Changes between Version 39 and Version 40 of Taskforces/FCS/FCS-Specification-Draft

Timestamp:

11/03/15 12:33:20 (9 years ago)

Author:

Oliver Schonefeld

Comment:

--

Taskforces/FCS/FCS-Specification-Draft

@@ -179 +179 @@
 || `diag`  || `http://docs.oasis-open.org/ns/search-ws/diagnostic`  || SRU Version 2.0 Diagnostics                              || prefixed ||
 || `zr`    || `http://explain.z3950.org/dtd/2.0/`                   || SRU/ZeeRex Explain                                       || prefixed ||
-|| `sru`   || `http://www.loc.gov/zing/srw/`                        || SRU Version 1.2, ''only comparability mode''             || prefixed ||
-|| `diag`  || `http://www.loc.gov/zing/srw/diagnostic/`             || SRU Version 1.2 Diagnostics, ''only comparability mode'' || prefixed ||
+|| `sru`   || `http://www.loc.gov/zing/srw/`                        || SRU Version 1.2, ''only compatibility mode''             || prefixed ||
+|| `diag`  || `http://www.loc.gov/zing/srw/diagnostic/`             || SRU Version 1.2 Diagnostics, ''only compatibility mode'' || prefixed ||
 
 = CLARIN-FCS Interface Specification
@@ -381 +381 @@
         <ed:Resource pid="http://hdl.handle.net/4711/0815">
             <ed:Title xml:lang="de">Goethe Korpus</ed:Title>
-            <ed:Title xml:lang="en">Goethe Corpus</ed:Title>
+            <ed:Title xml:lang="en">Goethe corpus</ed:Title>
             <ed:Description xml:lang="de">Der Goethe Korpus des IDS Mannheim.</ed:Description>
             <ed:Description xml:lang="en">The Goethe corpus of IDS Mannheim.</ed:Description>
@@ -596 +596 @@
  - organized in one or more annotation layers
  - annotation layer := annotations of a specific type, e.g. part-of-speech or orthographic transcription
- - annotations of two annotation layers may freely overlap; no self-overlap in an annotation layer
+ - annotations of two different annotation layers may freely overlap; no self-overlap in an annotation layer
  - Data View serialization in a stand-off like format, i.e annotations are ranges over the signal (= language resource as character, token or audio stream) are denoted by start and end offsets
  - layers alignable (through their offsets) and referable (trough their layer identifier)
@@ -602 +602 @@
    - a list of segments (= "inventory" of all ranges used to describe annotations")
      - units can be "items" (= offsets in character or token-stream) or "timestamp" (timestamps in audio-stream), timestamps may have a resolution of up to 1/1000 second.
+       - to come up with the correct offsets is up to the endpoint; it must do so in a consistent manner. a recommendation for character streams: character := Unicode codepoint, normalized to Unicode Normalization Form KC (NFKC; Compatibility Decomposition, followed by Canonical Composition)
      - segments may also have an endpoint specific reference (= URI); can be show in aggregator and if user clicks link can open a viewer (e.g. audio-player) at the endpoint
    - a list of layers, each has a type (e.g. "pos", "lemma", see Layer Type identifier in section Layers above) and an layer identifier (= URI)
      - a layer consists of one or more Spans. A span references a segment (and thus inherits the start- and en- offsets) and contains the actual annotation (e.g. the port-of-speech label) in it's content; MAY also contain alt-value (e.g. original annotation value)
+     - document order of layer elements define the view order in the Aggregator
+     - endpoints should return at least all layers that where referenced the query; they may return more
    - Hit Makers are added by marking Spans as hits (add `@highlight` attribute); multiple hit-makers are supported and Aggregator may display them visually distinct
+     - where to add hit markers is up to the endpoint; generally "things" that where referenced in the query should be marked.
+
 
 Example: a sentence interpreted as a character stream