ViewVC logotype

Contents of /trunk/projects/theory/snapdm/doc/note/SimDB-note.html

Parent Directory Parent Directory | Revision Log Revision Log

Revision 434 - (show annotations)
Fri May 9 16:16:54 2008 UTC (12 years, 11 months ago) by gerard.lemson
File MIME type: text/html
File size: 52882 byte(s)
Further updates
1 <?xml version="1.0" encoding="iso-8859-1"?>
2 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
3 <html>
4 <head>
5 <title>IVOA Working Group - Internal Draft</title>
6 <meta http-equiv="content-type" content="text/html; charset=iso-8859-1" />
7 <meta name="keywords" content="IVOA, International, Virtual, Observatory, Alliance" />
8 <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" />
9 <meta name="maintainedBy" content="IVOA Document Coordinator, ivoadoc@ivoa.net" />
10 <link rel="stylesheet" href="http://ivoa.net/misc/ivoa_wg.css" type="text/css" />
11 <link rel="stylesheet" href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/css/simdb-note.css" type="text/css">
12 </head>
14 <body>
15 <div class="head">
16 <a href="http://www.ivoa.net/"><img alt="IVOA" src="http://www.ivoa.net/pub/images/IVOA_wb_300.jpg" width="300" height="169"/></a>
17 <h1>Simulation Database (SimDB)<br/>
18 Version 0.x</h1>
19 <h2>IVOA Theory Interest Group <br />Internal Draft 2008 April 19 </h2>
22 <dt>This version:</dt>
23 <dd><a href="http://www.ivoa.net/Documents/...">
24 http://www.ivoa.net/Documents/...</a></dd>
26 <dt>Latest version:</dt>
28 <dd><a href="http://www.ivoa.net/Documents/latest/...">
29 http://www.ivoa.net/Documents/latest/...</a></dd>
31 <dt>Previous versions:</dt>
32 <dt>Interest Group:</dt>
33 <dd><a href="http://www.ivoa.net/twiki/bin/view/IVOA/IvoaTheory"> http://www.ivoa.net/twiki/bin/view/IVOA/IvoaTheory</a></dd>
34 <dt>Author(s):</dt>
35 <dd><a href="http://www.ivoa.net/twiki/bin/view/IVOA/GerardLemson">Gerard Lemson</a> (editor)<br /></dd>
36 <dd><a href="http://www.ivoa.net/twiki/bin/view/IVOA/LaurentBourges">Laurent Bourges</a><br /></dd>
37 <dd><a href="http://www.ivoa.net/twiki/bin/view/IVOA/PatriziaManzato">Patrizia Manzato</a><br /></dd>
38 <dd><a href="http://www.ivoa.net/twiki/bin/view/IVOA/RickWagner">Rick Wagner</a><br /></dd>
39 <dd>others?</dd>
40 <hr/></div>
42 <h2><a name="abstract" id="abstract">Abstract</a></h2>
43 <p>In this note we propose that the IVOA develop a standard protocol for discovering simulations.
44 We will call this protocol the <i>Simulation Database</i> (SimDB). Implementations of the SimDB will allow users to query for
45 results of simulations in quite some detail and will provide links to services for accessing these
46 simulations. </p>
47 <p>The results presented in this note, which form the core of the peoposed standard, are one half of a concerted effort of the theory Interest Group that originally went by the name
48 S<i>imple Numerical Access Protocol</i> (SNAP), and is now split up in two parts. The second part defines protocols
49 for accessing the simulations data products themselves. This part will be written up in a separate Note
50 (Gheller, Wagner et al, in preparation), under the name Simulation Data Access Protocol (SimDAP).
51 </p>
52 <p>The current proposal is built around a UML data model describing simulations, a representation (mapping) of this model as a relational
53 database schema and a mapping to an XML schema.
54 We propose the relational schema to be the outer facade of a SimDB-TAP implementation which is to be queried using
55 <a href="http://www.ivoa.net/internal/IVOA/IvoaVOQL/ADQL-20080415.pdf">ADQL</a> <em class="todo">.@@ TODO update the ADQL link to later versions @@</em>
56 The XML schema provides type definitions from
57 which a machine readable serialisations of the model may be constructed. The schema also defines root elements for documents
58 describing SimDB-resources. The SimDB should return such documents for identified SimDB-Resources upon request, as an
59 alternative to the tabular (VOTable) results of ADQL queries.
