Conference azur::mcc

> 5. In the Graph window, when there are three attributes being displayed 
>    simultaneously, the 'Time Minutes:Seconds' label on the X-axis overwrites
>    the time values of the X-axis.

Yes, this is in the release notes.  The workaround is to enlarge your window.
It is fixed now, and will available for field test update.  The time string will
 not cover time numbers or bars. Under crowded conditions, the time label string
will not be displayed.  Since experienced users will know the x axis is time.

>    In addition, when two attributes (not three) are selected simultaneously
>    for display, only the top attribute has a label on the Y-axis (its 
>    attribute name.)

I can not duplicate this problem.  What entity did you graph? Class and instance?
What attributes did you graph?  The graph picks up the attribute label from the
data dictionary.  If units are in  the data dictionary they are displayed, else
the attribute is repeated.  If there is null string that will be displayed.

Phil,

Thanks for all the feedback.

Re 3.,  could you describe more what you are doing. What attribute group, what
	operation, does the operation complete or is only the internal error
	box displayed.

Re 4., 	what entity class are you dealing with for Delete and default value.

Re 6., 	we are aware of the problem and are looking into ways to fix it.

Thanks,
Verna

Thanks for the replies.

rec .1

Problem #5.

Hi Al,


>    In addition, when two attributes (not three) are selected simultaneously
>    for display, only the top attribute has a label on the Y-axis (its 
>    attribute name.)

>>I can not duplicate this problem.  What entity did you graph? Class and instance?
>>What attributes did you graph?  The graph picks up the attribute label from the
>>data dictionary.  If units are in  the data dictionary they are displayed, else
>>the attribute is repeated.  If there is null string that will be displayed.


Turns out that the attribute name, MaxNumber Requests, was too long. increasing
the window size takes care of it. The other attributes had shorter names.


rec .2

Hi Verna,



>>Re 3.,  could you describe more what you are doing. What attribute group, what
>>	operation, does the operation complete or is only the internal error
>>	box displayed.


The operation I am doing is SHOW <entity> ALL <attr_group>. No the operation 
does NOT complete but I get a blank Operations window with the internal error 
box displayed. Here is the same operation from FCL:

MCC> show qms_cd stelth all state attr

QMS_CD TPEDNS_NS:.stelth
AT  6-JAN-1992 11:23:24 State Attributes

                      Operational State = Enabled
                            Usage State = Idle
                   Administrative State = Unlocked




>>Re 4., what entity class are you dealing with for Delete and default value.


My global entity classes QMS_CD and QMS_DATABASE and their Child entities. It
happens consistently for all classes. The operation completes (the default value
is fetched) but the internal error exception box is displayed. Here is the MSL 
definition of the DELETE argument and its default:


        REQUEST
                (* QMA_C_REGARDLESS *)
                ARGUMENT Regardless = 13 : Boolean
                   ECHO    = TRUE,
                   DISPLAY = TRUE,
                   REQUIRED= FALSE,
                   DEFAULT = FALSE,
                   SYMBOL  = REGARDLESS
                END ARGUMENT Regardless ;

        END REQUEST ;



What about problems #1  and #2  ?


Thanks,
Phil

1. The graph real-time display increment seems to be set at 1 minute. Is there 
a way to modify that time increment. In order to monitor local system resources,
I need to be able to set it at something like 3-5 seconds.

2. In the same way, in the schedule time window, the repeat interval can't be 
lower that 10 seconds.  This is also too long to monitor local system resources.
Can't it also be lowered to 3-5 seconds.

3. The table format is quite useful but the operation Time option on the 
operations menu is grayed out. Are there plans to support it before V1.2 ships ?


One observation about the table format. The Time column and the first attribute
are too close to each other. Perhaps the first attribute can be indexed a bit
forward or the time column should be wider.

Thanks,
Phil

2025.4 question 1

The real time graph default for statistics is 70 seconds and other partitions 
it is 30 seconds.  It can be set as low as 10 seconds.  The graph inherits the
time set in MAP window, so it can be set from the map operation pulldown.
Look at the map window help on operations graph for detailed help.  We know
that this is not convient or intutive way set time.

2025.4  question 2

The limit is 10 seconds. imposed to keep mcc running.  It is a set constant.
When we get alpha vaxes, we should performance test with lower limits.  
RISC 5000  is snappier but has its own burdens.

Al Moore

Phil,

For V1.2, the only operation Time option for Table mode is time=now.
Start Operation can be pressed as often as desired.

