KICKS - The CICS compatible transaction monitor for VM/370 - M204
HTML-код
- Опубликовано: 17 окт 2024
- a CICS environment for VM/370 CE by Rene' Ferland
KICKS was developed by this amazing gentleman: www.kicksfortso...
Obtain KICKS here: www.kicksfortso...
Discord channel to talk to like-minded folks: / discord
This was great! I found all the PDF's for these books thank for mentioning them. I am finally understanding CICS and almost to writing my own cobol program using KICKS! Great fun.
beautiful! i already watched your previous video about kicks, and this is very useful
Thank you Moshix and Professor Ferland for another amazing video!
Our pleasure!
That was fun! I was a developer and later architect on The team that developed CICS for CMS, VM/PC (on the XT370) and CICS VM. We were a small team (~12 people) in a separate location from the main CICS laboratory - the idea being to prove that you didn’t need 200 developers and 100 testers to make a CICS release! We did succeed although none of these products got the same level of sales as the mainstream VSE and MVS products - partly I think because the sales incentives weren’t well thought through. Anyhow we had a fun time proving that a small team is better than a big team - no surprise since back in the famous days of the OS360 fiasco!
Wow. I am honored to meet you ! I have been looking for CIC/VM for years… and I have VM/PC running here at home. And yes, small teams far outperform large teams. There is no comparison
From the IBM field support side (NOT sales!), I think you are probably right about the incentives at the time. From our perspective, the priority seemed to be CICS on satellite systems - the PC and the AS/400 as feeders to the mainframe. I worked briefly for a customer to write an interface from AS/400 COBOL programs to pass a data area through my (COBOL) program to send the data through the standard CICS DPL to the customer's mainframe, and return the response data to the caller. This needed an interesting combination of AS/400 pipes and CICS functionality to support multiple concurrent users. That particular facility wasn't adopted for production, but it demonstrated the compatibility between the CICS systems on the different architectures. However, the use of PCs driving CICS for PC to talk to the mainframe was used extensively.
Thanks for the trip down memory lane. The last time I admin’d CICS was 1999. The version of CICS we used allowed dynamic updating of the tables. My first exposure to CICS was release 1.3, MANY years ago.
Great video.
Thanks for sharing!
I love the look and feel of your intros. Great work!
Thanks!
Thank you professor Ferland and Moshix. You are the main man. Great video. Appreciate it. Don't stop. Can't get enough of mainframes. 👍
Wow! CICS is almost exactly two months older than I am, myself!
same with me
Hi moshix, thanks for share, your videos are amazing. I have a question, i'm a beginner on mainframe and cobol, i want to know if we have a image of docker like z/os system with kicks(opensource of cics) and cobol installer and db2 too, if this is possible? like a environment corporate locally to test and study. Thanks for your help.
Maybe possible but not legal. You would have to ask IBM
@@moshixmainframechannel thanks, i will stay search some like this, i can run tk4 on docker, but i dont know if this will be similar with i had expose to you. thanks for your help and reply
New to your channel, can you run IBM VM on apple mac?
Yes
You might want to define earlier on - for non-mainframe people - what you mean by VM, as we have no context. generally VM means virtual machine, but which one?
In the mainframe world it means VM/370 or descendants. The installation of KICKS is really meant for people who are already familiar with mainframes.
@@moshixmainframechannel - Ok. I dont know mainframes, so I wont watch your videos any more.
Z/VM was the original virtualizing software. VMware, for X86, was architected from the mainframe Z/VM. We still use Z/VM for virtualizing on mainframes. Today, you can run hundreds of LINUXs on 1 mainframe using Z/VM.@@bobdobalina838
People may wonder why IBM didn't design a simple CALL interface for CICS for High Level Languages. I suspect that the very first implementation (for OS MFT and MVT) was for Assembler programs only. It was natural to build macros like " DFHFC TYPE=GET ,,, ,,," right in line in the code. Adding COBOL and PL/I was a different matter. The decision seems to have been made not to use a visible CALL interface. Instead, the programs included the Assembler macros in-line exactly as already implemented, and went through a pre-processor that punched out an Assembler REPRO statement in front of every COBOL or PL/I line of code, and passed the macros through unchanged. The output from this mix was passed to the Assembler, and this punched out the REPROed cards untouched, and converted the Macros to COBOL or PL/I code using a different MACLIB for each language, which did indeed call the internals of CICS. Why? All before my time, but it did mean that the programmer's manual only had to describe one format for each CICS function.
An independent software company then did write and market a CICS COBOL CALL Interface product, which meant that only the Compile and link Edit steps were needed. The Anti-Trust legislation at the time was extremely strict, and IBM couldn't do anything that appeared to "hit the little guy" by producing one itself. Command Level applied to all languages, so was not seen as an attack on this rival. It was a necessary step towards the functional extensions like EDF, and applied to the Assembler and PL/I users as well. The design work for CL was done in Hursley UK after the product was (as some rumoured) "shipped to Hursley to die". When customer demand for CICS meant that this clearly wasn't going to happen, some of the Hursley people who subsequently worked on CICS development came from the group who had written one of the PL/I compilers. The work of Bohme and Jacopini which criticised the use of GOTO because it "inevitably led to spaghetti code", rather threw out the baby with the bathwater. In fact some languages could only make decisions within their code by using a branch instruction to implement IF / ELSE / ENDIF, and CASE structures. Assembler was the most obvious example, with COBOL close behind. The PL/I language did much to reduce the need to use GOTO, and the CICS HANDLE CONDITION was a reflection of this.