60 In case updates are supported by a SimDB implementation, such documents may be sent
61 </p>
62 <p>
63 This Note describes use cases and requirements and the approach we have taken to define a specification
64 that and current state of the results. We feel that the results are
65 sufficiently far evolved that they can start following the formal IVOA standardisation track.
66 To this end it could be turned over to one of the existing working groups. If that is the decisions we feel
67 that the data modelling WG is closest to its scope, but there exist very strong links to Registry, Semantics, ADQL
68 and DAL as well. One might argue that a targeted WG for this effort alone might be as appropriate.
69 We leave the decision about this to the IVOA exec.
70 </p>
74 <div class="status">
75 <h2><a name="status" id="status">Status of this Document</a></h2>
76 This is a Note. The first release of this document was 2008 April 19.
77 <p></p><br />
79 <!-- Choose one of the following (and remove the rest)-->
80 <!--Note-->
81 <p>This is an IVOA Note expressing suggestions from and opinions of the authors.<br/>
82 It is intended to share best practices, possible approaches, or other perspectives on interoperability with the Virtual Observatory.
83 It should not be referenced or otherwise interpreted as a standard specification.</p>
86 A list of <a href="http://www.ivoa.net/Documents/">current IVOA Recommendations and other technical documents</a> can be found at http://www.ivoa.net/Documents/.
88 </div><br />
90 <h2><a name="acknowledgments" id="acknowledgments">Acknowledgments</a></h2>
91 <p>We thank various persons for useful discussions in the course of this work. First the participants of the
92 <a href="http://www.ivoa.net/cgi-bin/twiki/bin/view/IVOA/CambridgeTheoryWorkshopFeb06">Feb 2006 theory
93 workshop</a> in Cambridge, UK, where this work was started. Second the participants of the
94 <a href="http://www.ivoa.net/cgi-bin/twiki/bin/view/IVOA/GarchingSNAPWorkshop200704">April 2007 SNAP workshop</a> in
95 Garching, Germany, where the design started taking shape. Then we want to thank particularly the following persons
96 for useful discussions and feedback: Jeremy Blaizot, Klaus Dolag, Ray Plante, Volker Springel. We finally want to thank
97 participants to the theory sessions in the interoperability meetings in Victoria, Moscow, Beijing and Cambridge where parts
98 of this work was discussed.
99 </p>
100 <h2><a id="contents" name="contents">Contents</a></h2>
101 <div class="head">
102 <ul class="toc">
103 <li><a href="#abstract">Abstract</a></li>
104 <li><a href="#status">Status</a></li>
105 <li><a href="#acknowledgments">Acknowledgements</a></li>
106 <li><a href="#contents">Contents</a></li>
107 <li><a href="#sec1">1. Executive Summary</a></li>
109 <li><a href="#sec2">2. Overview</a></li>
110 <ul class="toc">
111 <li><a href="#sec2_1">2.1 SNAP &rArr; SimDB + SimDAP</a></li>
112 <li><a href="#sec2_3">2.3 Simulation Database: structure and interface</a></li>
113 <li><a href="#sec2_3">2.3 Registration</a></li>
114 <li><a href="#sec2_4">2.4 Technology: UML, XMI, XSLT</a></li>
115 <li><a href="#sec2_5">2.5 Reference implementations</a></li>
116 </ul>
119 <li><a href="#sec3">3 Usage scenarios</a></li>
120 <ul class="toc">
121 <li><a href="#sec3_1">3.1 "20 questions"</a></li>
122 <li><a href="#sec3_2">3.2 SimDB-standard implementation</a></li>
123 <li><a href="#sec3_3">3.3 Legacy database</a></li>
124 <li><a href="#sec3_4">3.4 Meta data production pipe line</a></li>
125 <li><a href="#sec3_5">3.5 Client tools</a></li>
126 </ul>
128 <li><a href="#sec4">4 Analysis model</a></li>
129 <ul class="toc">
130 <li><a href="#sec4_1">4.1 Universe of Discourse</a></li>
131 <li><a href="#sec4_2">4.2 <i>Domain Model for Astronomy</i></a></li>
132 <li><a href="#sec4_3">4.3 SimDB analysis model</a></li>
133 </ul>
135 <li><a href="#sec5">5 Logical model</a></li>
136 <ul class="toc">
137 <li><a href="#sec5_1">5.1 Overview</a></li>
138 <li><a href="#sec5_2">5.2 Normalisation</a></li>
139 <li><a href="#sec5_3">5.3 Target</a></li>
