Readers Riff on the 2008 System i Wish List
Published: January 21, 2008
by Timothy Prickett Morgan
In last week's issue of The Four Hundred, I did a stream-of-consciousness rendition of a wish list I had for the System i product line for 2008. Making my own list is fun--well, if you like banging your head against a wall, as most of us in the i5/OS and OS/400 ecosystem apparently do--but the real reason I do it is to try to solicit ideas from readers of this newsletter and publish them so one set of readers of this newsletter--the executives at IBM--can see what we are all thinking.
A number of you responded to my request for input, and I thank you, as always, for that. As for the rest of you who didn't provide input, Iíll give you a second chance to throw some ideas my way by hitting the Contact button above and sending me a message with your thoughts. You may not believe it, but this exercise actually has an effect on IBM--although I will admit that it seems to take five to 10 years to accomplish something that should take maybe six months. Big Blue did $21 billion in stock buybacks last year. If IBM made it $20 billion and committed $1 billion in hard, cold, cash to really revamping the System i line and doing real marketing, it would not have had an appreciable effect on IBM's short-term financials, but it sure would radically improve the System i line and perhaps give IBM a $4 billion business with maybe 60 percent margins.
The point is, customers need to apply pressure to IBM and to do so with their checkbooks.
Without further yammering on my part, here's a sample of your feedback and ideas on The Official 2008 TPM System i Wish List, starting with some criticism from a faithful reader and commentator.
While I will always be a loyal reader of The Four Hundred, I wish TPM would ask more challenging questions of the IBM execs he interviews. Marc Dupaquier said he was happy with the sales of the System i in 4Q 2007. Was IBM happy with sales of the System in the years since 2000? If not, why did it underperform?
Similarly, it would have been fascinating to read in the Jim Herring interview some numbers on IBM's level of investment in core i5/OS versus AIX. Then broaden the comparison to industry-wide investment in core Linux and Windows. How many programmers are working on the next version of Windows versus the next version of i5/OS? If the disparity is as large as I assume it is, how do IBM execs justify retaining ownership of i5/OS as it becomes less and less feature rich relative to the competition?
These questions are always asked privately, and sometimes I get hints about things. IBM execs will never answer them publicly, except when they are trying to make a point that they have been allowed to make by the higher-ups and the IBM PR machine.
This is what it is like to "interview" IBM. And just about every other IT vendor out there, too. Executives say things on "deep background" that they will never cop to publicly. In many cases, they would lose their jobs if they did. I have a lot of off the record conversations with IT execs who want to know what I think as much as I want to know what they plan to do and not do. This is the horse trading that goes on.
That said, in this case I asked Dupaquier about how the System i was doing and he promised to get back to me as soon as the fourth quarter was over. The irony is that he was moved out of the job running the Business Systems division before IBM announced its results last week. I can assure you Dupaquier was not aware of his impending job change when I talked to him only weeks before, and he was quite geared up to transform the product line. I did not get into relative investments by IBM in different platforms with Dupaquier. And you are right. No matter where the interview is going and what questions I think IBM will and will not answer, it is indeed my job to do the asking. Thanks for reminding me. But I am also mindful that there is a delicate balance between access to IT vendors and antagonizing them that I have to tread carefully, too.
Yeah, this is annoying, and they don't tell you about this in J-School, where they are pumping newbie journalists up with the ideals of objectivity. You have to learn this stuff on the street, like I did, and I believe that the approach I have with IBM gives me the most leverage to affect change on behalf of the customers I am advocating for.
I enjoyed reading this year's wish list. I hope that IBM takes it seriously and we see the IBM midrange product (with its name du jour) make large strides in increasing market share and popularity.
One item that I disagree with is Number 5, and I quote: "5. Make OS/400, DB2/400, and the RPG and COBOL compilers open source. I said it before, and I will say it again. If these products are as good as everyone says that are, and are merely limited from exposure in the market, then put some or all of the software out as an open source platform. If the software is tied as tightly as I suspect to Power-based iron, then very few people will have the skills to port it to other architectures, even if they have the desire."
The reason I disagree is summarized in the last sentence I quoted above. Open software works when there is a communal effort to improve and add to the software. The IBM midrange community, outside of a few notable exceptions, has shown that it can't or won't expend the necessary effort to participate in open source projects. This is true for utility and application software, where there are many qualified individuals available to participate. When it comes to specialty items like compilers, the number of available qualified minds shrinks to almost nil. The same lack of skill that prevents people from porting it to other platforms will prevent its future development and expansion in the first place.
IBM would gain nothing, but potentially lose a lot of intellectual investment made over many years.
I would like to add an item to the list. Whether IBM likes it or not, a fully functional GUI interface is a mandatory component of software today. The programming interface must be an extension of the current tools IN COMMON USE in our programming community. It can't be an add-on like screen scraping (which opens the door to code being out of sync and tends to be rather ugly) and it can't be another heavy programming layer (a la WebSphere), rather something bundled as a basic part of the OS (like DDS is bundled). That will encourage development and marketing of software that can compete in the market, which in turn sell more hardware peripherals and copies of the OS.
