RE: Why NOT the web?

["Nathan M. Andelin" <nandel@xxxxxxxxxxxxxxxxxxx>]

> Yes. Unless I misunderstood something up front, that's what
> the discussion was about. Updating data, and possibly updating
> batches of data.

Pete,

The following page depicts the flow of a header/lines maintenance program,
within the context of an application control system (all programs share a
common login, where menus and runtime environment are tailored according to
individual user roles).

We plan on implementing a lot of programs like this as part of a
broadly-scoped modernization effort.  The HTML screens show a level of
detail which is comparable, if not a little more robust than most 5250
interfaces, but more point & click oriented.

http://www.relational-data.com/ascent/xur120/hdr_lines.html

For controlling database updates, you might consider something like the
following, where "screen_rec" holds values returned from the screen,
"previous_rec" holds values previously read from the database file and
written to the screen, and where "current_rec" holds current values in the
database file at the moment the update operation is requested.

Comparing "current_rec" to "previous_rec" would indicate whether the record
had been changed since being written to the screen.

I'd suggest this as an alternative to locking a record while displayed on
the screen for an unknown duration.

*-----------------------------------------------------------------

D screen_rec    E DS                  Extname(xur120L1: XUR120R )
D                                     Prefix(s_:2) Export

D previous_rec  E DS                  Extname(xur120L1:XUR120R  )
D                                     Prefix(p_:2)

D current_rec   E DS                  Extname(xur120L1:XUR120R  )

*-----------------------------------------------------------------

HTH,

Nathan.