System iNetwork

This may be big news for big shops that develop web apps. For those like me who buy 'AS/400' applications off the shelf, the change is meaningless. And annoying, I haven't even put the 515 I bought last month into production, and it's obsolete already. By the way, when we still call it an 'AS/400' it is because the CEO, CFO and other execs who need to be on board with IT strategy cannot be expected to keep up with IBM's semi annual name change. But mostly we just call it by the apps it runs. We buy business tools, not hardware.

Unbelievable. While these changes do sound great I think many people familiar with the number of name changes to OS/400, will shake their heads in disbelief at yet another name change and further loss of brand identity. Of course this is most likely part of the plan. Bury legacy and hope that "i" rises from the ashes.

Excellent announcement by IBM. The SQL based reporting application I wrote in 2007 will benefit greatly from the performance boost of these new systems. Without exaggeration, half the code and design work I put into that app was to get an acceptable level of performance. With these faster systems, programmers can focus more of their efforts on the features of their apps.

That said, too bad there is nothing for developers. After you add client access and the language bundle, a programmer has to pay over $10K for a system of their own.

-Steve

Chris,

I am stoked at the announcement. This is a big thing and ports this system out to "everybody". No more excuses about why you don't have the most reliable system in the history of computers.

Now we can take off and run businesses in this new world of commerece and interaction.

way to go!

Mike

IBM misses the boat, again!

I heard IBM was considering renaming i5/OS back to OS/400. But instead they chose 'i' as the name of the OS.

What IBM should have done, was not ask their engineers, not ask their "marketing" team what to call the OS, but asked people like us. I've been an advocate of rebranding OS/400 as "Blue" or IBM Blue for at least 5 years now. And "i" is what we now get.

If you think about the whole eco-structure, I would have done it this way:

Rebrand the p and i as POWER as they have done today, but rebrand i5/OS as "POWER OS".

Then when you order a POWER system, you can get POWER OS or AIX or LINUX. But no, instead we get AIX, IBM i, or Linux.


Just wish somebody at IBM would watch any other company make a major announcement. I heard the audience was bored out of their minds in Nashville.

So this makes the new name of the "SystemiNetwork" what?

As my editor pointed out this morning, "This is the dumbest name change yet. i? It practically disappears in a sentence. And did anybody think about search engines?" If you search i on Google, the system doesn't show up anywhere near the top; it's way below Star Wars Episode I and the I Index entry for the Catholic Encyclopedia. The New York State Tourism people should like it though -- you'll find "I Love New York" on the first page.

It's no wonder the platform isn't as successful as it should be. Nobody knows what they're selling, buying, or using. Fortunately nearly everybody still calls it AS/400 and OS/400.

A simple "i" for naming an Operative System is ridiculous -- how you could find in Google software that runs in "i"?

"i" name is even worse than i5/OS.

A name as PowerOS (one not too small nor common word) or PowerBusiness or PowerBlue (as Cozzi suggested) would be great.

A name easy and no-confusion for using in searches as "payroll software for BluePower"

Stupid stupid stupid.

Yet another branding change and yet every customer I visit still calls it an AS/400. More pointless confusion for the midrange customer.

Why the large price difference for IBM i 6.1 on the JS12 vs JS22 blades? The price per core on the JS22 is $14,995. On the JS12 the price is $2,245.

The S chassis ( JS12 ) is $4,499. The H chassis used by the JS22 is $3,849. Why does the chassis for the more expensive JS22 cost less than the one for the JS12?

Do you have to use an IBM chassis for the JS12 and JS22 blades?

You drink too much of the IBM cool-aid.

I too am dissapointed with the fact that they change the name every year. Even those of us that have worked on the system the last two decades cannot keep track of what the "moniker du jour" is.

IBM seems to think that changing the name will remove the perception that it is an out-of-date legacy system. As customers, we know that it is not. But consultants and prospective buyers are aware of the facts. This is not because of the name. It is because of two things: 1) IBM never really attempted to get the platform out into educational environment where kids could learn something about it, and 2) IBM has never made an effort to differentiate it in the marketplace. I guess they felt their marketing dollars were better spent elsewhere.

No question in my mind that separating the Operating System from the platform is a good thing. But the question remains: Is it too little, too late?

Doug Pence,
CPU Medical Management Systems,
San Diego, California

i THINK ( - therefore I am?)

Maybe it is time for IBM to haul out some of those old desk signs and hand them out again.

Rebranding every year is not the solution. But if they have to rename it, how about the OS that just refuses to go away, in spite of our efforts?

Without a native fully integrated graphical development environment, it is still just an AS/400.

Thank you for the exceptional hardware. I guess without 3rd party software it is still just a green screen machine.

After all these years I just don't expect anything more...

"without third party software"..."green screen" blah blah blah

There is no truth to this stuff whatsoever. There is no 5250 unless you specifically order it and pay extra for it. You only see it if you want to, and speaking of third party software, only if you buy third party software written for 5250 and not front ended, in other words, your choice, or you write it yourself, again, your specific choice of what you complain about.

So don't choose it and quit complaining.

rd

Initial thoughts on impact of this announcement:

1. Impact on Platform: IBM POWER Platform promises to do well and may be much better. It also enables larger customers to consolidate their System i and System p workloads more optimally.

2. Impact on OS/400: It will still continue to serve existing customers. But, why would anyone buy i (OS/400) licenses to run new Java/Websphere applications on IBM POWER platform? Won't they prefer AIX or LINUX?

Naru

April Fool's Day or truth?

[*Note From Chris: All irony aside . . . truth.]

The announcement of Power Systems is definitely a step in the right direction. There is tremendous opportunity and possibility with the merging of System i and System p into Power Systems. And I am definitely excited about the announcement.

Ross and Mark specifically emphasized the importance of academic initiative and education during the Town Hall Meeting. That is very reassuring and heartening to the community. As some of the audience mentioned, the i professionals are aging rapidly. Unless there is new blood joining the i community continuously, finding IT professionals with skill in i will become a more and more critical issue over time. And that will be devastating to the i and the Power Systems communities.

Another point that is stressed over and over again by the audience is marketing and advertising, and making Power Systems known to business executives and managers. IMHO, that is extremely vital to the acceptance, approval, and adoption of Power Systems in the business community. How are business executives and managers going to accept Power Systems or i easily, readily, and with open arms if they have not even heard of Power Systems and i before?

One minor issue -- I thought Power Systems are going to target SMB also. On the other hand, the Power T-shirts given out at the entrance of the Town Hall Meeting come only in sizes L, XL, and XXL. What about S and M? :-)

[* Note From Chris: Keng! Say, didn't the University of Nebraska at Lincoln receive an IBM-COMMON innovation award this year? Nice work, by the way -- I heard that you and your colleagues . . . if I have this right, were delayed by airline flight issues on Sunday. In any event, kudos!]

Dr Johnson said that "patriotism is the last refuge of a scoundrel". Well re-branding must be the last refuge of a panicked IBM marketing executive.

We seem to have the same scene played out every year: "Oh my God, we don't know what to do with this AS/400 thing. I know! We'll change the name (again). That'll keep everyone so busy updating documents and webpages, they won't have time to worry about the future of the system they've worked on for decades."

If you want a new name, how about TITANiC (as in "rearranging the deckchairs on the...")

Why don't they do something truly exceptional and different, like a graphical subsystem? Maybe even a monitor and keyboard port on the back with a built in LPAR tool so I can admin my AS/400-iSeries-systemi-POWER server without having to buy another server/workstation with a different OS to configure it? How about some actual low-cost hardware that developers can buy and learn/develop on? How about making all the IOA's and hardware the same serial numbers and prices as the pSeries and stop charging me more for the same thing?

IBM needs to really step up to the plate and instead of changing the name every couple of years, put their money were their mouth is and get these machines in the hands of the schools, starting with high schools, so classes can be taught on them to the students taking business classes.

My hat is off to Gateway Technical College who received last year's Inovation Award for their ground breaking role in partnering with the LUG and also teaching a iSeries concept class to one of the technical high schools in Kenosha, WI.

This was the first time these students had heard of the AS/400 / iSeries / System i / i in their young lives and many really liked it once they found out that they could message their other classmates. :-)

My hat is off to Jim Buck and the rest of the teaching staff at Gateway Technical College in SouthEast Wisconsin. Keep up the great work in promoting this platform -- lord knows that IBM isn't doing the right job.

This from a person who, like most, works for a company that still calls the machines AS/400s because that is what they knew the company used since they converted from the mainframe. Even my title as "Manager of AS/400 Systems" will never be changed by HR.

The next renaming: just use the "dot" on the i for the OS name.

How do you Google for stuff about the i operating system now?

i's motto:

"i thinks, therefore i is"

The Power Equation announcements really emphasize that the "system" and base-level branding is defined first by hardware and virtualization technology, then second by operating system. You pick the server (or blade) first, then add one or more operating systems to one or more virtual partitions. Gone are the days when there was tight one-to-one coupling between hardware and operating system.

Also gone are the old gimmicks of geared down processors, and processors constrained by the absence of L2 cache, and adapters to activate "interactive" workloads, and higher-priced iSeries disks, and similar shenanigans that placed the iSeries at a competitive disadvantage.

I agree that this is a step toward mainstreaming IBM i. I liked Bob Cozzi's idea of calling it POWER/OS but that would have made AIX and Linux appear like disenfranchised step-children.

[*Note From Chris: Great point about throttling, speeds, and the like, Nathan. With consistent hardware pricing, the differences between AIX and i will be more readily seen -- if an application (and all the pieces and parts you need along with it) prices out less than its i counterpart, that disparity will be much more difficult to hide.

