[ibis-macro] Re: Questions from the Editorial Group
- From: "Walter Katz" <wkatz@xxxxxxxxxx>
- To: "'Bob Ross'" <bob@xxxxxxxxxxxxx>, <michael.mirmak@xxxxxxxxx>, "'IBIS-ATM'" <ibis-macro@xxxxxxxxxxxxx>
- Date: Tue, 20 Mar 2012 12:57:56 -0400 (EDT)
All,
I think I now understand the issue. In 5.1 Tables have a type for column,
and if only one Type then it applies to all columns. The exceptions are
the Tx_Jitter and Clock_PDF with Format Table. These Reserved Parameters
have 3 columns: ID, Time, Probability, and Type is either Float or UI.
These two types translate to the following 5.1 Table Types:
UI (Type Integer UI Float)
Float (Type Integer Float Float)
Does this resolve the issue?
Walter
From: Bob Ross [mailto:bob@xxxxxxxxxxxxx]
Sent: Tuesday, March 20, 2012 10:53 AM
To: wkatz@xxxxxxxxxx; michael.mirmak@xxxxxxxxx; 'IBIS-ATM'
Subject: RE: [ibis-macro] Re: Questions from the Editorial Group
Michael and Walter:
I think the issue is with using UI in tables as is shown
in V5.0. This is resolved in V5.1 by allowing a Type
for each column.
Otherwise, we would have had to interpret a single Type UI is a
generic Float in any Model_Specific table column, but with a specific
interpretation only in defined Reserved parameters where columns
are pre-defined.
Bob
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Walter Katz
Sent: Tuesday, March 20, 2012 6:11 AM
To: michael.mirmak@xxxxxxxxx; 'IBIS-ATM'
Subject: [ibis-macro] Re: Questions from the Editorial Group
MM,
I still do not understand what the issue is. UI applies to parameters that
are "Time" parameters. "Time" numbers are either in seconds (Float) or in
bit_times ("UI").
Am I missing something here?
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Mirmak, Michael
Sent: Tuesday, March 20, 2012 7:59 AM
To: 'IBIS-ATM'
Subject: [ibis-macro] Re: Questions from the Editorial Group
Walter,
I can assist on #3. A good example would be "UI": must it follow the Float
*Type* and/or Integer *Type* input format requirements, despite UI being
its own Type?
In our reading, "Type" really defines rules that apply more broadly than
just to the entries under "Type". The rules instead appear to apply to
data generally used in .ami files as strings, integers, floats, etc. If
this is true, we can reorder or rewrite the draft appropriately.
Thanks!
- MM
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Walter Katz
Sent: Monday, March 19, 2012 8:09 AM
To: Arpad_Muranyi@xxxxxxxxxx; 'IBIS-ATM'
Subject: [ibis-macro] Re: Questions from the Editorial Group
All,
1) The unit of UI are unitless. 1==bit_time, .5==bit_time/2, .
2) Four Gaussian, mean and sigma are both either Type Float (second)
or Type UI depending if Type is set to Float or UI respectively (see
answer to 1). We are allowing multiple Type fields for Table, but not for
Gaussian, DjRj or Dual-Dirac.
3) I do not understand the question, please clarify by an example
other than Description.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Sunday, March 18, 2012 12:10 PM
To: 'IBIS-ATM'
Subject: [ibis-macro] Questions from the Editorial Group
Hello everyone,
A few questions were raised in the Editorial Task Group
for I would like to get answers for from our experts.
1) What units are assumed for UI? Seconds? Picoseconds?
Percent? Other? We are assuming no suffixes are allowed.
2) For Gaussian and similar types, are we assuming "mean"
and "sigma" are floats?
3) Do the rules under "Type" apply to all of the fields
in an IBIS-AMI .ami file? In other words, do arguments
other than "Type" need to follow the same rules as their
corresponding type?
For example, is the argument to Description a string?
If so, does the argument follow the same rules as the
formal type "string"?
We only defined the rules for the *Types* string, Boolean,
Integer, etc., but other AMI file contents are apparently
of these types without being Type arguments.
Answers and comments are encouraged to help the Editorial
Group in finalizing the v5.1 specification.
Thanks,
Arpad
============================================================
Other related posts:
