Hello All Image/SQL and TurboIMAGE users:
(cross-posted to both hp3000-L and OpenMPE lists):

I talked to Tien-You Chen at the HP R&D Database Lab
at the end of last week:

(1)  The Lab is still in business.

(2)  The latest release of TurboIMAGE is now at a point
where it appears to be in good shape:  Stable, with no
known critical or serious bugs.  Version for MPE/iX 7.5
is:  C.10.05 (as compared to Vers C.10.03 just before
HPW-LA last fall).  This release incorporates 64 out of
65 items on the latest SIGIMAGE "Now / Soon Available in
Image/SQL" list:  All EXCEPT item # 59:  One-time make
DBUTIL DUMPING and DSEM flags default = ENABLED. The
Image/SQL Advisory Committee (ISAC) agreed with the HP
R&D Lab that with end-of-HP-sales fast approaching this
relatively minor, incremental "nice to have" was probably
not worth the always-present risk of messing something up
when changing user functionality.

SIDEBAR:  YIKES !!!..  Somehow I had it in my mind that I
posted the latest post-HPW-LA "Now / Soon" list to 3000-L
after the conference last fall, but now at least on quick
search I can't find any record that I did that. If that's
true (probably)...  well, slap me...:  sorry...  I'll try
and get that out in the next few days in text format; and
if a reasonable number of users would like to have their
very own heavily formatted version as a Word2000 ".doc",
I can send email direct on request;  or, alternatively:
If another ISAC member that maintains an externally
accessible web site is willing to put it up we can make
the formatted version available to everybody that way;
so that you can keep your very own archival copy if you
want..  But right now back to current business:

(3)  As we close in on end-of-HP-sales just over nine
months from now, the Lab is wondering if there are any
"last call" low / medium effort, low risk enhancements
that the user community would like them to do.  Up to now
I believe there was some consensus it might be prudent
not to make anymore changes to TurboIMAGE between now and
end-of-HP-sales unless it was absolutely necessary, but
after talking to Tien-You about this for a bit I have
changed my mind.  Rationale:  Even if the Lab comes out
with another TurboIMAGE release (C.10.06;  which we would
then expect to be the last one from the HP R&D Lab), that
would come out as a patch; and I understand the current
C.10.05 production release would continue to be available
AFTER end-of-HP-sales.  So any additional features would
be optional.  While not picking any firm date, Tien-You's
off-the-top opinion was that they would want to have any
potential C.10.06 TurboIMAGE patch done, tested, and out
the door no later than about the end of July 2003;  about
three months prior to end-of-HP-sales..  That would still
give them ~six months to work on a potential C.10.06; and
if they don't get much else to "clean up", they could
probably get a fair amount of stuff done in six months...
It is further my understanding that the Lab is willing
to consider incorporating a few more enhancements if we
want them to...


Speaking strictly for myself, there may be another good
reason for giving the HP R&D Database Lab some additional
enhancement work to do for the next six months:  If they
the R&D Lab have the current version of TurboIMAGE in a
"stable state" and there are no more calls from the user
base for any more enhancements, what are they going to do
for the next six months ??..  Perhaps more to the point,
what are they going to tell HP Corporate bean counters
they are doing;  i.e.:  I'm a little worried that if the
R&D Database Lab doesn't have a lot more to do, that said
bean counters will decide there is no reason to keep all
of them on for the next nine months...  Witness recent
jaw-dropper:  At the end of last week if you sent Walter
Murray a direct email to his long-time HP address, in
reply you got his "out of office" message:

"Friday, January 24, 2003, was my last day at Hewlett-
Packard.  For matters pertaining to HP business, please
contact Randy Roten.  ....."

WALTER MURRAY is NO LONGER WITH HP !!!!...  sheesh...


Anyway:  I went back to the last "complete" TurboIMAGE
Enhancement Ballot (not just the items that were sent
forward to the last Interex SIB) and took all "Low" and
"Medium" effort items (reality sez forget "High effort"
items).  Same with the five "Low effort" items from the
latest HPSQL Ballot (but please note not sure if even LOW
effort HPSQL items will be considered by HP).  Turned all
into text mode list and attached to this email.  Before
going any further would like to get comments from the
Image/SQL user community.  Since six month clock already
running, will leave this action item open for ONE WEEK:

(1)  Do you think having HP make additional changes to
TurboIMAGE beyond the current release C.10.05 is a good
idea, that would be available as OPTIONAL patch C.10.06
??
     YES       NO       ASK SOMEONE WHO CARES

(2)  IF answer to (1) above is "YES", what items in the
following list would your site most like to see done ??..
Note that at least for now am not conducting a "how many
votes" ballot;  just trying to get a feel for what if
any items would be of most interest. For the moment don't
worry about trying to assess "risk":  If it ends up that
anything more is done, IMO it would be prudent to leave
"risk assessment" mostly up to the HP R&D Lab.

