<<< OFFICE$DISK:[NOTES$LIBRARY]SWS_DEVELOPMENT.NOTE;1 >>>
-< Software Development & Tools >-
================================================================================
Note 1.0 Welcome ... 1 reply
OPNOFF::SYSTEM 5 lines 21-AUG-1992 12:01
--------------------------------------------------------------------------------
Welcome to VAX NOTES
This conference is for discussions relating to
Software Development & Tools
================================================================================
Note 1.1 Welcome ... 1 of 1
OPNNET::KELLEYT1 2 lines 24-AUG-1992 16:56
-< Test Reply >-
--------------------------------------------------------------------------------
Testing...
================================================================================
Note 2.0 DECforms futures? 10 replies
OPNNET::WALLACEK21 16 lines 24-AUG-1992 09:33
--------------------------------------------------------------------------------
What is the future of DECforms?
Is there any chance of a Windows client being available?
If it is only going to be for VT-style terminal access, why change from
FMS or TDMS. It seems too little, too late if this is all it will ever be.
Are other vendors seriously committed to IFDL?
Is All-In-1 likely to be moved to DECforms?
What about other DEC software?
I coming from the view of a programmer after using pure escape sequences,
SMG's and FMS for the user interface. I'd like to be able to use code on
VMS, Motif and MS-windows with 3rd generation languages.
================================================================================
Note 2.1 DECforms futures? 1 of 10
OPNNET::HAWLEYD13 "Dave Hawley, IT Manager, Tektroni" 5 lines 25-AUG-1992 09:38
-< Extended DECforms clients >-
--------------------------------------------------------------------------------
I too would like to second this move, and ask that it be extended not
only to Windows but also to X.11 and Macintosh desktop clients as well.
There are a number of 3rd party and DEC clients that already have this
functionality - X.400 email is one of them (known as A1MAIL). Perhaps
the client forms builders could be optional?
================================================================================
Note 2.2 DECforms futures? 2 of 10
OPNNET::CAMERONJ12 6 lines 25-AUG-1992 14:51
--------------------------------------------------------------------------------
I've heard only rumours of DECforms under X11... almost as if it is
planned but not yet done. I think it's real neat to have something
that will run on both VMS and that other OS.
James
(DEC)
================================================================================
Note 2.3 DECforms futures? 3 of 10
OPNNET::ANDERSONK31 21 lines 25-AUG-1992 17:56
--------------------------------------------------------------------------------
At this point, I'd suggest getting a DECforms PID!
There is some serious R&D in the areas of support on a PC, under Motif,
and client-server style distributed. AS to what will appear when and as
product, is too early to say. The Motif support would've been available
sooner at the cost of no Ultrix support - but we did Ultrix character
cell support first (V1.4).
Mary Tamir said that because of Nippon Telephone & Telegraph
Multivendor Integration Architecture choosing FIMS as one of its
application presentation methods, you can expect MIA vendors to supply
compliant products. To be an MIA vendor you must implement all
mandatory MIA interfaces - I don't know if FIMS is mandatory or
optional.
>I coming from the view of a programmer after using pure escape sequences,
>SMG's and FMS for the user interface. I'd like to be able to use code on
>VMS, Motif and MS-windows with 3rd generation languages.
Thats the DECforms goal.
================================================================================
Note 2.4 DECforms futures? 4 of 10
OPNNET::TAMIRM1 10 lines 26-AUG-1992 09:06
--------------------------------------------------------------------------------
DECforms is Digital's STRATEGIC multi-platform User Interface product.
We've just added ULTRIX support, as Keith said, and other platforms are
merely a matter of time. As for other vendors implmenting FIMS, Keith
also mentioned that NTT's MIA-TP specifies IFDL as its choice for UI.
That means that in order to sell a UI to NTT, the vendor must offer a
MIA compliant FIMS UI. FIMS is up for ANSI/ISO voting at the end of
this year.
Mary
================================================================================
Note 2.5 DECforms futures? 5 of 10
OPNNET::CAMERONJ12 9 lines 26-AUG-1992 09:27
--------------------------------------------------------------------------------
I'd also like to mention, that having used DECforms for the last year,
I would think that porting any TDMS or FMS application to it would be a
reasonably straightforward task.
