comments on beta9 (as of last night's cvs checkout)

[Peter McCartney]

HI . im in albuquerque at a workshop and so have limited access to cvs and
email to communicate. Bad timing, but so was your vacation :). Im sticking
in some comments on individual modules that i would like to see addressed.
most are things i pointed out in the set of files i distributed last week .
others are minor bugs ive spotted. I dont think these should be contentious
issues, but i would be very frustrated to see this beta come out today
without addressing these and unable to comment!!! Ill have email access
later tonight and maybe i can get in on irc later this afternoon.

So here are my comments. i just checked out a set and dont see any of them
fixed, but forgive me if any of these are now obsolete: 


	1.	eml
			a.	So.....eml is basically a wrapper to to
carry the identifier for the one resource plus any inline data or xml
documents that have been stuck in additionalMetadata . this is fine. At one
point we thought it would be convenient to put multiple documents (like
literature) in a single tree for bulk processing, but when we wrote
xanthoria, we found that left us wide open to choking on files of totally
unpredictable size!, so we simply accepted working with collections of
documents instead.
	2.	Coverage
			a.	Redundant repetition: Choice should not
repeat OR importing element should not repeat
	3.	Resource:
			a.	No reason for intellectual rights to repeat
	4.	Literature (this module is totally unacceptable as is)
			a.	Drop section from book
			b.	Drop editors from edited book
			c.	Add bookChapter or bookSection
			d.	Drop conference proceedings. If these are
published, then they are a book. The information about the event and venue
is part of the title.
			e.	Drop publicationPlace - the locational
information is already in publisher
			f.	Drop presentationPlace. Move the proceedings
information from conferenceProceedings to this module.  
			g.	Drop institution from report. Institutional
affiliation of authors is already in the RP information of the authors.
			h.	Make report number optional. This may be
part of the title or non-existent
			i.	Drop publisher from thesis. If it is
published then it is a book.
			j.	Drop software package. This is covered under
eml-software
			k.	Drop the unnecessary sequence element
containing access and project.
	5.	physical
			a.	there is no need to make format name
required if you are choosing asciidelimited, asciifixed, or rasterbinary.
FormatName is only needed if you anre NOT providing the parsing information
inline.  This structure is confusing because someone could enter ascii fixed
info but also enter dbase under format.
			b.	There in a need for something in the
binaryRaster section to tell the parser that the data are arranged rowfirst
or columnFirst, this is similar to dataTable/orientation but  not exactly
the same as it is meaningless at the logical level once the data are parsed.
			c.	Distribution element repeats and contains a
repeating choice.
	6.	attribute
			a.	import protocol* into attribute. This is
VITAL for many datasets which merge columns from different analyses that
share a common observation unit. 
			b.	Import my changes regarding enumeratedDomain
			c.	I assume that siUnit is going to get much
longer to include the common complex ones like "milligrams/liter", "parts
per million",  and "murder of crows"
			d.	We need to decide whether allowNull is to be
handles as 1) a check constraint in which we type the statement into
checkCondition in uncontrolled syntax, 2) as an entry under missingValueCode
( if they don't put it there then we assume the field wont have null
values), or 3) as an explicit element in attribute. Technically, disallowing
nulls is a table constraint, but I fear few people will think to add it when
they are editing the dataTable section and wont to back to it when they are
entering the attribute info. The advantage to the latter is that this is
most familiar to people based on how popular design tools are organized. 
			e.	JoinCondition looks good. 
			f.	There are some annotations that are in
documentation rather than appinfo
	7.	entity
			a.	
	8.		Distribution
		I have a lengthy comment on this connection issue because I
think is really important, but we are not going to address it in this beta
so I will send that later with a request that we plan on reviewing this
issue in  more detail prior to 2.0 final., possibly with a solution that is
consistent with how we deal with units and spatial reference which I see
share a similar problem of complexity.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/eml-dev/attachments/20020613/5564e104/attachment.htm