[Crm-sig] Call for Comments

Vladimir Ivanov nomemm at gmail.com
Tue Mar 22 09:41:44 EET 2011


Dear all,

First of all, I agree with both Maximilian's objections.

My comments:
1. Real use cases (regarding different syntax, encodings) are needed.
2. I would not use URIs like "http://www.russianmuseum.ru/Ж-91/1
(the last slash '/' is a part of the ivnentory ID)

3. Why not to propose more flexible solution, based on the following
(or similar) approach:
The main idea is to keep inv. IDs and URIs separately.

1) The curator should identify an object in the real world environment
as he/she describes the object in the real world (in museum).
(to attach inv. ID to the description) -- already done

2) The curator (or somebody else) should identify an object in the Web
(of Data, SW, etc.)
as he/she describes the object in the Web (to attach URI to the description).
That "URI" is not necessary equals to "Base URL" + "inv ID"
but

3) Any agent on the Web should be able to (automatically) access the:
- a trusted procedure or description of  matching ("inv. ID" <-> "URI") and
- its the source provided such a matching.

Such a procedure (step 3) should not necessary be
a concatenation of two strings ("Base URL" + "inv ID"),
and may be provided by external trusted service(s) or by museum itself.
Those "identity" services should only resolve URI into inv. ID and
vice versa (recall DNS).
I thought I've seen this idea somewhere (OKKAM project?).


--
Best,
Vladimir

2011/3/21 Maximilian Schich <maximilian at schich.info>:
> Dear all,
>
> While I think this initiative is long needed, I'd like to raise two
> objections:
>
> First, I think the simple procedure should be explained in a equally
> simple way, in a single paragraph, and then include the whole
> explanation, without having to open a Word document. The first paragraph
> should contain the CEO/president's summary:
>
> """
> We propose that all cultural heritage objects should be assigned with
> Uniform Resource Identifiers (URIs) by their stakeholders [find better
> word], in the form of a base URL and a unique identifier (for e.g.
> http://www.britishmuseum.org/inv/ME91678). A base URL is... A unique
> identifier is...
>
> * Yes it should say http://www.britishmuseum.org/inv/ME91678 not
> http://www.britishmuseum.org/explore/highlights/highlight_objects/me/s/stone_lions_head.aspx
> """
>
> Second, I disagree with the implications of point 10c in your Word
> document, as this inhibits the scientific process. Looking at reality,
> there often is no "best institution for preserving a particular objects
> identifier", and competition in coming up with the best identifiers is
> part of art research. Well known examples are Helbig's famous guide
> through the collections of Rome, or Artprice.com as a repository of
> auction events, with many "moving" objects, for which the museums or
> collections are too small, too heterogeneous, or too frequently
> changing, not to speak of many objects that are hard to define as such,
> or not even belong to a single entity.
> In any case, by nature, there will be a long tailed frequency
> distribution of the number of URIs assigned to any number of objects,
> with reconciliation (i.e. stating "x is {the same as|part of} y") a key
> problem to solve. While I agree that the proposal may help to contain
> this problem, we should not define an artificial hierarchy of authority
> of who may issue the best identifier. If some museums or collections
> don't do, we should encourage the Helbigs and Artprices of the world to
> publish their URIs as a fallback. In other words: Coherent identifiers
> for a large numbers of objects created by a properly published Ph.D.
> thesis or a secondary institution such as Artprice are better than no
> identifiers at all, and as such should be encouraged.
> The same is true for identifier reconciliation: It should also be
> encouraged; but I am pretty sure, this can not be done in a successful
> manner by excluding competition. On the contrary, competition might be
> the major force that will encourage scholars to produce better identifiers.
> As a consequence, in my opinion, the president's summary should end in
> the following way:
>
> """
>
> Primary stakeholders, such as museums or owners of large collections,
> should be obliged to create URIs for their objects. For objects where
> this is not viable for a variety of reasons, or for museums where the
> creation of URIs simply does not happen or unclear, secondary entities,
> such as scholars, auction houses, online-encyclopedias, etc. are
> encouraged to provide alternative identifiers to which the community can
> refer to in a better way. Reconciliation of these identifiers should be
> a common goal of the community wherever reasonable.
> """
>
> Best regards,
> Max
>
> Dr. Maximilian Schich
> http://www.schich.info
>
>
>
> Am 21.03.11 12:02, schrieb martin:
>> Dear All,
>>
>> Your comments on http://www.cidoc-crm.org/URIs_and_Linked_Open_Data.html
>> will be most welcome!
>>
>> Best,
>>
>> Martin
> _______________________________________________
> Crm-sig mailing list
> Crm-sig at ics.forth.gr
> http://lists.ics.forth.gr/mailman/listinfo/crm-sig
>



-- 
С уважением,
Владимир Иванов

-----------------------------------------------------------------------
Центр оцифровки культурного наследия, ЦИТ РТ
+7 (843) 264-72-54
+7 (917) 270-03-12
e-mail: vvi at tatar.ru

-----------------------------------------------------------------------
Лаборатория математической и компьютерной
лингвистики НИИММ им. Н.Г. Чеботарева -
cll.niimm.ksu.ru

-----------------------------------------------------------------------



More information about the Crm-sig mailing list