Judge for yourself; get access to DECforms or someone who can convert a
portion of a TDMS application for you...
James
(DEC)
================================================================================
Note 2.6 DECforms futures? 6 of 10
WPOPTH::Longney 6 lines 26-AUG-1992 16:47
-< ALL-IN-1 for DECforms - simple! >-
--------------------------------------------------------------------------------
Re: .-1
James is probably the last word on this subject, I understand he
is moving to the ALL-IN-1 for DECforms conversion team and will have
product out at years end. :-)
================================================================================
Note 2.7 DECforms futures? 7 of 10
OPNNET::OLSENK1 3 lines 26-AUG-1992 16:49
--------------------------------------------------------------------------------
Please don't talk about unannounced products.
Ken (no last names,please)
================================================================================
Note 2.8 DECforms futures? 8 of 10
OPNNET::CAMERONJ12 14 lines 26-AUG-1992 18:19
--------------------------------------------------------------------------------
Re: Note 2.6 by WPOPTH::Longney
> James is probably the last word on this subject, I understand he
>is moving to the ALL-IN-1 for DECforms conversion team and will have
>product out at years end. :-)
;-)
Years end when?
Seriously, the conversion of so many FMS forms could be done in a
programmed manner; they are well defined.
James
================================================================================
Note 2.9 DECforms futures? 9 of 10
OPNNET::FOLEYT24 9 lines 27-AUG-1992 11:25
-< Screen layout not the problem but rather function >-
--------------------------------------------------------------------------------
The problem with the conversion of ALL-IN-1 from using FMS to forms
product XYZ is not that simple. James is correct is saying that
converting the forms would be relatively simple.
The difficulty with ALL-IN-1 is it's usage of FMS named data which
provides that functionality of the form.
Trevor
================================================================================
Note 2.10 DECforms futures? 10 of 10
OPNNET::CAMERONJ12 10 lines 27-AUG-1992 12:31
--------------------------------------------------------------------------------
Re: Note 2.9 by OPNNET::FOLEYT24
> The difficulty with ALL-IN-1 is it's usage of FMS named data which
> provides that functionality of the form.
Yes. I agree. The translation of this named data would prove very
"interesting". The syntax is perfectly suited to FMS named data; and
may prove difficult to translate.
James
================================================================================
Note 3.0 RDB New Features Questions 1 reply
OPNNET::WILMOREA31 18 lines 24-AUG-1992 16:41
--------------------------------------------------------------------------------
A few questions on the new features announced for RDB V4.1:
Global buffering - will this be available on a database by database
basis ? (IE if one is attached to one or more databases can a mixture
of global and local buffering be used between these ?)
The availability to run more than one RDB version on a node ? Does
this also mean that more than one RDB version can be used on a single
database ?
(More generally the problem we have is that we run RDB, on various
nodes, 24 hours a day, seven days a week. This means that RDB upgrades
are difficult to schedule and acheive. Are there ways of upgrading all
nodes to minimize down time ?)
We are also considering volume shadowing our disks. We have some disks
that contain RDB databases only. These have AIJ enabled. Does
shadowing these disks give us much of an advantage in terms of
recoverability.
================================================================================
Note 3.1 RDB New Features Questions 1 of 1
OPNNET::MURRAYJ2 41 lines 24-AUG-1992 18:20
--------------------------------------------------------------------------------
>> Global buffering - will this be available on a database by database
>> basis ? (IE if one is attached to one or more databases can a mixture
>> of global and local buffering be used between these ?)
yes it is enabled or disabled per database.
>> The availability to run more than one RDB version on a node ? Does
>> this also mean that more than one RDB version can be used on a single
>> database ?
no - each version can only access a database that was create or upgraded to
its own version, hence you cant attach to v4.0 from v4.1.
>> (More generally the problem we have is that we run RDB, on various
>> nodes, 24 hours a day, seven days a week. This means that RDB upgrades
>> are difficult to schedule and acheive. Are there ways of upgrading all
>> nodes to minimize down time ?)
