Talk:IBM System i

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing (Rated B-class, Low-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 Low  This article has been rated as Low-importance on the project's importance scale.
WikiProject Technology (Rated C-class)
WikiProject icon This article is within the scope of WikiProject Technology, a collaborative effort to improve the coverage of technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
Checklist icon

Precursor to Smalltalk[edit]

The article suggests that the AS400's virtual intruction set might be a precursor to a programming language whose standard version, Smalltalk-80, was released EIGHT YEARS before the AS400 line was released. Smalltalk itself had been around for some considerable time bbefore that. This is something of a reach, to say the least. Anyone care to justify this claim, before I remove it?

Liam Proven 15:54, 20 Apr 2005 (UTC)

Agreed, remove it. The entire comparison of the S/38 / AS/400 / iSeries MI to platforms like Java and .NET (which are in themselves fairly different) is pretty weak; I'll try to think of some better phrasing for that section. -- uberpenguin 16:16, 2005 Apr 20 (UTC)
It is done. I've reworded the entire paragraph to remove the claim of primacy but retain the association of ideas. Liam Proven 13:09, 23 Apr 2005 (UTC)

"Said to be" means that it is up for discussion. It is not. I have seen OLD programs run over time. This is not a sales pitch. If you are a competent programmer and have source you WILL recompile programs to make use of the latest compiler improvements. Removing Mr Soltis as the source of the statement of the programmers working on OS/400 and licensed programs is silly as without it makes no sense anymore. What is the point of explaining about verbs etc without mentioning its benefit? The point that all software is included and integrated, is what differentiates it from all other systems. So it is relevant. The kernel and the hardware related code are essentially different; this allows IBM competitors to create an OS on the bare machine. This is a result of lawsuits relating to mainframe OS clones.

iSeries redirect[edit]

I changed iSeries to redirect here once more. Upon reading through the article carefully (and posting comments on that article's talk page), I found no information that wasn't covered better and more thoroughly in the AS/400 article. The iSeries article was basically just a list of features and a short section on criticisms of the system, which is a good idea but was poorly executed on that page. "iSeries" is nothing more than a rebranding of the AS/400 line. It doesn't suggest any major architectural changes (that is, any changes more significant than the AS/400 has seen in the past; the TIMI is what really makes the AS/400 / iSeries unique), so I can't see any reason why iSeries shouldn't redirect here. Please post your comments if you feel differently. -- uberpenguin 21:11, 2005 Apr 18 (UTC

Just my opinion[edit]

As I am new to using wikipedia, but not the AS/400 or iSeries, I may have not laid out my ideas in the best way. I have tried to be factual and ub-biased, except the misunderstood section and have tried to correct any errors in what I originally thought, but it bothered me that there wasn't an iSeries entry. I am aware that iSeries is just a rename of AS/400 and most people who have used it still refer to it as AS/400. If you look at the entries for Windows, there are entries for win95, xp, 2000, etc. the argument could be there is no need for more than a single Microsoft entry or Linux entry. I mean the different entries for those operating systems are only different versions as well. In fact it hasn't been OS/400 since V4R5, 5 years ago, i5/OS is now at V5R3.

I don't believe this argument holds up very well. Win9x and WinNT are two VERY distinct OSes; one based on an aged MS-DOS code base, and the other based on Dave Cutler's work in designing the NT kernel and API. Windows 2000 and XP are arguably very similar (their kernel revision numbers are 5.0 and 5.1, respectively), but have different articles because their target is very different (desktop rather than server) and a lot of things were changed to reflect that new target. If you wanted to apply this example to why there should be an iSeries article, you might as well create an article for every revision of OS/400... -- uberpenguin 05:45, 2005 Apr 22 (UTC)

Actaully I started creating the iSeries article, based off the layout of the Windows 2000 article. My feeling was that if someone had a interest or heard about this system somehow, it wouldn't say much if all it did was redirect it to the AS/400 article. Their impression would be "Oh that old system". From a PR stand point the rename, didn't change the system, but there has been substantial additions to its capabilities in the last 5 years. It is why I tried to add a lot of links, so someone could find out more, if they wanted to, not just the TIMI, which is important, but not the only thing which makes iSeries different than other operating systems.

