System iNetwork

I think it is already all here in System i. WDSc 7.0 is vastly improved, yet has an image of being incomplete.

I would ask IBM to work with COMMON, User groups and forums to handle such silly long-standing problems such as the slow download time initially for SEUPDM members and get rid of the dorky first screen/form that displays.

Also make sure that customizing colors is intuitive and not through the convoluted means described by Bob Cozzi.

I would like very much a real dialogue with end-users (public) to confront the issue of the missing GUI.

Please make all aspects of DB2/UDB are up to snuff with the other versions of it that IBM has out there. Microsoft's juggernaut, SQL Server 2005 is knocking on both IBM's and Oracle's doors. Let's address that!

.NET ! IBM, Please buy or partner with ASNA's RPG.NET. I wish this had been done before they were bought out.

Instead of IBM's scattershot approach, please give us a unified management that will complete each one of these goals, one by one. Warning: this will require a long-attention span!

Thanks!

--John

In a requirements demo, where the decision is made on what software to build or buy for an application, no one sees the server, just the client. Whether it is a company is buying the million dollar ERP, building the next departmental in-house application, or a college is teaching the next generation of students, the winner is the best interface. Happens over and over, right?

Isn't it critical then, that developers using IBM systems have the advantages to enable them to build exciting and visually compelling applications? Isn't that why the System i needs to give its developers a standard integrated out of the box GUI environment. The interface IS the face of the box, right? So it is needed and vital that both new customers of the System i world as well as those with 30+ years of existing business logic have a way to build, buy, modernize and enhance cost-effectively.

When a program generates 5250 data stream today, at your PC, the Telnet client in iSeries Access translates that into the GUI we see, right? So should the new interface but enhance that 5250 with whatever a GUI needs. Want to base it on Eclipse - go ahead. On a client with the proper Eclipse plug-ins, let the system generate the most beautiful Windows-like or Browser interface ever seen. Run it on native Windows, browsers, phones, Macs, iPods, iPhones - whatever the customer wants. All of it powered by System i. That would help the System i ecosystem thrive!

Maybe IBM would have to extend DDS or UIM. So what? If delievered, it would do more than all the advertising to-date in the System i world to say the box is NOT old and outdated. Enables IBM, ISVs or in-house teams to more often win the battle of the interface.

What are the answers? Supposedly RPG has a great future. Cobol, where does it stand. But what is the screen presentation; the UI to be? To name a few: HATS, Webfacing, JSF, java, CGIDEV2, your own (after go thru all the API's).

What is the 'native' GUI for the i5 (System i, AS/400, iSeries, S/38)? WDSC has something called JSF (replaces the iSeries Web Components) but try to find anything more about using JSF-RPG than the accompanying tutorial. No mention of EGL (is that truly an old mainframe tool) and Rational. Which was unbundled from WDSC. And WDSC comes in an Advanced and non-Advanced versions. Then there's either HTTP (IBM or Apache) server or WAS.

Finally, it appears that it's the third party companies that are attempting to bring some type of gui to this platform. Lastly, depending upon your needs, your choice of WDSC, Ops Nav or Client Access gets launched. All of which are pc tools.

Conclusion, there is no target direction from IBM on either the UI nor the platform; lots of technologies are being put on the playing field and it is up to the players to pick, choose and develop what they want/need. It would appear that the talents of the labs are not delivering what is being asked for.

Send me an email and I'll send you a link to an RPG based graphical iSeries Web site. Seek and you will find.

...Just get started on what's available.

There's IBM's HATS, WDSC, CGIDEV2, Java, WebSmart, PHP-WebSmart, and others.. Pick one..

Someday soon I think somebody may come up with a good client-based GUI-rendering utility, in some open-source format, which will serve up fast binaries too...

I hope that comes in fast, it'll probably make it all lots easier, plus yield up some good impetus, we'll see.

But my philosophy is that we can't wait. Period.

--Alan

If IBM would have invested in DDS, they could have done the perfect GUI.

How many PFs, LFs, DSPFs and PRTFs are there on this planet ya think?

Probably more than web sites (sites not pages).

A small but efficient file maintenance can be made in 2 DAYS with DDS not 2 months (+++) with ANY other solutions.

That's why the 400 is (was) so popular.

How about a new protocol like HTTP5250? ONLY available on the 400 and ONLY supported on Client Access and ALL the hard work of GUI on the PC (zillion of megahertz waisted everyday here) not the server.

