Additional Documentation

inventory_instances

• id (ID)
• cataloged_date (Cataloged Date)
• discovery_suppress (Discovery Suppress)
• previously_held (Previously Held)
• status_updated_date (Status Update Date)
• status_id (Instance Status)
  • inventory_instance_statuses: id, name
• instance_type_id (Instance Type)
  • inventory_instance_types: id, name
• mode_of_issuance_id (Mode Of Issuance)
  • inventory_modes_of_issuance: id, name

  data (JSON):

• statisticalCodeIds (Statistical Codes)
  • inventory_statistical_codes: id, name
• natureOfContentTermIds (Nature Of Content Terms)
  • inventory_nature_of_content_terms: id, name
• instanceFormatIds (Instance Format)
  • inventory_instance_formats: id, name
• publication->dateOfPublication (Date of Publication)

inventory_holdings

• id (ID)
• instance_id (Instance ID)
• callNumber (Call Number)
• call_number_type_id (Call Number Type)
  • inventory_call_number_types: id, name
• permanent_location_id (Permanent Location)
  • inventory_locations: id, campus_id, institution_id, library_id
    • inventory_campuses: id, name
    • inventory_institutions: id, name
    • inventory_libraries: id, name
• holdings_type_id (Holdings Type)
  • inventory_holdings_types: id, name

data (JSON):

•   statisticalCodeIds (Statistical Codes)
  • inventory_statistical_codes: id, name

Not found:

Discovery Suppress (was originally included in the LDP but not present as of 20 May 2020 )  11/12/20: now there for instances but not holdings.

inventory_items

• id (ID)
• chronology (Chronology)
• holdings_record_id (Holdings Record ID)
• number_of_pieces (Number Of Pieces)
• material_type_id (Material Type)
  • inventory_material_types: id, name

  data (JSON):

• statisticalCodeIds (Statistical Codes)
  • inventory_statistical_codes: id, name
• metadata→createdDate (Date created)
• circulationNotes→descriptionOfPieces (Description Of Pieces)
• status→date (Status Date)
• status→name (Status Name)

Inventory diagramm

Records structure

• Laura verified that in FOLIO if there are multiple copies of the same thing, there is only one instance record (one instance record with multiple holdings records).
• Laura said holdings records must have Inventory records, and item records must have holdings records.  There can be inventory records without holdings record, but it's just a possibility. Some possible uses Laura could think of where there might be just instance records: using to order things, eg., ordering everything from a publisher (but maybe this should be containers which isn't implemented yet; when received would be cataloged separately); to keep instance for something no longer held. Laura said it sounded reasonable only to count titles that have holdings.
• In Inventory, units are generally defined as what the circulating unit is.  Item records can have piece counts.
• Some institutions do create item records for e-resources.  It has something to do with how information is displayed in OPACS.  It is not common, but you can it if you want to.
• Local Cornell (from JL in March):  Withdrawing records is a sequential process that involves actions on separate record components (item, holdings, bib).  BIB: if the whole title is being withdrawn (it’s gone and there are no copies) the BIB will be suppressed (BIB_MASTER.SUPPRESS_IN_OPAC = “Y”). At some point after that, it’s deleted entirely.  MFHD: If the BIB record is still valid (that is, one copy still exists, but one or more copies are being withdrawn), the MFHD record for the withdrawn copy will be suppressed (MFHD_MASTER.SUPPRESS_IN_OPAC = “Y”). Also, you will often (not always) find a “x” note (saying “withdrawn”) in the MFHD 852 field. The 852$x subfield is for staff notes, so it can contain any text! Also, you will *usually* not find an item record. I say “usually” because sometimes the item records aren’t deleted on suppressed MFHDs because of oversight/error/some other reason that I can’t guess.  ITEM: Circ staff will often apply a “withdrawn” item status to an item record when it has been Lost or Missing for a given amount of time and has completed the search process. From that point, selectors are forwarded a list of item records that have the withdrawn status. At some point, the item record gets deleted but it can exist for quite some time with the “Withdrawn” status on it. In Folio, excluded Items and Instances will show under "DiscoverySuppress' as True.

ERM

• A suite of apps (organizations are part of Acquisitions)
• ERM will only show what we are paying for.  At Cornell, may include OA items if the EBSCO knowledge base includes.
• eholdings can come in through knowledgebases
• Nothing in ERM identifies serials as only being accessible through a database (i.e., not in full)
• Nothing right now identifies backfiles?
• Peter not sure if anything identifies formats in ERM.  Some packages have more than one.  Perhaps through the holdings info?
• Agreements: Defines what is the resource
• Licenses: Defines what you can do with the resource
• OA: 08/20/20: As a library community, we haven’t addressed this.  So standards would be great to address.  One example of why: so users will be able to limit by OA.  Perhaps the ERM folks have addressed for electronic items?  There are different models of OA.  We know probably more about what is not OA than OA.

Where resources are tracked

• Libraries will use the apps very differently.  You don't have to have title level info for e-resources in the Inventory; you may not even be using ERM or the Inventory.  For example, right now, Cornell only has electronic  titles CU is paying something for in the ERM.  Over time, if the knowledge base tracks OA titles, will bring those titles into the FOLIO ERM.  Right now the ERM does not have any government documents (Laura said we process these as government documents, not as e-resources; they generally don't cost anything; not in packages; would generally track in the Inventory).  Laura said there is no logical place to put OA materials.  The Inventory is to track holdings and items. Maybe some resources will only be  represented through the cataloging Source records?  Some will only be in non-FOLIO applications like institutional repositories. Each institution will need to determine where its titles are tracked and report through all of those locations, avoiding duplication if possible.  As Peter noted, we don't know best practices yet.  Need to take stats into consideration when deciding where to track resources.

Adding general classification, language and geographic info from bib info?

• I see in MARC format https://www.loc.gov/marc/bibliographic/bd01x09x.html :
  
  • 041 Language code; in instance records; there is also a fixed field for language; 041 is repeatable and displays the full name of the language; Language is in the Instance record.
  • 044 Country of publication  (Laura says she has never seen this used; code used instead; see next line)
  • 008 position 15-17 = country of publication (not in Inventory?; also in the place of publication (MARC260 or 264), but won't be standardized – it's to be as transcribed from the piece; may just be city; place of publication is up to a 3 letter code (e.g., NYU is New York, USA)
  • 050 $a or "Classification" in Instance Library of Congress call number (see other classification schemes as well).  Should use call number instead; Instance record can have classification data, but it will not always be there; it is repeatable, can have multiple types; in MARC has 2 subfields - |a=classification, |b=cutter or shelflist.    There are no subfields in the call numbers in holdings in FOLIO.  One could populate your instance records with bib 050 |a from your holdings records when loading records in from Voyager for example.  At Cornell, there was a time when we actually stripped call number info from electronic bibs; we aren't ensuring they are there now, but we are not removing them any more.
• 08/20/2020:
  • Languages:

    It is repeatable in the JSON array.  There may not be a primary language for some titles.  But if there is a primary language, it would go first.  If you want all in a particular language, using all of the repeatable information is fine. If you want no overlap in the count, you would use the first language. MARC:  The primary language, if there is one, is recorded in the 008, positions 35-37 (a 3 letter code).    There has to be something in this spot.  If there are more languages, all are entered in a 041.  The 008 language is in first 041 subfield a. (https://www.loc.gov/marc/bibliographic/bd041.html_)

Status Updated Date

• We currently get this information by joining to the inventory_statuses table and looking at the JSON data ('metadata', 'updated date'). MM has noted how difficult it is to pull a lot of the data needed from JSON.  Christie Thomas  from Chicago is creating a list of elements needed routinely.  MM will suggest building a set of tables to pull this data out to make it more easily accessible.  Chicago is interested in status update date to do something in their OPAC.  

Item Created Date

• The only way to get this information through JSON data ('metadata', 'createdDate').