I appreciate your good intentions, but as I stated before, the format left something to be desired and wasn't a very good representation of Wikipedia style and convention. In its last state, the iSeries article really resembled a product datasheet or bullet-point list of features, not a well-fleshed out discussion of the relevance of those features, how they compare with other OSes, etc. Also, this isn't the appropriate place to be endorsing a 'PR stand point.' People who have a bias about 'that old system' aren't necessarily going to be instantly swayed by a fancy new title (I do realize the irony in this, as it's exactly what IBM has done). However, there is a commonly thrown-around and often agonized over editing concept here called NPOV, which basically means that articles should try to be accurate and factual, not supporting one opinion or another. In other words, the iSeries article didn't cover much material that this article does not, but this one follows writing and style conventions better. -- uberpenguin 05:45, 2005 Apr 22 (UTC)

My personal opinion after the last 10 years of working with AS/400/iSeries, as well as the alternatives, is there is no better OS or computer system. The point of the misunderstood system section, if not well done, was to highlight what is positive about this system, which has often been ignored, even by IBM, but usually by people who have never used it, or assume it is only a text only system and GUI's just automatically make alternatives better. My feeling is the iSeries article did have value. I don't consider myself some iSeries expert, just enthusiastic and I would have appreciated any help to make it better, including adding different pictures, but I think over the last couple of weeks I saw one change made by someone else. Whether the article was changed from what I did, I really don't care, but not having an iSeries entry does a disservice to this system and to people who may be interested in finding out more about it. nstelmack

As another long-time user of IBM machines (System/34 on), I do understand what the purpose of the 'Misunderstood system' section was, and do think it's a good idea. However, whenever something like that is included in an article, it is good practice to also include a section of criticisms to counterpoint the praise and try to keep a non-biased tone in the article. I'm not at all against bringing the section back, but I think more should be added to it in the way of valid criticism (most of the criticisms you brought up were trivial, though very common).
Right now, I still don't really see much justification in having an iSeries article simply because there have not been any enormous changes to the system since its rebranding. Of course there have been noteworthy updates to the hardware architecture (which is covered in the IBM POWER article since the AS/400 has been on a more or less microcomputer architecture for a while) and OS/400 itself, but they aren't major enough (in my mind) to merit another article when they could live just as happily here. Thanks for your remarks, and I hope you'll stick around Wikipedia and continue making edits! I'll try to get some other folks to weigh in as well. -- uberpenguin 05:45, 2005 Apr 22 (UTC)


I think what would be most constructive for Wiki, would be a section or related article that puts in perspective how the AS/400 is radically different from other kinds of computer platforms.

I think there is a split in viewpoint here that is conceptually similar to that of computer security and computer insecurity articles, where one talks about how to achieve good security, while the other talks about in the real world where most people must function with computer systems whose security is absolute crud, and what you can then do about it.

Likewise, with AS/400 iSeries whatever, there are applications and industries for which it is the absolute best, and other areas where it is not a good solution. So a few paragraphs explaining that, followed by the notion that often times a computer solution gets used, that is not the optimal, if the people involved had only known about the larger trade-offs.

As it currently stands, the article talks about what the AS/400 is, not how it fits into a larger computing world. AlMac|(talk) 19:25, 14 August 2005 (UTC)

In the passage: It also supports Unix-like file directories, industry-leading support for Java, client-server technologies and a native Apache web server, for "modern" and GUI-style applications. can we remove the apostrophe please? Also I think the wording "industry-leading" is not suitable for an encyclopedia. It's nothing but a personal opinion. Tnx

first 64 bit?[edit]

The Alpha was introduced in 1992. How is AS/400 "in 1995 was first to employ a 64-bit processor and operating system."?

And the MIPS R4000 was introduced in 1991 (I think; it could have been 1990)... Hmm... I've never noticed this in the article before; it indeed seems spurious. Let me do some research and I'll change it. -- uberpenguin 02:02, 2005 May 28 (UTC)
The alpha was NOT both 64-bit processor, 64-bit OS, and 64-bit software applications, all at the same time. AlMac|(talk) 19:14, 14 August 2005 (UTC)

I'm going to need some clarification on this discrepancy. There's a lot of water between 1992 and 1995. If the Alpha didn't have 64 bit across the board out of the gate, it wasn't long after. It certainly didn't take them 3 years. Can someone firm up the details on this? (posted by )

I was around when this was going on. Now in some areas of computing, stuff comes out pretty fast, while in other areas, deployment does in fact take several years between releases of different things & there's various reasons for this.

