RPG improvements

[Guest.Visitor]

RPG improvements

Ok, then I'll have to politely disagree since the programming process that I use seldom include RPG. chuck Opinions expressed are not necessarily those of my employer. "David Abramowitz" wrote in message news:6aec9022.28@WebX.WawyahGHajS... > Chuck Ackerman asked: To whom are you referring when you say "existing programming process."? > > I wasn't referring to a "whom" at all, I was referring to a "what". > > Dave

[J.Pluta]

RPG improvements

From a technical perspective, I think it would be difficult to make a stateful RPG program interact well with the stateless HTTP protocol. I beg to differ. It's quite simple, especially since HTTP stopped being solely stateless with the introduction of persistent connections in HTTP 1.1, circa 1996. Tie an HTTP session to a submitted job, connect with data queues, and you have a stateful connection with a browser UI and an RPG back end. And it's not just good for browsers, either. You can also use the same approach to connect to a thick client, if that's the UI you need. I give labs on this every year at iSeries DevCon. It isn't as simple as switching out 5250 for HTML, since 5250 is designed for synchronous dumb terminals and HTML is designed for asynchronous browsers. Once again, I disagree. It is exactly that simple, and I laid the process out in detail in my e-Deployment book nearly four years ago. Not only is it relatively easy, the process can be automated, as PSC/400 does. "Easy" is a relative term - it depends on how many DDS keywords you need to support - but the general architecture is available today for the price of my book. It doesn't turn your order entry application into Amazon.com. You get a stateful connection to an RPG program, which means you won't have a meaningful back button, for example, but that's not what we're talking about. We're talking about an improved interface for OLTP applications. Frankly, it wouldn't take more than a few DDS keywords to make the transition completely transparent to the programmer. Joe Pluta www.plutabrothers.com Online PSC/400 demos

[buck.calabro@commsoft.net]

RPG improvements

Hi Joe! >> From a technical perspective, I think it would >> be difficult to make a stateful RPG program >> interact well with the stateless HTTP protocol. > > I beg to differ. It's quite simple, especially since HTTP > stopped being solely stateless with the introduction of > persistent connections in HTTP 1.1, circa 1996. > Tie an HTTP session to a submitted job, connect with data > queues, and you have a stateful connection with a browser > UI and an RPG back end. How easy it is for me to forget persistent connections! In my own defense, when I made my reply, I was thinking along the lines of some new DDS keywords to add HTML capability, giving us the traditional externally described display file (UI) and the RPG program that uses it. I think this is what Dave was thinking about. Under the covers, IBM could certainly connect HTTP and RPG (see WebFacing), so I suppose I should reconsider my statement, hm? > It isn't as simple as switching out 5250 for HTML, > since 5250 is designed for synchronous dumb > terminals and HTML is designed for > asynchronous browsers. > > Once again, I disagree. It is exactly that simple, and I > laid the process out in detail in my e-Deployment > book nearly four years ago. A copy of which I own, have read and used. Nice work, by the way. It is very possible to exchange EXFMT for an API call, like you do, and it is very possible to automate that. But 5250 does some things that HTML cannot, and HTML does some things that 5250 cannot, and there's no way around that. We can emulate workstation controller data validation via Javascript and so on, but at the bottom-most layer, our hypothetical DDS which allows an HTML UI will look and behave differently on a 5251 model 11 and a browser. Thinking some more about the actual 5250 datastream, it seems to me that part of the issue is the missing formatting. That is, when the actual 5250 sends back a screenload of data, the fields are enclosed in formatting tokens. Outbound has the same problem, but HTML can emulate most of that. The inbound issue has problems emulating modified data tags, doesn't it? I haven't got a PCS/400 install here, but how does it handle READC? I'm certainly willing to have my eyes opened... Do note that I'm sticking closely to the 5250 issues, and not straying onto the fact that a web designer would never do things the way a green screener would and vice versa. I'm very much aware of the differences and perhaps that coloured my earlier responses. So on the face of it, it doesn't look too technically daunting to add HTML-aware DDS keywords, but I still feel that the nitty-gritty details are more significant than trivial (integrate a self-configuring HTML server in place of a workstation controller?) On the other hand, there are plenty of smart people working at IBM, and what I think is hard may be trivial to them, eh? In any event, I tried to end on a positive note by giving the link to the DCR, so that folks could tell 'official' IBM that this is something that they want to have. Respectfully, --buck

[David Abramowitz]

RPG improvements

Starbuck5250 wrote: I think this is what Dave was thinking about. Nope..... I don't equate web pages with DDS. OTOH I do believe that web pages lend themselves to transaction processing (as opposed to event driven programming). And, if that is the case, then you should be able to read and write a web page as you would any other file. In point of fact, SYSIN and SYSOUT are files! My point of contention is that the processing of SYSIN and SYSOUT could be made easier. Dave

[buck.calabro@commsoft.net]

RPG improvements

Dave wrote: > I don't equate web pages with DDS. OTOH I do believe > that web pages lend themselves to transaction processing > (as opposed to event driven programming). And, if that is > the case, then you should be able to read and write a web > page as you would any other file. In point of fact, SYSIN > and SYSOUT are files! My point of contention is that the > processing of SYSIN and SYSOUT could be made easier. Sorry for putting words into your mouth. I have great regard for your posts. I guess I carried it one step too far. Are you looking to have STDIN and STDOUT act like any other external file, defined with DDS, or just the ability to directly write and streams without having to use open() read() write() and friends? I guess I thought you wanted some formatting to happen under the covers, as it does today with 5250. --buck

