RE: Sockets client in RPGIV

Thanks for the advice, everyone.  I'm leaning toward implementing this, with
reservations.  The programmer on the other end wants to formally acknowledge
each record ("ok" or "rt"), while I feel that this additional overhead is
unnecessary.  We already had a spec for "send next", which implies that the
record was valid and accepted.  RT would be sent only when we detect a bad
record.  I don't feel that this added complexity adds to the protocol we've
established.

Guess I'd better get busy..... Gotta get this thing delivered Monday. Thanks
again.......

Eric DeLong
Sally Beauty Company
MIS-Sr. Programmer/Analyst
940-898-7863 or ext. 1863



-----Original Message-----
From: Scott Klement [mailto:klemscot@klements.com]
Sent: Wednesday, January 30, 2002 12:44 AM
To: 'rpg400-l@midrange.com'
Subject: Re: Sockets client in RPGIV



TCP ensures that every packet sent is received in the same order that it
was sent.  At a lower level, IP performs a checksum on the data to make
sure that the data within each packet is not garbled.

So, yes...  TCP/IP will make sure that same data that was sent is also
received.

Unfortunately, this is not really "data validation".   Just because the
sent data and the received data matches doesn't mean that this data is
"correct".  The sending program could have a bug.  The receiving program
could have a bug.   The data could have gotten corrupt on disk on one
side or another...   reasons why the data could be invalid are many :)

So, I would never advise against a check-digit.   It's not nearly as
robust a mechanism for ensuring data integrity as that used by TCP, but it
may validate your record layout in a way that TCP can't even understand...

Of course, it's your call...

On Tue, 29 Jan 2002, DeLong, Eric wrote:
>
> Group,
>
> I got a q+d project this morning to write a socket client app.  The
> programmer at the other end wants to include a check digit to validate the
> record before processing.  I always thought TCP handled all data
validation,
> eliminating the need to add this layer of validation.  He claims that bad
> data and/or garbled records will still sometimes be returned to the
> application, making this additional testing and process control necessary.
> I have doubts.  Can anyone confirm that this is necessary, or even useful?
> TIA.
>
> Eric DeLong
> Sally Beauty Company
> MIS-Sr. Programmer/Analyst
> 940-898-7863 or ext. 1863
>

_______________________________________________
This is the RPG programming on the AS400 / iSeries (RPG400-L) mailing list
To post a message email: RPG400-L@midrange.com
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/cgi-bin/listinfo/rpg400-l
or email: RPG400-L-request@midrange.com
Before posting, please take a moment to review the archives
at http://archive.midrange.com/rpg400-l.