  • Obviously various companies thought various 64 bit things were important to get developed and to marketplace, so we end up with who's the first to have a 64 bit operating system that runs on a particular kind of computer, but the hardware has to be ready for it, ether first, or at same time release.
    • Likewise application software that runs 64 bit programs on 64 bit operating system on 64 bit hardware, needs to either have the 64 bit OS in place in advance, or at same time release, and that's what IBM did with their 64 bit. They developed the hardware, but did not release it. They developed the OS to run on that hardware, but did not release it. They contracted with thousands of software houses to develop upgrades to use the new 64 bit, which IBM called RISC to replace the previous which was called CISC, then they had a big marketing splash to tell the world of people interested in IBM computer stuff that here was this 64 bit hardware, with 64 bit OS, with thousands of applications to run on it.
    • Other companies came out with the parts piecemeal, because their 3rd party service companies are less integrated than the IBM business partner market.
  • Then there is the concept of maximizing revenue from existing technology before releasing something that will wipe out your own market. I understand that IBM has perfected 128 bit ... the hardware, the OS needs, what can be done with the software, but they deliberately not releasing it until they soak their market for what it will deliver on 64 bit.

AlMac|(talk) 15:10, 8 October 2005 (UTC)

The Alpha most certainly did have 64 bits "across the board" at introduction. If you wanted to compile programs that weren't "64-bit clean", you had to give special compiler options that resulted in less efficient code but used 32-bit pointers. But the Alpha missed being the first 64-bit computer by at least 30 years. The IBM 7030 Data Processing System (also known as Stretch) was a full 64-bit system, and was introduced in 1961. --Brouhaha 06:01, 27 October 2005 (UTC)


I disagree with User:Uberpenguin on server vs. mini-computer since there has been an evolution of general usage of terminology within IBM and within the larger computing world.

Once upon a time, the precursors to AS/400 were in fact mini-computers, then IBM provided the terminology of midrange computer which meant more powerful than a mini, but less powerful than a mainframe, but we found this terminology darn few places outside of the IBM world.

At the time of the change in naming from AS/400 to iSeries, IBM used eServer for the naming of the 4 categories of computers, of which iSeries was one category ... the others being new name for mainframe S/390, scientific RS/6000, and what is now the xSeries.

I think best terminology here is to emphasize that the AS/400 is like multiple computers in one box ... you can choose just one, or on one platform have many different OS & server technologies, whether using different LPAR or being on an older model of AS/400 which supported M36 objects.

I think that to label this as a mini computer is to call it something a lot smaller and less powerful than it really is, while server is in fact the terminology of the computer world today, where the AS/400 can replace a server farm, because it is like a super-server, except that terminology is not in widespread usage. AlMac|(talk) 19:14, 14 August 2005 (UTC)


this screenshot, showing debian running on some of the 1000 supported architectures, is rather irrelevant to as/400. I'd suggest putting some pic of the outside of a typical as/400 there. I mean like hardware, not software. sorry, I don't have one.

Fit for ordinary consumers?[edit]

At the science ref desk someone spoke of the superiority of AS/400 (when it comes to safety). So I looked it up to see if it is a viable option for an ordinary consumer. But I see no information on that. As far as prices go, I only see the 9000 vs 2 million $ (assuming USD) comparison. Does the cheapest one cost 9000 USD? DirkvdM 10:32, 5 February 2006 (UTC)

Speaking as a person with over a decade experience working on a mid-sized AS/400 (cost under $50,000.00), in my opinion, the AS/400 is not appropriate for the kinds of applications desired by an ordinary consumer, nor is it appropriate for science applications. The AS/400, rebranded as iSeries for integrated business, is scaled to all sizes of businesses from the mom and pop up to the Fortune 500, or whatever the multi-national equivalent is in other nations, and alwo can be used by government agencies. For science applications, IBM's answer is the RS/6000, rebranded as I forget .... basically All of IBM's computers were rebranded as eServers with a line for each kind, in which the xSeries is what is aimed at the consumer market. For example, you can buy an ISP in a box, which does all the stuff a consumer might want from what people normally associate with being an ISP ASP OSP what have you. But that IBM line is quite different from the AS/400. User:AlMac|(talk) 09:16, 6 February 2006 (UTC)
Thanks for the info. You might add this to the article. I won't because it'd be second hand info coming from me. DirkvdM 09:43, 6 February 2006 (UTC)
Quite. Even on an architectural level, AS/400 hardware has always been designed for bandwidth moving (transaction processing), not number crunching. I have seen 400s used successfully in several small business account and transaction processing related tasks, but never for personal or "ordinary consumer" usage. -- uberpenguin 16:34, 6 February 2006 (UTC)
I added a bit of text on this at the end of the first paragraph. We will see how it holds up. User:AlMac|(talk) 04:55, 9 February 2006 (UTC)

