[Babase] groups doc

Karl O. Pinc kop at meme.com
Thu Feb 8 18:26:26 EST 2007


On 02/07/2007 03:51:26 PM, kfenn wrote:
> Hi Karl,
> 
> I briefly reviewed the Groups documentation and I have the following  
> comments (below each excerpt, in bold).

We can resolve what's left unresolved in our next conference call.

> 
> 
>      "Special Values
> 
> Group 9.0, Unknown, has a special meaning. Individuals are placed in  
> this group by Interpolation  
> <http://papio.biology.duke.edu/babase_system_html/ch04s06.html> when  
> an their whereabouts are unknown. Another group code for unknown  
> whereabouts should not be created."
> 
> *Typo above in red.*

I could set this off in a big WARNING section, or whatever.  There's
two reasons why not.  First, the consensus is that Babase is pretty
easy to fold, spindle, and mulitate and we're just assuming that
everybody's read the directions and knows what they're doing.
Second, there's a lot of other "Special Values" sections in
the section documenting the support tables with similar
kinds of warnings and they all look like this one.

There's a good argument to be made for putting the GROUPS table
in the support table section.  That might address this issue.
I suppose the reason I didn't was because I have the GROUPS
table on some of the ER diagrams, where support tables never
show up.  And the reason GROUPS does, besides it being so
important, is that it contains references to itself and is
thereby "worthy of diagramming".

> 
> 
>        "From_group
> 
> The Gid  
> <http://papio.biology.duke.edu/babase_system_html/ch03s01.html#Groups-Gid>  
> of the group from which this group split off. This column may be  
> NULL, depending on the value of the Permanent  
> <http://papio.biology.duke.edu/babase_system_html/ch03s01.html#Groups-Permanent>  
> column."
> 
> *I don't intuitively understand how the Permanent column might change  
> the Gid value.  Can it be briefly explained?*

This sentence "Temporary groups (those with Permanent of FALSE)
must have a non-NULL From_group value and must not have their
own Gid value as their Supergroup." shows up in the description
of the table.  I put stuff like that there rather than duplicating
the same information in the documentation for both the Temporary
and the Permanent columns.

So, Permanent does not change the Temporary value, but it does
constrain it.

Which does not mean there's not a better way to clarify the
documentation.

> 
> 
>        "Permanent
> 
> This column indicates whether the group is a permanent group or not.  
> The value TRUE in this column means that the group is permanent, the  
> value FALSE means the group is temporary. This column may not be  
> NULL."
> 
> *Is there a formal definition of permanent versus temporary as it is  
> used regarding groups?  The distinction would probably be helpful to  
> babase users.*

No.  The users of Babase get to make it up as they go along when
group fissions are encountered.  I want to document the capabilities
of Babase in the Technical Specifications document and leave
questions like that to either the Protocol for Data Management
or some other such policy guide.  The policy will probably
vary with different group fissions.

Basically, according to Babase, a temporary group is one with
a FALSE value in GROUPS.Permanent, and the other values thereby
constrained as documented.  That's the Zen of a temporary group,
everything else is what you make of it.  (Grasshopper.  ;-)

> 
> *
> Why are there no links to the From_group definition? (gid, permanent,  
> supergroup all have links to a definition) *

Fixed.  From_group is now linked.

One policy I once had was that, when writing a table definition, the
first occurence of a column name, of the table being defined, would
be a link and subseqeuent occurences would not.  This broke down
as I edited and moved things around and the GROUPS documentation
sure does not follow this rule.  I'm not sure if any of the
table definitions do.  In any case it does not seem to be bad
to link all the time so I'll do that for GROUPS and we can worry
about slaveish consistency later.

Karl <kop at meme.com>
Free Software:  "You don't pay back, you pay forward."
                  -- Robert A. Heinlein



More information about the Babase mailing list