SCOUG-Programming Mailing List Archives

<< Previous Message <<

Date:	Sat, 3 Jan 2004 23:50:26 PST8
From:	"Lynn H. Maxson" <lmaxson@pacbell.net >
Reply-To:	scoug-programming@scoug.com
To:	< "scoug-programming@scoug.com" > scoug-programming@scoug.com >
Subject:	SCOUG-Programming: QA equals testing, Part One:Detection

Content Type: text/plain

Gregory Smith writes:
"Peter's comments about organization still hold. Assume that
IBM gave the complete OS/2 source code to the open source
community. My opinion is that the existing source code would
be easier to grok than what you have described so far. As I
understand what you have described, your equivalent for the
OS/2 system would be something like: ...

...OK, so the system can be made to produce something
understandable
to the uninitiated. But I still do not grok how this unordered
way of doing things helps me as a software "author." ..."

Greg,

I wish that you had written your code example using PL/I
syntax which SL/I uses. At this point I have considerable
difficulty with deciding whether you don't believe SQL works
or you don't understand how it works. In an SQL query with
the general three-clause form of "SELECT column(s) FROM
table(s) WHERE condition(s);", only in the SELECT clause does
the order of the columns matter and there only to the author.
The tables in the FROM clause and the conditions in the
WHERE clause can appear in any order.

Now I challenge you to find the same freedom in any file
management system or the same writing or rewriting ease.
You certainly won't find it in any imperative language like C,
PL/I, or COBOL in the code supporting embedded SQL queries.
The writing ease you find in a relational database manager
which relies on logic programming you do not find in either
network or hierarchical database managers.

The fact that you don't have to impose a
physical-equals-logical order on the source code means you
can write it as it occurs to you in stream of consciousness
mode. As long as it contains all the necessary specifications
the software will construct the necessary and optimal
ordering. That means that changes can occur and get entered
in any order without the software grumbling or worrying
about the cost of tossing out the previous organization, which
in terms of cost and time approaches zero in the limit. It
simply again at near zero cost and time generates a new
organization regardless of the extent or logical shifts brought
about by the changes.

Now you may enjoy excess coding and recoding. If you are
getting paid to do it, then it's no big deal. If you are paying to
have it done and faced with a 200 to 1 cost difference, it gets
to be a big deal. If you are doing it for yourself and faced
with a 50 to 1 time difference, then I would think it a big deal
also.

It would seem that that you resent the fact that logic
programming works, that SQL works, that Prolog works, that
AI works, that they go directly from specs to executable with
unordered input...except for the order of the SELECT, FROM,
and WHERE clauses in SQL. They all use an HLL. They all
simultaneously function as specification and programming
languages.

They only require a single instance of reusuable code in input
regardless of the use instances occurring in the logical
organization. That not only reduces the amount of writing
required, but it reduces the number of instances of rewriting
required. It absolutely eliminates any manual ordering or
reordering of the source.

You still get ordered source output, i.e. logically organized
source. You don't get the need to manually reorder it based
on changes. You don't need to manually order it in the first
place.

Now I suggest that you take a look at how you spend your
time in coding, the time repeating logic segments, the time
ordering and reordering your source, the time making multiple
changes to a single logic segment, the time thinking about the
effect a change will have on the logical organization, and so
on.

In truth you get everything you now get. You just get it in
less time and effort.

As to IBM giving access to OS/2 source code through open
source you have to ask yourself if bureaucratic excess lay at
the cause of allowing 18 months for a fixpak release and to
release one every six months meant having three full teams
working fulltime concurrently. Did IBM have excess people
resources because it engaged in some form of social welfare?
Or did somehow the number arise from the time constraints on
release dates?

You can grok to your hearts content with that source code,
Unless you also have that excess in skilled people resources
you can forget about meaningful, i.e. timely, progress in terms
of enhancements or improvements. If you have them fine, you
can waste them to your heart's content. Personally I would
take the time to translate that set of specifications into a
different set where, one, I don't need that number of people,
and, two, I don't suffer the time constraints. Along the way I
eliminate essentially 90% of the current time spent in testing.
I just need 10% of that time to validate the automated testing.

You're welcome to grok. I prefer to glide.

=====================================================

To unsubscribe from this list, send an email message
to "steward@scoug.com". In the body of the message,
put the command "unsubscribe scoug-programming".

For problems, contact the list owner at
"rollin@scoug.com".

=====================================================

Return to [ 03 | January | 2004 ]

SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group. OS/2, Workplace Shell, and IBM are registered trademarks of International Business Machines Corporation. All other trademarks remain the property of their respective owners.