We have been using Newlook from Looksoftware as our GUI for 3 years and are very happy with the results. This approach supplies a robust interface with great tools and the ability to create macros to bring data from other Servers/Applications to be integrated with our existing screens when we wish to add new functions that require data from PC Based Servers.

IBM would be very wise to purchase this company and add this product as an integrated offering with the i5.

I don't know why IBM keeps tripping over itself to announce and then bury Web and GUI products for the AS/400. Remember Net.Data, VisualAge, Java, JSP, WebSphere, WebFacing, HATS, iSeries Web Access, PHP, and CGIDEV2?

Frankly I have tried almost every one and the only product that stuck with me was CGIDEV2.

All the other major industry players, Microsoft, Oracle, and Sun Microsystems for instance, have worked through the GUI/Web issue years ago and have been releasing generations of coherent evolutionary solutions/upgrades with strong support from user community. IBM just keeps us scratching our heads and asking nervously if we are on the right track. Companies do not want to evaluate 12 competing products from IBM. They want a single strategy and a long term comittment.

IBM has absolutely fumbled here and is paying the price in market share.

--MC

Why don't use the event driven GUI language features of IBM VARPG native at System i.
And produce an client independent GUI (HTML, XML, WML, Rich-Client etc.).

MC, I couldn't agree more.

Ibm has arrived at a point that I and probably many other System i professionals predicted 3 years ago. Ibm is no longer a leader in the IT industry, not by a long shot in the developer arena and now also as a hardware manufacturer, being second after HP/Compaq.

I would advise Ibm and especially Rochester to acknowledge its downsized/lowered ranking in the IT industry and act accordingly, to drop its high attitude and life in splendid isolation, to open its eyes and ears to what contemporary IT industry leaders are saying and doing. Learn from them fullheartily (no, Microsoft is not the devil, nor are they kids playing in a sandbox), and then, come up with a plan to:

1. Present a coherent roadmap plan for the future of the System i product line and stick to it.

2. Convince ISV's that with this roadmap and support from Ibm, they will be able to develop once again successful products for System i.

There's no doubt it is very noble of Chris Maxcer and may developers working in company's home grown applications, to experiment and looking for ways out of this mess, but the real test is the fight for the ISV's. If Ibm doesn't succeed in winning back the ISV's, then it's game over Imho.

Personally, I need to focus my development on "business applications".

To me IBM needs to step up to the plate and write something "SIMPLE" so I can write one screen (probably as a DSPF) and have some IBM supplied interface use something as simple as a HTML Style Sheet to create the GUI Screen "directly from my DSPF" for me.

I just don't have time, or the brain cells, for the "solution of the week club" they have been throwing at us for the past ump-teen years.

I don't want to learn a whole sub-development system or buy some third party software that is going to be eaten up by some other entity next month, only to have them tell me the following month, to scrap that now do it this way.

All I want is for my tried and true green screen RPG applications to run using a GUI screen, with NOTHING in between.

How about a new protocol like HTTP5250 or HTML5250 (instead of or in addition to a DSPF). I want to be able to call one RPGLE program and have the system know if it needs to load the 5250 DSPF or a GUI display depending on what is calling the application. I simply don't want to have to think about it, that is why we have a great system like the iSeries, "set it and forget it", not clone it, tweak it, run it through this, run it through that, get a degree in this today and learn that tomorrow, and hope it works next week.

Sheez, I am a RPG business application developer who improves my skills as I need to, but I have my hands full learning the enhancements to RPG, I have a (more than) full time job, I am not a full time student with nothing to do but learn the latest and greatest every other day. Also, my company doesn't have pockets deep enough to invest in this technology this week and another one next month.

Come on IBM, take a step back, then take a step up (up to the plate that is), and give us (your iSeries fans) something that works for us (remember us the ones who have been with you since the System 34/36/38 days) who are buried in the trenches writing Business Applications (doing real work).

Let's make the iSeries the most popular box on the planet (like we all know it should be). But let's get back to the grass roots, and keep it consistent, simple, yet functional, and reliable.

MC hit the nerve with the '12 products' sentence.

12 different products, suddenly changing license prices, sudden end of support from products, new 12 products...

And of course no simple examples, no '4dummies'-guides, no 'get&set up and go'-guides... just 1000 page redbooks and some technology previews of beta versions.

• WDHT needs to be put inside Client Access license now. No separate license.
  
• After that a new website (or maybe there is already, haven't seen?) where you can download WF and HATS project ZIPs, EARs and WARs with simple 4dummies guides to get quickly a demo for decision makers and test the possibilities. Usually there are only few ppl supporting some System i RPG/COBOL ERP software and they don't have time and possibilities to investigate WebFacing, WebSmart or PHP for weeks.
  