140 <li><a href="#sec5_4">5.4 Characterisation</a></li>
141 <li><a href="#sec5_5">5.5 Semantics</a></li>
142 </ul>
144 <li><a href="#sec6">6 Physical models</a></li>
145 <ul class="toc">
146 <li><a href="#sec6_1">6.1 Identifiers and references</a></li>
147 <li><a href="#sec6_2">6.2 RDBM Schema</a></li>
148 <li><a href="#sec6_3">6.3 XML Schema</a></li>
149 <li><a href="#sec6_4">6.4 Identifiers</a></li>
150 <li><a href="#sec6_5">6.5 JAVA/JPA+JAXB (non-normative)</a></li>
151 </ul>
153 <li><a href="#sec7">7. Query protocols</a></li>
154 <ul class="toc">
155 <li><a href="#sec7_1">7.1 ADQL</a></li>
156 <li><a href="#sec7_2">7.3 REST</a></li>
157 <li><a href="#sec7_3">7.2 TAP?</a></li>
158 </ul>
160 <li><a href="#sec8">8. Next steps</a></li>
161 <ul class="toc">
162 <li><a href="#sec8_1">8.1 Reference implementations</a></li>
163 <ul class="toc">
164 <li><a href="#sec8_1_1">8.1.1 France</a></li>
165 <li><a href="#sec8_1_2">8.1.2 Germany</a></li>
166 <li><a href="#sec8_1_3">8.1.3 Italy</a></li>
167 <li><a href="#sec8_1_4">8.1.4 USA</a></li>
168 </ul>
169 <li><a href="#sec8_2">8.2 SimDAP services</a></li>
170 </ul>
171 <br/>
172 <li><a href="#appA">Appendix A: Data modelling specifics</a></li>
173 <li><a href="#appB">Appendix B: XSLT pipe line</a></li>
174 <li><a href="#glossary">Glossary and Acronyms</a></li>
176 <li><a href="#references">References</a></li>
177 </ul>
178 </div>
179 <hr/>
182 <br/>
183 <h2><a name="sec1">1. Executive summary</a></h2>
184 <em class="todo">@@ TODO Modify this text, which was originally an email to be sent to THEORY, TCG, DM, maybe EXEC @@</em>
185 <p>
186 We propose to derive two WG projects from what was so far the
187 SNAP project of the theory interest group: SimDB and SimDAP.
188 In this note we discuss the first of these, SimDB, in some detail.
190 </p>
191 <h3> Simulation Database (SimDB)</h3>
192 <p>We propose to developa standard specification project, called the "Simulation Database" (SimDB).
193 It is based on the description+discovery part of the old
194 SNAP project. Its normative deliverables are
195 <ul>
196 <li> A logical data model for describing simulations.<br/>
197 Following SNAP we keep concentrating
198 on 3+1D simulations, with which we mean simulations modelling a
199 space-time sub-volume of the universe OF ANY SIZE, so not only large
200 scale structure, galaxy clusters, but everything down to asteroid collisions etc.
201 As the model <i>describes</i> simulations, it may be called a meta-data model.
202 It will be a logical model in the sense of standard data modelling approaches <em class="todo">@@TODO add some references@@</em>,
203 and is based on an analysis, or domain model which is presented but not normative.
204 The logical model is presented in fully detailed and documented UML2, serialised
205 to XMI 2.1, created using the MagicDraw 12.1 Community edition tool.
206 The data model is using a small subset of UML2 and has some UML profile
207 extensions added. Together this can be seen as a domain specific language,
208 and this can be formalised in a UML Profile. We will propose using such a profile
209 to the DM working group as a general approach for DM efforts.
210 </li>
211 <li>A query protocol based on the logical model.
212 <br />We propose this to have at least an ADQL version.
213 To this end we will provide a relational mapping.
214 This physical model is completely derived from the SimDB logical model using rules
215 implemented as a pipe-line of XSLT2 scripts working on the XMI representation of
216 the UML. The scripts will produce relational database DDL scripts defining the
217 database schema. That schema itself is not normative, instead we will define the
218 replies to TAP metadata queries. We provide implementaiton scenarios in the text below,
219 for the case of someone using the results from this project completely and for the
220 case of someone implementing a SimDB on top of a legacy database.
221 </li>
222 <li> a messaging format for sending instances of the various components
223 in the data model around.
