Hi Saašha,

Isn't it rather the case that if you wish to do anything with an ISBN number, say use it as a lookup value, you have to reduce it to a string of digits (plus any final 'x') anyway, in effect "deformatting" it?

If the schema was to check ISBN numbers, I think the validity of the string of digits (plus any final 'x') should be checked, not the format (prefixes, hyphenation). (I am not saying that the schema should only accept valid ISBN numbers (run the checksum test), only that existence of the required number of digits (plus any final 'x') might be checked.)

Since the publisher codes are not freely available, it would be difficult to require the use of any of the hyphenated formats, however "meaningful" they may be to librarians and book sellers. Without a freely available list of publisher codes, I don't think everyone would be in a position to parse a format like "123456789X" - <http://pcn.loc.gov/isbncnvt.html> e.g. cannot hyphenate all ISBNs.

The identifier type attribute could be standardised, but whether a 10- or 13-digit format is used can be determined automatically, so would not this just mean extra work on the part of the cataloguer?

Cheers,

Jens

PS: In <identifier> I think we see an example of the over-use of language attributes in MODS. What meaning is there in requiring a full MODS implementation to allow users to specify language, script and transliteration for values that are meant to distinctively identify a resource universally?

On Jul 14, 2012, at 5:25 PM, Saašha Metsärantala wrote:

> Hello!
> 
> The MODS schema is very permissive and validates idioms such as (I removed
> the greater-than and less-than characters to avoid possible
> misinterpretations by some mail-software)
> 
> identifier type="isbn" 123-456-789-101-2
> 
> identifier type="isbn 13" urn:isbn:123-456-789-101-2
> 
> identifier type="isbn10" 123456789X
> 
> identifier type="ISBN_13" URN:ISBN:123-456-789-101-2
> 
> identifier type="isbn-10" 1-2-3-4-5-6-7-8-9-x
> 
> and much more. The MODS guidelines at
> http://www.loc.gov/standards/mods/v3/mods-userguide-elements.html#identifier
> state that:
> 
>> There is no controlled list of identifier
>> types. Suggested values include [...] isbn
> 
> This does not specify anything about the different ways to write an ISBN (10
> vs. 13, etc.) and I wonder whether the value of the @type attribute is
> subject to any "common practice" to improve the consistency and portability
> issues created by the permissiveness of the MODS xsd schema.
> 
> Regards!
> 
> Saašha,