  Right now at Cornell, we count the format by the bib record format.  Laura from MM thinks we should continue to do so at the Instance level; she thinks it’s ok to count accompanying as the format of the main item.

• Linda asked whether we will have the equivalent of Voyager’s net id to know who created an item record.  Laura from MM said we will (e.g., this login made this change).  She mentioned something about “change tracker.”  There are no development resources on that right now.  But trying to make sure we have this info.  We should raise this issue.  Should also be able to turn it off if it is a privacy issue. (08/20/20)

Date of Publication

• This is found in JSON ('data', 'publication', dateOfPublication) but messy data.

• MARC: Date info is in the 008 and the 260/264 (repeatable). Text data.  For Cornell mapping to FOLIO JSON, the first place to check for date info is the 008 (places 7-10 for date 1, and 11-14 for date 2).  If not there, the 260 |c is checked, and if not there, the date in 264 |c is used.  This is what Theo Tolstoy (EBSCO developer) used for Cornell’s test migration; this is what the FOLIO community recommended/MM hashed out.  But some institutions may map this differently.

  May use “u”s when year unknown in the 008 (https://www.loc.gov/marc/bibliographic/bd008a.html ).  “9999” means continuing indefinitely.  In the 260, when a date was unknown, blanks were used (could  be confusing to users), possibly with a dash afterwards, or question marks?  The spaces somehow need to be preserved.  May have brackets around it if the date of publication is not actually on the pieces (maybe just copyright date, e.g.).  May have “c” for copyright, etc., etc., etc.  (https://www.loc.gov/marc/bibliographic/bd260.html ). 

  Laura is afraid that this will just be problematic. Challenge even in MARC.

• 08/21/20: Can SQL normalize this field?  We want 4 numeric digits, clean, without brackets, parentheses, dashes etc., but blanks should somehow be preserved

Record suppression

• Use “suppress from discovery”. “Staff suppress” is only to be used in odd cases (like no longer have an item, but have a PO attached so can’t get rid of);  “suppress from discovery” would also be set.
• As of 08/21/2020, Discovery Suppress in Inventory_holdings is not implemented yet.
• In Folio, if you suppress an instance, everything related to that record will be suppressed (also true of holdings to item – it is inherited).  (In Voyager at Cornell, if you suppress the bib, you are also supposed to suppress the holdings record.)

Enumeration/chronology

• Only need it if we have to go looking to see what was added retrospectively, or to remove bound withs (if not done through the instance containers instance relationships)? And would need a different query because shouldn’t be grouped for that need.

• Containers do not exist yet. Not ranked as go-live, so not a priority. Couldn’t say it’s a priority since we don’t have that functionality now. Container functionality won’t be used for bound withs.

  Bound with: This is in development.  A change to the data model is being proposed by MM today. Asking if doable.  They would like the functionality of being able to link a given item record to multiple holdings(?) records.  Would have a check box to suppress from discovery and a check box for bound with.  Would have accordion showing all titles linked to.  Currently at Cornell, Linda is excluding item records for bound with items with “bound with” in the enum/chrons.

Added new in a timeframe

• What is best way to determine if an item was added new in a timeframe?  (e.g., item records sometimes created before cataloged?) Is there a date for when added to stacks? Date cataloged will not work for ongoing titles.  Is there any way we will indicate item records created retrospectively?
• Answer: Date cataloged is from Instance Status.  But many titles have added volumes after the cataloging date.  For counting items added new and added retrospectively you would have to do something different.  For retro(?) possibly item statistical codes; but dependent on people remembering. 

Items withdrawn

• For items withdrawn, there will also be an item status, but you lose that info if you delete the item record.  This has been a problem and will likely continue to be a problem.  Our group could point that out; we won’t be able to meet these requests. This will continue to be a problem for everyone.

Getting only items cataloged (made ready for use) / Excluding PDA/DDA unpurchased items

• Use Instance status (not repeatable). There will be a set list of values (cataloged, uncataloged, batch loaded, temporary, other, not yet assigned).  However, each institution will be able to remove values from it and add values to it.  (The same is true for items – a set original list that people can modifiy as they want; may influence loan rules.  Cornell’s list of instance statuses in the CU sandbox?  (have to get an account).)  
• Of the set list, at lease these would represent cataloged: cataloged, batch loaded (see also below).
• We asked Laura if there is anything that would allow us to exclude items that have not yet been purchased in PDA or DDA agreements (but that for which there are instance records in FOLIO). She said that maybe we can have an instance status of “pda unpurchased.” 
• Laura says to add the status update date to get the date that status started? (3/13/20)
• 08/21/20 update: Possibility of using instance status code or statistical code. We need to ask the acquisitions folks as it is acquisitions data. It’s more an Acquisition issue.  Now the data is in a 9XX field.  Would be easy to map to an instance status.  If institutions have it at all in the MARC data it would be easy to map to the instance status.

Record formats  (from Laura Wright on 2/27/20 and 3/2/20)

• In general/what’s what:
  • Laura said that the RESOURCE TYPE and the FORMAT TYPE in the Inventory App are based on three fields from the bib record: 336 (Content type), 337 (Media type) and 338 (Carrier type).  336 is used for RESOURCE TYPE, and a combination of 337 and 338 are used for FORMAT TYPE (in that order).    She said that in the LDP, these fields are called INSTANCE TYPE and INSTANCE FORMAT.
    • So Inventory resource type = LDP Instance type
    • And Inventory format type = LDP Instance format
  • She said that these fields were used rather than the coded fields because they were more consistently used across all FOLIO members.
  • 336 is: MARC 21 content type: “form of communication through which a work is expressed”. https://www.loc.gov/marc/bibliographic/bd336.html
  • 337 is: MARC 21 “media type”: “reflects the general type of intermediation device required to view, play, run, etc., the content of a resource.” https://www.loc.gov/marc/bibliographic/bd337.html
  • 338 is: MARC 21 “carrier type”: “reflects the format of the storage medium and housing of a carrier in combination with the media type (which indicates the intermediation device required to view, play, run, etc., the content of a resource).” https://www.loc.gov/marc/bibliographic/bd338.html
• For items like umbrellas there will be no MARC record, just instance records that are not “cataloged.” (In Cornell’s Voyager, these are suppressed.)
• For Report 61: The report asks for any audio or visual digital item (physical or virtual). Using RDA content type, (Laura got the following through apps/setting…):
  • The Instance types of interest would be: performed music; sounds; spoken word; 3-dimentional moving image; OR 2-dimensional moving image.
  • AND
  • The Format type/Instance format would be: audio disc; computer disc; computer online resource; OR video – videodisc (need to see list to see if got text correct).
• How to define an ebook (2/27/20 Laura Wright):
  • Instance: Mode of issuance: from the default suggested terms (may also be defined locally)  either multipart monograph or single unit  (full list: integrating resource, multi-part monograph, serial, single unit, unspecified) 
  • AND
  • Instance format: Computer- Online resource (may be broken up into Format Category and Format term)
• German speaking countries will use “Nature of content” in FOLIO (in MARC 655?). RDA (replaces AACR2). This is the list they came up with previously: https://docs.google.com/document/d/15b5ZLN3ChRy8RKqTOpfmYLZ5sZtmJJ62RZ0u-B7ZWd4/edit?usp=sharing .  Laura doesn’t know of any US libraries using it; but they could.  Should be included in appropriate reports so can be used.  Complete list: https://wiki.dnb.de/download/attachments/106042227/AH-007.pdf   I believe Laura said we have MARC fields for all this data, but from different spots.
• How to define an ejournal (2/27/20 Laura Wright):
  • See that for ebooks. But use “serial” and serial only for ejournals for mode of issuance.  “Integrating resource” is used mostly for web pages and databases.  Could make case that NYT would be cataloged as a serial or as an integrating resource.  If a set thing, like the printed NYT, serial.  If online, integrating resource.
• “Unmediated” means anything you don’t need equipment to use. g., physical book or serial.  Could we used “not” something to get at it?
• Item type: Material type, term; Material type, category; Material type, source: Item record format codes will be locally defined.  Had a list that they all whittled down?  https://docs.google.com/spreadsheets/d/1scRQl09jroOy-c_emITk3EQ6lkj7XPRlkupPNuL-FfM/edit#gid=1928495227
• Laura says Statistical codes could be our new best friend. We should include this, because at least some institutions (like Chicago) are using this for various types of counts (e.g., ARL and serials maintenance (including way to tell if a serial is currently received or not)).    The MM folks decided to have a statistical code for Instance, holdings and item records.  Locally defined.  In general, Laura noted that it would be good to get data out of local 9XX fields as many people that use them now will not be using MARC in the future – instead they will use the Inventory.
• (3/13/20) Statistical codes should be added for the holdings records too. Print v.s e? 
• All statistical codes are repeatable.
• Instance mode of issuance: how published on instance record vs. how received on holdings.  "Integrating resources" mostly for webpages, databases.  "multipart monographs" and "single unit" for ebooks.  "Serial" for serials. Etc.
• Serial VS. not serial: Should use Instance Mode of Issuance.  This should cover all records in the set; equivalent to using bib format (e.g. “as” for text-based items) in Voyager.  Serialness is not indicated at all in the instance type or instance format coding.

Subscriptions

(4/1/20):  At least for physical, this information would be in Orders.  The Holdings records indicate what we have.  Receiving information is in the Orders app.  Info needed: if current subscription; if subscription added or cancelled; last checked in

Resources of interest?

• Is this the latest version of the Inventory Beta – Metadata Elements?: https://docs.google.com/spreadsheets/d/1RCZyXUA5rK47wZqfFPbiRM0xnw8WnMCcmlttT7B3VlI/edit#gid=952741439 (ask Laura; maintained by Charlotte Whitt)  Yes,
• Material type, term; Material type, category; Material type, source: Item record format codes will be locally defined.  Had a list that they all whittled down?  https://docs.google.com/spreadsheets/d/1scRQl09jroOy-c_emITk3EQ6lkj7XPRlkupPNuL-FfM/edit#gid=1928495227 ??
• Reporting SIG master spreadsheet: https://docs.google.com/spreadsheets/d/1svUM74Dkg4KvTXLzKZK_2k_SxeukX-87NnYf8CaTrYQ/edit#gid=312878932
• LDP-QURY project: https://issues.folio.org/browse/LDPQURY-30?jql=project%20%3D%20%22LDPQURY%22
• API Documentation: https://dev.folio.org/reference/api/
• Axel’s file: https://docs.google.com/spreadsheets/d/15KTHJ7H7P7x1CEX8ycA6tz0jZg44qEddGSL1yso-nsE/edit#gid=1459821778

Open Access

• Laura from the Metadata team noted there is no standard way to do OA naturally.  Should be set so that it can be used in discovery: some users only want OA materials.  No set place to look for.  Could be in 5XX notes - but not consistently.  There could possibly be something that indicates it in the 856 on the bib in the URL.  May not be noted at all.  OA is a publishing model; cataloging records are provider neutral.  Could be different versions.  Holdings level? Not doing cataloging for most e-resources; may be defined in EDS's index? OA will depends on how each institution will represent it to the public.
  

Adding language and geographic info from bib info?