Also, about the step-children . . . yeah, I totally see your point. IBM might have been better served by going with Power-i 6.1 as the operating system, with Power-AIX and Power-Linux as the nomenclature for the others. Seems pretty obvious to me, but then again, IBM didn't ask me. I'm just a guy who spends much of his day typing and talking on the phone. Ha ha. But even with Power, that would require that IBM would be interested in keeping the "Power" name for as long as it develops its POWER processors. Shift to a new processor technology, and IBM would have to be willing to change names again.]

"my name is long and important"
"my name is long"
"my name is"
"my name"
"my"
"m"
...oops, I'm gone!

I can't help but wonder why this announcement is being touted as huge. Everyone can go on and on about what the name should have been. At the end of the day, the real question to IBM is why is this better?

To me, this name just makes it even less likely that a Google search will provide anything useful. The combining of System i and p was the next obvious step to anyone who has been paying attention. So, I ask again, what is the big deal? Is this name change all IBM has in store to make the System i more mainstream? The spin here is making me dizzy....

I just listened to a summary of the announcement for business partners. The Google complaint was addressed in the Q&A after the presentation. Just Google "IBM i" and the first page is all relevant.

The announcement merely reflects reality.

Everybody's comments must be from business users or ISV's. Two weeks ago we held a career fair at our college. In the IT portion, one of our instructors asked how many students (16-18 years old) had heard of IBM. Less than 20% raised their hands. If the name IBM isn't recognized by potential college students, how is rebranding the best hardware platform on the planet and so simplifing the name of a great operating system to something that no one understands going to help IBM gain recognition for this great product?

I mean, frequent name changes are fine, if it's for a person who is a criminal running from the law with several aliases. IBM, try listening to your customers and encouraging the IT professionals that are going to be needed after the baby boomers retire. OK

What a surprise! Another year, another name. IBM's big announcement?! Maybe next year IBM can rename it to " ". That might help the system disappear even more. Great job IBM.

With consolidated hardware, at least we will no longer pay more for the same components.

The big question in my mind is, how much is the i OS going to cost?

And of course, I still dream of a developer's machine for under $2K.

Bill mentioned "...one of our instructors asked how many students (16-18 years old) had heard of IBM. Less than 20% raised their hands."

I agreed with Bill that this is a critical issue -- exposure and awareness of IBM, and in this case, informed about IBM Power Systems. And that is the reason I am thrilled that both Ross Mauri and Mark Shearer stressed the importance of academic initiative and education at COMMON.

“The UNL-IBM Global Innovation Hub Program” at University of Nebraska – Lincoln (UNL) aims to inform, expose, educate, and train business students and business executives on Power Systems. This is a niche and unique program as it complements and supplements existing IBM Power Systems academic programs in many other schools that teach more hands-on courses in Power Systems or IBM i. Making business managers and top management executives aware and knowledgeable in Power System is vital. First, major investment decisions are made by business executives or have the intimate involvement of top business executives. Second, with the movement to make Power Systems appealing to SMB, it is critical that business executives have basic knowledge, awareness, and appreciation of Power Systems. Thirds, exposing and educating business students and executives on Power Systems will have both immediate and long-term impacts on sales and market shares.

We are honored and excited about winning the Education Innovation Award at COMMON this year (as Chris mentioned earlier – Chris: Thanks for the kind words). We look forward to working with IBM, COMMON, IBM partners and clients, and other academic institutions to further advance the Power Systems and IBM i academic programs.

Not many people are familiar with the University of Nebraska-Lincoln (UNL). We are best known for football (i.e., the Cornhuskers). Other less known facts are: (i) we are ranked as one of the Top 50 Public Universities by the US News and World Report; (ii) we are ranked 6th in the Most Popular Colleges category by the US News and World Report in 2008 (behind colleges such as Harvard, Princeton, Stanford, and MIT).

If you have ideas and suggestions on collaborating and working with us on Power Systems and IBM i academic programs, please contact us and we would be most happy to work with you. Thanks.

Consolidation, what means IBM with consolidation.

IBM System i and p 520 Express versus IBM Power 520 Express AIX Edition (8203-E4A1)/ i Edition (9407-M151).

And wy not IBM System i/p 570 as IBM Power 570 with all processor features for IBM i (i5/OS).

Who could understand this (re)naming rules.

Who cares about the damn name?!

The announcement is the convergence on one machine - Power.

IBM is in the relatively unique position in that it still has its own OS'es on its own boxes. So in comparison to every other wannabe vendor you have both a machine and multiple OS'es to "Google" for. Jeesh, much whining about nothing.

I for one am glad that i can run multiple Os'es on the POWER box, that is a huge marketplace advantage. Try doing that on Wintel and WinDoze!

Whoever said that this announcement has nothing to do with developers is insane. This has everything to do with product development.

For all those complaining about not being able to google an i....
try--- ibm i ---- works every time...:-)

I have read all of the above criticisms and I disagree with only those that suggest that IBM is doing right by what once was the AS/400 platform. It is no longer a platform so IBM cannot have helped the platform by eliminating the platform? IBM committed identity theft by eliminating its own brand -- a trick that Microsoft would have liked to accomplish through the business process but could not. How does this help anything for the System i or AS/400 community? The vestiges of the AS/400 are identifiable only in an operating system with a lousy name that only those who must, will remember. As bad as it is, instead of IBM i, methinks iIBM would have even been better -- though lousy also.

This "new" operating system is not universally available on all platforms. That would be nice for the AS/400 community and that's why it will not happen. iIBM is only available on a hardware system whose home operating system has always ended in an x (Unix - ala AIX and Linux, ala Linux). By the way, a freshman in our business program at Marywood University would fail Marketing 101 for playing with a successful brand name and would fail a second time for calling anything IBM i or iIBM. But IBM knows better than its customer set and so we are left with just an OS that is now nothing more than a guest OS on the formerly System p platform boxes. These boxes as we all know were enriched last week to be able to fully swallow the full identity of the once mighty AS/400.

All of the good, including pricing cited above in the responses of a mostly faithful community could have been achieved by IBM at any time in the past if it had wanted to market, rather than milk its system. IBM owned what this once successful system platform and it could have chosen to make it even more successful. But, like so many things in business, IBM would have had to invest in the platform over the years, rather than planning its demise. That identity removal is the only way that the pricing anomalies could have been worked out would again warrant our freshman Marketing 101 student a failing grade.

IBM had the most unique machine ever invented and now it no longer exists. Don't try to sugar coat it. Since 2000 and somewhat before, but surely with the homogenization scheme of the eServer family, IBM has chosen not to differentiate this fully proprietary system as the best for Business. The fact is that a best system cannot sell when the word best is not permitted in the marketing material. IBM's goal for many years was to make the AS/400 the same as all other systems.

IBM chose not to market the AS/400 or iSeries or System i in any meaningful way. This was not lost on the AS/400 community. Burying it inside the System p carcass will certainly not help bring the knowledge of such architectural elegance and excellence to the forefront. There will be no new AS/400 bigots created from this new announcement. Nobody, other than we, the once faithful, even know something happened last week. Our neighbors don't know. Students in Universities don't know. Even System p people don't know because it is their box and their system that the new Power Systems are based upon. Nobody but us will know because when markets the new system, it will say that the IBM i operating system also runs on the new box. Your neighbors, the students, and the former System p crowd will way, "what is that?" Nobody will answer. IBm i also runs on it. So, IBM i is by definition in this scenario, if you will pardon this logical conclusion, an "also ran" in the Power Systems OS race.

Is this what Microsoft would do? Considering that Microsoft's engineers had happily christened their first GUI OS, with the name "Interface Manager," Bill Gates surely thanks his lucky coins each day that his marketing department said "No Way," and they coined the sheer-genius brand name "Windows." The name matters since in branding, that's about all you have.

IBM just doesn't get it. Many do not know that even Bill Gates did not know how precious that name was to be. He offered IBM the name "Windows," when the two companies were playing ball back in the late 1980's but IBM knew better. Big Blue insisted on calling its then mighty PC operating system "OS/2" -- one of the all-time worst names for anything by anybody. IBM i comes right before it on the list. History is already written on OS/2. The lack of sales and the multi-billion dollar write-off showed the brilliance of IBM's naming fiasco. Using the "Windows" name as offered by Gates would have saved the day and countless billions of dollars wasted on OS/2 salvation attempts. That's how important a name actually is.

To make OS/2 a real marketing winner, the IBM Naming Society was not finished. Since the GUI came out a year or so late for the original OS/2, the namers chose to separately market the GUI from the black screen base product much like DOS and the early Windows were actually two separate products. This gave the IBM Naming Society a chance to create a new name for the GUI.

If you didn't live through this, you'd think I was kidding. Apparently impressed by Microsoft's original name for Windows, the "Interface Manager," IBM's namers wanted something just as good. Of course, they did not want to completely steal Microsoft's original name. So they decided to use just one of the two words.

In the same spirit that prompted IBM i, because the names System i and i5/OS weren't making it, the namers decided to give the folks at home a name they would never forget. It would be completely unlike Interface Manager but similar enough to make it a real snappy name. They chose "Presentation Manager." Can you believe that -- for a consumer product. Our Marketing 101 failing freshman student would again wonder why that naming trick did not turn the whole PC Division around.

Call it whatever you like but even IBM must call OS/2 a huge marketing failure. Presentation manager did not rescue it from its fate. How can a superior hardware and software solution not be marketed successfully. The idea that IBM would have a superior hardware and software with its PS/2 and its OS/2 (including Presentation Manager) is not lost on the AS/400 community.