Replies to 3000-L encouraged (since per consensus that is
the officially designated discussion list for this SIG).
Please put me on direct copy;  or send "direct only" if
for some reason you don't want to go public...

Ken Sletten
SIGImage/SQL Chair
======================================= Attachment:


From TurboIMAGE Ballot:

SIB ID:  97I01
IMAGE multi-record / multi-set DBGET, DBPUT, & DBUDPATE
(process entire blocks of data instead of just individual
records).
Lab Effort:  MEDIUM

SIB ID:  97I02
New IMAGE GETUPDATE intrinsic, to retrieve / update record
in one call without DBGET. Save overhead; simplify coding.
Lab Effort:  MEDIUM

SIB ID:  97I13
Better DBUPDATE performance for IMAGE Sort Fields with
CIUPDATE:  Compare new value to old, then search for new
position in the "right" direction, instead of start at
end of chain.
Lab Effort:  LOW

SIB ID:  97I15
Ability to set max IMAGE chain lengths, with unique key
option:
Lab Effort:  MEDIUM

SIB ID:  98I17
Allow >16 paths to DETAIL dataset ( >16 to Master now).
Lab Effort:  MEDIUM

SIB ID:  01I42
Increase IMAGE DBOPEN concurrency.  Significant reduction
in delay; if many instances of one app launched at same
time:
Lab Effort:  MEDIUM

SIB ID:  01I43
$CONTROL option to allow >= 2047 items in IMAGE database.
Disallowing Remote File Access (RFA) for > 1200 items is
O.K.
Lab Effort:  MEDIUM

SIB ID:  00I20
Drop the need for IMAGE BlockLength to be a multiple of
256B  (saves space):
Lab Effort:  LOW

SIB ID:  01I24
Add new DBINFO modes (probably '5xx') to return all root
file flag settings (several flags are missing from current
mode-list).  Avoid need for child process & parse of
DBUTIL-FLAGS output.
Lab Effort:  LOW

SIB ID:  01I44
Allow a single IMAGE character field to be as large as
the 2378-Word maximum entry length.
Lab Effort:  ?? (Not evaluated yet;  threw in anyway)

SIB ID:  97I04
DBUTIL: Non-destructive creator access across groups:
Lab Effort:  LOW

SIB ID:  97I05
DBUTIL: Ignore file equations (can't use / do not apply
anyway):
Lab Effort:  LOW  (but backward compatibility risk ??)

SIB ID:  97I07
DBSCHEMA:  Update so can be single tool to CREATE IMAGE
database: Be able to specify all DBUTIL flags / settings
(has not kept up with recent changes). Make it fully SQL-
aware.  Flag ALLBASE reserved words with WARNing to catch
conflicts up-front, instead of only later when ATTACH
database to DBE:
Lab Effort:  MEDIUM

SIB ID:  98I08
DBSCHEMA:  Allow Masters to be declared after Details:
Lab Effort:  MEDIUM

SIB ID:  98I09
DBSCHEMA option:  Allow symbolic expressions for capacity:
Lab Effort:  LOW

SIB ID:  98I10
DBSCHEMA option:  Optionally auto calculate PATH count:
Lab Effort:  MEDIUM

SIB ID:  00I22
Make the DBQUIESCE intrinsic used by TURBOSTORE user-
callable;  and provide external documentation for it,
including appropriate end-user cautions:
Lab Effort:  MEDIUM

SIB ID:  97I30
Allow default values for FIELDs not specified at DBPUT
time, with equivalent to SQL/Server "AUTOINCREMENT"
feature:
Lab Effort:  MEDIUM

SIB ID:  97I31
Allow tracking files to be associated with an IMAGE
database (all associated files;  such as schemas, 3rd-
party indexes, etc.):
Lab Effort:  MEDIUM

SIB ID:  97I32
Tool "memory area" associated with a database (allow
logging of tool-specific IMAGE data on structural
problems, etc.):
Lab Effort:  MEDIUM


From HPSQL Ballot:

SIB ID:  97I53
Enable / Disable end-user SQL access without Attach /
Detach:
Lab Effort:  LOW

SIB ID:  97I54
Allow IMAGESQL ATTACH by DBA user without password
(requiring password in this case is actually LESS secure):
Lab Effort:  LOW

SIB ID:  99I61
Do not require IMAGE Maintenance Word in IMAGESQL if none
is specified in the IMAGE database (potential security
issue ??):
Lab Effort:  LOW

SIB ID:  99I62
Let TurboIMAGE keys split by IMAGESQL during ATTACH be
accessed as a full IMAGE key via SQL at run time:
Lab Effort:  LOW

SIB ID:  98I68
Bundle all of ALLBASE with Image/SQL (need 9th module);
at least on the new A-Class and N-Class boxes:
Lab Effort:  LOW
===============================================

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *