Print

Print


Il 13/06/2016 23:19, Martin Holmes ha scritto:
> Hi Paolo,
>
> "urn:" is an existing scheme:
>
> <https://www.w3.org/Addressing/URL/URI_URN.html>
>
> so you don't want to use it for your private URI scheme. In addition,
> the presence of multiple colons in your existing referencing scheme is
> probably not such a good idea -- I think the structure comes from
> Perseus, doesn't it?

Hi Martin,

the colons come from the CTS (Canonical Text Service) / CITE 
architecture: see
https://cite-architecture.github.io/
and
http://cite-architecture.github.io/ctsurn/overview/

In a URN like
urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.12
the second block ("cts") refers to this. Shall I then leave the URN as 
it is or still substitute the colons with another character?

In fact, the URNs I'm building are based on the work-in-progress made by 
Perseus (but not only) to
1. create CTS/CITE URNs to identify late-antiquity authors (such as 
Priscian and Donatus), that currently don't have such identifiers, and
2. create a 'new version' of CTS.

So my URNs... currently don't map to any website as they are.
They could, if processed into an URL of a website publishing Priscian & 
company (but not all late latin grammarians are published online).
At this time, they are simply meant to be a formal way (CTS protocol) to 
encode "Priscian, Ars grammatica, 2.28.67".

Paolo

> I might be inclined to choose another character to represent the colons
> in the eventual output, and then in your prefixDef, convert them to the
> colons:
>
> perseus:latinLit|stoa0234a.stoa001|2.53.8-2.53.12
>
> then your prefixDef can reconstruct the complete URL to access the
> relevant part of the target document.
>
> Does that make sense?
> Martin
>
>
> On 2016-06-13 02:04 PM, Paolo Monella wrote:
>> Il 13/06/2016 10:16, Lou Burnard ha scritto:
>>> At the risk of staying the obvious, the @source and ,@target attributes
>>> both take Uri values. So you could use them directly to supply your CTS
>>> values. No need for cref.
>>
>> Thank you, Lou -- it was useful to me. So I could just replace
>>
>> <ref cRef="urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.12"
>> type="source">
>>
>> with
>>
>> <seg source="urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.12"
>> type="source"> ?
>>
>> The current segmentation of my text (of my an unpublished IX century
>> manuscript) is:
>>
>> <body>
>>    <ab>
>>
>>      <ref>
>>        <w/>...
>>      </ref>
>>
>>      <ref>
>>        <w/>...
>>      </ref>
>>
>>      ...
>>    </ab>
>>    <ab>
>>
>>      <ref>
>>        <w/>...
>>      </ref>
>>
>>      ...
>>    </ab>
>> ...
>> </body>
>>
>> Where <ab>s mark thematically distinct portion of the text.
>>
>> (I tought of not using any <p> because there is no typographical
>> 'paragraph' in my text. So I can't use <div> either, because a hierarchy
>> like <body> / <div> / <w> would not be allowed.)
>>
>> Right now, below the <ab>-level, the whole text is segmented into a
>> number of <ref> elements, some with URNs like
>>
>> <ref cRef="urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.12"
>> type="source">
>>
>> and some which I am temporarily (illegally) marking as
>>
>> <ref cRef="unknown" type="source">
>>
>> Thus, I mark that I know the late-antiquity sources of some portions of
>> the text, but that I don't know the sources of others.
>>
>> I am now thinking that I could use this hierarchy instead:
>>
>> <body>
>>    <ab>
>>
>>      <seg source="urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.12">
>>        <w/>...
>>      </seg>
>>
>>      <seg source="unkown">
>>        <w/>...
>>      </seg>
>>
>>    ...
>>    </ab>
>>    <ab>
>>
>>      <seg source="urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.23">
>>        <w/>...
>>      </seg>
>>
>>    ...
>>    </ab>
>>    ...
>> </body>
>>
>> Going back in-topic on @cRef and @source:
>>
>> 1) Would <seg> plus @source be appropriate here?
>>
>> 2) Specifically, is @source appropriate to mark the relation between
>>   a) the text of a IX century Latin epitome (and mash-up) of
>> late-antiquity grammarians like Priscian and Donatus
>>   b) and the text of those grammarians?
>> The use of @source in the Guidelines seems to be more specific (the
>> source of a quotation, or the like), but I hear that @source is going to
>> go global.
>>
>> 3) I still need a solution for those passages that have an unknown
>> source. Neither @cRef or @source can accept a value like "unknown" (<seg
>> source="unkown">), as it's no URN or URI. A solution might be not
>> marking those chunks at all, but for processing ease I'd still like to
>> mark them with a <seg> (or <ref>) element. Any ideas?
>>
>> Thank you,
>> Paolo
>>
>>
>>>
>>> Sent from my Honor Mobile
>>>
>>> -------- Original Message --------
>>> Subject: Re: cRef usage
>>> From: Paolo Monella
>>> To: [log in to unmask]
>>> CC:
>>>
>>> I am using @cRef with CITE/CTS URNs to encode that a passage from the
>>> medieval manuscript I am encoding has a specific passage of Priscian as
>>> source:
>>> <ref cRef="urn:cts:latinLit:stoa0234a.stoa001:2.53.8-2.53.12"
>>> type="source">
>>>
>>> Should I switch to the <prefixDef> mechanism too?
>>>
>>> Paolo
>>>
>>>
>>> Il 12/06/2016 14:29, Lou Burnard ha scritto:
>>>> Yes, and no.... it rather depends whether "#vergil-aenide" is a valid
>>>> URI in your document.
>>>>
>>>> Sent from my Honor Mobile
>>>>
>>>> -------- Original Message --------
>>>> Subject: Re: cRef usage
>>>> From: Thierry Pellé
>>>> To: [log in to unmask]
>>>> CC:
>>>>
>>>> Hello,
>>>>          short, fast and lousy but useful ! ;-)
>>>>
>>>> Hence, if I understand well, I may define
>>>> <prefixDef ident="cref:"
>>>>   matchPattern="(.+?), (.+?)"
>>>>   replacementPattern="#$1-$2">
>>>>   <p>
>>>>          Replaces cRef references.
>>>>   </p>
>>>> </prefixDef>
>>>>
>>>> and then use
>>>> <quote ref="cref:Virgile, Enéide">
>>>>   ...
>>>> </quote>
>>>>
>>>> That sounds good.
>>>>
>>>> Thank you!
>>>> Thierry
>>>>
>>>> Le dimanche 12 juin 2016 à 08:00 -0400, Syd Bauman a écrit :
>>>>> [Fast-and-lousy response from an airport.]
>>>>>
>>>>> The @cRef mechanism has essentially been superceded by the
>>>>> <prefixDef> mechanism, which can be used on any pointer.
>>>>>
>>>>> >     Guidelines specifies that @cRef "specifies the destination of
>>>>> >     the pointer by supplying a canonical reference". I wonder why
>>>>> >     only gloss, ptr, ref, term provide such a canonical forme for a
>>>>> >     pointer. For instance, why is it not possible to specifies the
>>>>> >     @souce attribute of a quote as a canonical pointer?