Was there not a superior hardware box and a superior (one of its kind) integrated operating system that comprised the latest System i and the AS/400s long before it.? With no visible marketing for the box, perhaps sales have been slipping so rapidly that IBM did not have to wait for the marketplace to defeat the System i. IBM took it out of the game last week.

Microsoft wins because it believes that failure is not an option. It took down Novell by investment, patience, and persistence. When Microsoft ran its business on 23 AS/400 boxes, AS/400 aficionados were laughing at the Gates Company since it was having a real struggle getting off the AS/400 platform. Using investment, patience, and persistence, during this time, Microsoft figured out how to run its business and how to take out the IBM's server divisions (all of them). This highly motivated company is well on its way to achieving its ultimate success. Last week's surrender of the AS/400 line helped Microsoft more than IBM. And when the mainframe comes onto the Power Chip, Microsoft will have just one server to beat, and maybe Intel will even help them. Though my heart is with IBM, my money is on Microsoft, and their not so friendly partner Intel.

The message for IBM is that by taking Intel's doink money and choosing to invest in shareholder stock prices rather than investing in product excellence, the IBM hardware product line, one server at a time, is being staged for defeat. Maybe Sam Palmisano ought to invest a few dollars in the IBM hardware and software product lines while its customers still care. Oh, and did I mention marketing?

and jus

I wrote my own feelings down on my own blog when I found out the news.

http://leedare-plex2e.blogspot.com/2008/04/damn-marketing-rebranding-machines.html

Lee.

What self-serving tech bigot would not claim that their system is the best -- HP, Sun, IBM, Intel, DEC? The "best system ever" is a perception much as marketing is. The top spot in world’s workloads would be IBM mainframes, not the System i, nor the AS/400 running OS/400. AS/400 is a terrible name, yet many people really want the name back. So how do you like your coffee, one lump or two?

Are we now going to argue that RPG is the best language ever and that ILE is the best runtime environment ever?

Trying to ridicule IBM into agreeing with your point of view is ridiculous much as trying to argue that Microsoft replaced the AS/400s with something other than a machine that can run very high workloads. So what point is there to argue that a manufacturer cannot consolidate many servers to achieve high workloads that competes with one AS/400, apparently either MS did this or replaced them with another server and OS brand. Either case, so what?!

My guess is that IBM spends millions of dollars analyzing data regarding their decisions and can justify the decision to consolidate certain OSes on the Power box. How is this a wrong decision? IT shops would rather have a box that runs AIX, Linux, and the i than just one, plus more in the future. In addition, how can you determine the reason? Maybe they’re opting out of building anymore OSes and want to start merging them into one; maybe they
want to phase out the i; maybe they decided that a bunch of farmers with pitchforks and really small businesses that don’t keep upgrading to the next AS/400 are boat anchors and should be treated as such. Does Wintel and Windoze do anything different? They all keep upgrading, new hardware, new OSes, new deals.

Although Windows certainly is a neat name, at what point will they catch up to the top system manufacturers in server workloads, if ever? One can only guess how all of this will play out, but IBM is not the only one in trouble. Many businesses chafe at paying more money for the same software suites from MS with little business advantages. MS has many irons in the fire and is just as much an un-wielding, slow moving ship as IBM. IBM has
never conceded the server to MS and neither has HP or Sun. There’s a whole other IT world that doesn’t run IBM or MS and would more easily shift into the IBM Unix boxes than Windows. The thought process isn’t necessarily pointed in MS's favor either. Why should we care if the "i" runs on Windows, how about Windows can’t run on anything but Wintel and nothing else? How about I can choose many manufacturers to run flavors of Unix and Linux
and thus port my products? How about when developing products, I have to spend that much more money and devotion to one machine versus many machines? Sure Windows has a lot of product but a lot more cheap crap, too. How many
shops have shifted development to languages that run on many OSes such as Java, PHP, Ruby, C and C++ and away from Windows-only development? This is happening more and more, including a shift to Open Office.

I do not have to wonder why we are moving all development from Windows products to open-source and select languages like Java or neutral C/C++. Why the current product used by the whole company that runs our business is being currently migrating by yours truly from MS Visual Foxpro to Java and why the database and server of choice will be a non-Windows server. The product is huge and the Wintel and Windows servers are dying and the
Desktop maintenance is too expensive. Nor do I have to wonder why a huge company that I used to work for never upgraded their Windows XP machines -- everyone’s looking for alternatives and finding them. Doesn’t MS see the writing on the wall with Open Office and Linux desktops fast on their heels? Why do you think IBM is running
Apache, PHP and MySQL on the i?

I for one am glad that IBM makes the decisions and not college students. Concerning future decisions,
it is a far more reasonable choice to select a box that runs may OSes, has a great legacy and company and is certainly a better choice than Windoze and (insert any manufacturer) Wintel box.

I think the shoe's on the other foot and IBM's glass is half full. This is a much welcome move by IBM, well thought out, and the overall perception will be positive. Way to go and thanks IBM!

Another stupid name change! QUIT RENAMING THINGS AND START WORKING AT IMPROVING THINGS!!!!!!

Jason

At least the name change is meaningful this time. The power of the hardware hasn't been a problem since the early 90s. The problem is legacy code and legacy databases make change prohibitively expensive.

Wouldn't it be great if .Net languages ran natively on the i5eries? I mean, most RPG programmers can't write a stored procedure, can't use SQL, can't write ILE procedures ... apparently, RPG died 10 or 12 years ago. The i5eries should have a language that modern developers can use.

If neither the "integrated system" approach, nor the operating system architecture and elegance matter, I concede. And that may in fact be the case, as less and less in the IT community care about the IBM i operating system, the System i or the AS/400. Not being able to convince new users to use the platform without spending a lot of marketing dollars was not lost on the IBM executive team.

As you note again correctly, "What self-serving tech bigot would not claim that their system is the best -- HP, Sun, IBM, Intel, DEC?" Everybody touts what they know. But, isn't it possible that just one system is actually best?

What user today is looking for single level storage, 128-bit instructions, object basis, capability based addressing, and a hardware instruction set that includes a build index function over relational database tables. (Can't be done without the relational database knowing about the hardware). I agree that a big part of the problem is that IBM no longer had the resources to explain those notions nor to sell potential users that they were essential enough to pay a premium for a real "system."

And while all that advanced technology was very good, how about display files integrated with high level languages and unmatched programmer productivity for decades. If IBM perfected this notion with a browser device, there are many who think this discussion would be unnecessary as the platform brand not only would exist, it would be thriving.

You're right though nobody cares about the techno-facilities that always made the AS/400 a far superior machine architecturally than IBM's mainframe boxes. For years, CISC AS/400s were at 48-bit hardware (128-bit firmware) while MVS and mainframes were at 31-bit. Only recently have mainframes achieved 64-bits.

Nobody cares and nobody differentiates computer architecture today. Few even debate it as we are doing. For the AS/400 to be successful to its potential, IBM would have to engage new minds about the technical capabilities of its box and it chose to not do that. IBM chose instead to let the marketplace dictate whether it would be successful, rather than by marketing its products to be successful. That is my opinion after watching events in this industry from the time the System/3 was announced.

The results are in. I don't know how one could consider the results as a success for IBM. From 1990 to present, IBM's hardware revenue went from roughly $60 billion to $30 billion as its services revenue grew from about $4 billion to over $50 billion. Software is about $15 billion. IBM transformed to a services and software company while many of us were thinking its game was really hardware. It is my humble opinion that the $30 billion did not have to be given away to Sun, HP, Microsoft/Intel, Dell, or anybody. Can you imagine if just one of the $30 billion (each year) that is lost and not coming back were invested in marketing the only unique server on earth, how much different the results might have been? For the AS/400 as a notion, a brand, and a separate entity, it is over. O-V-E-R. The ride is done.

And, as for Windows / Intel not being the go-to-guys. They do it quietly but they are winning big time while everybody else is marking time or like IBM, losing market share. BTW, Sun and IBM also sell Windows Servers.

In 2005, Windows Servers eclipsed UNIX servers for the first time in terms of revenue, by nearly $200 million. According to IDC, over $17.7 billion worth of Windows servers and $17.5 billion in UNIX servers were sold worldwide in 2005. Additionally, Linux took third place at $5.3 billion and mainframes slipped to $4.8 billion. With sales of approximately $2 Billion, IBM’s System i fell short of making the report.

I admit I do not have apples for apples current numbers handy but I would bet a dollar to a donut as AS/400 revenues were falling in the recent past that the box did not win back any market share. For the full year 2007, worldwide x86 server revenue grew 10.4% to $28.7 billion. It is ironic that the $28.7 billion is so close to the $30 billion market share IBM lost. These numbers also show just how small a bit player in this huge marketplace, the AS/400 as a box with any name had become.

Once a lot of the cash had melted from the AS/400 cow, trying to maintain the notion of an AS/400 brand, no matter what it was or is called became too expensive for IBM. With no differentiated, targeted marketing campaign, there was no opportunity for a resurgence of AS/400 popularity. So, IBM did what it had to do, it killed the brand.

Positive spin suggests that it merged two brands. The fact is that at least in the announced models, the merge has been to what was the System p type approach not the System i or AS/400. System p users should not experience a change other than a faster system. The change, both physical and emotional is for the former System i user set.

IBM is a services and software company first today. There is no room for a rogue system that does not need lots of software pieces and lots of services.

I don't know if it was intentional, or it just happened. I don't know whether it will be good for IBM or not. I do know that unlike the PC platform, IBM did not sell the brand because another company would eat up the industry, including IBM if it had this box in its arsenal (opinion). Just as in 1988, when the AS/400 was fresh and relevant and it was marketed well, it eliminated the entire minicomputer market. Please note that was not by accident. The box was so much better, it was an easy sell.