(Many people wrote "Assembler Structure Programming macros". I did so myself as an exercise, and found them useful. Decades later, working for an IBM customer at the Y2K time, I found no less than three sets of such macros all in use in the same installation in different applicationareas.)
Wow! So much history explained. Thanks !
Thanks to you! I haven't had the courage (yet) to try to run what you do on one of my laptops, but if I'd been told in 1995 that I would be able to run MVS or VM on a laptop ... ... ... I'd have asked what my informant had been drinking!
I don't want to hog your excellent talks. Is there any way we could get in touch offline? I'm open to discuss ancient history of CICS - that is usuallly the reason for some of the apparently strange development decisions.
I was born in the same year with CICS birth. First time used CICS in 1993
subscribed and liked. Thank you!
The cross platform versions of CICS only supported the command level API. Macro level applications were only supported on MVS and DOS/VSE I believe.
Yes. But there Cobol compilers on every platform wheee cics ran.
@@moshixmainframechannel ,
Yes, but "macro level" CICS applications (available to ASM, COBOL or PL/1 applications) were only supported on S/360/370 platforms (MVS and DOS).
In the 1980s, IBM ended support for CICS macro level. All CICS applications calling CICS services via macros had to change to EXEC CICS commands (command level programs) - replacing every CICS macro call with the equivalent EXEC CICS statement, re-compiling and testing. Today CICS macro level programs will not run on CICS - at least not without some third party emulation software.
Yes. Of course. But command command interface also existed on many other platforms. AIX, OS/2, MS-DOS, and now Linux
I taught CICS Macro Level for IBM UK from 1972 to mid 1974, and then moved to the UK central support group. (Up to that time CICS ran on pre-Virtual Storage 360 machines, and was compatible at the application code level with the early CICS/VS versions on the 370.) I remember attending the pre-announcement workshop that included Command Level (?1975?). This led to considerable new facilities (e,g, EDF), but with hindsight had one major problem. ML issued CICS commands and then tested for return codes; CL was heavily influenced by the PL/I language approach of setting the points before a command was issued, to branch to code in the program that would handle errors or exceptions if and when they arose. In other words, the HANDLE CONDITION command was modelled on the P/I "ON" construct. There was a laudable desire to achieve GOTO-free code, especially in COBOL, which was the language of choice for most installations, but this just hid the transfer of control under a different name. This was aggravated by the lack of information in the control block which got the actual return code details (the EIB), which made it hard to know what you were dealing with when you got there. This was not addressed until a later release. Whatever the "purity of design" advantages, the practical effect was to make it impossible to convert a program from ML to CL without a total rewrite, until enough information was exposed in the EIB to allow the programmer to dispense with the HANDLE command and do a one-for-one conversion, By the time that happened, a huge amount of application code continued to be written in Macro Level, to keep the house style for on-line applications, and to save on re-training programmers who were very familiar with CICS already. It was particularly irritating for customers who used DL/I, which was the strategic IBM data base software at the time, which was committed to the "test after" approach. (In the DOS arena, DL/I also offered an EXEC DLI interface later, but this was rejected wholesale because it didn't feed back all the information required; also it was terribly clear that the OS version, which was part of IMS, was never going to go that way.) I worked with a team in 1978 whose mission was to provide a gentle way into CICS with DL/I by offering a subset which was for VDUs only, and DOS/VS only, with DL/I, and a completely new set of documentation. The result was manuals which were about 30% the size of the originals, and covered only what so many prospective customers actually wanted. We went for EXEC CICS, which had matured sufficiently to offer testing after the command, but stayed with the DL/I Call interface. I was back in education for 1983 and 1984, in time to reap the whirlwind when Macro Level support was finally withdrawn, and we had to teach a specially built course to help people handle that. Many customers said on that course that more than 50% of their ML code was actually written after Command Level had been delivered, and usually gave the reasons of "house style" and "existing programmer expertise" as the reasons.
I may be wrong, but i think that all the versions of CICS for other platforms were announced some time after CL had been firmly established as the way forward, and it would have been daft to create a ML version on systems which didn't have an Assembler to interpret the macros - though VM/370 most certainly did!
@johnwade1325 very educational feedback! Thank you very much. The CICS arena is a world unto itself.
Hi Moshix , Any KICKS tutorial series on MVS 3.8 - windows 10
Yes it’s already there
CICS is like Apache Server for the 3270 Terminal. Apache is for HTTP/HTML Web Browsers.
Ok
Very strange seeing CICS in VM. I have always programed and worked with CICS in mvs/zos. We use ZVM only for Linux now.
CICS in VM - as opposed to CICS running in a VM environment on a guest MVS or VSE system, never really took off. Have a look at the CICS article in Wikipedia for a bit more on two attempts in the 1980s that didn't last long. By that time, all MVS installations and most VSE installations probably had enough machine resource to have at least one "real" CICS region for testing that could be completely segregated from production. And then would either have TSO for MVS or VM CMS with their VSE system, or a mini interactive development system under VSE itself.
@johnwade1325 all true
Its not working for me
Try again
Im a total newbie when it comes to mainframes, I got kicks working but how do I access and edit the datasets of kicks on vm/370
Join our discord. This is not a good medium for support
@@moshixmainframechannel what is the discord called
Check description below newest video. It has a link to discord there
Shame common Western European accents do not seem to be available for video overlays. Poor René.
Well, I learned to live with Rene or Rene' instead of René :-)
We Can Make a CICS VM/370
Ok