I take it the nodes are not on the same cluster.
I quess you will have to decide when you wish to cut over to V4.1 and at that
time do your conversions of the database.
The point about multi-versions is that you can install 4.1 and test it out
with your applications while 4.0 is still running on your production system.
Unfortunately this does not overcome the problem of 24 hour shops. We are
certainly considering these in our plans for the product.
>> We are also considering volume shadowing our disks. We have some disks
>> that contain RDB databases only. These have AIJ enabled. Does
>> shadowing these disks give us much of an advantage in terms of
>> recoverability.
Mainly Speed.
shadowing will enable rapid recovery of the files as it will with any file
not juts Rdb files. With shadowing you would be able to switch over to the
shadow disk rapidly, unlike if you have just journalling in which case there
will be the neccesary time take for the AIJ recovery process.
================================================================================
Note 4.0 Postscript Viewer 1 reply
OPNNET::KELLEYT1 4 lines 24-AUG-1992 17:00
--------------------------------------------------------------------------------
Does anyone know if there is a Postscript Viewer for
MS-Windows or X Terminals running on a VMS system.
CDA doesn't work (NCD X Terminals)
================================================================================
Note 4.1 Postscript Viewer 1 of 1
OPNNET::CAMERONJ12 8 lines 25-AUG-1992 14:53
--------------------------------------------------------------------------------
The PostScript viewer for VMS is implemented in the DECwindows Server
process on the VAXstation end of things. That's why CDA can view
PostScript.
It doesn't work on just X Terminals (e.g. VT1200, VT1300) because they
don't have the appropriate layer in their server code.
James
================================================================================
Note 5.0 What do we do with TDMS and Alpha? 10 replies
OPNNET::MCINTOSHR2 10 lines 24-AUG-1992 17:47
--------------------------------------------------------------------------------
Apparently TDMS will not be available on the new Alpha VMS.
We have about 100 applications using TDMS most of which will are no
prepared to rewrite.
Does this mean we won't buy any Alpha systems?
Does this mean if we go to the trouble of rewriting to get rid of TDMS
then we might rewrite for another platform altogether?
================================================================================
Note 5.1 What do we do with TDMS and Alpha? 1 of 10
OPNNET::CAMERONJ12 7 lines 25-AUG-1992 14:55
--------------------------------------------------------------------------------
Would you really throw out all those good VAXen?
I'd *guess* that merely converting the product image files might
probably work. Why not ask for some porting time on an Alpha and try
it out?
James
================================================================================
Note 5.2 What do we do with TDMS and Alpha? 2 of 10
OPNNET::MCINTOSHR2 18 lines 25-AUG-1992 15:32
-< I'm not doing the job for Digital! >-
--------------------------------------------------------------------------------
Re: .1
"Would you really throw out all those good VAXen?"
No, I'm not saying I'd through out any VAX, I'm just wondering why I'd
buy any Alpha.
"I'd *guess* that merely converting the product image files might
probably work. Why not ask for some porting time on an Alpha and try
it out?"
I don't want to convert TDMS for Digital - they should do it for me.
I assume that since they are reluctant to do it there must be be a lot
of work involved in porting TDMS. So I'd rather port my applications,
and if I have to do a major rewrite I might as well consider other
vendor's platforms.
================================================================================
Note 5.3 What do we do with TDMS and Alpha? 3 of 10
OPNNET::CAMERONJ12 18 lines 25-AUG-1992 15:58
--------------------------------------------------------------------------------
> I don't want to convert TDMS for Digital - they should do it for me.
> I assume that since they are reluctant to do it there must be be a lot
> of work involved in porting TDMS. So I'd rather port my applications,
> and if I have to do a major rewrite I might as well consider other
> vendor's platforms.
I agree with the last bit; but I'm not so sure that Digital is actually
reluctant to do it. I've heard only one person voice such an opinion,
and I don't think he was the TDMS product manager. Seek an official
opinion in writing from Digital.
What I said before was from the hacker's point of view. "If it works,
do it". It was not meant to be taken as any reliable, business
sensible method of achieving a supported configuration...
