Jeff Vance let us know:
> Vikram's HP World database slides are available on Jazz at:
> http://jazz.external.hp.com/papers/HPWorld_01/HPSQL/index.html
I would encourage all users of TurboIMAGE, Image/SQL, and
Allbase/SQL to take the time to go through this slide show. Vikram
gave us a very nice presentation during the SIGImage/SQL meeting
in Chicago; lots of detail on recently completed and in-process
enhancements; along with a number of interesting slides on results
of HP's investigation into making TurboIMAGE "thread safe / thread
aware".
WRT the SIGImage/SQL meeting last week: I thought we had one
of our better meetings (of course I may not be an objective observer).
At the high point I counted 47 people in the room. I'll try and provide
some additional follow-on info when I return from leave in two weeks;
right now I've got yet another Transact program I need to finish
before I can escape. Summary / overview (I won't repeat what is
already well-presented in Vikram's PowerPoint slides):
(1) We are now up to 65 items on the "Now / Soon Available in
Image/SQL" list. Nine new items have been added just in CY 2001
(so far).
(2) After technical presentations by Lund on their "Image Stats"
enhancement to their SOS/3000 product; by HP on SPT/iX; and
mention of T.RenzSoft's DBGauge product: There were zero
objections to moving the long-standing enhancement request for an
IMAGE performance data utility to the "Now / Soon" list; i.e.: It has
now been addressed by two reputable 3rd-party vendors *and* HP;
i.e.: It's DONE.
(3) There was agreement among attendees: For the enhancement
request added in 2001 asking for > 1200 items in one TurboIMAGE
database, it is acceptable to require that Remote Database Access
(RDA) be DISABLED for > 1200 items. HP has indicated that this
restriction will make the enhancement significantly easier to do. The
enhancement request will be modified to so indicate; and to ask for
"Allow >= 2047 items in one TurboIMAGE database". Once RDA is
no longer required, HP has informally estimated that it should be
*relatively* easy to go to at least that many items (and maybe more)..
.... and; no: I can't specifically tell you "relative to what" yet....
:-)
(4) Added new item to request list: Allow a single TurboIMAGE
CHARACTER field to be as large as the current 2378-word max entry
length (minus four words for forward and backward chain pointers);
i.e.: WITHOUT having to use IMAGE compound items.
(5) Added new item to request list: Allow TurboIMAGE user logging
to function with more than the current "1140 processes per Log ID"
limit.
(6) Since IMAGE QUERY has now been "shared sourced" by HP,
and for reasons of other more urgent competing priorities, there was
unanimous agreement to move the remaining "gang of six" IMAGE
QUERY enhancements to a "B" list; i.e.: We won't lose track of
them, but they will not appear on the 2002 TurboIMAGE Ballot that
will be voted on in (we hope) a couple months.
(7) Representatives for the three TurboIMAGE database tools
(Adager, DB General, and Flexibase) were asked to comment on
status of their support for all the recent and ongoing enhancements
to TurboIMAGE. Adager and DB General can support all of the new
IMAGE features that are General Release now (but you may want
to verify that you have the latest version of your favorite database
tool before launching into use of some of the most recently released
IMAGE features). They will also (as expected) continue to support
additional new IMAGE features as they are released. Flexibase
supports DDX and does version checking for some additional new
features, but it does not support database transformations with those
other new features (and there are no current plans to do so).
(8) We hope to have the updated "Now / Soon" list and the 2002
TurboIMAGE and HPSQL Enhancement Ballots up on the web for
READ-ONLY review, discussion, and comment by about the end of
October. After a "decent interval" for open discussion, the comment
period will be closed and we will proceed to a vote on the web
sometime in late 2001 or early 2002; the "Top 10" or so from the
SIGImage/SQL ballot process will then be forwarded to the Interex
SIB process.
**** ALL PLEASE NOTE ****:
All TurboIMAGE, Image/SQL, and Allbase/SQL users have been able
to vote on the SIG ballots for several years, whether or not they were
members of Interex. For the first time in 2001 (I'm pretty sure it was),
Interex did not require voters on the Interex SIB to be members of
Interex. We are expecting that the same rules will apply in 2002 as
applied in 2001. Given that, it is important that all users remember to
vote on BOTH the SIGImage/SQL *and* the Interex SIB in the next
few months.... and for that matter the SIGMPE ballot and other SIG
ballots in your area(s) of interest....
(9) Last (for now) but certainly not least (drum roll): All be advised
that THE ENCHILADA LIVES !!!!.... Special thanks to Ted Ashton,
who as he mentioned in his recent "HPWorld report 1" brought up
the subject of the ENCHILADA at the Image/SQL Database Round-
table. HP's answer to Ted's question generally indicated that HP was
still kind of waiting for further guidance from the SIG. As you might
expect, this lead to some additional (but surprisingly short) discussion
at the subsequent SIGImage/SQL meeting......
HANG IN THERE WITH ME NOW:
FIRST, for those who might have forgotten (surely not...):
ENCHILADA = ENhancement for CacHIng of Limited Authorized DAta.
Credits for the acronym to Steve Cooper of Allegro Consultants.
Most of y'all may remember that the Image/SQL Advisory Committee
(ISAC) carried on an extended and "animated" technical discussion
on the merits of four competing ENCHILADA proposals; since the
concept was first sprung on the world at the December 1998 ISAC
meeting in Las Cruces, NM by Mr. SCooper; plus there was much
discussion of the concept at several SIGIMAGE meetings and on
3000-L.... DON'T WORRY: I'm not going to repeat any of that now.
The ISAC did not reach unanimity on a recommendation to HP; but
did vote as follows on implementation methods; after which all four
competing proposals were passed on to the HP R&D Database Lab:
[a] A separate database or databases: 2 votes.
[b] An external-to-database PRIVileged file: 1 vote.
[c] A "veiled" / partially hidden dataset: 4 votes.
[d] Expand the current IMAGE root file: 10 votes.
Given HP's indication that they were still looking for guidance from
the SIG, the above choices were put to a vote of the attending
membership in Chicago.... Results:
[a] Zero votes.
[b] Zero votes.
[c] One vote.
[d] Everybody else (about 30 people left at time of this vote).
.... Now....: Some might say that above attendees were a pretty
small sample...... But this has already been languishing for way too
long: A decision needed to be made and we the attendees at HPW
collectively made one. Baring any tremendous and wide-spread
out-cry of violent disagreement, last above will be passed on to HP
as the official recommendation / request of SIGImage/SQL.
How did we manage to resolve and conclude the ENCHILADA issue
in relative short order in Chicago ??.... Believe I know one big
reason: NONE of the authors of the four competing proposals were
in attendance at the time.... Moral of the story: What can I say:
All else being close to equal, Government tends to go to those who
show up and vote.... ;-)
Thanks to all at the HP R&D Database Lab in both Cupertino and
Bangalore, for their recent and ongoing efforts to improve and
enhance TurboIMAGE, Image/SQL, and Allbase/SQL.... Don't
worry: We still have lots more work for you.... :-)
Ken Sletten
SIGImage/SQL Chair
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
