Basics of good MIF modeling
A good MIF file contains groups that are standard,
consistent, modular, scalable, flexible and reusable. To design
a well-formed MIF definition, follow these principles:
-
Make a detailed list of the manageable
attributes for your product.
At first, do not worry about classifying the attributes.
After creating this initial list, look at MIF files
for similar products or in related Standard Groups
MIF files for groups and attributes that suit your
product well.
-
Use the DMTF-approved Standard Groups
MIF definitions whenever you can,
rather than creating a private, non-standard group.
When you use a Standard Group, change only the attribute
values for all attributes in that group. If you do not
want to use one or more of its attributes, use the
unsupported keyword in that attribute's Value statement.
Do not change any other items in the standard attributes,
such as AttributeID or type. Do not renumber the attributes
within a Standard Group.
-
Do not add your own private attributes to Standard Groups.
Use the Standard Group and add a separate, private
group of your own definition to add attributes to the
MIF.
-
Make private groups reusable, if possible.
Keep all of your similar products in mind, even
if you are only writing a MIF for one of the products.
Standardizing your private group definitions reaps the same
manageability benefits for your products that using the
DMTF-approved Standard Groups definitions brings
to general interoperability of DMI systems.
-
Make private groups small.
The accepted practice is to use 20 or fewer attributes per
group. As long as the group makes sense, it is even reasonable for
it to have only one or two attributes. Several small groups are
easier to manage, more modular, and easier to reuse than a few
large groups.
-
Use consistent naming and terminology
for groups and attributes so they are easier to reuse.
-
Organize your MIF files consistently.
If you produce more than one kind of component,
this enables you to reuse proprietary MIF groups
from one product to another, saving time and effort.
-
Use enumerated lists liberally.
Enumeration can improve the readability of your MIF
data in DMI browsers, simplify efforts to translate
your MIF data into other languages, and enhance
your key lists and attribute values for data in keyed groups
(tables).
-
Use meaningful keylists.
Use index tables with attributes that make sense for the component
rather than an arbitrary indexing value. Use more than one key, if
necessary. Additional keys allow multiple views of the same data
and provide more flexibility for management applications.
-
Pack as much meaningful information into your descriptions as possible.
Description statements can be a very powerful help tool;
use as many lines as needed.
-
Include minimum, maximum and default
values in Description statements when possible.
-
Try to identify the possible audiences for your
MIFs and address their needs in
your descriptions.
Remember that different audiences will be reading the descriptions;
you want to make your product as easy to manage as possible.
Next topic:
Questions to answer during the modeling phase
Previous topic:
Defining groups to classify attributes
© 2004 The SCO Group, Inc. All rights reserved.
UnixWare 7 Release 7.1.4 - 27 April 2004