• And inside that website a forum to discuss about problems and development and possibility to share ZIP/EAR/WAR-files and necessary JavaScript's
  
• And while doing that, IBM needs to improve WebFacing and HATS.. a lot... just for one example WebFacing has to have FireFox support and better IE7 support
  

Why? There shouldn't even be that question... System i community needs this kind of support and help to develop GUIs (BUIs) now to save IBM's System i business and to protect IBM's investment on the platform.

Without this kind of development, System i with it's CUI (Character UI) is slowly just slipping away from companies and slowly these so called legacy systems are replaced with other GUI/BUI softwares.

And yes, I know. I'm repeating myself like the rest of the System i professionals/community. :)

We have been using Lansa and most of their offerings for years.

IBM would do well to buy this company and make their tools part of System i.

My company, unfortunately, decided years ago to replace our iSeries with HP/Dell servers because of one major reason . . . the interface.

Even though the iSeries application was superior and more complete in many ways (also in-house for easy modification), it was green screen and at the time of the decision we had only touched on the Lansa web tools and Visual Lansa interface.
The replacing application on the HP servers is all GUI.

Thank you Chris! Thank you, for asking for this feedback. It helps me get my thought process laser-like in what direction is needed!

I liked best MC's clear summary:

"Companies do not want to evaluate 12 competing products from IBM. They want a single strategy and a long term commitment.

IBM has absolutely fumbled here and is paying the price in market share."

Chris, do you know Mark Shearer's Email and Dr. Soltis' Email? Who are the shakers and movers all the way up to the CEO (Samuel Palmisano) -- and what are their Emails?

I have read some "right on" comments, including IMHO my own! They need to read these. I just don't think the System i is in trouble but that there is a lack of direction in varying degrees in all IBM products.

I think what these posts point out to is a crying lack of LEADERSHIP.

I am so glad that System iNetwork has joined the chorus from just about all publications. I want very, very badly some Serious Dialogue with the IBM Higher-ups.

--John deCoville
--Pinal County Government

--It's time -- the window of opportunity
may not stay open.

[*Note from Chris: Obviously, everything we've been talking about makes for some massive topics. Even just the GUI focus here blows up with competing degrees of concern, solutions, and needs. Taking just the GUI issues to IBM should be reasonably thought out -- on the surface, saying, "You've given your customers too many choices. We don't want options when it comes to an i5/OS GUI interface." . . . makes total sense. But from the flipside, imagine hearing from your customers that they don't want the abililty to customize their systems, they want to be limited. . . . Choice is generally good, but too much can be a problem. If you offer a child three types of Kool-Aid to drink, you might find yourself waiting all day for a decision. A kid can just lock up on you. But if you offer two choices, boom, they make a decision and off you go. So what I'm saying is this: I'm game for going to IBM, and I'll definitely end up asking IBM about their GUI strategy, but we need to have some serious solutions -- which would have sort of clear return on investment for IBM -- before presenting IBM with a dozen flavors of Kool-Aid. Part of the problem with the System i is how very diverse its customers are, so in that respect, I'm very interested to learn what IBM is thinking and doing as it reorganizes. In the meantime, most IBMers can be tracked down through IBM's web site: http://www.ibm.com/contact/employees/us/ ]

To bad for IBM...

If only they would have included a "free GUI" from the start, they would be selling more iSeries than they could manufacture!

Yes, there would be an initial loss of revenue for the GUI but just think about the increased sales of the box!!!

They have AIX, they have XWindows, they know the technology people need! They simply dont feel they need to give the i5 that technology. Perhaps doing so will kill off the pSeries? or even worse create a world beating solution that beats the pants off Windows and all the other OS's out there!

I was thinking about how to develop a GUI interface for our products and just gave up in disgust! There is a need to have a single solution and IBM has to provide it, not a 3rd party. It must come from the core of the OS just like other OS's do. No more Java please! I dont want to have to build all that code! The OS should be able to send and receive the data, not rely on me to build a client to do that!

Today I simply create UIM code which IBM interprets for me and displays the Green Screen, why can't they take that same code and build a GUI for me in a browser? It's structured and will not compile unless it fits the screen layout.

I have looked at the GUI generators and none of them work with UIM! plus to be honest the output looks like it came from an old Sinclair PC200 or worse, thats not going to take the platform into the next generation!