Major rewrite[edit]

This article badly needs a major rewrite. At the moment, it's a "garbage bag" of disconnect factlets blended together like a pizza. How about some structure, people? Kelly Martin (talk) 05:38, 8 March 2006 (UTC)

A brief look at the history of activity on this article ought to let you know how likely such a rewrite is to occur. -- uberpenguin 06:00, 8 March 2006 (UTC)
I dunno, I might get ambitious. I'd also like to get a little more in about the AS/400's relationship to System/38, and the fact that OS/400 V3R6 and later are 100 percent pure C++ down to bare metal in the emulator layer. (I vass dere, I helped.) --- Charlie (Colorado) 01:22, 13 March 2006 (UTC)
I agree, and nothing has happened to improve it in 18 months.
1. Terminology needs to be consistent. The same platform is referred to by many different names
2. There are a number of superstitious statements made on the page. For example "everything is an object"
3. A huge amount of space is dedicated to what appears to be a parts list table.
--Scott1328 14:16, 17 September 2007 (UTC)

East or West, COBOL is best![edit]

>It is not a suitable platform for scientific number crunching or consumer applications.

That depends on where you are located on the map. In the former communist block a lot of big AS400 were purchased between 1989-1994 (until the end of COCOM restrictions), because you could buy 4x as big mips-power in AS400 box compared to unix boxes, as americans thought as400 platform is no good for math (industrial and weapons design obviously). But it was used by many companies for number crunching in the east, illegally obtained Fortran/400 compiler was very popular. Some of them liked the platform for reliability and cheap maintenance so much that they got new black-box AS400es for upgrade when the time to upgrade arrived, instead of some Unix box. 18:27, 13 March 2006 (UTC)


Per the request at the Wikiwork Brigade, I've restructured this articles as best I can. I've preserved as much of the original text as possible, reorganising the existing content. Some of this could do with a rewrite, especially to reduce some of the technical talk which can be confusing even to the computer-literate. I did remove the following text, because it seemed like it had been cut from a press release and was unsourced:

The AS/400 is well suited to a broad range of business, non-profit, and government applications, including back office, manufacturing, retail wholesale transportation, database management, being an ISP. IBM claims superior total cost of ownership due to high system availability, which averages 99.98% uptime, and lower costs of administration and maintenance, when compared to Unix, Linux and Microsoft-based implementations.

If someone can find a source for this (particularly IBM's claim of superior uptime) it would be good to add back into the article. --bainer (talk) 01:23, 18 March 2006 (UTC)

is this acurate?[edit]

I know of so many companies using some very old, dos like software - and they call it the IBM AS/400 — Preceding unsigned comment added by (talkcontribs) 13:30, 27 July 2006‎ (UTC)

?POV I'm reading through the comments; it sounds like an ad. My experience with the AS/400 is that it's a clunky dinosaur. What's up. — Preceding unsigned comment added by (talkcontribs) 13:35, 27 July 2006‎ (UTC)

The article is mostly accurate. What's your actual experience on the AS/400? Is it actually programming for it and administering it, or is it just using some app that your company uses? It isn't wise to dismiss a system just because of inexperience or because its different from what you're accustomed to. There is nothing in common between OS/400 and DOS except that they both use text-based user interfaces (even then, OS/400 has a consistant menu and dialogue system which DOSes do not). -- uberpenguin @ 2006-07-27 20:42Z
Clunky Dinosaur? IBM i runs on Watson technology, IBM POWER Systems with POWER7. Comparing this to an AS/400, would be like saying PC's are clunky dinosaurs based on the PC-XT. Or like saying Windows is a dinosaur because 3.1 came out in what year? The object-based paradigm used in IBM i operating system provides integrity and security that other operating systems cannot achieve.
1988. Says it all really - what on earth is this article all about? I got a joke through with AS/400 on it and came here hoping for some enlightenment - who is using this ting at the moment - — Preceding unsigned comment added by (talkcontribs) 17:49, 27 July 2006‎ (UTC)
What is this article about? This line of computers.
Who's using it? These people are using it. Guy Harris 00:55, 28 July 2006 (UTC)
Huh? The Intel 80386, which more or less introduced the first form of what we know as 32-bit x86, came around in 1986. Windows NT was introduced in 1993. The modern incarnations of both those technologies are obviously more capable and advanced than their ancestors. Anyway, Guy answered your questions. -- uberpenguin @ 2006-07-28 02:48Z
There are more of these systems in use than any other family of computers from any other vendor. It is a designed top to bottom for business applications, and it has been frequently sold via Independent Software Vendors with applications for particular industries. This success is due to the software vendors believing in the system as the best platform for their package and for their customers. — Preceding unsigned comment added by WarrenTea (talkcontribs) 21:38, 16 May 2012‎ (UTC)

Ref 4 does not exist (404)[edit]

FYI Reference 4 ( V4R3 Questions and Answers. Reference # 8625668200695613. Retrieved on 2007-04-04. ) of the article does not exist (404). —Preceding unsigned comment added by (talk) 13:19, 7 March 2008 (UTC)

System/36 did not descend from System/38[edit]

The Summary section states that the System/38 "...underwent several rebrandings ("System/36" in 1983; "AS/400" in 1988...". This isn't correct. The System/36 was a successor to the System/34, which was of a completely different architecture. IBM introduced the System/36 because the System/38 had proved to be too revolutionary a step for many System/34 users.

Yampy (talk) 09:08, 4 September 2009 (UTC)

I concur with the above: The Sys/38 was not an evolution of the Sys/32, /34, /36 [nor were they an evolution of the Sys/3]. Sys/38 was [according to all the launch presentation I attended] a clean sheet design and should probably be viewed as a prototype AS/400. (talk) 07:28, 8 September 2009 (UTC)

Too few names![edit]

The poor old AS/400 had as least two other names 'System i5' and 'eServer i5'. Source: I'd change the entry but I don't know the dates - anyone else? -- John Muir (talk) 13:01, 30 July 2010 (UTC)

Platform with a weird, but admirable bug[edit]

During the middle of the 48-bit AS/400 CISC era, there was one curious period when several major customers reported their largest 80 or 9x models having halted suddenly and unexpectedly.

One must understand that there is no garbage collection for addressing in the CISC variant of the AS/400 system. The object master table is only cleared up during a full system restart, it doesn't matter if you erase objects during run-time.

It turned out those customers suffering from halting AS/400 boxes did never restart their machines in the previous 18 months, not a single time! Those machines were the biggest models in the AS/400 product line and most worked 7x24x365 at near full utilization. Therefore the addressing table was completely exhausted via such intensive work in 18 months' time and the systems halted with a memory full error.

No person at the design table ever considered such a situation could occur, since all AS/400 machines were supposed to restart at least once in every 4-6 month, as IBM regularly supplied system software update and bugfix tapes. IBM then released a memo asking customers to reboot the AS/400 at least once a year, even if they ignored or did not subscribe the software updates.

Such incidents say something about the workhorse nature of the AS/400 platform! (IBM says there is no need for auto-clearup in 64 bit RISC based AS/400 at all, because the larger sized bject table cannot fill up faster than 20k years even with the largest possible cluster of the largest available iSeries servers.)

One must also mention the case of a US hotel, which was renovated during the mid-1990s. When time came to turn off the reception desk computer, everybody was suprised as that turned out to be a dumb remote terminal. Its data cable led to a small hole in a single-layer brick wall and behind it was a big UPS and a small, ancient IBM S/36 server that had been running for at least 7 years unattended, after being such "incarcerated" in a doorless, window-less section of a former machine room, whose majority was mis-appropriated for a new luggage storage shed. Of course the hotel was lucky they never suffered a lasting power outage, as the backup was infinitely overwriting itself. Still, the S/36 performed flawlessly in such an extreme situation. (talk) 17:00, 28 October 2010 (UTC)

AS/400: A punchcard legacy with lipstick[edit]

For all the talk of support for the latest and greatest languages, at its core, the AS/400 is basically a punchcard processor.

It's a laugh looking at a paper "COBOL Coding Form" for the AS/400, and the Source Entry Utility. It's a highly restricted code editing environment, with just 64 characters of editing space per 80-column line. The leading 7 characters and trailing 7 characters of every line is reserved to allow for per-line encoding of a program.... onto IBM's punched cards, if necessary. Also it's fun to note that even as of 1993 books about the AS/400 were still talking about programmers having to use paper COBOL Coding Forms.