Without major differentiation, the AS/400 cannot sell. There is just too much to talk about. So, now it is no more. The brand is gone. I don't think that is good for me as a consultant and a player in this arena nor do I think it is good for anybody relying on a thriving AS/400 platform. It is just my opinion and I mean no disrespect to IBM for having that opinion. I do think IBM could have made the outcome much different if it chose to. The verdict, however is in. The AS/400 is no more. We'll just have to see what the future brings.

Brian, I am convinced you are totally correct in saying "it is over. O-V-E-R. The ride is done."

When watching the evolution in this platform on a daily basis, things happen unnoticably, but when you look over a longer period and compare for example the market penetration and prestige this product format once had in 1998 and where we are today in 2008, the writing is all over the wall.

I personally think the o/s will linger on until the average age of the as/400 developer approaches 55-60, that gives us still one decade to go assuming we're all around 45-50.
So comes the question what's the work asignment on the i going to be? Babysitting 20 year old not-always-well-structured rpg and cobol code may probably be the main dish. As a consultant, there may be some opportunities if the outflow of developers outpaces the migration of as/400 workloads to other platforms, giving you a last chance to shine and earn some big bucks. Now there may be some regions in the world where the as/400 format is actually thriving and gaining market share (Ibm should tell us, why they keep so silent!?), but Brian, let's not kid ourselves.

As you've said, "it is over. O-V-E-R. The ride is done."

PS: anybody noticed AIX and Linux run natively on the Power hardware, while i runs only under a virtualized i/o system (VM io)? Ibm didn't bother to rewrite the i drivers for the newer hardware. Another babystep in the down-the-drain evolution of the i maybe?

[*Note From Chris: About the virtualization, I noticed but haven't looked into it. Good catch; I'm glad you brought it up. Doesn't seem like it would create much of a performance issue for the vast majority of System i owners, though, since the newer boxes are so fast. In terms of your thoughts on the demise of System i, I'm not yet convinced this is particularly relevant. The concept is annoying, most certainly, but I suspect that the unification had to go one way or another and the obvious direction was toward Unix where high-end performance is critically relevant to the Unix-focused market. But there may also be sound technical reasons, such as the need for a way to deal with disk differences between previous i and p systems.]

"For the AS/400 as a notion, a brand, and a separate entity, it is over. O-V-E-R. The ride is done. - Brian Kelly"

It's actually a relief to hear that finality expressed so strongly and repeated. Maybe the healing process will begin.

Since about 1990 there has been too much competition between the AS/400 and RS/6000 and too little market differentiation, except for price.

In 1991 I worked for an ISV and was assigned to create an AIX version of our credit union software, based on our AS/400 version. The effort nearly bankrupted the company, and I was eventually reassigned to support the AS/400 version. The RS/6000 version was dropped entirely - money down the drain.

Marketing just got simpler. Whether AIX or IBM i, you just say your package runs on Power!

Chris,

The demographic situation of the as/400 developer population will become a very important factor imho. Let's assume the average age in the US Army is 50 years with 70% of the soldier population in the 45-55 age bracket. Good news, we would live in a very peacefull world or otherwise the army isn't appealling to the youth anymore. Assume the only in or outflow of the workforce is happening in the same age bracket, then come back in 10 years. Right, no US Army anymore.

Ok, there are presumably 500,000 as/400 developers in the world, will the 1,000 or so young college students that IBM boasts about help to change the massive, on average 50 year, age pyramid?

So let's focus on the next 10 years. Are companies shifting in a panic their AS/400 workloads to other platforms? I don't think so, roughly estimated 5% of iseries workloads is moved each year while 15% is undergoing a modernization transformation. The rest remains in a dormant state [the panic will probably set in with these dormant companies who won't find any personel in 10 years].

I give these figures without any thorough investigation, but I would suggest somebody would write an impartial and accurate article about this which would give us some insight in "i"s future as it relates to people and labor.

[*Note From Chris: Nate Viall has been tracking this better than anyone I know . . . and while there will be jobs for RPG development for a long time going forward, I'm hard-pressed to believe that there will be much in the way of "new" jobs. I think the challenge for organizations with custom code in flavors of RPG will be long-term strategy -- how does the company want to utilize it in 5, 10, and 15 years? As a magazine, we've been thinking about this issue as well. IBM is seeming to be pushing toward EGL, with it being "the newest business language" as George Farr as said. There's also the idea of using PHP with RPG, and that would give a company an opportunity to bring in newbie talent that has PHP experience or education and expose them to the business of RPG. But yes, demand for talent . . . would be on my mind if I had serious custom investments.]

“the ride is over…”

If you mean the AS/400 brand name, that ended years ago. Brand name and the constant arguing the same is thankfully over. If you also mean that the legacy code-base may be in jeopardy via that RPG is not an Internet language and there is likely no 5250 follow-on then I concede. BUT the System i ride is hardly over.

However, the perception that the AS/400 was somehow architecturally superior to all the other IBM OS’es seems a bit of presumption considering that most of the base architecture of the /34, /36, /38 and /400 all post-dated mainframe developments. The /400's approach was to integrate the /36 and /38 boxes into one and add the integrated DB. MI/TIMI and single-level store were both /38 implementations. However, the main communication architecture was SNA which had an added protocol – 5250(LU7)(originally /34), an easy step since the mainframe had SNA with 3270(LU2). The main system indexing mechanism and used extensively was originally based on the ISAM architecture. The current TCP architecture was borrowed from AIX. There was no 64 bit architecture, so the AS/400 started with the /38's 48 bit architecture. So the point is, many if not all of the technologically superior implementations that run on the /400 did not start there and were indeed borrowed.

I see no point in lamenting the fate of the AS/400 as it like its predecessors before it lives on in the System i. The System i is not going anywhere anytime soon. Its current transformation is certainly very welcome and includes robust Internet capabilities. After all we all survive by constantly morphing ourselves to some degree. And sometimes that may require a revolutionary step. It is a combination of considerable engineering over the decades which will continue to improve. The IBM and nix OSes not only can run monster batch workloads but also lead the Internet revolution. So before one concedes defeat I would take a look around and ask where are they and what's the competition doing? A far as I can tell, chasing IBM and Sun. Then ask yourself if MS is so comfortable in its own skin how come it just has to have Yahoo? Why do they put huge resources into server development? I would think that it is they who hear footsteps and see the writing on the wall.

Then ask yourself, if server based workloads, server applications, internet and browser based interfaces are the revolution, then what do you mean the ride is over, the ride is just B-E-G-I-N-N-I-N-G! GO P-O-W-E-R!

Mr Iyeeee,
Why buy proprietary hardware if you're doing open source development? Why, if you're doing Microsoft .Net development, buy hardware that is incompatible? I hope you're not holding your breath waiting for someone to buy an IBM midrange box so they can run internet applications.

System i, i5eries, AS400, whatever-they're-calling-it-today is a batch box in an interactive world. It has monolithic applications when everyone else designs components. Its databases are designed around primary keys when open source frameworks require foreign key relationships. It's green-screen, not browser. The dozens of RPG programmers I know want to "learn" the Websphere IDE and continue to ignore ILE and SQL.

The IBM midrange systems competed against Burroughs, Wang, DEC. They're gone because the midrange market consolidated as generic hardware got cheaper and more powerful. Sun's stock price has fallen from over $100 to $15 because no one needs their hardware. Generic hardware comes with easy to use, point-and-click management systems which everyone in the world knows how to use.

When I go my start, the Sys/38 I worked on had almost a GB of disk and half a MB of RAM. That company's accounting needs haven't changed that much in twenty years. Their apps and their website now run on Microsoft. What would compel them to go back to a proprietary midrange system?

Mr Greg,

All machines are proprietary. Open is not something one would apply to a Windoze OS, nor the Intel boxes they run on. Most of the open source movement and software runs on a variety of OS'es not just Windoze and applies more to a NIX OS using either Java or C/C++ (not VC++). Stock prices come and go and are not really germane to anything discussed, albeit pointless. The Power box runs Linux, AIX and i. The i machine can also run Windoze.

Running internet apps on the i is done all the time. Unless you've completely missed the point, the Power machine runs all three OS'es named above and will run more in the future. So I can breathe very easy.

The standard Windoze OS will only run interactive apps and generally poorly.

Most machines even running XP have to be re-booted, re-started frequently. The Java framework (invented by Sun) is far more extensive and robust than .NET. Both MFC and NET are a plague on programming and thank God they are just stuck on Windoze.

I have no idea what "batch box in an interactive world" is supposed to mean. Does not Windoze have server editions, albeit poorly designed and yes crash all the time? I have worked in shops with hundreds of Windoze "servers" that take an army of people to administer and yes do crash on a daily basis and yes the internet runs like mud and daily some critical service and machine requires re-booting. The number of Windoze network admins and servers per user is ridiculously high.

So interactive Windoze OS doesn't run so well in batch, otherwise there wouldn't be all these problems. Interactive can mean a variety of devices attached to a “batch” machine include NPT’s and telephones, etc. I would expand your thinking beyond Windoze.

"It has monolithic applications when everyone else designs components."

What components are you talking about and/or designed? Most developers build a series of modules in an application, this hasn't changed much. Vendors have tried to spin the notion of components in order to create a "plug and play" environment which I have not seen used that effectively except in the vendors products. On Windoze that generally being a COM "component". So instead of shipping DLL's they now ship ActiveX or COM components. Just more Windoze crap and not widely developed outside of the vendor community.

"Its databases are designed around primary keys when open source frameworks require foreign key relationships."