Some have said buy LANSA or whoever! No buy Apple or someone who has a real sexy interface... Dont worry about the old applications and how to convert them, the owners of those products should be able to do that for themselves, thats why they charge customers for the software and maintenance! Another reason for them to bring out a chargeable upgrade!

Give us an XWindows server in the AIX core! then we have something to work with... They have the technology they can rebuild it! ($6million dollar man)

Chris...

http://www.shield.on.ca/Blog

Somebody put ugeerts as head of IBM development please!!! ;-)

The ONLY thing making the AS/400 seem old school is the green screen. And why IBM hasn't provided a simple means to translate a DSPF and menus to a web-based GUI interface that really works is beyond me.

I should be able to name a menu, XYZ, compile it as a window-based menu, and the user on our network should be able to access http:/myas400name/menu/XYZ and voila that menu appears & functions as a web display.

Then, every program on that menu should be a simple click away.

Stanleys and others commented upon the interface of VARPG and the commands that drive that GUI. VARPG was/is awesome and easy to develop in. Why hasn't IBM eveolved that into the browser world as well? I agree with the folks that say IBM has drpped the ball a little on this and should begin to aggressively play catch-up. Go Sox!

The reason this platform (S/38, AS/400, iSeries, i5) became popular is because businesses could easily, quickly, and cheaply build BUSINESS applications (not games). Pure and simple economics. If the business needs changed the software could too.

It had a unified development environment where all the components were in use by ALL the developers (eg. data base, screen definition, application language RPG & COBOL, and OS language CL). IBM, Vendors, and in house programmers all used the same tools to build applications. Everything they built was interchangeable. You could use a system command to display information inside your own application, or just as easily call a vendors program, or call something you built in house.

The tools were easy to learn and produced high quality, scalable, stable business applications.

The tools were stable in the long term making the platform more inviting to customers.

Just to beat the point home, the real key was that IBM was using the same tools and UI to build the OS and tools and apps as we the customers were using. We were able to leverage IBMs investments in the OS for ourselves. This also created community, since we were all doing pretty much the same thing.

COMMUNITY SIDE BAR: We have lost community for several specific reasons. One is that many have found a 3rd party tool to make programming easier (Lansa, BCD, etc). For the rest of us its the many many incarnations of RPG which splinter our community on account that they are incompatible. Somebody writes an article, you try to use the code, but can't because its two releases ahead of your shop. IBM needs to focus more resources on RPG and be done with it. Put out a new and final version and not touch it (syntax wise) for 5 years. Then make it available via PTF to prior OS releases.

Back on point, I am beginning to see the light at the end of the tunnel when IBM is trying to build new tools and applications using the same tools they gave us. I am referring to their efforts at building applications on top of Eclipse. Like the new Notes release and Office suite Symphony.

Unfortunately the light at the end of the tunnel is dimmed by the drab totally uncool appearance of Eclipse. It looks like and smells like a Java app because it is. Grey, boring, not sexy. It needs skinning capability. Want to know what it should look like, look at any "modern" app built on Windows. Ouch. Hey, it looks new it must be great. NALM (Not A Laughing Matter)

But the war is on and IBM appears to be finally trying to take a serious "byte" out of Microsoft. But Windows isn't standing still either, although Vista does give that appearance.

Code on!!

To me it seems like we are at a cross roads. In the past, developers have started with some business function code and added screens to it that would display or accept additional data. Screens were present but they were extensions of the business code which was the real driver.

I think we are entering a new era, where the development will start at the web with a screen, and the business code that processes that data will then be attached to the screen. We complain about the lack of a single tool, and that is definitely a valid criticism, but I am not sure most developers are ready to accept the fact that switching to GUI also means switching to a browser page centric methodology.

I would sum up what has been said and what I believe as follows:

We originally bought the AS400 because:

1. You got more hardware for the money.

2. You got a programming environment where interfaces (files, display, printer, SNA communication) was simple, consistent, reliable, and responsive.

The whole AS400 community has been slow to transform user interfaces because all of the options listed in previous posts failed to have at least one of those four characteristics.

The Java based options had the problem with simplicity, and responsiveness. You were left with the option of buying capacity that you could get cheaper on any other OS platform.

The third party products here have all of the characteristics except consistency. We just don't know if the OS features that they built their product on are going to continue to be supported. We also don't know whether the changes required in apps using the third party tools are going to be locked out of using new features because of their required design.

The CGIDEV2 options have all the advantages except simplicity. Why should business application programmers have to build style sheets, JavaScript for basic business field validation, or management for user spaces to maintain state for browser applications. All of the third party tools that build CGI applications handle these issues.