The same is true of RPG. The coding forms for RPG are, once again, basically intended to be printed on a stack of cards and fed into a card reader. (Oh, I seem to be out of blank cards. Guess I'll have to use a 64 gig flash drive.)

The punchcard legacy capabilities (hindrances) of the system should be mentioned in this article somewhere. DMahalko (talk) 04:11, 8 July 2011 (UTC)

You apparently know very little about the AS/400. While it does have RPG as it's primary language, and RPG did originally appear as a quick way to generate reports on machines that used punch cards, neither the AS/400 nor any of it's successors have ever used punch cards. Oh, and SEU has long been replaced by a fairly modern GUI IDE built on top of Eclipse. That IDE is currently called Rational Developer for Power (RDp) version 8.5, and has been viable since WDSC version 6.0. In fact before you wrote this comment, IBM had "stabilized" SEU. That is IBM speak for software that you can still get if you really want it, but IBM isn't going to be updating any more as there is a better replacement available. And concerning those 1993 books, they were way out of touch, even when they were written. Of course someone who doesn't know any better could even use your comments here as evidence that some folks were still writing about having to use coding forms for the AS/400 even as late as 2012. Doesn't make it true, just means that the writer is uninformed. SEU, which is itself now obsolete, made the coding form obsolete. — Preceding unsigned comment added by (talk) 18:59, 2 April 2012 (UTC)
Well, its predecessor, the IBM System/38, did support punched cards (the same 96-column cards that the IBM System/3 introduced), but I suspect the AS/400 is about as tied to its punch-card heritage as the System z is, i.e. not very much any more. Guy Harris (talk) 21:03, 2 April 2012 (UTC)

Time for update[edit]

It is no longer the System i. The hardware is officially POWER Systems. The OS is officially IBM i. — Preceding unsigned comment added by WarrenTea (talkcontribs) 04:44, 17 May 2012 (UTC)

And, in fact, the article already says, in the lede:
The IBM System i is IBM's previous generation of midrange computer systems for IBM i users, and was subsequently replaced by the IBM Power Systems in April 2008.
so, to that extent, it's already been updated. The AS/400-and-successors systems' significant characteristics are 1) the operating system (both the OS/400-and-successors level and the "licensed internal code" level) and 2) the hardware features (such as tag bits) that support it. The IBM Power Systems are replacements for both the AS/400-and-successors hardware and the RS/6000-and-successors hardware; whether this page should be "updated" by being merged into the IBM Power Systems page or not is another matter. Guy Harris (talk) 06:02, 17 May 2012 (UTC)

Proposed merger[edit]

I suggest this article be merged with IBM i. That discusses the OS and this the hardware, but I think they might better be considered together as a system. Peter Flass (talk) 12:56, 10 December 2012 (UTC)

IBM i also runs on IBM Power Systems machines, and I don't think IBM are making machines branded "System i" any more; I'm not sure what should be done, given that. Guy Harris (talk) 19:14, 10 December 2012 (UTC)
Perhaps all of the stuff about this article that discusses software (and "licensed internal code", and even "vertical microcode", is software; AS/400 "vertical microcode" was software, but was produced by a group that was under a hardware manager, for legal, not technical, reasons; see Frank Soltis's books) should be merge into the IBM i page. (Or, alternatively, maybe there should be a separate page about the licensed internal code for AS/400, IBM System i, and IBM Power Systems when running IBM i.) Guy Harris (talk) 08:16, 12 May 2013 (UTC)
IBM i now runs on IBM Pure Systems. An OS page should not be merged into the page for only one hardware product line when it applies to two hardware product lines. WarrenTea (talk) 18:10, 7 July 2013 (UTC)

The As/400 and the IBM i are separated by around 10 years of other machines such as iSeries, System I. The last AS/400 was produced in the 20th century. Merging this archaic system in would be a lot like merging an early bagphone with an android or iPhone.jacona fire (talk) 16:37, 16 November 2013 (UTC)