224 <br />This format will be based on a number of XML
225 schema documents (XSDs), one of which contains the root elements defining valid SimDB resources.
226 This requires a mapping from the UML to XSD.
227 This mapping will take the form of one or more XSLT documents.
228 </li>
229 <li> An IVOA working draft document describing these components.
230 <br />This will be based on the current document.</li></ul>
231 </p>
232 <p>
233 We introduce some non-normative solutions that can be taken over for generic
234 data models (this is ofcourse also true for the UML/XMI+XSLT approach for the
235 normative standards).
236 <ul>
237 <li> The XSLT scripts we propose above do not work on the XMI itself, but on
238 an intermediate representation of the UML data model. This is an XML dialect
239 based on a schema we define and which captures the UML profile more directly.
240 XMI is very generic and rather cumbersome to work with. The representation of
241 the UML in our intermediate XML form is much more readable and XSLT based on it
242 is much simpler. It also allows easier adaptation to future modifications in UML,
243 or to tools whose XMI representation is different from the standard. We only need
244 to update the XMI->Intermediate XSLT transformation scripts. Not the more complex
245 transformations to the other official representations.
246 We will propose a similar approach to the DM WG.
247 </li>
248 <li> We will provide XMI->Java+JPA+JAXB transformation scripts in XSLT (properly, intermediate->Java).
249 These scripts generate Java classes corresponding to the types (Class, DataType, Enumeration)
250 in UML. These classes are annotated with Java Persistence Architecture (JPA)
251 and Java Architecture for XML Binding (JAXB) attributes to assist in the transformation
252 between relational database and XML representations.
253 Similar scripts can be written for C#. C# allows the same annotations as Java 5 supports
254 already for longer. For persistence we will likely use Linq, which seems similar to JPA.
255 </li>
256 <li>We propose an approach for including application specific and legacy simulation databases
257 in this framework. This approach follows the "global-as-view" approach to information
258 integration (see for example http://www.deg.byu.edu/papers/PODS.integration.pdf;
259 Leonid Kalinichenko from the RVO is an expert in this field).
260 Implementors with an existing relational database schema may be able to define database
261 views which implement the relational representatiopn of the SimDB data model,
262 and in this way provide a simple way to support querying of their database using ADQL.
263 </li></ul></p>
264 <h4>organisation</h4>
265 <p>
266 The SimDB is ready to be transferred to the DM WG.
267 <br />We propose that Gerard Lemson keeps leading this effort (as main editor), also when it is moved
268 to that WG. The DM WG's chair (Mireille Louys) will be responsible all WG-chair
269 issues associated with moving a specification through the document process.
270 The people at the bottom will be part of a "tiger team" to push the standard to RFC.
271 We may want to expand this group with an expert from each of the WGs mentioned below.
272 </p>
273 <p>
274 We have been discussing the data model for some time now.
275 Various projects (Italy, USA, France and Germany) have implementations that are similar
276 to the envisioned SimDB. We believe that by autumn 2008 it can go to RFC.
277 Patriza Manzato and Rick Wagner will have reference implementations based on existing DBs,
278 so will various projects in France (Lyon: Jeremy Blaizot and Laurent Bourges;
279 Galmer database: Igor Chillingarian) and GAVO.
280 </p>
281 <p>
282 Other relevant working groups for this process are Registry, ADQL and Semantics, possibly DAL.
283 Registry because the simulation database is similar to a registry. We can
284 learn from implementations and the registry interface. Also, we (think we) may need an
285 extension to the IVO Identifier in the implementation of references in SimDB.
286 ADQL because we propose it to be the standard (main) query interface to a SimDB implementation.
287 Semantics because our model includes usage of semantic vocabularies, maybe full ontologies
288 DAL because we our proposal for using ADQL in the query phase requirs a version of
289 the TAP protocol for defining the interface.
290 We would like to include a person from each of these WGs in the tiger team.
291 Our wishes are: Ray Plante (Registry), ? (ADQL), Norman Gray (Semantics), (?) TAP.
292 Ray and Norm have contributed to early discussions about SNAP.
293 </p>
294 <p>
295 Of these other efforts it seems TAP offers the main risk for the SimDB standard to go to
296 RFC by the Autumn. What may help us is that we do not need all the details of TAP.
297 In particular the information_schema approach allowing users to
298 query for the data model is not required as it is part of SimDB specification.
299 We mainly need a prescription for sending ADQL queries to the SimDB, and what the
300 format of results should be.