James
================================================================================
Note 5.4 What do we do with TDMS and Alpha? 4 of 10
OPNNET::COLDRICKD6 12 lines 25-AUG-1992 16:44
-< talk to Mary Tamir >-
--------------------------------------------------------------------------------
There has been a lot of discussion in this area. Talk with Mary Tamir
from Engineering.
Basically I believe the answer you'll get is "We've been emphasizing
for four years, every opportunity we have, that TDMS will *not* be
ported to any new platforms. This is still true."
The way Mary described it, the code is really impossible to port (for
which read, a complete rewrite!).
Regards,
David
================================================================================
Note 5.5 What do we do with TDMS and Alpha? 5 of 10
OPNNET::CAMERONJ12 1 line 25-AUG-1992 17:32
--------------------------------------------------------------------------------
I think I'll take back 5.3 ...
================================================================================
Note 5.6 What do we do with TDMS and Alpha? 6 of 10
OPNNET::ANDERSONK31 2 lines 25-AUG-1992 17:58
--------------------------------------------------------------------------------
The binary converter cannot convert TDMS and have it work. The internals
of TDMS just won't work on Alpha VMS.
================================================================================
Note 5.7 What do we do with TDMS and Alpha? 7 of 10
OPNNET::ANDERSONK31 1 line 25-AUG-1992 17:59
--------------------------------------------------------------------------------
Note: there is a session for TDMS users on Thurs at 5:30pm.
================================================================================
Note 5.8 What do we do with TDMS and Alpha? 8 of 10
OPNNET::CAMERONJ12 1 line 25-AUG-1992 18:07
--------------------------------------------------------------------------------
Thanks Keith...
================================================================================
Note 5.9 What do we do with TDMS and Alpha? 9 of 10
OPNNET::GARSOND11 8 lines 25-AUG-1992 19:01
--------------------------------------------------------------------------------
re .0
Does your application use ACMS? Any scope to do the terminal interface
stuff, keeping it using TDMS, on a VAX (input from users won't benefit as
much from the faster CPU anyway) and do the job steps on an Alpha? 'Course
this means waiting for mixed clusters I suppose.
Derek (not an ACMS guru!)
================================================================================
Note 5.10 What do we do with TDMS and Alpha? 10 of 10
OPNNET::MCINTOSHR2 16 lines 27-AUG-1992 15:37
-< No ACMS, but what about an emulator? >-
--------------------------------------------------------------------------------
Re: .9
Yes, it may work if we had used ACMS. But we didn't and recoding the
applications to use ACMS would take even more work than recoding for
DECforms.
Maybe there is a market there for an enterprising third party to do a
TDMS emulator for Alpha. Some restrictions would be acceptable (if
documented), eg. no async would be OK. The 80:20 rule should apply. The
remaining 20% that cant be ported we'd (cross fingers) be happy to
leave on ageing VAXes.
The main point is that if we could port a substantial portion of our
applications to Alpha we'd be buying Alpha, while most can't we ponder
where we go from here.
================================================================================
Note 6.0 Joining Rdb tables 2 replies
OPNNET::LANGMAIDS17 4 lines 24-AUG-1992 17:55
--------------------------------------------------------------------------------
Does anybody know if the ability join tables across different Rdb
databases will be forthcoming in a future version, and if so, when ?
If not, why not ?
================================================================================
Note 6.1 Joining Rdb tables 1 of 2
OPNNET::ANSONS10 3 lines 25-AUG-1992 11:05
-< sql4.1 may be your answer >-
--------------------------------------------------------------------------------
STEVE,
This is Johnsy & Sheryl from PA saying that SQL 4.1 should be
able to do this for you. Check session notes from 041.
================================================================================
Note 6.2 Joining Rdb tables 2 of 2
OPNNET::MURRAYJ2 8 lines 25-AUG-1992 12:23
--------------------------------------------------------------------------------
You can join across two different schemas in v4.1 , however both
schemas must be within the same database.
as for full cross db joins, not in the forseeable future as this will
require an large change to the way we carry out our optimization and
data processing steps. We are currently looking at this problem ( and
have been for a while ). no promises as yet.