The last system IBM sold under the name "AS/400" was produced in the 20th century. IBM didn't stop making "AS/400s" because they discontinued making machines like that and started making something else, they stopped making "AS/400s" because they decided to name all their servers "eServer [ipxz]Series", and thus the machines they made that used to be called "AS/400s" were all of a sudden called "eServer iSeries", and then all the "eServer [ipxz]Series" were renamed "System [ipxz]", and then "System p" and "System i", which no longer differed that much in hardware characteristics, were renamed to IBM Power Systems. So the IBM Power Systems are just the latest incarnations of the machines that started out as the IBM RS/6000, as well as the latest incarnations of the machines that started out as the AS/400 or, at least, the PowerPC-based AS/400.
And, in fact, the AS/400 page just redirects, as is entirely appropriate, to IBM System i, so that "archaic system" is already merged in with the page for the generation of those machines prior to the introduction of the IBM Power Systems.
The reason to oppose a merge is that the OS has run on machines with names ranging from "AS/400" (back when the OS was called "OS/400") through the IBM Power Systems (which has no lower-case "i" in the name), and the IBM Power Systems are, at the hardware level, as much the descendants of the RS/6000/.../System p as of the AS/400/.../System i, and thus should not be merged with any of the IBM RS/6000, IBM System p, or IBM System i pages, so there are, and should continue to be, two separate pages for hardware running IBM i, so there's no single hardware page to merge the IBM i page with. Guy Harris (talk) 19:34, 16 November 2013 (UTC)

I oppose a merger on the grounds that AS/400 is a past, ie historical, system, much like the System/360 is. IBM spent a great deal of effort to make the current operating system backward compatible all the way back 25 years and more. This backward compatibility has lead many people to simply assume that, since the new machine does every thing the old one did, the only thing that changed was the name badge and paint job. Nothing could be further from the truth: the current hardware is Power 8 RISC, AS/400 had a custom CISC processor. The new operating system, IBM i, can host virtual machines of IBM i: OS/400 can't do that. The new DB2 supports a huge subset of DB2 for z 10 functions: OS/400 had very primitive SQL abilities. I can and do invoke Java methods from RPG programs on IBM i 7.2: OS/400 had no Java. The list goes on and on. The fact is that the Power hardware line isn't simply rebadged AS/400 hardware, and IBM i isn't simply re-branded OS/400. The current hardware and software is a distinctly different superset of the older generations. If anything, we should be splitting out 'current' information (like the current Power processor) from the historical systems - AS/400 doesn't have a Power 8 processor option. The AS/400, iSeries, and System i are machines from the past. Frozen in time. Historical. It makes perfect sense to keep these machines separate from the current, living Power Systems running IBM i. Buck (talk) 12:09, 6 October 2015 (UTC)