[J.Pluta]

RPG improvements

That is, when the actual 5250 sends back a screenload of data, the fields are enclosed in formatting tokens. Outbound has the same problem, but HTML can emulate most of that. The inbound issue has problems emulating modified data tags, doesn't it? I haven't got a PCS/400 install here, but how does it handle READC? I'm certainly willing to have my eyes opened... The goal is not to emulate the 5250 data stream, but to get data in the appropriate fields. Except for some VERY specific hardware issues (for example, not being able to map your keyboard however you want), a browser/servlet approach can emulate just about everything using a combination of HTML and JavaScript. MDT? It's a little trickier. I only set on MDT if the new value of the field is different than the old. That means you will not get a hit on READC (or the CHANGE indicator) if you blank out an already blank field. But other than that, PSC/400 (note the name, PSC, not PCS [grin]) handles READC just fine. As it does command keys. And auto-tab from field to field, just like a real 5250 controller. Automatic conversion of lower-case to upper-case. Numeric editing. Even INFDS fields, such as cursor positioning, can be emulated to a reasonable degree. In fact, there's really very little a 5250 can do that JavaScript (and thus PSC/400) can't do. Some features aren't implemented in PSC/400 because they're so rarely used, but that doesn't mean they couldn't be. Joe

[David Abramowitz]

RPG improvements

Are you looking to have STDIN and STDOUT act like any other external file, More, perhaps as an internally defined file. Believe it or not these two were a fundamental part of the original RPG II language. Using Sysin and Sysout was the normal method of processing on the System/32, and while not needed on the System/34 it was still part of RPG II. Support was dropped when the System/36 was introduced, and TTBOMK it was never a part of RPG III. COBOL programmers can use Sysin and Sysout with the Display and Accept verbs. Even in ILE COBOL this is possible. There are possibilities here. Dave

[TBarstow]

RPG improvements

I agree with Chuck. True, 5250 won't sell boxes, but neither will IBM's advertising. I've had people ask me "Who does IBM's commercials - Microsoft?" I have to wonder .....

[K.Forsythe]

RPG improvements

"I know - but what computers does microsoft run their business on?" Guess what - last I heard its iSeries - they (Microsoft) would be happy to sell you PC solutions - but they use iSeries - because that is what realy works. But feel free to buy into their marketing machine and spend yourself into oblivion.

[Guest.Visitor]

RPG improvements

I think your information is many years old. Microsoft did use AS/400s at one time but the last one, in a distribution center in Ireland, was retired years ago as I recall. chuck Opinions expressed are not necessarily those of my employer. "kforsythe" wrote in message news:6aec9022.38@WebX.WawyahGHajS... > "I know - but what computers does microsoft run their business on?" > > Guess what - last I heard its iSeries - they (Microsoft) would be happy to sell you PC solutions - but they use iSeries - because that is what realy works. But feel free to buy into their marketing machine and spend yourself into oblivion.

[K.Forsythe]

RPG improvements

As I recall, Frank Soltis made that statement at a COMMON conference post Y2K - not sure which one. I am not at all certain about this as it is heresay and Frank Soltis has some (OK lots) of motivation to stretch the truth. But i have no idea if he did. MS also has motive to stretch the truth on this issue. If anyone actually knows for sure about this - I would love to hear it.

[J.Pluta]

RPG improvements

There is no published information to back up either statement. The last reported story was in the 90's when Microsoft attempted to move their business processing to NT servers and failed miserably. The most reliable sources indicate that at this time Microsoft still outsources a significant part of its business operations to an iSeries provider, but neither IBM nor Microsoft confirm or deny those reports. Joe

[K.Forsythe]

RPG improvements

If Microsoft isn't denying those allegations then wouldn't that imply that it is true? They work through a third party so as to be able to say "we" don't use iSeries - "we" just farm the work out to someone that does. IBM has to watch what it says - probably to avoid violating a non disclosure agreement. What would keep Microsoft from denying the claim - except the fact that its true? Is it that hard to see Bill Gates deciding to continue running MS business apps on a non windows platform? Not that long ago they had an Anti-unix web site hosted on.... a UNIX platform! http://www.midrangeserver.com/mid/mi...2-story03.html Is it so hard to see MS using one thing and selling another? Its all about the marketing!

[David Abramowitz]

RPG improvements

Joe Pluta wrote: The most reliable sources indicate that at this time Microsoft still outsources a significant part of its business operations to an iSeries provider AFAIK, this is correct. On the MS web site a few years back, MS had pictures of their servers, and lo and behold the published pictures were of AS400s. Just a few months later an MS spokesperson claimed that MS no longer ran anything on AS400s. That statement may be true, but all other sources indicate that MS farmed out its basic financial functions to a firm using AS400s and Iseries. This is political doublespeak. MS can claim not to own or run any AS400s, but the hardware that serves up the info that MS runs on, may very well be an AS400. You can't blame MS for using the AS400. There are some very smart people in Redmond. Dave

[Guest.Visitor]

RPG improvements

I've caught Soltis in a few errors over the years so I wouldn't be surprised if his info on this is out of date. chuck Opinions expressed are not necessarily those of my employer. "kforsythe" wrote in message news:6aec9022.40@WebX.WawyahGHajS... > As I recall, Frank Soltis made that statement at a COMMON conference post Y2K - not sure which one. > > I am not at all certain about this as it is heresay and Frank Soltis has some (OK lots) of motivation to stretch the truth. But i have no idea if he did. MS also has motive to stretch the truth on this issue. If anyone actually knows for sure about this - I would love to hear it.