You need a database 101 course. RDBMS’s have been around quite awhile, long before SQL server. The i integrated database has been around since the /38 and is basically an VSAM technology with the capability to support logical views, ie keyed relationships among various indexes. DB2 is the relational (SQL capable) version on the i box, so it has both types of DB's. Open source frameworks have nothing to do with foreign keys. I'm starting to get a clearer picture of why you enjoy Windoze so much.

"It's green-screen, not browser."

The built in green screen features of the i box also exist on every major server except Windoze. Windoze has console apps, but they really aren't used for much actual interactive design. Even today you will find developers programming "green screen" apps on every major platform but obviously not as much as before. Browser based apps are out-pacing desktop apps. Which means server based applications (straight CGI, Java, PHP, Ruby, AJAX, etc) favoring the major server vendors, ie that's not Windoze. Again, that's why MS realizes the Windoze desktop OS has a finite life. Develop a browser app and now the choice is which server is the best to run those apps on, not Windoze. Further most all major Java applications that I have been technical lead or participated in never, ever deploy on Windows, almost always a NIX box. These are large government institutions.

Basically, Windoze has been regulated to whatever toy apps they can be used for but major institutional deployments? Nope, it's an IBM or NIX box. Also, the browser runs on a myriad of devices and OS'es not just Windoze including Linux desktop. The MS faithful keep trying all sorts of psycho-babble to convince the IT world that the a Windoze server can compete.

"The dozens of RPG programmers I know want to "learn" the Websphere IDE and continue to ignore ILE and SQL."

No programmer expecting to be considered competent can ignore SQL; it is fundamental. The i has had SQL for years via DB2. I don't know what RPG has to do with anything, it is not what the POWER server machine is about nor has been for a long time. It happens to have a esteemed legacy much like COBOL. RPG tends to be an easy target for those less capable to make fodder of.

The IT market has always been consolidating, all stocks have suffered including MS's.

The reason Windoze continues to exist is because so-called engineers who point and click their way to some nirvana like state where they are completely oblivious to the facts, much like the Matrix.

You can bank on the fact that the POWER will be purchased to run internet and open-source applications by those institutions needing the horsepower.

The company I consult with now has probably 40 some Windoze/Intel boxes. There is always some service interruption or outage. SQL Server runs like mud and is re-booted at least twice weekly. DB2/Oracle/MySQL/Postgres anyone? If the impetus is to create server applications such as all internet applications then Windoze becomes a moot point and Intel will have to look for greener pastures as the Wintel workloads start to be run on more server capable OSes, such as IBM and Sun and Power and Nix variants. The company I work for now hired me to convert a huge desktop application from Visual Foxpro to Java. They are the atypical Windoze customer who cannot wait to move all apps off of Windoze. It's happening as we speak all across the world. How long do you think it will be before Open Office starts to over-take MS Office, the final straw? The DOJ had a very hard time swallowing MS upgrades and many companies run very outdated versions of MS Office. It just a matter of time. It's back to the server baby!

Mr Iyeee,
Your rantswer is really too long to bother with. I'm Java certified and RPG certified so accusing me of prejudice in favor of MS-SQL is a non-starter.

I believe that Sun's old motto, "The network is the computer" has been borne out. Big box solutions are designed for a single system; they are designed so they do not scale. Networked applications running on smaller systems necessarily are built to scale. I don't know a thing about open office and the DOJ (Mr Iyeee's area of expertise) but I do know about layered and distributed architecture. Have you known a Java developer to jump up and down when Intel upgrades their processors? I haven't. Java and Windows developers scale their apps across several servers, adding more when needed. Big box developers have to wait for CPU upgrades to get a performance increase which leaves them waiting for the next version of the POWER chip.

Five, six seven years ago, I thought there would be a lot of work wrapping legacy AS400 code and data with Java front ends but it never materialized and I don't think it ever will. AS400 (i, i5eries, etc) shops just don't have the technical expertise to do anything new.

We are in agreement - the IBM midrange is a great system. I, however, think it is the most glaring example of hardware being crippled by its users, in this case, the "waiting for retirement" crowd of developers. The internet runs just fine on *nix and Windows and when a POWER chip is used, it generally is to run AIX not i5/OS.

Mr Greg,

You baited the conversation in a number of ways. My purpose was to support the announcement and IBM. But since your purpose like others that have preceded you is to be a Windoze troll then you get what you deserve. In other words, you used me to wiggle your way into the discussion to trumpet Windoze instead of making your own independent comments. If you want to say IBM sucks and .NET is leading the way, then just say it instead of hiding behind me.

You are what’s called a “Concern troll”. To quote “The goal is to sow fear, uncertainty and doubt within the group”.

“so accusing me of prejudice in favor of MS-SQL is a non-starter”
Try again, I never accused you of anything, on the contrary it's an open discussion. I merely pointed out the weaknesses in your statements.

Such as:

“Big box solutions are designed for a single system; they are designed so they do not scale.” This means exactly what? So I take it that mainframes and mini’s could only handle the most trivial workloads? Please indicate clearly the name of any major institutions, the application and details that have ported their applications to Windows .NET that actually required horsepower. How about the reverse, those Infinite software ports that failed and had to return to the i.

and this

“Networked applications running on smaller systems necessarily are built to scale.” This means exactly what? An attempted trumpet of Windoze.

and this

“don't know a thing about open office and the DOJ “ Obviously you don’t perform research in your chosen field.

and this

“The internet runs just fine on *nix and Windows”. The internet does not “run” on Windows, never has. More psycho-babble. Attempt to trumpet Windoze again.

and then this

“Java and Windows developers scale their apps across several servers”. ROFL! Attempt to trumpet Windoze again by aligning it with Java.

Most mainstream Java deployments generally involve 2 servers(app and db), 3 if you count the HTTP server and then the client tool of choice per user. This is not “distributed” nor “scaled” but a basic 3 tier application. If you mean something along the lines of Google or other major search vendors that is very rare. If you mean EJB, RMI, SOA, etc. that is pretty much left to the major application server vendors such as Websphere, BEA, Sun and the like whereby developers are hidden all the dirty details. Before I left my last job one of the final projects I worked on was primarily a BEA Weblogic product including RMI and SOA servers. That was designed as distributed application from the ground up involving a multitude of servers and developers. It was deployed to a large government agency using Unix servers and was expected to scale upwards of 1000 concurrent users.

Most Windoze deployments are desktop built as proprietary one horse applications which happen to use a DB. I know of very few that are actually deployed as server based applications, even among the .NET crowd. Doesn’t mean it’s not done just not mainstream. Back in the client/server day we used to call these applications networked applications but they were really desktop applications that accessed a DB. There was no application server involved. I now work in a mixed use shop, the only network applications are Java based, and the rest are standalone desktop Windows based that access a DB, currently being re-engineered to Java based web and server applications.

I think the best thing is to most definitely stick with Windoze and your .NET. Leave network and serving computing to competent, albeit mostly non-certified experienced developers. Clearly the meaning of IBM and POWER is lost on you.

In the future I would also stop trying to use Windoze in the same breath as IBM and Sun since their achievements are considerable. Microsoft has either stolen or copied most if not all of the ideas that ended up in Windoze. To top that off you are comparing a company who does not even have a machine to their name to the likes of current leading vendors such as IBM, Sun and HP.

For the record, I and most people I know have no real problem with Windoze other than when people keep trying to make this constant comparison to the legacy’s of major vendors who have led and continue to lead IT. Microsoft has it’s contribution as the desktop GUI system, that is all and it is fading fast. If it were otherwise then MS’s contributions would have spread across all of IT. This has not happened, on the contrary it’s the other way around, many of IT’s achievements are ported in some fashion to Windoze from where they were invented.

My start on IBM came in the form of the S/36 writing thousands of lines of assembler. I’ve been involved in Java development for almost a decade. My first Java project involved developing a servlet and database framework for a cost accounting system for a government agency. They refused to use any other application server but their IBM Domino Go Webserver. At the time it was primarily a HTTP server but fortunately supported Java servlets but not database connection pooling. Therefore I had to write a multi-threaded database connection pool for us to use. To this day I have no language certifications and probably never will, I don’t need them nor would feel the need to parade them, nor hide behind them. So next time you want to screw around then just be prepared to defend yourself and your statements.

Now we’re done.

After reading the previous post, I am appalled by the comments -or better the ranting and flaming- of 'Iyeeeeeeee' directed at another poster who happens not to agree with his opinion.

Here are some examples of the tone used:
- since your purpose like others that have preceded you is to be a Windoze troll.... (continues raging)
- You are what’s called a “Concern troll”....
- Obviously you don’t perform research in your chosen field...
- More psycho-babble. Attempt to trumpet Windoze again...
- I think the best thing is to most definitely stick with Windoze and your .NET. Leave network and serving computing to competent, albeit mostly non-certified experienced developers...
- Clearly the meaning of IBM and POWER is lost on you...
- Microsoft has either stolen or copied most if not all of the ideas that ended up in Windoze... [strong statement which in other circumstances could mean court]
- So next time you want to screw around...

And the last statement, to "finish off" the previous poster, is "Now we’re done"
(end quote)

I find it very regrettable these kind of agressive posts directed to a person himself make it tru the screening. Some of us may chuckle at the agressive blows dealt here, but to most readers they offer no added value and should be banned all together.

Mr Ugeerts, thanks for sharing your wisdom. I shouldn't have disturbed the slumbering giant ....