================================================================================
Note 7.0 Public Interface to DECplan data 2 replies
OPNNET::BINGHAMP12 14 lines 25-AUG-1992 16:10
--------------------------------------------------------------------------------
There was a question today at the DECplan talk about public interfaces
to DECplan data. I have sent questions to the DECplan engineering
team about this and I will post any information I receive.
I don't believe there is currently a supported public interface to
DECplan data. However, one interim solution might be to define your
own DECplan report to dump the data you need in the format you need.
DECplan reports can be placed in textual files which can then be
read by any other tools necessary. I don't know that DECplan would
guarantee that report formats would stay consistent across releases,
but I believe this is a better bet than trying to read the exported
.dtif files.
Paula Bingham
================================================================================
Note 7.1 Public Interface to DECplan data 1 of 2
OPNNET::CAMERONJ12 12 lines 25-AUG-1992 17:31
-< Also try Import >-
--------------------------------------------------------------------------------
Also, to get data *IN* to DECplan, use the import function. I've used
it successfully and I'm very pleased with it's abilities.
It lets you use DECplan on an ordinary text terminal - in a way. You
can prepare the data using a text editor or other tools rather than put
up with using one of these newfangled workstations.
It's also a way to avoid the data entry delays when you have
insufficient memory to properly buffer the DECplan RDB database.
James
(DEC)
================================================================================
Note 7.2 Public Interface to DECplan data 2 of 2
OPNNET::BINGHAMP12 15 lines 26-AUG-1992 12:13
-< Information from DECplan Engineering >-
--------------------------------------------------------------------------------
I have received answers from DECplan engineering. They say you can use
SQL to look at DECplan data but they do not recommend using SQL to
change DECplan data.
They also endorsing using import as a way to get data into DECplan.
They have said that there is an export capability but that it is
currently limited. They have plans to improve export in future
releases.
You can also contact Digital Services if you have specialized needs
for DECplan data. They can contract to do some of this work for you
or to provide consulting.
Paula Bingham
================================================================================
Note 8.0 Answers to DECplan questions from yesterday's talk No replies
OPNNET::BINGHAMP12 19 lines 26-AUG-1992 12:18
--------------------------------------------------------------------------------
DECplan V1.2 performance enhancements - there has been a lot of work
in this area, including work to speed up table editing operations,
making notifications not take time from user operations, making
negotiated time charges, meetings and assignments faster, making
whole planner operations (e.g. delete, copy) faster and in preventing
long operations like scheduling and rollup from blocking interactive
user operations.
Licensing - DECplan client and server licenses are sold separately.
A typical configuration would have a single server license and multiple
client licenses. The server license is sold based on the size of the
machine. The client license is a personal use license that doesn't
depend on the size of the machine. The DECplan team is interested in
customer input on their licensing scheme, so feel free to reply here
or contact your Digital rep. if you have feedback.
All-in-one integration - there are no plans at this time.
Paula Bingham
================================================================================
Note 9.0 RDB SQLMOD - A useful tool? 1 reply
OPNNET::MULLERB28 14 lines 27-AUG-1992 15:19
--------------------------------------------------------------------------------
Is there anyone out there who uses SQLMODule Language on RDB to avoid
precompilers/embedded SQL?
Any useful tricks and tips?
How do you find interfacing to the CDD? (Plus and/or Repository)
What uses /advantages/disadvantages does it have?
Why should I use it?
What should it NOT be used for?
================================================================================
Note 9.1 RDB SQLMOD - A useful tool? 1 of 1
OPNNET::DUMBRELLG22 6 lines 28-AUG-1992 12:06
-< Sort of... >-
--------------------------------------------------------------------------------
Yes we use it at ASX (Australian Stock Exchange). Unfortunately I
personly have no experience. Call me on (02) 3696459 next week and I'll
try to put you onto somebody who can help.
Greg Dumbrell
================================================================================
Note 10.0 SYSMAN STARTUP or SYSTARTUP_V5.COM? 6 replies
OPNNET::MULLERB28 13 lines 27-AUG-1992 15:29
--------------------------------------------------------------------------------
Can someone out there in DIGIT land clarify the support/advisability of
using SYSMAN startup databases instead of (or with)
sys$startup:startup_v5.com etc.