• I see in MARC format https://www.loc.gov/marc/bibliographic/bd01x09x.html :
  
  • 041 Language code; in instance records; there is also a fixed field for language; 041 is repeatable and displays the full name of the language; Language is in the Instance record.
  • 043 Geographic area code (Laura says this is not in the Inventory). 08/20/20:
  • No plans to map that info to instance records, so would have to go to source records (so we will have to wait on this).  It is related to the subject (6XX subfield zs; e.g., if have a book about snakes in New York, New York would be a geographic area code).    [something about there is a way on instance, but can’t tell what role because you lose the subfield?]  The source records will be used for the OPAC. 
  • 044 Country of publication  (Laura says she has never seen this used; code used instead; see next line)
  • 008 position 15-17 = country of publication (not in Inventory?; also in the place of publication (MARC260 or 264), but won't be standardized – it's to be as transcribed from the piece; may just be city; place of publication is up to a 3 letter code (e.g., NYU is New York, USA)

Instance notes (from 5XXes)

• Repeatable
• Should include; will need notes and note types (paired value?)  May include, e.g., OA information.
• (ignore option to staff only for reporting)

Previously held

• Laura doesn’t think implemented yet. Will be a yes/no check box.  (in the administrative data of the instance record)  On 11/13/20 see was implemented at some point.

Data Migration - Axel

• How is the relationship between instance, holding and item meant to be?
• Will there be a best practice how to migrate your own storage into this data model? How are different kinds of units stored? (ex. a drawer full of microfilms)
• Might there be competitions between the types?  holdings type vs mods of issuance; instances formats vs material type?
• How will 948 |f data be transferred?

Title Count

• Not counting earlier titles in title changes (is this even possible??  requested by ACRL)

What defines added new?  What defines withdrawn? Transferred?

• This will be so library-specific.  Date stamp on item record.  Cornell bib-948-like statistics?  In the Inventory, there is a statistics type for every record type.  Possibly for item record creation, we could have a template that encourages a retrospective stat type if appropriate?  U. of Chicago uses a stat code this way?  Paired values: type of code (e.g. ACRL) and code itself (what count as).  Will Cornell map 948 stats to that?  Will need to decide; Laura created a code in the local sandbox?  Maybe PO for those items that are paid for?  Mainz is doing dummy purchase orders for gifts or group orders on one holdings?  Withdrawals and transfers are the same.  Noted it is currently difficult/impossible to get summary counts.

Counting things not represented through item records?

• Pieces of microforms.
  
• Linear/cubic feet of archival materials. As of 08/21/20, this is still a 'not urgent' matter. (At Cornell we have been basing this on bib format and a particular call number format.)
• Unbound serials (e.g., some libraries keep file of estimated number of volumes - as they are not bound)
• unbarcoded rare items?
• Pieces: Item record piece count. Physical description if item is complete and we own all (300|a). Statisitical codes would hever have counts.  Put in note?  Very institution specific. Statistical doe on item - pls. count piece count.

How to defines rare items in general?  Is it only locations?

• Can we ensure all rare locations have the world "rare" or something like it in them?  Makes sense for each institution to plan to do so.  This is the only way Laura can think of to identify all rare items easily.  At Cornell, for locations review in preparation for move, they started by asking which locations we could get rid of because they are not used any more, and how we could make codes more consistent.
• 08/21/20: We are still assuming that location is still the only way to get to rare books.  Martina, from RM group, had someone tell her they get at this in part by call number also?  German libraries also use Material Type for rare books.

Containers records

• Not implemented as of 08/21/20
• Q: Will container records let us get at ACRL’s request to not count earlier titles held in title changes? A: No. Nothing in FOLIO will allow you to do this.   What  Linda was thinking about was “Instance relationships” and that won’t help either.  Still can have referral to an earlier/later title and not own.  (Axel has put instance relationships in the query already.)  Bound with was an option for “instance relationships” but that shouldn’t be there.  See earlier info here on bound withs.
• Q: Is there any reason we should not count all titles in cases of analyzed titles? A: Probably we should stick with how we have always been doing it – which is to count all records that meet the normal criteria.  But if you want to exclude one or the other, you can use the “instance relationships” info.  MARC: 773.

MARC records

• Geographic area: Will this be only available in MARC records? This is 043s.  No plans to map that info to instance records, so would have to go to source records (so we will have to wait on this).  It is related to the subject (6XX subfield zs; e.g., if have a book about snakes in New York, New York would be a geographic area code).    [something about there is a way on instance, but can’t tell what role because you lose the subfield?]  The source records will be used for the OPAC.  