As the Power chip and Intel's have adopted a multi-core architecture (i.e., a server farm on a chip), a couple of my friends have picked up the new language Erlang in order to design programs that take advantage of all the CPUs networked on a single chip. Apparently, i5/OS keeps each process on a single processor of a multicore chip. Erlang programs have the capability of distributing processing in lightweight threads across several processors. Like it or not, the problems and promise of distributed processing is encroaching on the i5, AS400, etc world with the multicore Power chip.

I think my cell phone has more processing power than the first S/38 I worked on. My 80GB iPod blows it away. Yet, so many AS400 developers I worked with write programs today with the same design they used 20 years ago. I presume that is a good thing? Seems strange to me.

Hi Greg,

No problem.
On the debate what the future will bring for "i", I think the demographic situation of the "i" developer crowd is a crucial factor; today we're all 50 so to speak, in ten years we're all retired and so is "i".
You just can't fight time.

How we've come to this situation is more or less a chicken-and-egg debate.

Either IBM is to blame to not enhance the system and keeping it competitive with other technology platforms (I usually refer to SAP which transformed from ancient mainframe cobol technology to a mouth watering I-want-that-too by all TOP-500 company ceo's) -or- "i" developers were too lazy to adapt to new technologies.

The chicken-and-egg debate will probably live on for a while; at least it is welcomed by blogs in case we've ran out of other subjects :)

PS: I only refer to the original S/38-AS/400 ecosystem (CMD DDS CLP DFU SEU RPG COBOL languages). Newer introduced frameworks like Java, PHP, Linux, Websphere... will undoubtly survive but not likely under a green screen "i" O/S. The "i" value is found only in it's RPG-COBOL application programs, stripped of it, these newer frameworks run much faster and much cheaper in for example AIX or Linux or under Oracle.

I heard "PragDave" Thomas speak at an OO seminar and he asked "Why do we do object oriented development?". He said it is because OO allows us to manage our code. That made sense to me. Where I used to work, they have 15 million lines of RPG and a simple field expansion, adding two digits to the account number, cost $7M. Did management and the developers learn a lesson and look for a better way to do things? RPG becomes viable again when developers start to use the ILE features so that change becomes less expensive.

I agree with your points. The AS400's benefit (ease of use) is its Achille's heel, it didn't attract enough geeks. Secondly, if IBM had evangelized RPG-CGI ten years ago, we could have at least gotten away from the green screen.

I doubt much has to do with attracting geeks, after all, unlike PC home computers, the demand for work on this system came only from Small and Medium Business (SMB) which in many cases gave onsite training or sent their IT people to training class at IBM.

SMB's in the years 1975-1985 really had not much of an alternative for IT than to use IBM Midrange. Mainframe was out of reach because too costly, the competion at that time (Bull, Wang,...) was way too inferior to IBM and PC's and Windows were still in their infancy. Most hiring happened also in that time. From the 90'ties, the hiring of new! AS/400 jobs declined sharply except for a small spike up in 1998 for year-2000 projects. After 2000, there was no more hiring, on the contrary, most of the 1998 kids had to leave the as/400 scene (lucky for them maybe).

I have about 150 addresses of colleagues, ages ranging from 36 to 67. The 67jr guy was retired but called back into service (he's as happy as can be). The 36 guy is frustrated as hell and trying to pursue other directions outside as/400.

I have no doubt to most of us this is very recognizable and we probably all know colleagues who are in the same situation.

There have been quite a few allegations and concerns expressed in this blog against RPG and RPG programming practices that don't seem fair to me. So in order to counter-balance the negativity, I've linked my name in this message to a couple screen shots of an application I've been working on recently.

The screen-shots depict a utility for uploading and attaching files to any type of record in a database. In this particular case the file-upload utility is accessed from a tab in a "person" maintenance screen, but it could be accessed from any maintenance screen and used to attach files to any record in my database.

This is a good example of a robust and modern user interface, where all the screen I/O, database I/O, and application control logic is handled by RPG code. And it's not monolithic code! Each tab is controlled by a separate application. Each application is further divided into screen I/O, database I/O, and application control modules.

A number of us RPG programmers have taken component-based design methodologies to heart and are implementing them in free-form ILE RPG. And we're coupling RPG with robust browser-based user interfaces. We're not stuck in the past. And in our case, RPG use is not declining.

If Mr. Iyeee's proposition is correct about the return to running apps on a single, dedicated server then, we can expect the market for web service developers to morph into one for server services developers (i.e., legacy programmers). I don't think web service developers will be asking themselves where their jobs went though.

Seriously, given Mr. Iyeee's resume, posted for us in detail above, he must know, at least by reputation, the power of JBoss on clusters of cheap, generic hardware.

Airlines have found that no matter how much leg room they offer, they'll lose customers to any cattle car operation that will save the customer a dollar. Hardware is the same. Apple sold machines that ran the Power cpu. Power may have been better but Intel was perceived by most as cheaper and/or better. Apple gained customers by switching to Intel.

Power6 simply is not going to bring new customers to IBM's doorstep. The perception of value has to be changed for that to happen. PHP and MySQL are small steps in that direction but it would need to run .Net and SQL-Server for a meaningful change to take place.

ugeerts

“I usually refer to SAP” You always refer to SAP which has not been a germaine topic ever as far as I can remember concering the i. Let's talk about Windoze lack of virtualization technologies. No relevance either.

“I only refer to the original S/38-AS/400 ecosystem (CMD DDS CLP DFU SEU RPG COBOL languages).” The i is far away a different machine and can run a myriad of current technologies. It has no relevance regarding Power unless one is citing history and legacy. If it is on topic then let's start a discussion on Windows 3.1 and 95. What's the relevance here?

“ Newer introduced frameworks like Java, PHP, Linux, Websphere... will undoubtly survive but not likely under a green screen "i" O/S.”
I would not considered any of these frameworks. A framework would be something like ICEFaces a AJAX based Java framework. Neither Java, Websphere or Linux can be considered new. Green screen has no relevance to any of the technologies listed. Green screen has nothing to do with either Java, PHP, Linux or Websphere, what's the relevance here?

“The "i" value is found only in it's RPG-COBOL application programs”
Another swipe at the i. The value in any machine and OS is the value the users and business place on the applications that run on them. RPG/COBOL is off-topic and a constant target by the Windoze faithful. What's the relevance here?

“stripped of it, these newer frameworks run much faster and much cheaper in for example AIX or Linux or under Oracle.”
Hmmm, now we've jumped to other OS'es. The i is more in line with both Linux and AIX. Since AIX is IBM's, the i and just about every other IBM machine runs Linux. I can't really respond to this because a large mix of concepts are convoluted into one statement. It is unclear what is meant here. Nothing runs in "Oracle". Oracle products run on other machines, like well IBM's. LOL!

Please show examples where these "other" "frameworks" run "faster". Please cite specifics, applications you have deployed to more than 10 concurrent users, what hardware configurations, costs, how you know they run "faster", etc. I think rather than waving a saber one might back up their statements a bit without hyperbole. Please show relevant experiences not things grabbed off the Internet. You're welcome to show off your resume.

Just to be clear I made this true statement:
“Microsoft has either stolen or copied most if not all of the ideas that ended up in Windoze.”

Verified by those who are not intent on re-writing history. It has been stated a million times, documented and proven in a variety of ways. Sure, sure I'm going to court. You might want to watch "Pirates of Silicon Valley" if not actually reading historical facts. I don't recall anyone going to court over telling the truth regarding M$. By the way, just how did Gates get the OS for IBM's first PC if not by deception?

The topic is the IBM Power announcement not "IBM considers running .NET on Power to delay the inevitable".

I'm strongly convinced running a .Net framework and Sql server in the "i" is totally out of the question for a very simple reason.
Suppose IBM allows these MS apps on his system (which would take an enormeous amount of time and money to port), and since we're doing this, let's trow in Windows itself also in the mix.

What will be the end result?
A system that could do everything a Wintel server could do. Does this help the (once) highly esteemed personality of the "i"?
Not at all. The "i" just becomes a plain Wintel platform.
That would be completely illogical from IBM's business point of view.

Moreover, IBM serves his customers already with a Wintel platform, called the "x Series". BTW a very succesfull platform (and I personally own one of them of which I'm very satisfied) along with the "p Series".

greg

"He said it is because OO allows us to manage our code. That made sense to me. Where I used to work, they have 15 million lines of RPG and a simple field expansion, adding two digits to the account number, cost $7M. Did management and the developers learn a lesson..."
There is nothing to learn. OO is not a panacea. Anyone who has done extensive OO generally via Java or C++(not .NET) knows that it stops outside of the neatly architected solution, stops dead at any RDBMS and a considerable number of other interfaces. In addition the cost to even attempt to make all things OO(not just the base solution) is enormous with questionable benefits. However, looking at your former companies problem from a different perspective, how much would the cost have been to build it using an OO language like Java or C++, probably triple if not more. So in reality, they saved money.

"if IBM had evangelized RPG-CGI ten years ago, we could have at least gotten away from the green screen."
Anyone could have gotten "away" from the green-screen with a little bit of effort and creativity years ago. However, the green-screen thrives and seems to be chased by just about every vendor on the planet. Funny, come to think of it M$ touts SNA server or Host Integration Server as the best 5250 solution. tee-hee.

ugeerts

"SMB's in the years 1975-1985 really had not much of an alternative for IT than to use IBM Midrange."
Nonsense. HP was very strong as were a number of other vendors. There were many choices. IBM just happen to be the best choice for many.

"The 36 guy is frustrated as hell and trying to pursue other directions outside as/400."
I in fact spent my happiest days in IT on the midrange. I have no regrets regarding RPG or green-screen and my resume would clean your clock in any modern technology.

You always spin everything you state about the IBM midrange into something negative or to be ashamed of. Your condescending attitude towards the i and it's followers is really something to be "appalled" of. If you're so smart and the i is so bad, why are you still on it, or you're not and just enjoy messing around with people. Even when indirectly discussing the midrange your negativity is apparent throughout.

Thanks, Nathan, for that positive post. We'll just have to show by example to counter the "can't be done on the iseries by RPG programmers" posts.

As for the constant harping about RPG programmer availability from a few, when I came on board the AS/400 in 1989 it was from PC programming, and everyone else in training came from other various backgrounds. When there's a need in our IBM midrange industry, we ramp up and train for hire. We always have.

I'm not concerned about more manpower needs whatsoever. I was concerned about people losing work until a couple of years ago when things picked up again. But I'm not worried about being able to bring people up to speed to work on the iseries and RPG at all. Seen it done too many times.

rd

greg

"If Mr. Iyeee's proposition is correct about the return to running apps on a single, dedicated server"
I did not say single dedicated server. Futher my comments were directed at your .NET and Windoze rant. If in fact you did develop a distributed app that runs n-way servers, then you would know .NET and Windoze has little to do with any of it.

"Seriously, given Mr. Iyeee's resume, posted for us in detail above, he must know, at least by reputation, the power of JBoss on clusters of cheap, generic hardware."
Again, I did not post my resume in any detail. Gave a few real-life examples. However, if you would like we can compare any time. Actually, the company I now work for recently decided to move off JBoss for performance reasons. I guess they weren't clustering enough, maybe you can consult?

"Apple sold machines that ran the Power cpu. Power may have been better but Intel was perceived by most as cheaper and/or better. Apple gained customers by switching to Intel."
Power chips are everywhere. Intel chips are everywhere. AMD chips are everywhere. They are many many uses for these chips, not just in servers. Get IBM Developer newsletter.

"Power6 simply is not going to bring new customers to IBM's doorstep. The perception of value has to be changed for that to happen. PHP and MySQL are small steps in that direction but it would need to run .Net and SQL-Server for a meaningful change to take place."
The annoucement was a convergence on Power. It will most assuredly bring in new customers. There are many, many products that run on a variety of OS'es including the i, where does .NET/SQL-Server/MS Word/Ms Office/Exchange/IE, etc,etc run? Your beloved M$ is a one huge proprietary bloated sinking ship.

"Big box solutions are designed for a single system; they are designed so they do not scale." and it's companion
“Networked applications running on smaller systems necessarily are built to scale.”
But really, are you ever going to back up ANY of your statements? The ones above are atypical. So instead of ducking and hiding, explain them.

WHOA!

We've got some people here who have strong opinions, lots of experience, and strong personalities. When it comes to approving comments, I'm all for a robust discussion over the issues at hand, but some of the posts are slipping into personal attacks, laden with anger and derision -- and there's a couple of comments I haven't even approved for publication -- and I won't be approving them.

I think everyone here can take a few bangs and hits when we make comments that might not be specific or backed up well enough -- I've been nailed, too, and expect I'll take a few hits again. Sometimes we miss things, make typos, and get stuck looking at a tree but can't see the forest. Or vice versa.

What we're looking for here is comments, opinions, and advice that might be helpful to System i-focused professionals who might be trying to a) understand the industry better, b) understand their strategies and roles in their careers, and c) consider business solutions for their companies.

Ideally a, b, and c are going to be all about the System i . . . and now Power and i-focused solutions. And while IBM doesn't always provide the tools, support, marketing, action, and direction that our industry wants, the company does a great job of offering choice and not pushing customers down dead-end paths -- V6R1 can run applications that were written how many years ago?

Let's try to keep the focus on comments that provide assistance or lend insight in at least a mostly polite way. Concrete examples are good. Personal experiences are good. Musings are good. Ideas are good.

We all get off track every now and then and sometimes even enjoy a little righteous anger. But let's get back on track, say, starting now.

I agree with the WHOA! but some cleanup is in order. ugeerts has haunted the i blogs forever. If you read from the top down it was inevitable as night and day as the pattern is that almost anyone who ends up with a strong differing opinion ends up on his hit list. That is, the inevitable spin on something totally unrelated that just somehow manages to raise a pointed question about .NET and Windows. Seriously, haven't you noticed the pattern that it's almost laughable that these blogs end up as IBM vs Microsoft? If you look at ugeerts pattern from years before, he has not changed much. That's one reason I stay away because instead of an educational and conversational tool it dissipates into two camps and then name calling. Maybe requiring full names and emails to participate and not pseudonyms might help, particularly for anyone busy running an "anti-i" campaign.

If you'll note at the end of the blog it's just the two MS pundits chatting. In fact, reading the comments, the phrase Concern Troll is right on. That's what happens, they take over the blogs by waiting for an opening, acting sincere and then bam we're right back into the Windows vs i -- that's what a concern troll is. If I was wrong then they would have addressed anyone's comments to them such as Nathan Andelin, RD's and Iyeeeee's. They have not.

[*Note From Chris: The trouble with putting the brakes on the name calling is always who gets the last laugh or the last jab. I'm letting this comment through for one key reason: Ted makes a point about degrading into IBM vs. Microsoft, and while I'm fine with IBM vs. Microsoft, we've got a lack of useful advice or insight as to what a System i professional could be doing or should be considering. Sure, it's a gray area, but my point is this: MS vs. IBM should focus on practical advice for now, moving forward, into the future, on Power and i. And seriously, our entire site is for System i enthusiasts, so if you're not a System i enthusiast, why bother crashing the party?]

Mr Iyeee,
You can see what I mean by "scaling" in the wikipedia entry on "scalability", specifically, their explanation of "horizontal scaling" per distributed software applications. MQ and ICF may be used occasionally but not in a distributed transaction... as a matter of fact, often transactions (commitment control) are completely omitted from AS400 apps. On occasion, I have included AS400s in a web service but I don't do so willingly. With AS400s (i5eries,etc) you get too much legacy baggage - see the numerous complaints on this site about the adoption rate of ILE including Scott Klement's article "ILE Basics: Its All About the Call" and my previous complaints about the non-relational design of AS400 databases. (In other words, RPG and DB2/UDB are good but poorly used.)

FYI, RedHat owns JBoss, not Microsoft. It confuses me when I'm accused of liking Microsoft after I've advocated JBoss. I do recognize that, for many, Microsoft is perceived as a competitive solution and I respect that. I also respect the IBM i platform but not its legacy baggage.

OO is no panacea but, *principled* OO development (see the writings of Bob Martin of ObjectMentor, Dave Thomas of Pragmatic Programmer and Martin Fowler) is close. Many lessons from OO could be applied to ILE but code reuse, layered architecture, etc are beyond what those developers want to do.

You don't get anything with an iseries except what you want. If a company is developing a new system, they can develop with J2EE if they want. IBM showcases it as their preferred solution and provides the EGL language for it.

I'm not sure what the discussion is about, but I know lots of great iseries programmers doing everything from /free to Websphere to legacy enhancements to clients and browsers and Web Services and on and on. I've been doing this with multi-billion dollar companies since 1994, worked for the largest AS/400 consulting company and an insurance company for five years before that.

I assure you, none have time to give any attention to a generalized rant against them, to be sure. If there's a point, they're too busy using technology of every type to provide solutions to listen to it.

rd

An earlier response got lost. It was stated RPG and RPG programmers are taking some heat and an example was given of a modularized RPG application. That is the point I've been making - RPG is a good language, it has the capablities of C. That most RPG developers don't utilize what RPG can do is my complaint.

Compare the percent of Java developers who can write a web service component that participates in a distributed service (including commitment control) versus the number of system i programmers who can. What percent of system i programmers write automated tests for their code? How many i5eries/400 databases use commitment control, how many are usable by a Java/Hibernate app? Databases are still being defined in DDS with 6 char field names and no foreign keys aren't they? Integer primary keys as Identity would be nice too. The world has moved on but not for system i development.

Greg,

"Horizontal-scaling" sounds like the new catch-phrase for "server-farm". If you really want one, you can have it! I'll stick with a single integrated server.

Most server farms are a result of software that's so unreliable or so bloated that it destabilizes any single server, so you need two or more to run it.

In other cases the runtime environments scale so poorly that you're forced to use a distributed architecture and rely on an external load balancer, just to manage workloads.

I admit that some data centers (like Google) need that type of architecture, but most companies would save money and have fewer headaches if they were to consolidate their workloads and run them natively under IBM i.

Regarding "legacy baggage", it's not clear what you mean, but most ILE developers who I'm acquainted with are implementing modular components that support code reuse, as time and resources permit.

On a recent Java project, we had as much test code as application code. My PL ensured we had very high test coverage on the code. He said, "Our team won't get paged tonight. Our tests prove our code works in all circumstances". I never saw a production bug in that app. The story on the AS400 side of the shop was very different.

Robert Martin wrote a very interesting article about modular development and top-down design in 1996. It was for C++ Magazine I think. Google "objectmentor DIP" to find the article. The example starting near the bottom of page 3 shows why modularity alone is inadequate to ensure reusability. Later he shows how to decouple the business logic from the IO functionality by use of a controller procedure. (Perhaps you already do this.)

15 yrs ago, I worked at an AS400 shop that had only two production bugs in two years. Their strategy was they fired the programmer and the QA person who put a bug into production. If the QA person caught the bug in QA, the programmer would be fired and the QA person would keep her job. That was the only AS400 shop where I respected management.

The buggy software, twenty thousand line programs and a database I can't use with Java & Hibernate is the baggage I want to avoid.