Even amongst the companies that are "midrange faithful," I'm constantly having to justify staying with the "older" technology versus adding a PC server for new projects. IBM, it's a very frustrating position you have put us in, and we're tired of it!
Anyway, that's my opinion--thanks for listening!
Hi there, Mark,
Yeah, I forgot to mention the whole GUI thing in my list. I have said it so many times, as have other Gurus at IT Jungle--particularly Brian Kelly, heaven love him--that I have gone hoarse from talking. Or maybe it is horse. . . . You know what I mean.
At this point in the computer industry, integrated GUI interfaces make as much sense as an integrated relational database management system did back in 1980. The whole brilliant point was to mask complexity. Maybe IBM doesn't know how to do that any more, or has learned there is more money to be made from selling software and support to cope with unnecessary complexity? Call me cynical.
That said, I think there are a mix of technologies available out there today that will allow someone else to take the lessons IBM has learned and create a great product. IBM better hope that no one does this. It seems unlikely Microsoft will, and with its buy of MySQL last week for $1 billion, Sun Microsystems could take a second go at a boom creating an integrated platform like the one we should already have in the System i line.
Well there is the System i want, how about the System U need and that "I" will stand behind. There is MYSQL, how about MYEYEOS or MYI5OS.
Also, the 70 days--I think that is what it still is--that you go through testing Licensed Program Products; maybe we should bump that up to like 120 days? Three months sounds good to me.
The IBM channel that markets the product is so loaded with bureaucracy and BS that it eliminates the small partners that can actually survive selling to the SMB customers. There are so many opportunities in the SMB space that are simply given over to the Wintel bunch that it is sickening.
In order to be a partner, IBM requires so much overhead that the partner will go out of business if there is any slack in sales. I was a partner for around three years and was successful with SMB sales, but could not pass an audit that required me to show the W-2s for 10 employees. We had done over $3 million in SMB customer accounts and there were only three of us--all very productive and skilled.
That mentality is what keeps IBM from success in SMB among others for sure.
I would just like to comment on a few items on your list.
1. IBM...just give us a name we can use...please. Google "i" and you will see my point.
4. "Anything perceived as real, is real in its consequences." (That's "perception is reality" to you and me.) If it looks high-tech than it must be.
5. Best of luck on this one.
6. This might actually happen.
7. Great idea, and most of it could probably be done online, which would reduce the cost.
8. This one is my main System i dream. Make sure it includes affordable software upgrades.
As a RPG programmer and RPG lover from Korea, I wish IBM would make RPG a true Web Programming Language as well. RPG has 24x80 and 27x132 screens. Why not have RPG have resolutions of Web page size, such as 800x600, 1024x768? And what about Web controls of a graphical user interface!
The AS/400 is the best server, I think. If we can get the GUI with RPG for the Web, then IBM can have many new customers as well and make an AS/400 Renaissance again! Also, IBM can make RPG as open source for this!
In Korea, we lost many AS/400 users and more RPG and COBOL programmers. It's so sad!
Hope it is not too harsh to say, but I wish IBM would sell i5/OS to a company (Oracle?, Zend?) that has the necessary get up and go to make the improvement to the OS that it needs to compete.
i5/OS needs long object name support, an increased pointer size from 16 to 64 bytes, ILE service programs need to support reflection like .NET assemblies do.
SQL procedures are by far my preferred way of programming the System i. Too bad the integration of SQL procedures into ILE and i5/OS remains as clunky as ever. They are a chore to debug, are difficult to call from CL code, and provide minimal diagnostics when your code bombs.
The biggest wish of all is that IBM finally understand the devastating impact of making CPW so expensive. SQL views can be used to great effect to resurface legacy AS400 databases. A common practice is to embed arrays of info (for example, 12 months of sales) in a single database record. An SQL CASE statement can be used in an SQL view to crack that single row containing 12 months of data into 12 SQL friendly rows that can be joined, selected, and ordered just like a properly built database table.
Another problem with legacy databases is tables can't be joined because of duplicate rows and keys that don't completely match. Using GROUP BY and user defined functions, SQL views can again transform an unusable database, as far as SQL is concerned, into a whatever normalized form first-class citizen. All the programmer needs to make this work is the horsepower of a modern computer system. As in a quad-core, rack-mounted, Power6 with 100,000 CPW. That would make for fun programming!
Just a note: Power6 is a dual-core chip. We don't expect quad-core Power chips from IBM until around 2010 or 2011 when Power7 might get out of the door.
The Official 2008 TPM System i Wish List
The CDW IT Holiday Wish List, Budget Forecasts
Q&A with Jim Herring: The View from the Top
The System iWant, 2007 Edition
The eServer i6 and i7 Wish List
iSeries and OS/400 Wish List
The iDeal iSeries, Part 4
I Want, I Want, I Want: An AS/400 Wish List
Post this story to del.icio.us
Post this story to Digg
Post this story to Slashdot