As a result of splitting one large cluster into a number of smaller
ones out startup procedures are under review. Any suggestions or
comments on a way to manage the startup procedures for
muliple clusters with similar applications and a disaster recovery
intention would be most welcome...
Brigette Muller - HEC TAS
================================================================================
Note 10.1 SYSMAN STARTUP or SYSTARTUP_V5.COM? 1 of 6
OPNNET::GARNERR30 "Robin Garner" 10 lines 27-AUG-1992 17:03
-< Simple, but ... >-
--------------------------------------------------------------------------------
My advice is not to use it unless you have a very simple layered
product dependency tree. I use it on my systems, but the above is true
there - there's no way to resequence startup without tearing it all
down and building it up from scratch. The phases provided (I'm sure
you could create more if you tried real hard) are not sufficient for a
complex system.
The best idea I've heard is using DECScheduler for system startup -
create a STARTUP class and put the dependency information in there, and
then reduce SYSTARTUP down to half a page.
================================================================================
Note 10.2 SYSMAN STARTUP or SYSTARTUP_V5.COM? 2 of 6
OPNNET::HAWLEYD13 "Dave Hawley, IT Manager, Tektron" 39 lines 27-AUG-1992 18:26
-< Simple possible solution? >-
--------------------------------------------------------------------------------
We have been using a simple trick for having a common SYSTARTUP_V5.COM
for a cluster (and a lot of other .COMs as well). Try this.
1. Ensure that the SYSGEN parameter SCSNODE is set to some unique value
for each VAX that will be using the system. In our case we use the
DECnet node names to prevent head-scratching later.
2. Create a simple command procedure that puts the current value of
SCSNODE into a DCL symbol, preferably a global one. I think I used
F$GETSYI to do this. This command procedure then sets more symbols
based on this value. Short example follows.
$ n1 == "!"
$ n2 == "!"
$ this_node == f$getsyi(..."SCSNODE"...)
$ gosub do_'this_node'
$ exit
$do_node1:
$ n1 == ""
$ return
$do_node2:
$ n2 == ""
$ return
What you have are some global symbols that can selectively "edit in"
those parts of SYSTARTUP you want. Eg:
.
.
.
$'n1' mount $1$dua0: node1_disk0 /cluster
$'n2' mount $2$dua0: node2_disk0 /cluster
.
.
.
There are a number of extensions to this basic procedure.
Hope this helps.
================================================================================
Note 10.3 SYSMAN STARTUP or SYSTARTUP_V5.COM? 3 of 6
OPNNET::HAWLEYD13 "Dave Hawley, IT Manager, Tektroni" 7 lines 27-AUG-1992 18:30
-< More on point 10.2 >-
--------------------------------------------------------------------------------
Oh, I forgot to add some points 10.2.
1. ENSURE the node lookup procedure is run right at the top of
SYSTARTUP_V5, and preferably by SYSTEM at login.
2. Make sure SYSTARTUP_V5 is reachable by all cluster nodes by being in
SYS$COMMON:[SYSMGR] (not SYS$STARTUP either).
================================================================================
Note 10.4 SYSMAN STARTUP or SYSTARTUP_V5.COM? 4 of 6
OPNNET::KEOGH 6 lines 28-AUG-1992 09:24
-< DECscheduler ... try it, you'll like it! >-
--------------------------------------------------------------------------------
> The best idea I've heard is using DECScheduler for system startup -
> create a STARTUP class and put the dependency information in there, and
> then reduce SYSTARTUP down to half a page.
If you want to see this, drop by the OpenVMS stand, and see how I have it
set up on SHIRAZ::
================================================================================
Note 10.5 SYSMAN STARTUP or SYSTARTUP_V5.COM? 5 of 6
OPNNET::BEISSELC15 36 lines 28-AUG-1992 11:54
-< We've tried SYSMAN's startup >-
--------------------------------------------------------------------------------
We have gone through the whole process of moving our system startup
procedures into the SYSMAN startup database.
