Re: Adoption of RPG OA (Summary) -- RPG400-L

Mihael

I suspect we are agreeing violently - that was what someone said when 2 people were saying similar things but did not know it!!

One of the issues raised by you and Joe is over using opcodes as you expect them to. I think this is a matter of style - there is the sense of purity, too, but IBM already gave us SPECIAL files, where a READ does not have to be done against only a PF or LF. I don't find that to be a problem, actually, I see it as an opportunity for more flexibility and reach to sources of data that are less traditional, which the present world not puts before us.

So I guess if you don't want to use "standard" opcodes in somewhat "unusual" ways, I can see that. Again, for me, the use of CHAIN to retrieve the temperature in a city from a web service is so simple to code, and it is still asking for information from a data source, based on a "key" value. Our definitions will vary here, I think, and we will find a way to live together in this - I've not exactly helped in that way, perhaps!

I find the pseudo-code as you presented it to be more complex. It is something I am used to, for it is very much like SQL CLI, where you declare handles - not handlers, right? - for connections and statements, etc. That's fine, and for many, that is the way to go. For others, they'd like a solution with fewer lines of code and no nested function calls.

So if one extends the definition of a CHAIN opcode, then OA is not a problem. And one DOES need to allow a broader, more inclusive definition for a CHAIN, now. Or perhaps one broadens the definition of a "file". I see a similar thing in file descriptors for IFS files - the same pool of descriptors is also used for sockets. Data is available from both.

You can see that I like to find precedence for something new. Perhaps I operate in a more general way, while you like to work with more specific things.

As to learning new things - this is a good thing. I was just talking to another developer here today. He mentioned a fellow who was very intelligent and would have liked very much to study the newest tools and techniques. He also had to get a job done. For him, it was a choice - take some time that he did not have, or use tools he knows, and be sure to satisfy a customer or client. I have the luxury of working where I am supposed to look into new ways of doing things. I also don't have a family that needs much of my attention. So my free time goes to getting a book on Visual Basic in 21 Days or the like. Not everyone has that luxury.

The man I was talking to - he has a list of things he wants to learn - one is HTML. There are many other topics on that list. He has not been able to get to them all yet. So he uses what he knows for now, and will get to other things as he can.

This is why I just can't criticize these developers too much, and I guess I try to defend them. Should they learn new things? Yes, of course. Do they want to? Many do, some don't. Can they? Not always, and for many reasons.

I wish life were as simple as I'd like to make it at times. It's easier, for example, to think that you and I can't agree on things here. That is just not the case, and I look forward to more conversations, maybe with some heat, but helpful in the end. We both feel strongly about our positions, maybe that is as it should be for the good of us all.

Regards
Vern

On 3/21/2011 8:43 AM, Schmidt, Mihael wrote:

Perhaps you have misunderstood me. The concept of OA is not bad. Not at all. I really like the concept of handlers and all that but the one thing I don't like is that existing opcodes has been used (abused was probably a bit hard =) for this. I would like to have it clear what the code does and I don't expect a chain to make something different than to go to a db and try to retrieve one record.

If the code would be something like:

<pseudo-code>
handler = vendor_handler_create();

data = ibm_io_proxy_read(handler);
</pseudo-code>

Then all would be clear.

... and yes ... RPG III devs can't used it as (IIRC) only ILE (service)programs can use it. That was a little emotional statement, I know.

... and to make it clear: I don't like overrides.

BUT ... IMO IBM has given those stated developers a new excuse for not learning something new (which they probably should have learned since V5R1). I can't understand people who don't WANT to learn new good things. It seems to be outside of my mental reach.

... and I definitely didn't want to insult anyone ... perhaps my english is not good enough for this discussion and I should take a language course =)

Mihael

-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Vern Hamberg
Sent: Monday, March 21, 2011 1:50 PM
To: RPG programming on the IBM i / System i
Subject: Re: Adoption of RPG OA (Summary)

Mihael, Mihael

You continue to misrepresent or misunderstand things regarding OA - this
one statement must be challenged -

"that would have missed their point/goal in providing something
that RPG III developers can use without learning something new"

That is your severely biased interpretation of the original requirement.
And to be very specific, OA is NOT usable at all in RPG III

As to abuse of the CHAIN opcode - let me see - once upon a time there
was a READ opcode - hmmm - but if you called the file a SPECIAL type,
instead of a DISK type, hey you could READ IFS files. Oh really? Did IBM
abuse the READ opcode in that scenario? I don't think so. You might
think so. But truth is, it has been a part of the language for decades.

Oh, let me see again - is there anything else in the system where you
redirect what is being processed? Hmm, overrides - hey, I can override a
PF with a printer file and get a simple report quite easily. Oh, but I
would be doing something a PF is not designed for. Ah, but that is the
whole point.

I ask you to do the work you do - it is good work. Stop implicitly
insulting developers who may not have your knowledge. Promote what you
do without tearing down alternatives - if your product is good, it will
be enough to say what it does, not what others do NOT do!

At any rate, I expect that if I've said anything with validity here, it
is for others in this list to consider.

Vern

On 3/21/2011 2:00 AM, Schmidt, Mihael wrote:

But then again that would have missed their point/goal in providing something that RPG III developers can use without learning something new.