Re: the ASSET issue again

Hi ... (sorry didn't see your name in the latest reply),

I usually don't like to enter discussions like this, but your logic used to
justify why you dislike AS/SET requires it.

Saying that 'Who cares wht the Action diagram looks like, it can be perfect and
AS/SET will still generate bloated and unreadble RPG' is, as per my opinion,
very comparable to the following logic:

>>>I don't care what the RPG (or C, or Visual Basic, or any other high level
language) looks like, it can be perfect, but it still generates bloated and
unreadable MI (Machine Instructions, i.e. something like assembler for AS/400)
code.<<<

Yes, if you have dealt only with assembler programming language in the past (but
you've seen more than enough of RPG (or C, or COBOL, or VB, or...) code to feel
confident to comment on it...) and you get code in assembler, with comments
inserted so that for each group of assembler instructions you also can see the
original RPG (or C, or COBOL, or VB...) instruction this group of MI
anstructions was generated from (just because you chose such generation/compile
option), of course that you won't like 'annoying' comments and in addition all
that extranious and non-efficient MI/assembler code generated from RPG (or
C,...) code, which an assembler guru would write much more efficient directly in
assembler. For the same reason, I can understand you wouldn't like to trade
analyzing such 'bloated and unreadable' assembler code with analysis of RPG (or
C,...) code directly, but try to explain that to programmers who have started
programming with assembler and have accepted to learn RPG (or C,...) to the
level that they do not need to look into assembler code anymore when analyzing
the code! Let alone those programmers who have never had the honour of learning
assembler....

I am sure that Genyphyr was not trying to say that ADK and AS/SET are perfect
and that RPG code generated from AS/SET is ideal and optimal and does not have
space for any improvement (neither was I), but only that if you haven't tried to
work with AS/SET (but seriously, in the meaning to learn AS/SET language and its
features and advantages, first, then gain some experience with it and then
judge), you cannot really expect that you are objective in comparing it with
RPG, or whatever else. Somebody on the list said that they would like to hear
more opinions from those who used both languages in question here and we
replied, honestly as programmers - not customer relationship representatives,
which still does not mean that we are right, but if I were a user asking a
question about efficiency of coding in AS/SET or RPG, I too would really like to
hear opinion of those who really used both languages.

And if AS/SET is such a poor, flawed tool, how comes that it had generated such
a reasonably well (if not more) performing version of BPCS? If I still had seen
here a lot of complains about the performance with V6.1.01, I would have thought
you had a case, but if with all that useless and not efficient code, latest
version of BPCS looks pretty good, then I would tend to vote for the other
candidate :-)
(please, again,  don't take this as my claim that there are no empty subroutines
or other things left in this version, but we can talk again after next BPCS
release and yes, it WILL be developed in AS/SET).

Regards,
Goran





"L. S. Russell" <leslier@datrek.com> on 09/19/2000 02:19:54 PM

Please respond to BPCS-L@midrange.com

To:   BPCS-L@midrange.com
cc:    (bcc: Goran Mihajlovic/SSA/US)

Subject:  Re: the ASSET issue again



I am basing my opinons on reading the source code on our system, would
you like some samples?
You seem to think AS/SET is god's gift to software design, it really is
just a flawed tool that generates what can only be described as a
pitiful excuse for RPG.

Who cares wht the Action diagram looks like, it can be perfect and
AS/SET will still generate bloated and unreadble RPG.

Genyphyr Novak wrote:
>
> The "No I don't use AS/SET" comment says it all for me. Not sure what you
> are basing your opinions on here, if you are just looking at the RPG source,
> rather than what it is generated from (ie, the Action Diagram).
>
> The generator does not generate empty subroutines from a well-written
> written action diagram. This is the fault of someone commenting out only a
> portion of the action diagram, probably in the process of fixing a BMR, and
> not getting references to the subroutine properly removed in the rest of the
> code.
>
> By the way, I never said there were not empty subroutines in the RPG source.
> My point is that the reason for them existing has nothing to do with some
> flaw in the AS/SET tool itself.
>
> Thanks
> Genyphyr Novak
> SSA
>
> -----Original Message-----
> From: L. S. Russell <leslier@datrek.com>
> To: BPCS-L@midrange.com <BPCS-L@midrange.com>
> Date: Tuesday, September 19, 2000 8:33 AM
> Subject: Re: the ASSET issue again
>
> >No, I don't use AS/SET.  But I have looked at more than enough AS/SET
> >code to know that, yes Virginia there are empty sub routines.
> >
> >Genyphyr Novak wrote:
> >>
> >> Hello,
> >>
> >> Subroutines getting executed with no code to run inside them is not from
> the
> >> tool itself generating this, but is due to the person who did the
> coding -
> >> most likely some BMR had a subroutine commented out, but not all
> references
> >> to it were removed elsewhere, so it is called and does nothing. Sorry to
> >> say, this is not the fault of the tool. Are you are using AS/SET? It
> sounds
> >> like you have never looked at the action diagrams to check to see what
> they
> >> look like.
> >>
> >> Thanks
> >>
> >> Genyphyr Novak
> >> SSA
> >>
> >> -----Original Message-----
> >> From: L. S. Russell <leslier@datrek.com>
> >> To: BPCS-L@midrange.com <BPCS-L@midrange.com>
> >> Date: Monday, September 18, 2000 2:54 PM
> >> Subject: Re: the ASSET issue again
> >>
> >> >A loop is a loop, but a loop executed 15 times unnecessarily in effect
> >> >slows down a program.
> >> >Things like statementless loops, and sub routines (with nothing but
> >> >BEGSR and ENDSR) are common in the AS/SET generated code we work in
> >> >here.
> >> >
> >> >Betty, back me up :) lol
> >> >
> >> >helen@dekko.com wrote:
> >> >>
> >> >> I do not agree that the ASSET generated programs take longer to
> execute.
> >> >> The development time has been reduced significantly using this tool
> and I
> >> >> believe that it is not all that difficult to debug using STRISDB.
> It's
> >> >> simply a matter of becoming familiar with the naming conventions and
> >> >> getting accustomed to reading ASSET source.
>



+---
| This is the BPCS Users Mailing List!
| To submit a new message, send your mail to BPCS-L@midrange.com.
| To subscribe to this list send email to BPCS-L-SUB@midrange.com.
| To unsubscribe from this list send email to BPCS-L-UNSUB@midrange.com.
| Questions should be directed to the list owner: dasmussen@aol.com
+---