301 Since we expect meta-data databases to be relatively small (compared to
302 say an SDSS or Millennium database), we expect fewer, if any problems with
303 performance and can stick to synchronous behaviour at first.
304 </p>
305 <p>
306 We may need some explicit registry-interface like features such as returning a
307 complete XML document according to the messaging format of the SimDB data model.
308 Other issues will come up during the next phase of the discussions.
309 </p>
311 <h3>Simulation Data Access Protocol (SimDAP)</h3>
312 <p>
313 The second spin-off of the SNAP project we propose we rename to <i>Simulation Data Access Protocol</i> (SimDAP).
314 It deals with accessing the data after discovery by some means,
315 likely trough an implementation of a Simulation Database.
316 It should handle special services such as cut-out, projection,
317 extraction (AMR-like cut-outs, produces regular grids), but also staging etc.
318 It should also deal with data formats. Claudio Gheller (Italy) is leading
319 this effort with close help of Rick Wagner (USA).
320 </p>
321 <p>
322 This project needs more fleshing out and is hopefully ready to be transmitted
323 to a WG, likely DAL by the Autumn interop.
324 </p>
325 <h3>Connections between SimDB and SimDAP</h3>
326 <p>
327 The two projects are connected as follows:
328 The meta-data formats to be included in SimDAP messages are derived from
329 the data model of the SimDB.
330 Vice versa, the SimDB will include a component describing
331 which SimDAP services are applicable/available for a given simulation.
332 </p>
334 <!-- ++++++++++++++++++++++++ -->
335 <h2><a name="sec2"/> </a>2 Overview</h2>
337 <h3><a name="sec2_1"/>2.1 SNAP &rArr; SimDB + SimDAP</h3>
338 <p>This document presents a model for describing certain types of numerical computer simulations
339 and certain types of simulation post-processing products. The model was oringinally envisioned to
340 be used in the query part of the <i>Simple Numerical Access Protocol</i> (SNAP),
341 and in discovery of interesting SNAP services in the first place.
342 After investigating the application domain carefully, we have decided to leave the concept of
343 designing a DAL-like SxAP protocol for simulations. Instead we have split up the effort into
344 two separate efforts that can be used each in their own right, though their is a clear link between them.
345 This document discusses the firsts of these, which we have named the <i>Simulation Database</i>, and
346 will have the acronym <i>SimDB</i>. The second will be developed further in a separate effort amd is
347 called the <i>Simulation Data Access Protocol</i> (SimDAP, "Sim" stands for "Simulation", <i>not</i> "Simple"!).
348 </p>
349 <p>
350 Following SNAP, SimDB only explicitly considers simulations for systems that represent a space-time
351 sub-volume of the universe and (part of) its material contents. Examples of such simulations are
352 cosmological, pure dark matter N-body simulations of the large-scale structure of the universe;
353 adaptive mesh refinement (AMR) simulations following the evolution of a galaxy cluster using full hydrodynamics;
354 a simulation of the evolution of a globular cluster using a combination of tools, together simulating
355 the various types of physics <em class="todo">@@ TODO reference to MODEST-like activities</em>; or
356 simulations calculating the few seconds of a super nova explosion in full 3D.
357 </p
358 <p>
359 In general these simulations will evolve this system forward
360 in time and are able to produce <i>snapshots</i>, representing the state of the system, a 3D volume of space,
361 at a number of discrete times (though there are alternatives: light cone simulations, individual particle orbits).
362 These direct, raw results of simulations we call Level-0 products, following
363 similar terminology for observations.
364 SimDB also covers Level-1 products, which consist of the results of certain types of post-processing
365 of simulations, namely those products that in some form create an alternative representation of
366 a spatial sub-volume of the universe. For example a density field calculated on a regular grid, derived
367 created from an N-body or an AMR simulation; a cluster catalogue derived using some group finder applied
368 to a cosmological simulaiton, or a synthetic galaxy catalogue derived from the cluster catalogue using
369 halo occupation distribution models (HODs) or semi-analytical models (SAMs).
370 </p>
371 We do not make any restrictions on the type of systems being simulated, or the size of the
372 simulation, or the way the system is represented in the simulation code and results. We also
373 make no restrictions on the type of "observables" produced by the simulations.
