KMLM List

["Kim, Tae" <Tae.Kim@xxxxxxxxxxxxxxxxxxxx>]

The idea is the pdf and html counts should be split up monthly as well. Of 
course, if the sushi service doesn't store that data at that granular level and 
instead just as a YTD total, that data can't be returned in a sushi call anyway 
because the requestor can request stats for an arbitrary length of months, for 
example Nov 2007 to Feb 2008.

-Tae Kim


On 4/20/09 3:35 PM, "Michael T. Puff" <mtpuff@xxxxxxxxxxxx> wrote:

It would seem to me that multi-month would be required for a proper COUNTER JR1 
report.  The monthly counts represent ft_total which can be a combination html, 
pdf, etc.  While the JR1 "YTD TOTAL" column is a simple sum of the monthly 
columns (and admittedly redundant), the JR1 columns for "YTD HTML" and "YTD 
PDF" cannot be derived from the monthly counts.  I think you would need 
multi-month instance to report these two figures as ft_html and ft_pdf else the 
report would not meet the JR1 specification.

-Michael

On Apr 20, 2009, at 9:09 AM, Dennis Vaux wrote:


At Innovative, we have coded our SUSHI client to ignore any period counts for 
periods that span more than one month.   I agree that no one needs this 
"summary" data, but it isn't explicitly prohibited by the xsd, and we were 
unsuccessful when we tried to get Scholarly Stats to stop sending these summary 
counts in their SUSHI response when we first tested SUSHI 1.0 with them.


-----Original Message-----
From: Kim, Tae   [mailto:Tae.Kim@xxxxxxxxxxxxxxxxxxxx]
Sent: Friday, April 17, 2009   9:30 AM
To: sushidevelopers@xxxxxxxxxxxxx
Subject:   [sushidevelopers] Counter XML Question

I'm   working with Proquest sushi service and they return the stats for every 
month   and a total for all the months.

For example if I ask for Jan-March, it   returns something like:
2008-01-01 to 2008-01-31: 3 requests
2008-02-01   to 2008-02-28: 4 requests
2008-03-01 to 2008-03-31: 1   requests
2008-01-01 to 2008-03-31: 8 requests

I'd LIKE to tell them   that really, the last is redundant and spans multiple 
months so it shouldn't   be in the result.
But really, I can't find anything in the SUSHI or COUNTER   standard that 
dictates that each data point should be exactly one month in   duration.

The xsd of course can't really validate  anything like   this.

Can anybody point me to something in the standard? Usage stats   spanning 
multiple months is kinda problematic for me and I imagine for other   clients 
as well.

Tae Kim
Software Developer
Serials   Solutions



---------------------
Michael T. Puff

Systems Specialist,
HighWire Press, Stanford University
Phone: 650-723-8348; Fax: 650-725-9335
mtpuff@xxxxxxxxxxxx