What do you mean be the Time column and first attribute are too close together..
Do you want more of the timestamp to be displayed without MB2ing on it.  We
felt that since it can be expanded, we wanted to leave most of the space for
attribute values.  

Verna

RE: .0

> 1. ... By the way, why is the Historian/Exporter directives showing up
>    for every entity irregardless of wheteher that entity has been augmented 
>    (mine has not and they still show up) ?

  The following information, which will hopefully answer your question, were
  meant to be included in the v1.2 EFT release notes, but, regretably, never 
  made it.  I apoligize for this mistake.  These notes will be entered into 
  the next set of release notes.

  New funtionality:

    Starting with the v1.2 EFT kit, the historian and export directives no
    longer have to be AUGMENTed to each and every entity class in order for
    them to operate properly.  These directives have been inserted under a new
    global entity class called GENERAL.

    The dictionary routines have been changed to automatically "inherit"
    these definitions.  Thus, these directives appear to be part of each
    global and child class defined in the DECmcc dictionary.  This capability
    has been named "pseudo" or "poor-man's" inheritance.

    The dictionary file (mcc_fdictionary.dat) is now much smaller than it
    would have been had we AUGMENTed every entity class with the 13 Historian
    and Export directives.  The v1.2 EFT dictionary is ~38500 blocks large now
    (on VMS), but it would have been at LEAST twice that if every class was
    agumented.

    After much discussion it was decided that v1.2 of DAP should not be
    effected by these changes to the dictionary routines.  So, when you run
    DAP and execute the SHOW and DUMP commands, you will not see the inherited
    directive objects and definitions appearing under each global and child
    class.  They will just appear under class GENERAL.  In other words, DAP
    will show the real and not the inherited view of the data in the
    dictionary.  We will revisit this issue in some later version of DECmcc
    when a more complete design for inheritance is implemented.  At that time
    maybe DAP will have the ability to display either view.

  Side effects:

    Hopefully, the effect of these changes will be transparent to everyone,
    but there is one exception of which I am currently aware:

    The global entity class GENERAL is not a real class in the same sense
    as NODE, NODE4, BRIDGE, etc.  It should be treated like the COMMON global
    class in that it may have to be ignored by programs that read through
    the dictionary to retrieve a list of global classes.  For example, PTB now
    ignores the COMMON and GENERAL classes when building the parse tables.
    The IMPM has changed the map to ignore these classes also.  Other 
    MM's will have to do the same.  The constant MCC_K_CLASS_GENERAL is now 
    available in the mcc_interface_def.h include file if your MM needs to 
    ignore (or manipulate) this special dictionary class.

  Caveats:

    The class GENERAL was created for the express purpose of the pseudo-
    inheritance implementation.  It's future is somewhat questionable since
    much depends on a proper design for an inheritance implementation.  Please
    do not plan on adding anything to this class!

                                                                       - Jerry

  P.S. - 

    For more information on this change, feel free to contact me by mail with
    specific questions (MOLAR::PLOUFFE).

#4. Delete with optional argument not working has been QARd #2076

Thank you for all the replies.

re .5 
>The real time graph default for statistics is 70 seconds and other partitions 
>it is 30 seconds.  It can be set as low as 10 seconds.  The graph inherits the
>time set in MAP window, so it can be set from the map operation pulldown.
>Look at the map window help on operations graph for detailed help.  We know
>that this is not convient or intutive way set time.


Sorry, I missed it in the release notes.  It works fine but what is inconvenient
about it is that after setting the repeat interval in the MAP window and 
selecting the graph window, I have to go back and reset the time interval since all
other operations in the MAP have their repeat interval ON ! 

re .6,

>What do you mean be the Time column and first attribute are too close together..
>Do you want more of the timestamp to be displayed without MB2ing on it.  We
>felt that since it can be expanded, we wanted to leave most of the space for
>attribute values.  

MB2 is nice! As the release notes says, constructor datatypes do not 
always work correctly.  What I meant is that the leftmost attribute value 
starts too close to the line separating the time column and the attribute value
window. I thought It could be formatted to start a bit righward.

By the way, clicking MB1 on the right scroll arrow in the attribute columns area
does NOT scroll the attribute area. The scroll bar has to be dragged for the
area to scroll. Clicking on the same right scroll arrow in the Time column 
DOES scroll the Time column.  

.7

I understand why the Historian/Exporter directives are showing up. Thanks for 
your reply. BUT, the Operations menu gets too long. It basically adds 13 
directives to the operations menu of all entities. I have an entity with its 
operations menu running across 80% of the 19-inch Monitor. It is ugly.
It would have been much nicer to have something like two items in the
Operations menu such as

Historian   [->
Exporter    [->

where the other pop menu will list Show recording, Modify recording, etc.
      for the Historian directives.
and   Export, Show Exporting, etc. for the Exporter directives, respectively.