If IBM built a file type BROWSER, that handled session state, building the HTML, CSS, and JS files that need to be output, they would find that AS400 applications were being converted at an amazing rate. It would be a cheap fix for both in-house programmers, and software vendors. After that, if programmers want to pay the performance penalty to make the applications portable, WDSC is still there with WebFacing, HATS, or Servlets.

IBM should focus on the core i5/OS features that would enable 3rd party developers and end users to create their own GUI.

Consider what it would take for a 3rd party to extend existing DSPF and 5250 to include more GUI components. You would need control codes in the 5250 data stream defined for user extensions. The telnet server and/or twinax controller in i5/OS has to be able to recognize those extensions to the extent it allows the GUI extension data to pass thru to the device. (This all might already be in i5/OS. just an example of the core work IBM should focus on.) IBM would have to release the code of the client access 5250 client or add hooks to allow user code to run depending on the GUI traffic in the extended 5250 data stream. In the RPG program you might need more low level access to the operations which send and receive the DSPF records.

My point is 3rd party developers exist which can provide competing GUI addons to DSPF applications. As long as IBM controls i5/OS, ( my hope is that IBM sells i5/OS to a dedicated startup company ) only it can improve the core functionality of the OS which makes it possible to build a decent GUI.

-Steve

Back in the 70's when IBM was developing the 5250 interface and integrating it with native languages and other tools, they desperately needed in-house developers and small ISVs to get systems into every industry.

Today however, IBM has relationships in every industry, and has evolved from a systems company to a services company which in many respects competes against in-house developers and ISVs for application development and support services.

Notwithstanding IBM's developer tools business, they don't have the same incentives to provide tools as they did in the past. Unfortunately we have to face facts; in-house developers, independent consultants, and small ISVs are losers in IBM's strategy to evolve into a global services company.

Additionally, IBM simply doesn't carry the same weight it did in the past with respect to user interface technology. Microsoft, Mozilla, Macromedia, Palm, and others are the new leaders in this area. It doesn't appear to me that IBM is able to define new UI standards the way they did in the past. The competition is too stiff. The market is too segmented, which is even evidenced by the lack of unity in this blog.

I'd suggest that in-house developers, independent consultants, and small ISV's learn HTML, CSS, Java Script, the browser Document Object Model, CGI (or PHP, or J2EE) and stop expecting IBM to provide a native interface comparable to 5250 in terms of performance, integration, and ease of use.

In the previous thread I wrote:
So what is lacking from the System i is a WDSc for our business users that we can communicate with in a standard way from our System i ILE programs, an interface even more powerful than our 5250 interface.

The browser is built in to Eclipse WDSc, it's not that a browser web page isn't useful. But it's just another tool in a complex desktop.

IBM should dot the i in integration and provide a sSystem i desktop client that our business uses don't have to leave to use System i served data in a spreadsheet amidst complex arrangements of screen data.

We can either shoot for the bottom with lowest common denominator web pages, or we could dare to be great. We have the system to do it. IBM should dare to make a client interface worthy of the System i.

But in calling for an integrated i5 desktop client that is more powerful than what other desktop clients offer, I also recognize the client foundation that IBM has created, the Eclipse desktop client used for such products as Websphere Development Studio Client and Notes, among others.

We also should acknowledge that the default system administrator interface is Navigator, and that the framework that IBM has laid out has allowed such innovations as Softlanding's WDSc extensions for i5 print spool files and message handling.

In fact, we should acknowledge that it is no longer accurate to say that 5250 is the native interface for the i5 and currently supported iseries systems.

Another current SystemiNetwork news article concerns VAI's S2K i5 software for small and medium sized businesses or departments of large corporations, such as where I was when I evaluated VAI for a department nearly ten years ago. During that evaluation, I did not see a 5250 screen unless I wanted to and toggled over to it. VAI used one of the many 5250 driven GUI solutions developed for the i5 in those days and I'm sure has added one of the many web based solutions since.

All these available solutions show that IBM provided the necessary API hooks to integrate with the i5. And the fact that everyone here who called for IBM to buy one of those products and include it with i5/OS named a different product shows that there is no commonly acknowledged superior third party GUI interface solution. If there were a cost effective superior solution, we should buy it along with the other products we choose for i5, but we all don't buy the same one there either. Some are closer to a default than others, but I've seen different database file utilities, change management, and job scheduling products, from i5 based to custom to commercial, in different shops throughout the years. And there are far more choices in GUI interface products.