374 </p>
375 <p>
376 The SimDAP
377 specification will includes protocols for services that process level-0 or level-1 results and produce
378 other level-1 results. The allowed services deal with selecting the results in a
379 sub-volume of the complete result, sampling a regular 3-dimensional grid, etc. SimDAP also allows for
380 services, that do not produce SimDB-like, level-0 or 1 products. Examples are projections, 1D or 2D samplings.
381 But also custom services will be allowed, for example calculating statistical properties such as correlation
382 functions or power spectra in cosmological simulations. A more detailed description of SimDAP
383 is outside of the main scope of this note.
384 </p>
385 <h3><a name="sec2_2"/>2.2 Simulation Database: structure, interface and applicable services</h3>
386 <p>
387 SimDB is a specification that defines the interface to a database containing meta data describing
388 simulations. To this end it contains two main parts, one is a model for the meta data, the other
389 a protocol for interacting with the database. The model is the core of the specification.
390 It describes the structure of individual data products in the database. We have chosen UML
391 as modelling language, as prescribed by the data modelling working group in the interoperability meeting
392 in Cambridge, UK, May 2003.
393 </p>
394 <p>
395 The UML model is a logical model (see [..] <em class="todo">@@ TODO add reference @@</em>) and
396 forms the basis for physical representations of the data products in the standard
397 language that the IVOA has chosen for such purposes, XML. We derive an XML schema defining valid
398 XML documents directly from the logical model. The SimDB interface will include functions for insetting
399 SimDB data products using such documents, and for retrieving individual, identified data products.
400 </p>
401 <p>
402 The logical model also forms the basis for a physical representation supporting formulation of queries.
403 For various reasons explained below we have chosen ADQL to be the query language and accordingly we derive
404 from the model a relational schema that defines the tables and columns that can be used in ADQL queries sent
405 to a SimDB implementation. The result of ADQL queries is supposed to be a VOTable, and this will in general
406 not represent a complete SimDB data product. However it can be used to browse the database, finally identifying
407 resources and possibly requesting these from the SimDB as XML documents.
408 </p>
409 <p>
410 We make very limited assumptions on <em>how</em> a data product discovered in a SimDB can actually be accessed.
411 We only assume there is a web-based service available, identified by a base URL and tagged with a service type.
412 The range of service types will be defined by SimDAP, but it will at least include "download" and "custom".
413 The data model contains an explicit element for indicating which services are available for a given data product,
414 and users may, if they wish, retrieve this information through ADQL queries and follow the links directly.
415 SimDB implementations can and likely will eventually provide SimDAP related functionality, but this is not part
416 of this specification.
417 </p>
418 <h3><a name="sec2_3"/>2.3 Registration</h3>
419 <p>
420 It must be possible to find SimDB instances in an IVOA Resource Registry <am class="todo">@@TODO add references&&</am>.
421 This implies we need a corresponding resource type, and we have to design its structure.
422 We also assume that one may define resources in the sense of [...]
423 <em class="todo">@@ TODO add reference to Resource data model document @@</em>
424 from within the contents of a SimDB. We take this into account explicitly in the model.
425 The SimDB will have a "getIVOAResource" function, which will execute the appropriate transformation from
426 the internal representation of the SimDB data products to the Resource model's XML representation [...]
427 <em class="todo">@@ TODO link to Resource XML schema document@@</em>.
428 This will likely put more requirements on the Registry model itself, maybe requiring extensions to its schema.
429 Possibly a SimDB itself can be an extension registry. This we think can be postponed to a future version of the
430 specification.
431 </p>
432 <h3><a name="sec2_4"/>2.4 Technology: UML, XMI, XSLT</h3>
433 <p>
434 We
435 </p>
436 <h3><a name="sec2_5"/>2.5 Reference implementations</h3>
437 <!-- ++++++++++++++++++++++++ -->
439 <h2><a name="sec3"/>3 Usage scenarios</h2>
440 <em class="todo">@@ TODO needs severe editing @@</em>
441 We have assembled a list of explicit use cases and scenarios from which we derive
442 requirements for the current model and the SNAP protocol.
443 <h4><a name="sec3_1"/>3.1 "20 questions"</h4>
444 <p>
445 SimDB defines a common data model for simulations.
446 Following the good practice for database design initiated in [], we here provide a number of
447 scientific questions one might want to ask such a database. The data model and associated data
448 access protocol need to be sufficiently rich that they can support such questions.
