Download presentation
Presentation is loading. Please wait.
Published byBrice Patterson Modified over 9 years ago
1
VOTable agenda Current VOTable status Current VOTable status News from Applications News from Applications Questions about VOTable schema Questions about VOTable schema Question of empty columns Question of empty columns Utype & STC Utype & STC Late topics Late topics Actions & Future Actions & Future
2
Where are we ? VOTable1.1 is 2 years old VOTable1.1 is 2 years old quite generally used, many tools quite generally used, many tools What’s missing: What’s missing: general referencing to external data models, especially STC general referencing to external data models, especially STC Final schema which satisfies the applications Final schema which satisfies the applications
3
News from Applications Heard about: Heard about: VOConvert v1.0 (successor of ConVOT) VOConvert v1.0 (successor of ConVOT) VOStat (beta version) (on Friday) VOStat (beta version) (on Friday) VOMegaPlot VOMegaPlot TOPCAT v2.2 TOPCAT v2.2 more applications support the different formats (TABLEDATA / FITS / BINARY) more applications support the different formats (TABLEDATA / FITS / BINARY) Some problems heard about input/output of VOTables apparently not fixed Some problems heard about input/output of VOTables apparently not fixed conservation of GROUP structures (TOPcat) conservation of GROUP structures (TOPcat)... ?... ?
4
VOTable schema changes Embedded spaces recommended as &32; or in the next document Embedded spaces recommended as &32; or in the next document Problem of code generators (xsd) where the class generated by TABLEDATA consists of TD[][] instead of a class TR[] where TR consists in TD[] Problem of code generators (xsd) where the class generated by TABLEDATA consists of TD[][] instead of a class TR[] where TR consists in TD[] Easy to solve if an attribute is added to Easy to solve if an attribute is added to acceptable ? acceptable ?
5
VOTable schema changes Question of recursivity (already discussed) Question of recursivity (already discussed) allowed in RESOURCE GROUP and OPTION allowed in RESOURCE GROUP and OPTION is useful, can describe trees instead of simple arrays, actually used in some applications is useful, can describe trees instead of simple arrays, actually used in some applications Accept ? Accept ?
6
Question of empty Question of empty How correct is (or ) ? How correct is (or ) ? Is a single representration (null or NaN) enough ? (FITS knows only one) Is a single representration (null or NaN) enough ? (FITS knows only one) In the TABLEDATA data representation, the default representation of a "null" value is an empty column (i.e. ); for fields containing arrays, individual "null" elements of the array can be specified either by the value specified in the null attribute, or by the "NaN" or "nan" text in place of the expected number.
7
(cont’d) (cont’d) Empty tag: Empty tag: concerns only the serialisation concerns only the serialisation Must be compatible with other serializations (FITS BINARY) Must be compatible with other serializations (FITS BINARY) The problem with FITS compatibility: The problem with FITS compatibility: NaN Floating-point values always possible (no possibility to force a column to contain actual values) NaN Floating-point values always possible (no possibility to force a column to contain actual values) No equivalent exists for integer (byte, short, int, long) i.e. the choice of a "special“ value is required in these cases No equivalent exists for integer (byte, short, int, long) i.e. the choice of a "special“ value is required in these cases
8
(cont’d) (cont’d) The VOTable document specifies that empty TDs are acceptable to represent NULLs – does not work for integer representations as well as arrays. Two possibilities: The VOTable document specifies that empty TDs are acceptable to represent NULLs – does not work for integer representations as well as arrays. Two possibilities: emtpy not allowed, use NaN emtpy not allowed, use NaN Accept widely used convention of null representation as the smallest integer (without positive counterpart): -32678=0X8000 (short) ; similarly int =0x80000000 long=0x8000000000000000L ? Accept widely used convention of null representation as the smallest integer (without positive counterpart): -32678=0X8000 (short) ; similarly int =0x80000000 long=0x8000000000000000L ?
9
Utypes: what is it ? Non-mandatory attribute of RESOURCE TABLE FIELD PARAM GROUP Non-mandatory attribute of RESOURCE TABLE FIELD PARAM GROUP originally created for DAL needs originally created for DAL needs is a possible attribute in VOTable1.1 wherever the ucd is an acceptable attribute is a possible attribute in VOTable1.1 wherever the ucd is an acceptable attribute contrary to the ucd, gives a fully detailed meaning of the field, parameter or group contrary to the ucd, gives a fully detailed meaning of the field, parameter or group ucd = broad semantics, typically used for data mining ucd = broad semantics, typically used for data mining utype = detailed semantics, refers to a data model utype = detailed semantics, refers to a data model
10
Utype: its usage in VOTable can supply an exact description of the column contents can supply an exact description of the column contents immediate application for a referencing of parameters which exact meaning is crucial immediate application for a referencing of parameters which exact meaning is crucial systems of coordinates: celestial, terrestrial, solar,... (connection with STC) systems of coordinates: celestial, terrestrial, solar,... (connection with STC) time definitions (connection with STC) time definitions (connection with STC) photometric systems & bandpasses photometric systems & bandpasses more generally any parameter in a model, simulation... more generally any parameter in a model, simulation...
11
Status after Victoria meeting At El Escorial: proposal of using in the utype attribute an XPath analogy was refused At El Escorial: proposal of using in the utype attribute an XPath analogy was refused DAL way ssa:Sky.Resolution DAL way ssa:Sky.Resolution a simple rule is defined to derive the utype contents (concatenation of elements' names with dots) a simple rule is defined to derive the utype contents (concatenation of elements' names with dots) define the recommended utype in the model ? define the recommended utype in the model ? should converge to common solutions across VO components should converge to common solutions across VO components
12
Scenario #1 for STC Use a with eters to define the reference system, as in this example (following Jonathan’s document “Cordinates Data Model“ 2006-09-12)
13
Scenario #2 for STC As in scenario #1, define the reference system in a with eters, but use the ref= mechanism to specify the frame – see this examplethis example (following Jonathan’s document “Cordinates Data Model“ 2006-09-12)
14
Scenario #3 for STC As in scenario #2, but closer to Arnold’s definition of STC – see the exampleexample
15
…conclusion ? #1 = GROUP of PARAM’s + FIELDref’s describing the coordinates #1 = GROUP of PARAM’s + FIELDref’s describing the coordinates #2 = GROUP defining the coordinate system; closer to the former ; FIELDs representing position components use the ref mechanism to indicate the coordinate system #2 = GROUP defining the coordinate system; closer to the former ; FIELDs representing position components use the ref mechanism to indicate the coordinate system #3 = similar to #2, closer to original STC definitions #3 = similar to #2, closer to original STC definitions
16
Sexagesimal representations Came out again during this week-end! Came out again during this week-end! Definition of a sexagesimal number: number edited with 1 or 2 `decimals’ in base 60; further decimals in base 10 if needed. Definition of a sexagesimal number: number edited with 1 or 2 `decimals’ in base 60; further decimals in base 10 if needed. Solution adopted at CDS: "quoted" units Solution adopted at CDS: "quoted" units (apparently adoped in TOPcat) Accept or discourage this multi-millenia usage ? Accept or discourage this multi-millenia usage ? should applications consuming VOTables be able to deal with sexagesimal representations? Or should VOTable promote decimal degrees (or radians) for angles, decimal seconds, hours, or days for time ? should applications consuming VOTables be able to deal with sexagesimal representations? Or should VOTable promote decimal degrees (or radians) for angles, decimal seconds, hours, or days for time ?
17
ref attribute to TABLE Current definition: Current definition: A TABLE may have a ref attribute referencing the ID of another table previously described, which is interpreted as defining a table having a structure identical to the one referenced: this facility avoids a repetition of the definition of tables which may be present many times in a VOTable document.
18
ref attribute to TABLE (cont’d) Question raised: should the ref attribute quote the table structure or the full table ? Question raised: should the ref attribute quote the table structure or the full table ? 1 2 1.1 2.1
19
Actions & Future Toward a VOTable1.2: Toward a VOTable1.2: Write down the utype conventions, and deprecate COOSYS Write down the utype conventions, and deprecate COOSYS clarify the NULL/NaN conventions clarify the NULL/NaN conventions clarify other possible controversies (ref attribute) clarify other possible controversies (ref attribute) VOTable1.2 schema – testing volunteers ? VOTable1.2 schema – testing volunteers ? And after adoption of VOTable1.2, continuation of this WG ? And after adoption of VOTable1.2, continuation of this WG ?
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.