Among those is XWindows, as someone suggested. XWindows is available in the PASE subsystem of i5/OS, and the PASE subsystem is integrated with ILE, so an XWindows client should be available of we wanted to use it. It also should make the i5 just as "graphical" as a Unix system with XWindows.

While I think that IBM could and should take Eclipse WDSc engineering, the Notes Eclipse client, the Symphony Eclipse office components, and their Webfacing product and provide a desktop portal that integrates directly with data in the i5/OS DB2 database, I also think that IBM has now provided the pieces needed for an i5 community driven open source interface to i5/OS.

And by open source I mean no license, just source code placed in public domain with a "For educational purposes only" tag so business would have no qualms using it. The i5 server code should be in RPG /free to provide reference code for RPG programmers as ERP source code acquired by a company did in the past.

Certainly interacting with Java JSP code is one of the interface routes to support, along with sockets and/or HTTP tunneling to an Eclipse plugin.

With a service program that provided 5250 to and from conversions from different protocols like HTML, XML, XUL, XWindows, and whatever else is desired, called from a dispatch program that is opening virtual terminal interactive sessions or submitting batch jobs as called for, the i5 would have the best default interface, an open source interface.

rd

Here is my 2-cent.

IBM needs to dress System i up for success. System i has a great personality and all the necessary qualities to be the champion. But it is not going to win the beauty pageant with a 1970’s or 1980’s or even 1990’s costume. To the business executives and senior management, the System i’s interface is the System i. The first impression is vital and should not be fatal.

One of my students was making a presentation last week and he uttered something along the line that “… you know, the old-fashion green screen that is not too user friend.� And this is a student who works as a technical support and is excited about computers and programming.

I think IBM needs to be proactive and take the lead on this before System i hits the point of no return. Too many cooks spoil the broth. IBM should decide on the best GUI for System i (native and/or applications), bite the bullet, and go for it. It may not be the GUI that some members are looking for, hoping for, paying for, or praying for. Most probably than not, there will be complaints and outcries. To alleviate the negative feeling, the best thing is to have a truly stunning and impressive GUI. In any case, it is impossible to satisfy everyone. But in the long run, a decisive move by IBM on GUI will be good for System i and its whole ecosystem.

There will be some who insists that the green screen and text input are good enough for a server. I do not dispute that it may be true. I think most members of the System i community will agree with that too. Nevertheless, to win the beauty pageant, internal quality is often less important than external beauty. Business executives and senior management, like judges in a beauty pageant, will only have a few hours (or a few minutes) to spare and will jump to conclusions very quickly. Moreover, the new generation of IT professionals and business executives will grow up using icons and avatars. To them, green screen is simply not acceptable and not presentable irrespective of how effective and efficient it may be. Perception, unfortunately, is important.

System i already has the necessary internal quality. To loss out because of dressing is really regrettable, unfortunate, and pitiful. Besides, it is much easier to change dressing than to build up the internal quality. One quick and easy way out is to go shopping and spend some money. I think it will be money well spent!

I have been follwing the comments that have been submitted. The one from Scott Mc "All I want is for my tried and true green screen RPG applications to run using a GUI screen" is the one that sums up the issue. Also another comment, "needs to be simple, consistant, reliable and responsive" is the other comment that needs attention. IBM has not developed any GUI that has addequately addressed these two comments.

There are over 180,000 iSeries shops out there that have been using the machine for a number years. They are predominantly RPG shops using 5250 screens that have evolved over the years. Many do not not have staff, time or the desire to learn new skills and IBM tools to modernize their applications.

They would like to have a GUI but they do not wish to reprogram their applications. What they need is a simple, reliable and noninvasive solution. This represents the the status of the majority of the 180,000 users mentioned above.

We have been working with IBM systems since the advent of the System/3 in 1968. Over the years we have seen excellent products, operating systems and support. They have not done a very good job addressing the User Interfaces or Printing requirements. At times they seem to lose track of the KISS principle (keep it simple stupid).

About 4 years ago we decided to begin modernizing the front end and the printing requirements for our systems. We have limited staff and resources. When we looked at the offerings/tools from IBM it became apparent that we did not have the staff, resources or time to explore their solutions (Websphere etc.). We were looking for simplicity and reliability. We wanted minimal affect on our applications (over 600,000 lines of code).

IBM was not the answer then (nor do they have a good answer today). They are very good at developing machines and operating systems, but not so good at comprehensive user solutions.

After examining various products/solutions, we chose the products from Looksoftware (with 12 years experience in the market) for developing our GUI. For our printing we have been using a product called Formsprint from ICS (full color laser printing without IPDS). Our experience has been very painless and highly productive.

