How to Get Out from Under an Obsolete Legacy Technology

[J.Pluta]

How to Get Out from Under an Obsolete Legacy Technology

Well, actually I misspoke with the best of intentions. I really don't like to quote you as saying "I can't get my head around Java." But I'm glad you're willing to set the story straight. I do have to admit, though, I am confused about why you consider PHP more comfortable than Java. The basic non-OO portion of PHP uses the same C-style syntax as Java. The OO parts of PHP are nearly exactly equivalent to Java. So... why are you more comfortable in PHP? Or are you talking about PHP vs. JSP? Because that's a completely different kettle of fish. Joe P.S. Are you promoting RPG-CGI?

[Guest.Visitor]

How to Get Out from Under an Obsolete Legacy Technology

Jon, Thanks for pointing this out. I stand corrected. However, I do have experience with clients of ours who have compared the performance of CGIDEV2 with that of WebSmart and found ours much, much better. That happened just recently with an application ISV. I don't know if this is because they were not coding CGIDEV2 optimally, though. I also read your comments on PHP versus Java and agree with you. I've been going through a similar experience with JavaScript- writing in both procedural and OO style- and still find there's just not much reason to use OO in many cases. BTW - has anybody corrected the RPG entry on Wikipedia that claims it has 'object-oriented features' ? Duncan

[R.Daugherty]

How to Get Out from Under an Obsolete Legacy Technology

This is the current entry start: http://en.wikipedia.org/wiki/RPG_programming_language RPG is a programming language for business applications. Originally an initialism for Report Program Generator, it officially no longer stands for anything. Its latest incarnation is RPG IV (aka ILE RPG) on IBM's System i servers; it inherits the System i Integrated Language Environment’s OOP features such as prototyped functions and procedures, static and dynamic binding, access to C routine libraries, dynamic link libraries, and fully recursive and re-entrant modular code. end quote Not only OOP, but inherited them, wouldn't you know. I'm disappointed they didn't write that RPG CHAIN's to objects. Other than that mildly out of bounds enthusiastic attempt to placate OO types, a nicely done Wikipedia writeup. rd

[R.Daugherty]

How to Get Out from Under an Obsolete Legacy Technology

I re-wrote my 8000 line 8086 Double Deck Pinochle DOS game in Java. Web enabling it to serve it is getting closer to the top of my list. If I didn't say that and just said I wrote it in Java, here's the source, isn't OO neat, they'd agree and never have known the difference. I plan on rewriting from Java to RPG. I've already done it in my head, just haven't sat down and done it yet. rd

[don.brown@msd.net.au]

How to Get Out from Under an Obsolete Legacy Technology

I have been reading the web this and web that, CGI, Java, PHP and all have their pros and cons but the way I see it there are two main types of applications, ones that need to be browser based and ones that don't. I see no benefit putting the General Ledger portion of a total application on the web. I do though see benefit in putting a new, graphical interface to green screen applications, (Unfortunately I am from the old school and can actually get excited over a well designed green screen!), predominantly to make it easier for new (read younger) users to start using the application(s). Theres HATS and WebFacing that have their place for Web but not a lot for a Richer experience. I have just started playing with HATS for Eclipse, which seems to provide a SuperSet of capabillity over HATS for web, performance is great. Anybody else had any early experiences with this ? So while we have a discussion on all the tools and ways to bring an application to the web, how about a discussion on which rich client technologies can be used to provide a better (and stateful) interface to the good old RPG back end logic. Don C Brown

[R.Daugherty]

How to Get Out from Under an Obsolete Legacy Technology

(Unfortunately I am from the old school and can actually get excited over a well designed green screen!) So do users. The stuff about people finding it easier to use a web page instead of a 5250 screen is propaganda from people who have web page products to sell, starting with IBM. Integration? Of course, who doesn't want better integration, all of which can be done better in high speed stateful 5250 sessions. Can 5250 or a client session in general be better? Sure, with better layout control. The little goofy pictures web types show just to show they can display goofy little pictures make for pathetic displays. An enhanced 5250 could display images in position x, y if we wanted. Geez, that's stupid. What is useful is bringing up documents and such with client side image programs, not a browser, but again a real program designed for that task, which I have done as early as the early 90's on a project for public access to government records using IBM's image storage and retrieval product on the AS/400. Web access is suitable for the public, but dedicated software will always kick the crap out of general purpose lowest common denominator stuff like a web browser. So the most powerful software will win. That's why Microsoft has been kicking IBM's butt around the block and taking names. But maybe IBM has seen the light through the stars. They have done outstanding work with a native GUI interface for Eclipse, and an open client based on Eclipse recently developed for Notes. I'm sure a few useful tricks can be added to the 5250 session interface that adds unique value to an iseries customer, as long as it remains high speed. People still care most about their business, not IBM's business. rd