The List box remains the same.

Anything wrong with doing it this way ?


re .8 

Thank you. problem #4 is confirmed. What about problems #1 and #3  ?


Question:  The new feature of adding child entities to the MAP is very good.
           Is there a way to have ICONs represent those child entities? If
           so, what the naming convention for the Icon files of child entities ?
           Is it documented anywhere ?

           

Question:  It would be nice to be able to double click on an ICON in the 
           toolbox and be able to look down into its lower level. The reason
           is that to add a child entity to the MAP, The only way I could do it,
           is to navigate down to that level in the MAP, update the Toolbox,
           navigate up in the MAP, and then add the child entity to the MAP.
           Is there another way around it ?



Regards,

Phil

>Question:  It would be nice to be able to double click on an ICON in the 
>           toolbox and be able to look down into its lower level. The reason
>           is that to add a child entity to the MAP, The only way I could do it,
>           is to navigate down to that level in the MAP, update the Toolbox,
>           navigate up in the MAP, and then add the child entity to the MAP.
>           Is there another way around it ?


    Put the pointer over the icon inside the toolbox & hold MB2.  Select
    "look into" option off the pop-up window...

    /doug

Table format right arrow scrolling is a known bug which is fixed in next
baselevel.

Verna

rec .10 and .11

Thanks.

IMPM On T1.2.4/VMS

I have a setable  attribute that is defined as a SETOF Latin1String.

It supports the follwoing modify directives: ADD/REMOVE/SET 

When trying to modify the attribute, if I am adding/removing/seting one element,
the IMPM will return a validation error unless I use the double quotes.
If I am adding/removing/seting more than one element (two or more), the IMPM 
accepts it without the double quotes !



Phil

    re: -.1 entered as qar 2141; priority 3; MCC_INTERNAL database
    

Phil,

Could you describe more fully what the attribute for SETOF Latin1String looks
like - is it all on one line, is it on multiple lines, do you enter braces or
parens, etc.

Thanks,
Verna

re .-1

Verna,


Say I am doing an ADD directive. I pull the menu and ask for SHOW curent_value.
The set-valued attribute happens to have only one element.
The attribute value is displayed without the double quotes. Now I click on 
start, I get the validation error. I enclose the value in double quotes and it
gets accepetd. Note that I did not use any parans. or braces.


Phil

Phil,

Sorry for the delay - I was on vacation last week.

There is a bug which makes it necessary to have quotes for set of latin1strings
with only one member.  It will be fixed.

Verna

    Should have put it here...  Sorry.

    When an icon is selected and you view the rule of a domain, the list
    box does not update to include commands relevant to alarm rules, even
    though the focus is set to one of the rules.  Clicking on this "rule
    of focus" does not change the list box context, either.  
    
    The work around is to click on another rule to change the list box,
    then click on the rule original "rule of focus".
    
    -dan

    I have been able to repeat the problem of the Map Window overlaying a
    non-DECmcc window.
    
    1.  Click on the map window (LIST BOX is displayed, too.) giving 
    	everything a few seconds to settle down so focus is completely
    	on the map window.
    
    2.	Click on the non-DECmcc window to bring focus to it.
    
    3.	Double-click on any word of text in the non-DECmcc window and
    	the map window will overlay it, but focus will still be on the
    	non-DECmcc window.
    
    -Dan

    
    .20, .21
    
    Entered as qar 2319 in MCC_INTERNAL.

This is a suggestion for the list box. When a verb is selected, the partitions
that are not relevent are greyed out. The user has to click (select) on one 
of the visible partitions for the SHOW verb and for the modify verbs ADD, 
REMOVE, and SET before he/she can click on the OK/APPLY button. 
However, when there is only ONE relevent partition that can be selected 
(visible), it should be automatically selected when the verb
is.  This becomes apparent especialy with modify directives when most times
the only relevent partition that has settable atributes is the characteristic
partition. You have to keep remembering to click on the partition everytime 
after trying to do an ADD|REMOVE|SET or you get a pop window telling you to 
select the partition.
Of course some entities have also settable attributes in the 
INITIAL_ATTRS partition. In this latter case, no auto-selection should happen
if there is more than one selectable partitions.

I think a fix for this should be very easy to implement and will make the List
Box friendly, especially after the fix is in to make it stay where the user 
puts it. 

Regards,
Phil 

.23 entered as QAR 2366