RE: Why no header files for RPG?

Jon,
 
<snip>
What a truly apathetic bunch we must be! Hundreds of people are wiling to
bitch on forums like this but frankly why the hell should IBM bother if only
29 people give a .....
</snip>
 
Ouch!!! Is that smoke coming out of your nostrils? Are those flames? :-)
 
<clarification>
 
1) IBM should give a... because they're the vendor - It's their job! If I,
and all of my colleagues, can't do our jobs as effectively as we could
because the vendor that supplies the hardware AND software can't be bothered
to write supporting documentation then we can't blame companies like our own
for looking over the fence at the unix/oracle/java/C++/Win2K apps and
thinking the grass is greener. I'm sorry, but writing the code is only half
of the job - there MUST be sufficiently thorough supporting documentation.
We're running multi-billion dollar businesses on this hardware/software. We
can't afford to take chances.
 
2) Regarding the Redbook review process:
a) I didn't even see the review page. Is that MY fault? Am I appathetic
because I didn't get to page 453 out of 454 in "the sorcerers guide"???
b) The book was written in February 2000, I read it last year. Even had I
seen the review form I would never have bothered reviewing a 4-year old
programming document. It may be a "modern classic", but that would be like
reviewing the latest Beatles album.
c) If I thought reviewing 4-year old Redbooks would make a difference then I
would. Frankly, I am currently not holding my breath.   
 
3) My initial post was about the fact that forming an "Integrated Language
Environment" does not stop at the code. The API docs need expanding.
Something as simple as putting the RPG prototype definition alongside the c
function signature would be a start.
 
4) To say that the prototypes are "on the web" just gets IBM off the hook. I
would rather get a prototype for a IBM API from IBM. I should not need to
look elsewhere. Granted, I could look on the web for usage examples, but the
prototype definition for the API should be a minimum requirement from the
vendor. It isn't just the prototypes though: I started using the stream file
APIs to create XML files in the IFS. I had to create XML files in UTF-8
format. The IBM API doc for open() specifies that I can use the O_CCSID
flag, not O_CODEPAGE. This is what I need, it allows me to specify a CCSID
of 1208 as the fourth argument during file creation. As a RPG programmer I
wouldn't know that the value of O_CCSID is 32. C programmers don't "need" to
know because it is defined in the header file and they can use the O_CCSID
flag without needing to know it's numeric value. I had to search the C
header files manually for the value - I couldn't find a single example on
the web.
 
5) On this forum, and many others, the "experts" always suggest that you put
service program prototypes into copysource. This ensures that the "client"
of the software always uses the correctly defined prototype. This is good
programming practice, it's also common sense. Why should IBM be any
different? The moment IBM changed the environment to allow a RPG program to
call certain APIs directly then the correct RPG prototype should have been
placed in a copysource file. I can't see any reason why they shouldn't.
 
6) I spend most of my time convincing management that we can do ANYTHING on
the iSeries. I truly love working on it, I'm passionate about it. I do not
spend my time bitching about IBM. Sadly, I spend more time defending them. I
just want them to meet me half way. I'm committed to IBM, I'm not so sure
that they're committed to me. 
  
</clarification>
 
You're right. Next time I think I'll just write to IBM.
 
Thanks for your time.
 
Cheers
 
Larry