[Guest.Visitor]

How to Get Out from Under an Obsolete Legacy Technology

Well, Jon Paris is teaching a few classes on CGIDEV2 at COMMON 2007 so I'd say the answer is surely yes. Chris 409218 CGIDEV2 - The Basics - "Printing to the Web" 409219 CGIDEV2 - The Details

[J.Pluta]

How to Get Out from Under an Obsolete Legacy Technology

Well, sure, but I don't think he says it's the ONLY technology. Even I say that there are some situations where RPG-CGI is the best option. Just not all. There are others who insist that RPG-CGI is always the best choice. These people are wrong Joe

[Guest.Visitor]

How to Get Out from Under an Obsolete Legacy Technology

Joe, Well, sure, but I don't think he says it's the ONLY technology. True. Guess that's why he's also teaching a class on PHP. :-) Chris

[J.Pluta]

How to Get Out from Under an Obsolete Legacy Technology

Yup. Jon teaches CGI and PHP. I think I've made it clear that Jon isn't a JSP advocate. But I hope I've also made it clear that there are a LOT more people both inside and out of IBM who do advocate JSP, for lots and lots of reasons. As I've said, RPG-CGI is right for SOME people, even though it is a technically inferior solution. But this isn't a VHS/Betamax deal where the technically inferior technology will win out. The pool of RPG-CGI development talent is limited and is not growing. The pool of JSP talent, on the other hand, is virtually unlimited. We can talk about PHP vs. JSP another day. Joe

[Guest.Visitor]

How to Get Out from Under an Obsolete Legacy Technology

When I started programming, twenty (!) years ago, I was told that RPG was a dead language. Not wanting to be Betamaxed into obscurity, I have kept an ear open to the pundits for the successor to RPG. Countless times I have gone to my managers seeking authority (if not funding) to pursue the language du jour, only to be told that such revolutionary technology is not supported by the application vendor and therefore not supported by the organization. Even internally developed projects were tempered by the dictum that they must be maintainable by whoever might be sitting in my chair in the future. What TAGrove illustrates so succinctly is that the majority of us on this platform were hired to maintain vendor code, not blaze new trails. What's more, we tend to work in very small shops (with small budgets), so we don't have the luxury of having someone else manage the payroll run for the next month so can go learn something cool. Learn on my own time? Great idea! I did just that. I went back to school and got my business degree in hopes that I could transition into a career that wouldn't make me dependent on any particular platform or language to survive. I enjoy all of the very knowledgable debates about the advantages of one language over another, and sincerely wish I had an opportunity to test them for myself. But, until I am ready to make a very significant decision, I am at the whim of the vendors chosen by my employer. Thanks, TAGrove, for bringing this perspective to light.

[Guest.Visitor]

How to Get Out from Under an Obsolete Legacy Technology

[Guest.Visitor]

How to Get Out from Under an Obsolete Legacy Technology

Sorry , pressed Enter too quickly. Since we are talking about scripting languages , specificallly PHP , what about IBM's "native" scripting language NetData? I have used PHP in a LAMP environmnet and found it easy to pickup (Disclaimer :I wasn't doing anything too mission crtiical with it. Mostly personal homepage stuff) AND I used CGIDEV2 a LOT a few years ago in another company.(Current comapny has a Do NOT Go There approach to web UI's) Now I am about to change companies again and they are open to enabling NEW apps to have a web front end , and I find myself trying to figure out : Do I go back to CGI ? Do I implement and learn WEbsphere ? Or Do I look at Netdata ? As I understand it NET Data is free and just needs some configuration setup in the Apache Server. And it can call ILE RPG programs ! p.s I dont know java. Any comments out there ? (The new environment is a FMCG company with a popular ERP package)

[R.Daugherty]

How to Get Out from Under an Obsolete Legacy Technology

Aren't all the possibilities you mention free? (PHP, CGIDEV2, Websphere as ships on iseries, and NetData (still supported?)) What shocked me was the cost of external users (web) who use iseries software, if I came anywhere close to saying that right. In other words, not anonymous, signed in as in customers signing in over the web, and not interactive users. IIRC, $4000 on the new entry systems. Given the anonymous stuff is trivial vanilla pages, you can't do anything useful without them signing in, that rules out anything but serious revenue producing web serving on the iseries, does it not? Or is everyone considered one iseries user and the signing in and control of what data is served to them just handled any way you can figure out? I'm assuming the Websphere validation and session stuff is what we're talking about for the $4000. Can anyone shed any light on that as well? rd

[J.Pluta]

How to Get Out from Under an Obsolete Legacy Technology

Larry Bolhuis said that external users are EITHER $250 per authenticated user or $4000 for an unlimited license. So you do have options if you have a smaller operation, although $250 is a pretty high price to pay for one person accessing your site. Joe