"AS/400 had a custom CISC processor". No. Older AS/400s had a custom CISC processor (IMPI). Later AS/400s had custom RISC (PowerPC plus tag bits and, at least in some processors, decimal assist) processors.
FWIW, no AS/400 had a PowerPC chip. The first Power 4 was in 02. AS/400 was gone in 00. Unless (see below) we choose to disregard IBM's official branding as a simple case of 'slap on a new name badge'.
"The software named XXX can do AAA, the software named YYY can't do AAA" is not sufficient to establish that the software named XXX isn't just a later release of the software named YYY, with a new feature to do AAA added and with a name change courtesy of the supplier's marketing department. For example, there are plenty of things that the current version of Wireshark can do that "Ethereal" couldn't do, but that doesn't mean that "Wireshark" is a different piece of software from "Ethereal" (no, they're not, regardless of what some people might think).
True enough; sometimes a new name doesn't mean we should make a separate Wikipedia page. It's exactly right that backward compatibility alone isn't enough to distinguish 'new product' from 'new version'. But what is? Should the Windows 98 article be merged with the Windows 95 article because 98 is so similar to 95 that we can reasonably consider is a different version rather than a different product? Microsoft spent a fortune shedding the 95 brand, and IBM did the same with AS/400. When should Wikipedia differ from the manufacturer (ie the citation that anyone can verify)?
"OS/400 had no Java". IBM disagrees with you here.
As one who lived through this era, the Java of V4 was hideous compared to the J9 JVM we have today, but your point is taken. I shouldn't have said 'no Java' but 'practically no Java'. I should be more careful with my writing.
At least according to Frank Soltis' Inside the AS/400 (ISBN 1-882419-66-9), the low-level code below the bulk of OS/400 changed between the CISC and RISC AS/400s (completely rewritten - the older CISC version was written in some PL/I derivative called, if I remember correctly what the book said, "PL/AS", and the newer RISC version was written in C++), with the MI-to-native-machine-code translator generating PowerAS code rather than IMPI code - so, yes, there was a significant change to the OS, but that change happened well before AS/400 got renamed to "eServer iSeries", much less to "IBM i".
IBM realized that the hardware for the machines running IBM i and the machines running AIX was essentially the same, as they were even using the same microprocessors for them, so they decided to just sell IBM Power Systems machines and have them run either OS. Guy Harris (talk) 17:11, 6 October 2015 (UTC)
As editors, it's our job to add clarity. It's very clear and reasonable to say that the operating system, IBM i is a descendant of the operating system, OS/400. I could be convinced that IBM i should be on the same page as OS/400 - but I'd argue that said page should use the current name, IBM i, and not a 'formerly known as' moniker. It's a bit of a stretch to say that the current Power 6, 7, and 8 hardware descends from the CISC or even the RISC AS/400 hardware. They don't share adapter cards, memory cards, controller cards or power supplies; what makes a new Power 8 just a new name tag on an old System i?
The original proposal notes - accurately - that IBM treats the hardware/OS combination as an integrated platform. If one wants to run IBM i, one has to buy Power Systems hardware from IBM. Someone researching the platform might well be interested in the history from AS/400+OS/400 through System i+i/OS to Power Systems +IBM i. Is the argument that this combined, history-through-today article should be called IBM i because that's the name of the current OS? The more I think of it, the better that sounds. Buck (talk) 00:31, 7 October 2015 (UTC)
"FWIW, no AS/400 had a PowerPC chip." FWIW, the International Business Machines corporation disagrees with you. Guy Harris (talk) 05:14, 7 October 2015 (UTC)
This citation helps underscore the confusion surrounding this platform. This reference is for the operating system, not for the hardware. Compounding the confusion is that it's the iSeries Information Center, not AS/400. Be that as it may, the reference cites 'PowerPC AS' processors. In general, a better hardware reference is either the System Builder or the System Handbook - both confusingly tied to OS version because IBM routinely announced a new OS version to support newer hardware. The place to see these old System books is the AS/400 Online Library In particular, the AS/400 Road Map for Changing to PowerPC Technology is the canonical reference for the first PowerPC machines, which were the custom PowerPC AS chips. There's no point in starting another circular discussion on whether these are really PowerPC or not :-) my memory was that they were unique to our box, and that memory is correct. If IBM called them PowerPC chips, that's good enough for me to stand corrected. This edit added only to illustrate why clarity has been so hard to come by. Buck (talk) 17:56, 7 October 2015 (UTC)
The "AS" is for the "plus tag bits and, at least in some processors, decimal assist" stuff. Soltis mentions the former in his Inside the AS/400 book; the latter comes from a USENET posting to comp.arch. I suspect the tag bit support may have been continued in the POWER chips used in IBM i machines (I tried looking for hidden signs of tag bits in at least one POWER8 document, but didn't find anything obvious), but not documented by IBM; the decimal assist instructions could have been omitted in later processors in favor of using regular PowerPC instructions, as there's no requirement in IBM i for binary compatibility for translated code, just for MI code. IBM did use some of the RS64 processors in RS/6000s as well as in AS/400s. Guy Harris (talk) 18:40, 7 October 2015 (UTC)
"I could be convinced that IBM i should be on the same page as OS/400 - but I'd argue that said page should use the current name, IBM i" It does have that name.
I think the stuff in the Summary section that discusses the OS should be merged into the IBM i page, as should the entire Features section and at least some of the History section. The IBM System i page should discuss the history of the hardware, both IMPI and PowerPC, and mention the software at least to the extent of noting that 1) all machines from the first AS/400 through the IBM System i require IBM i (including the called-microcode-only-for-legal reasons "vertical microcode"/licensed internal code), the fact that changing the low-level machine code didn't break binary compatibility (unless you threw out the MI code), and note that it was replaced by the IBM Power Systems, which could run AIX, Linux, or IBM i. Guy Harris (talk) 06:08, 7 October 2015 (UTC)
I like the way your suggestion is shaping up. Put the OS information on the OS page, focus on the current generation while presenting previous generations for historical perspective in a 'history' section. Cleaning up the IBM System i page to focus on the hardware seems a very good idea. Buck (talk) 17:56, 7 October 2015 (UTC)