Wow, a lot of heated discussions. Our manufacturing business is what's important to the company I work for. If IBM keeps enhancing OpsNav, they'll have a decent GUI interface for most sys admin stuff. We're already using Rational and Zend IDE's for development. Using PHP on front end and RPG & CL (which both scream on the back end; moving away from Net.Data) along with SQL and DB2, wirh Apache, Tomcat and Zend's PASE servers, we've a pretty strong web environment that serves up at an extremely respectable (usually sub-second) performance level. We can use these tools to process data both from the shop floor (mostly in MS-SQL) and ERP (DB2 on the i5). Since the OS was written for 128 bits a long time ago, I am anxious to see those 128 bit processors :-). Wishing all the very best to you and yours, Tom

I guess I'm wrong and my detractors are right.

In a story from April 14, headlined "IBM Wants to get Youngsters Hooked on Power" The Register reports, "IBM has claimed that "the entire internet" could run on a single system" running on Power. http://www.channelregister.co.uk/2008/04/14/ibm_power_revitalize/

Of course, they're only going to sell the one machine (we don't need two internets, do we?) so I stick to my contention that a chip upgrade to the AS400 ain't gonna bring in a lot of new customers.

The internet will be going down every Friday night at 6:00PM so we can run Turnover.

Clearly, everybody has a different perception of the "RPG developer". My perception is along the lines of Greg (monolithic code, refusal to use procedures instead of obsolete subroutines etc) while I can understand that Nathan's perception, working in a software house selling RPG based Web apps who sees RPG developers eager to apply the latest state-of-the-art techniques.

The classic RPG developers are in NO way to blame, because they are mostly forced into this situation. I've worked for a Fortune 500 US company (bottler) who are undergoing a worldwide 100% migration to SAP. The iseries "legacy" team were all contractors with the sole aim to keep the iseries apps alive until SAP took over.

What I would like to see is a statistical study over a population of several thousands rpg programmers who are asked the question what methodology they are following.

The result may be that 90% uses classic techniques (other word for stuck in the past, mostly code maintainers) while 10% using new. Others might debate, not so, i'm seeing more of a 50/50 proportion..

So, can we have some really valuable insight reports on the state of the "i" industry in terms of workforce, demography, development and deployment strategies used etc... Would be interesting read, but probably never be publicized.

About a business converting to SAP . . . in no way would I consider consultants whose jobs are designed to keep "legacy" iSeries apps running a bad thing. Just to clarify. Organizations that have nothing to do with anything i move to SAP all the time, and anyone moving to SAP is going to want to keep previous apps running, no matter where or how they were built. Moving to SAP is a huge business-wide decision where the OS and hardware are secondary factors -- an SAP decision is not usually about fixing what a company perceives it can't do . . . but is instead a move toward what it perceives it can gain across its entire business by moving to SAP.

My point here is that moving to an ERP solution is much different to the i ecosystem than ditching the platform and RPG in favor of some other development methodology. Vastly different issues at play. So I would encourage System i enthusiasts to ignore loaded phrases like "legacy" and "keep the iSeries apps alive".

I believe that ugeerts' point here -- which I missed on the first read and which is why I'm posting this -- is this: the RPG developers who worked for the bottler moving to SAP are in a situation where they had no room to create more modern, powerful, and useful applications.

It's not because the i sucks or they couldn't do great things; it's that previous apps were being replaced by SAP and the consequence to our world is that these RPG experts were not able to expand their skills. A situational example.

Which brings up the survey ugeerts mentions . . . a bit of a can of worms. If you put two developers side-by-side in the same organization, you could easily find two extraordinarily different personalities, skills, inclinations, and career objectives. I hear this all the time -- one developer has more aptitude while another has little interest in change. In which case they may be working on different projects. And this is in the same company. How do you measure that?

Back to the topic of this blog.
It seems that with the new POWER hardware, the "IBM i" O/S will run on top of VIOS (Virtual I/O system - I may have the name spelled wrong). Only VIOS will access the hardware directly.
Lets elaborate on that.
Why would we need VIOS? The AS/400 productline has always been able to morph itself onto whichever hardware it runs on because of its inherent MI/TIMI layered architecture. By modifying MI, the OS in fact is shielded from the hardware which has allowed IBM fi. to move from 48bit Cisc to 64bit Risc without altering one line of application code.

Essentially, this is also what VIOS does. VIOS presents the OS with a set of virtual hardware pheriperhals that mimic real former AS/400 communication, memory, and disk hardware features.
My take is that IBM wants to stabilize the hardware image that "i" sees (and we all now what stabilize means). If not, they would have done it the proper way and modified the MI. That would have allowed "i" to access the hardware natively and run at full speed.

PS. I still have to get used to the name "i". I am driving a Ford i? Hmmm. Ibm i? I run on i? Maybe the stranger the better it will be remembered, we will see.

Chris,

".. the RPG developers who worked for the bottler moving to SAP are in a situation where they had no room to create more modern, powerful, and useful applications"

Indeed, that's exactly what I ment.

"Which brings up the survey ugeerts mentions..."

I ment the survey to be cross company of course. It would be very rare/virtually impossible to find an "i" shop where one developer is using state-of-the-art techniques while the other is printing 20000 line programs on a matrix printer and works in the old fashion way. Every "i" shop has something called "standards" which lays out a framework, methodlogy, covering all aspects from the very early development till the final deployment.

FYI, I'm working as a consultant mostly on the "i" systems in the EU mainland. In that capacity, what is my perspective on the "i" world for the past 5 years? All contracts are either about 1) migrating "i" apps to other platforms or 2) maintaining "i" apps until a final decision is made (the added value is technical and functional support while internal staff can focus on other tasks). Migration work is per definition finite, so I see about 2 to 3 shops per year.

"...the i sucks"

was never the point. I too saw buggy Windows and Java apps put into production with all the nasty implications for the apps developer.
My point is, all technology platforms, hardware or software, these days are good. They work as advertised (with some reservaton for the latest fancy MS or Open Source frameworks). That means the quality of the produced work relates to only one factor: the quality and talent of the developer in making the right decisions and testing the produced software.

Chris, hopes this clarifies things a bit.

You don't measure it. There are all kinds of programming and people who prefer different niches.

I was interviewing for a job three years ago (and had been interviewing unsuccessfully for almost a year) when I had the opportunity to interview for a job at a well known financial services company which in iseries circles is known for their Java development. This was to be for an RPG job.

Imagine to my surprise when in the interview I was handed a source code listing for a program using the cycle and level indicator logic and asked to tell him what it did. I was further surprised to see it had been converted to RPG IV as is.

I was honest with him. I just don't have the experience with level indicator logic. Certainly while I could work out what will do what when, I wasn't going to try to pass myself off as familiar with something I'd never worked with before.

But yet they wanted someone with that experience. And there are people that fit their bill. I was just one of several people interviewing.

I would throw two general principles up here. One is that I understand the managers who want to keep their code base somewhat homogeneous.

When I go into a system to make modifications I write in the same style as the existing code. Many programmers consider each line they write to be converting to the latest syntax one line at a time. I don't agree with the resulting jumble effect it has.

But some managers will also even have new programs written in existing style for conformity, and some programmers will like the uniformity of the system and others will chafe at not being able to develop skills in new syntax.

I think the happy medium is to take advantage of those who want to take full advantage of subprocedure modularity and /free syntax to write new programs in a consistent style and let those who prefer the existing code base to work on it.

Having said that, the second principle I believe in is that any of us can work on an existing code base, whatever syntax it's in, and add to it in that syntax. It's a matter of the patterns and methodology of say isolating logic to subprocedures and using local variables instead of all global that some would prefer to add to any program, while others would prefer to use working with a base of code written that way.

So there's nothing to measure, just decisions on running a business and cost effectiveness of running it with a maintainable code base. Some argue that every snippet inserted in /free or subprocedure added just made it more maintainable by definition, and others prefer a consistent code base that is plug and play across the spectrum of apps.

Whatever the decision, it's the effectiveness of supporting the business that is the criteria to keep your eye on. There should be room for those comfortable with the existing code base and those who want to write apps in new syntax that prove whether to be more solid in execution and more maintainable when it comes time to enhance them.

rd

VIOS is a technology to allow easier management of hardware resources across multiple partitions
in a virtualized environment. The technology was brought over from the System p and is directly aimed at getting SMBS to consolidate workloads onto Power instead of running multiple separate servers. It supports AIX, Linux and i partitions.

One of the benefits is to allow seamless integration of hardware resources across multiple partitions, which spells inter-operability. This is excellent news and means that we can expect more capabilities from the Power line.

It's all explained in more detail here:

http://www-03.ibm.com/systems/power/software/virtualization/

VIOS will allow improved access and speeds among hardware resources for all partitions be they i, AIX or LINUX on Power.

That brings up another interesting topic, I can now proudly state "i run on Power". Did I just copyright that?!

On the subject of morphing system i onto a Power 6 Blade or otherwise server, I understand Ibm is excited about this. To them it is the next logical step to take. After all, on the hardware front, we see an evolution towards a PC format architecture. Even when you look at the description of the latest Ibm Mainframe Z10, you see it is organized in what they call "books". A book is essentially a PC like motherboard equiped with multiple core Power CPUs of which some cores act as a i/o controller, surrounded with a set of standard DIMM memory.
The mainstreaming of hardware for i, p and z series into equal sized components with equal electrical characteristics is an important factor in the reduction of costs and in the increase of volume output.

All good for IBM, but to the customer, it doesn't matter a single bit.

What matters is the price/performance ratio. We'll have to see how far IBM is willing to reduce prices for the same power output in CPW or MIPS. And perticulary for i series, which already had an overpriced structure compared to unix midrange systems, how low will they go without aggravating the existing customers.