Call number:

• Is it important to have broad call number classification searchable? Call number info is brought into FOLIO instance records as text strings.

Retention commitments

• In MARC 583. There is a place to record this in Folio at the instance, holdings and item levels.
• Laura from MM thinks it should definitely be at the item level at least so we know not to withdraw this copy.  It is an action note.  MARC 853.  Also bib 008?  Could be a statistical code, but not ideal.

Government documents

• Coding in MARC not always accurate.  (008 has values for different types of gov docs; 086 for federal US docs only). How this was determined has been fuzzy over time.  Had to think about this for HathiTrust as they want to know specifically if items are US federal documents.  This matters for U. of Alabama, but they use a location as they receive these documents as part of a depository program.  Cornell is not a regional depository.  There are ways to pull.  We talked about using a stat code, but won’t work for earlier titles cataloged.  Presence of 086 (first indicator of zero for US federal(?)) (https://www.loc.gov/marc/bibliographic/bd086.html ). 

Excluding equipment and personal reserve copies

• Yes, suppress from discovery.

Obligatory copies

• What are DBS libraries using? Still need to be determined.

Counting separately multiple formats attached to same bib record

• The Tile Count query could take care of that.
• MulVers as Scott called them.  E.g., print records with print, and microform and/or electronic holdings records at the same time. 
• Laura from MM thinks it would work to count titles by unique instance ids and holdings format. But only if each holdings record gives format.  Laura suggested we bring this to Jenn Colt.  We can’t automatically load data that isn’t there.  Could we identify these records now and make sure they have all the info they need so they could be counted this way in FOLIO – i.e., prepare the records before final migration?  Presently, we are identifying microforms on print by seeing if bibs with microforms also have holdings with call numbers that don’t include text that indicates they are microforms.

E-Resource  duplication

• How can we make sure to not have duplication? Some institutions do create item records for e-resources

Consortial database concerns for counting (see also Harder issues?)

• Many German universities have shared catalogs.  Having a  bib alone does not equal ownership; for those with shared catalogs it's a bib plus at least one holdings record.   Need also to consider that subscriptions of different formats may be linked to the same instance record.  Will the holdings data allow us to determine different formats?  E.g., may not be able to count on call numbers to identify microforms (even if it normally would, there may be transfers in from other universities with different call number typs).   Some libraries without shared catalogs will also have some records with multiple formats attached to one instance record.

Look at format identification info to see if can define

• physical vs. virtual  (local CU practice: location code - serv,remo)
• digital or not digital
• general bibliographic formats (eg., media)
• very specific formats (e.g., newspaper, database, journals vs. serials?)
• serial (see instance mode of issuance)
• microform
• served remotely or locally? (URLS?  look at those.  Or a project.  Instance or holdings.)

Other breakouts?

• digital digitized and served locally; housed in institutional repository; when digitized;  (URLS or local bib fields might have project info?)
• digitized as part of a project (e.g., HT); print that was digitized to be part of HT; previously held? Local Cornell for Google project (JL March): There is a 903 field of the BIB record that says "google" for at least a subset of these records.  See for example https://newcatalog.library.cornell.edu/catalog/5089795/librarian_view
• perpetually owned vs. leased
• titles we have partial access to full-text, only through databases
• where to get specific notes: left by decedents; provenance; precious binding; copy notes; dedications; inscriptions (up to each institution to decide; on all records possible; specific notes; items record for special binding, provenance?;  could also use tags – but not yet implemented in Inventory, and not sure how consistently they would be applied; it is a separate application, that shows up when enabled; each institution defines, and defines who can use; like putting a colored sticky on something; less reliable than statistical codes; holdings and instance notes likely most reliable? 
• how will items that should not be counted as part of collections be coded for exclusion (e.g., equipment, personal copies on reserves, collections not owned by library but cataloged by library in catalog)?
• text searchable (nothing Laura or Peter could think of)
• item purchased at campus store

LikeBe the first to like this