IBM has developed excellent products, but they seem to have lost track of a basic marketing principle; "people buy what they see".

Check out VAI's S2K Enterprise Software Version 5.0. VAI's top notch GUI developers have created a world class Windows Interface using Seagull's JWalk product that is simple to develop and easy to maintain and install. The advanced features and capabilities that can be coded in JWalk are years ahead of HATS, Webfacing or other GUI products we have seen.

I find this whole thing rather funny in a few ways.

First off, where do we need this GUI? On the OS side for administration or on the client side for apps?

If you say apps then clearly you haven't heard of this thing called a browser. You can develop to it today, and you should be.

I don't know of a GUI for System/z and I don't see this argument over there. I also know very few AIX administrators who use the XWin interface. They all use that command prompt and rattle off endless strings of chaotic parameters to do what we can do with wrkactjob.

The app is what is important. If any CEO is out there making a platform decision, wow, that's scary. Leave that to IT, have the CEO look at the app itself and then IT can decide where best to run it.

So I guess I don't get it. Other than to say the System i's biggest problem isn't that doesn't have a native GUI, but rather that the dev community seems more willing to argue about it than use what they have to provide a rich interface, and that the industry analysts and in fact parts of IBM have endeared themselves to this "industry standard vs proprietary" servers moniker.

i5/OS isn't currently offered on x86, and that's a problem larger than the lack of any GUI.

Mike Downey, that was one flattering comment you made there, but I suppose I'll have to take it with a grain of salt ;).

I know it's harsh to say that I doubt if anybody will be able to upgrade iSeries succesfully with a contemporary GUI and a new modern personality. After so many efforts, we see Ibm is still nowhere. Now we have fallen 10 years behind the competition in terms of GUI development while they have caught up (and more) in terms of reliability and scalability. Oracle and SAP are certainly NALM (not a laughing matter) while Windows has become ANALM (almost not a laughing matter).

Maybe IBM should take a bold leap forward, and develop a whole new innovative computing platform. A new product format, as marketeers would say. If Rochester invites me in town for a week or two and books me a hotel with warm bath, I could shed some more light on this (and at the same time, they should invite all other posters in this blog too).

ugreets wrote:

"Maybe IBM should take a bold leap forward, and develop a whole new innovative computing platform."

Even better, IBM could begin with i5/OS but drop support for integrated System x servers, LPARS, Linux/AIX partitions, Java Virtual Machines, Webfacing, HATS, PASE, PHP, mySQL, System i Navigator, WDSC(Eclipse), ODBC/JDBC, and basically reduce it to the native virtual machine, languages, 5250 interface, and tools.

Then turn around and add support for browsers using consistent Web 2.0 UI standards that interface with ILE based servers. Simple, robust, reliable, responsive.

Wait, that would be too easy ;-)

One would be surprised at how many CEOs or top level business executives would like to have information and be consulted on a few hundred thousand or a few million dollars investment -- unless one is talking about businesses where a few million dollars are considered loose change. Most business executives are fairly technical savvy these days -- unlike the cluelessness of their predecessors.




It is also good PR for CIOs and IT managers to "consult" and inform senior business managers before making major IT expenditure.




Besides, an antique looking interface will turn off students and young IT professionals.




In short, there is really no reason why System i should compete at a disadvantage position because of its interface. If System i has to loss out to a competing platform, it should be for some good reasons and not for its interface.

This is the most commented on story I have seen by the i5 team! Shows just how important we feel the issue is. Having Unix/Linux /AiX or whatever being controlled by the terminal is not important, when the boss comes down and shows off his box its normally running XWindows or the like, the admin may go back to the terminal when he's gone but he will be just as adept at showing the XWindows views. Lets get IBM developers in a room and see what they have to say! I dont see any comment from them at all, and they are the only ones who can change anything!

Chris...

Well Kevin the reason the UI is so important on the System i and not so important on System p or z, is because System i applications are usually used directly by the business user.

System p boxes may be dedicated to file sharing, hosting relational databases, or web servers. System z is usually tied up in batch processing of data supplied by other systems.

So the CEO doesn't have to go to the data center to see the System i interface, because everyone in the company is using it. IBM's lack of a native GUI interface for a platform that is primarily used at the personal interface level is directly related to the sales statistics for the System i for the last decade.

Richard,

The application use model you discuss is indeed how things tend to work on System i, and THAT my friend is the problem.