We run most of the jobs in BATCH, giving logfiles and parrallel
execution.
Here are some gotchas:
1. By default P1 is passed as "FULL". You may need to specify a
blank value for P1.
2. You can't have the same procedure twice. E.g. with different
values of Pn on different nodes of a cluster.
3. By default batch jobs go to SYS$BATCH. We point SYS$BATCH to a
high quota STARTUP_node queue early on and redefine it back to
its normal value (node_BATCH for us) in the last phase.
4. BEWARE of procedures that exit with status!
A job that exits with an error status will terminate startup in
that phase! I.e. no more startup phases will run.
You may even need to edit DEC-supplied startup procedure to fix
this. Another solution is to jacket one or more DEC procedures
in your own and $ EXIT 1.
This also can be used when you need to serialize more procedures
than can fit in the available phases.
5. Log files always go to SYS$SPECIFIC:[SYS$STARTUP], named
STARTUP_n.LOG. You have to search all the logs to find the one
you want.
By the way, try editing SYS$SYSTEM:STARTUP.COM and setting the Verbose
symbol TRUE and the Debug symbol TRUE. You can then run STARTUP.COM and
see what it would have done.
The four supplied phases have been enough for us.
================================================================================
Note 10.6 SYSMAN STARTUP or SYSTARTUP_V5.COM? 6 of 6
OPNNET::CAMERONJ12 12 lines 28-AUG-1992 12:13
-< Parallelism vs Single stream >-
--------------------------------------------------------------------------------
I've found that on slow disk systems, making system startup highly
parallel sometimes slows it down. Instead of one disk I/O stream,
there are many, causing more head movement.
For personal workstations, I actually set up rarely used products (such
as DECplan, RDB & co) as manually started. The command to start
DECplan then becomes a startup invokation if it is not already started.
I also defer most layered products to be started up *after* I am able
to log in to a windowing session.
James
================================================================================
Note 11.0 Answer for DTM client/server No replies
OPNNET::BINGHAMP12 6 lines 28-AUG-1992 08:44
--------------------------------------------------------------------------------
A DEC Test Manager (DTM) developer has said that the purpose of
the client/server line in the slide was to point out the client/
server architecture of DTM. DTM is designed so that its UI, test
library and test execution can all be happening on different nodes.
There is nothing special about client/server on ultrix, so "(ultrix)"
shouldn't have been on the slide.
================================================================================
Note 12.0 Answers to DEBUG questions No replies
OPNNET::BINGHAMP12 19 lines 28-AUG-1992 08:50
--------------------------------------------------------------------------------
The buttons in the new DEBUG graphical UI are totally configurable.
There is a way through debugger commands (not the graphical UI) to
set "temporary" breakpoints - ones that only stop you once. The
graphical UI also has the concept of activated and deactivated
breakpoints, which might also help you. You can deactivate a
breakpoint so that you don't lose its definition but it doesn't
stop your program execution until you activate it again.
MB3 is configurable through the normal X customization capabilities.
DEBUG doesn't provide anything special but also does not prevent you
from using X customizations.
And, a probable correction. Some people I checked with are under the
impresssion that the new DEBUG graphical UI will be free with VMS.
I haven't been able to definitely confirm this, but my sources are
more likely to be up-to-date on this issue than me.
Paula Bingham
================================================================================
Note 13.0 Answers to Thursday DECdesign PTECH questions No replies
OPNNET::BINGHAMP12 16 lines 28-AUG-1992 12:16
--------------------------------------------------------------------------------
Some customers had some questions about DECdesign PTECH on Thursday.
I have received more information from the engineering team.
Use of DECthreads - The DECdesign team has not tested DECthreads as
an external class library, but there is no technical reason why it
shouldn't work. You can hook it into your design in the same way you
would bring in any other external class library.
Maintenance of design - If there are changes to your application,
you should change the design in DECdesign and regenerate your code.
Since you express your hooks to external code in DECdesign, these
will get regenerated too. A regeneration will currently regenerate
all the code in your application; a future consideration is to
support partial regeneration.
Paula