and so forth. IBM just claimed to know a lot more about that stuff than the end users. IBM's viewpoint was for platform programmers to write everything in RPG or COBOL and be quiet about it!

This of course, alienated all of the "old school" mainframe people, because on an mainframe, everything is configurable. System Programmers (the equivalent of AS/400 Operators I suppose) routinely write operating systems level code in assembler.

So what happened again? IBM made the gear so good and priced it high enough that small business owners were absolutely loath to upgrade.

Yet they hid the internals of the system so well that really good programmers disliked the machine because they had no opportunity to tune it or the software to their liking. Argh! Double Whammy!

And for the coupe de grace, IBM made the OS and programming tool licenses so restrictive, there were few-to-no people who could afford to have one and run the OS on it legally. (They changed all that a few years ago though.)

Now the one good thing to come out of the latest dustup is that the platform is perhaps about to be opened up a bit. It will be probably be a case of too little too late, but who knows? IBM is, altogether pretty smart! They may know something I don't! I hope they do. :)

-Paul

What is IBM's leading product? It has been a while since I've read their financial statement, but I seem to recall that most of their revenue comes from Global Services which defies definition, so perhaps the leading product is the IBM name itself. And Power is becoming a leading name too.

One thing that IBM is not, is a consumer products company, and particularly not a consumable products company, so I'm not sure the burger and hot dog analogy applies, but there has been too much competition and too little differentiation (except price) between "i" and "p" for too long. It's better to merge the two, and let IBM i join AIX and Linux as a mainstream offerings.

The merger dashed hopes of a number of programmers holding out for IBM to release a native GUI with support comparable to 5250 which would be a clear differentiator and presumably propel "i" back to its former glory. Oh, Well. Que sera, sera. What will be, will be. It's actually a relief to hear the refrain "it's over, O-V-E-R, it's done" repeated with regards to the separate servers. Now let the healing begin.

For disenfranchised RPG programmers, my advice is to find ways for RPG to participate in a Web services world.

[*Note From Chris: Re "programmers holding out for IBM to release a native GUI with support comparable to 5250 which would be a clear differentiator" . . . this is a particularly apt statement. At COMMON, Ian Jarman made an offhand comment on a tangential subject to me and several of my System iNEWS colleagues -- I don't have an exact quote, unfortunately, but basically my take away is that it's going to be less and less likely that we'll see any truly i-specific application innovations out of IBM . . . that couldn't be translated across IBM's product lines. IBM will certainly continue to make some investments in RPG, but I think there's a reason IBM has resurrected EGL . . . as "IBM's newest business language." Any significant new investment out of IBM will likely be cross-OS functional. That's my take. Not sure how that will shake out in reality this year, next year, and the year after, though. . . .]

There has been discussion on this pre-and-post the IBM announcement on Power systems at COMMON-Nashvile.

With the continual name changes that plague tracing the systems and applications, IBM still insists on "i".
Someone or more than one person mentioned Power/OS. I think power/OS would be the best choice as the Power chip has been in this system for years. Power could have much better marketing slogans than "i." They could market something like "I want Power for business." and it can run AIX and Linux too.

Although there is truth to the fact that IBM did themselves in by making such a reliable platform as S36/S38/, AS/400 and i, and this could be oversimplification, but if they get the hardware/OS/DB2 to (they have approached it but not yet equaled Windows boxes) a more affordable level, why would you order a Windows/Intel/AMD/Citrix box that will crash, when the AS/400, i, and now Power systems can run mulitple OS's, have 15,000 applicationas available for them, and ISVs that are working on more, and the system is rock-solid reliable for the price. IBM could have the majority of the market in mainframes and power servers.