Re: Product Enhancement by Voting?

Richard wrote:
>It is my opinion that language features should be
>selected by a small group of people thinking carefully
>about the problem, not by a large group of people
>thinking carelessly about it.

That's more or less how things work today.  We have a small
group of RPG compiler writers and language planners deciding
on release content guided by customer input.  Our semi-
regular enhancement poll is used to provide input to our
decisions, but we certainly don't consider the poll results
to be binding.

>
>I assert that there should be two categories of new
>features: features needed today by the general programming
>population and stuff needed to take the product into the
future.  [Or, in the case of Don's multidimensional arrays,
>into the brave new world of the 1950's :)]  Both the present
>and future are critical.  It is perilous to ignore either
>one.

I've sometimes said that RPG IV is our attempt to drag
RPG kicking and screaming into the 1970's!

We tend to classify possible enhancements into two groups:
Those needed to support other O/S changes, and others to
grow the language.

I would argue that your category "needed today by the
general programming population" is meaningless since we
cannot provide instant gratification.  For example, V4R4
has been available for about a year now, but many programmers
still have to support V3R2 or V3R7 systems.  For many, the
latest RPG enhancements can't be used until all of their
supported systems are at the current release level.  So, in a
sense, ALL RPG enhancements are intended to take the language
into the future!

>
>We cannot leave the decisions to the language writers because,
>in almost every case, they have a lot of experience creating a
>compiler but very little experience using the target language
>to solve business problems. This isn't an insult to their
>intelligence or yours.  Those are different skill sets.

To some extent, I agree with you.  Compiler-writing skills
are indeed different from application-development skills.

But an important part of our job is to try to understand
the needs of the people who use our product.  And so we
talk with our customers.  We participate in newsgroups
and mailing lists (like this one).  We visit user groups
and conferences.  We sometimes invite customers in to our
offices for more detailed discussions.  And sometimes we
invite people to rate possible enhancements via a poll.

For example, two enhancements we added in V4R4 can be
directly traced back to a discussion I had some years
back with a well-known AS/400 personality.  He argued
the need for two things:  A *NEXT option on the OVERLAY
keyword, and the ability to implicitly define the
overlayed field based on the total lengths of the
overlaying subfields.  It took a while to figure out how
to do it (particularly the latter), but we finally did
and it made it into V4R4.

As another example, if you were here about a year ago,
you would have seen how discussion on this mailing list
helped greatly to influence the design of one of the V5R1
enhancements.  In addition, other discussions here have
led directly to other small enhancements.

Also consider that cost of development is an important
factor in choosing release content.  Only a compiler
developer can decide that.

Cheers!  Hans

Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com


+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---