Too many apps are simply not being modernized for Web 2.0 or anything else.

As you mention, one will find few if any other platforms where the user signs on directly to the box with any kind of textual terminal emulator.

So, how are those applications used? Full desktop install? Sometimes. Web browser? Certainly. Sytsem i can do this as well.

Our company has AIX, they have mainframe and of course Windows. For ages the mainframe was used directly with a 3270 session, and in fact still is in many cases. The apps are terrible to use.

So, what did IT do? They created WEB interfaces to those apps. Yes the mainframe is acting as a back end data server - and so can System i.

I am not saying the interface doesn't matter. It does. I am also not saying I would reject having a native GUI. I wouldn't. I am saying the application interface is what is important here, not the OS looks like. If it weren't, then the mainframe and AIX would have been dead ages ago. I would also challenge someone with telling me how this magical native GUI would be presented to the business user directly? Terminal session? Not likely. If your answer is browser . . . guess what, you can do that today.

Kevin wrote:

"i5/OS isn't currently offered on x86, and that's a problem larger than the lack of any GUI."

That's a problem? There are advantages to running on the POWER architecture (aside from performance arguments). One of the features of the POWER processor aids in the tremendous security and stability of the System i (and all its predecessors). Any pointer modified by anything other than the processor is rendered invalid. This eliminates so many of the security wholes other systems suffer from. If you want the details, read Fortress Rochester by Frank Soltis.

I think arguments could be made regarding better performance due the architecture. I don't believe the systems could scale anywhere as well on x86 as they do on POWER. (Again you can read the book--or attend one of Dr. Soltis' Vision sessions at COMMON or anywhere else he's speaking. You can ask him about the comparisons of scalability on Intel vs. POWER.) They don't call it POWER for nothing.

I also think that price of the architecture shouldn't be an issue. Look at the game consoles that run on POWER. They sell for less than many PCs and they've got a lot more power.

More than two (2) years ago IBM announced that they were studying the feasibility of offering a native GUI for RPG developers but since that time the announcements have centered around PHP, Rational Tools (like EGL), J2EE and standards based GUI solutions. It doesn't look like IBM is going ahead with a native GUI. Even if they did, do you think they'd convert all I5/OS menus and commands to the new browser device file, or the new HTML display file?

The more I read what people are asking for in a native GUI the more it sounds like what IBM is already offering via Webfacing and HATS (except for better performance and less complexity). However the performance problem can be solved by buying new hardware and the complexity problem can be solved by embedding a small-footprint J2EE application server into I5/OS, which is what IBM has announced in V6R1.

IBM has also announced a browser based version of System i Navigator in V6R1, and they currently offer iSeries Access for the Web if you prefer getting to traditional I5/OS menus and commands via a browser.

If you can't make up your own mind on your own GUI development road-map, is it because IBM is giving you too many choices, or because you've never learned to think and act for yourself - outside the box?

Yes, Michael I see it as a problem. And here's why:

Simple market perception. Those who get it, get it and we use the full blown architecture, i.e. a "System i" branded server with i5/OS.

But, System i falls into the same industry category with System p because they're based on the same hardware which is labeled "proprietary server" by the industry analysts. Anything x86 based is considered an "industry standard server."

This is how it is and what we have to deal with. Proprietary equals bad and industry standard equals good. It really is that simple.

However, what are we really talking about here? Do we really need a branded System i or do we need the operating environment called i5/OS. I would argue that i5/OS is more important than the hardware under it.

If we had i5/OS on x86 you could toss that out like Sun has with Solaris to all kinds of folks who could then learn what this environment can do for them. If we're so interested in getting younger folks into the platform then there's a reason.

Don't take this that I don't like POWER, because that isn't accurate. I think POWER is hands down the best thing out there. BUT - that doesn't necessarily sell servers today.

I have been around this platform for quite some time and am fully familiar with Dr. Frank's body of work. Trust me it isn't that I don't get what the full package can do. I get that, always have. My point here though is that you have to look at where it is market-wise with the rest of the business.

Some chalk this up to the fact there is no native GUI, and I don't think that's the entirety of the story. Neither is price. This standard vs proprietary thing isn't about the OS but what it runs on.

Regardless of how killer POWER is, run i5/OS on x86 or x64 and you open doors. Lots of them. The OS is the future we need to embrace, and I think it can happen on BOTH POWER and x86/x64. When I look at what I need to do, it isn't POWER that gives me what I need, it is i5/OS.

I know not everyone agrees and I know this is somewhat contentious, but that's OK. Debate is good. : )