449 </p>
450 <ul>
451 <li> Scientific goal: investigate baryon wiggles in the evolved density field<br/>
452 Query: Return all cosmological, pure dark matter, N-body simulations with WMAP 3 initial
453 conditions and a box size of at least 1000 Mpc comoving, containing snapshots at about
454 10 redshifts between 3 and 0.
455 </li>
456 <li> Scientific goal: investigate whether observed structures in X-ray cluster that seem to
457 indicate turbulence, can truly be that.<br> Query: return all hydro-dynamical simulations of
458 galaxy clusters of mass at least 1o<sup>14</sup> M<sub>sun</sub>,
459 that have a model for viscosity included in the simulation.
460 Moreover, return only those simulations that have associated to them an online visualisation
461 service that can produce projected temperature and pressure maps.
462 </li>
463 <li> Scientific goal: interpret the possible histories of an observed galaxy merger to calculate
464 possible star formation episodes and compare these to the observed stellar populations.<br>
465 Query: Return all simulations of galaxy mergers where the component galaxies have a particular
466 mass ratio and where there are enough snapshots to follow the evolution over a few Gyr.
467 </li>
469 <li> Scientific goal: compare the luminosity function of galaxies in the SDSS survey with those
470 in synthetic catalogues.<br>Query: Select all cosmological simulations that have produced as
471 secondary product synthetic galaxy catalogues on a light-cone and provide those via an SQL (ADQL?)
472 query interface.
473 </li>
474 <li> ...
475 </li>
476 </ul>
477 <p>
478 In the design of the model it is useful to think about the steps a user might go through
479 when querying a database system in various "drilling down" steps. For example the following
480 questions might be asked :
481 </p>
482 <ul>
483 <li>What system/object is being simulated?</li>
484 <li>What physical processes are included?</li>
485 <li>How is the system being represented in the simulation
486 (particles (Langrangian), (adaptive) mesh (Eulerian)), both, other?</li>
487 <li>Per process:<ul>
488 <li>How are the physical processes implemented ?</li>
489 <li>Characterise the numerical approximations (.e.g. resolution, softening parameter)</li></ul></li>
490 <li>What observables are available for the system/object, possibly as function of time?
491 As it is a spatial system, at least size, center-of-mass position.</li>
492 <li>What observables are available for the constituents, i.e. what is the schema of the atomic objects?</li>
493 <li>Per snapshot, per atomic object type, per variable:
494 <ul>
495 <li>Characterise the possible values</li>
496 <li>Characterise the result</li></ul></li>
497 <li>Are post-processing results available?</li>
498 <li>Are services/applications available working on the results?</li>
499 <li>Which code ran the simulation?</li>
500 <li>What were values of physical parameters?</li>
501 <li>How were initial conditions created, what parameters?</li>
502 </ul>
503 </p>
505 <h4><a name="sec3_2"/>3.2 SimDB-standard implementation</h4>
506 We foresee a simple implementation scenario based directly on products developed
507 in the course of the SimDB effort. We believe that from the data model to be developed
508 in this effort we should be able to derive physical representations that
509 can be used directly in implementations. We envisions that with only a little custom infrastructure code
510 it should be possible to
511 <ul>
512 <li>fill a relational database with tables and views representing the SimDB data model from
513 DDL scripts generated from the UML</li>
514 <li>create a web-based service that accept XML documents for inserting new simulation results
515 and translates these, using generated code with JAXB annotations, to in memory Java objects</li>
516 <li>flush these objects to a relational database using the Java Persistence Architecture (JPA) implementation,
517 structured using the JPA annotations generated on the Java classes.
518 It should be not too hard to support other languages as well if they provide similar simple XML binding and
519 OR-mapping capabilities. Python+Django and C#+LINQ or NHibernate come to mind.<em class="todo">
520 @@ TODO check with people knowing more about these technologies @@</em></li>
521 <li>accept ADQL queries that are translated to the appropriate vendor specific SQL
522 (using modules defined by the ADQL effort?) and return a VOTable</li>
523 <li>accept requests for identified SimDB resources (using an IVO or implementation specific identifier),
524 translate this into a JPA query to retrieve the object form the database, which is translated to
525 the appropriate XML using the JAXB layer and sent back to the user.</li>
526 </ul>
528 <h4><a name="sec3_3"/>3.3 Legacy database</h4>
529 Although by no means as common as similar efforts in the observational domain,
530 databases have been developed containing the meta data of simulations.
531 How could a SimDB be implemented around such a database.
532 Our ideas are inspired by (what we understand from) the "global-as-view" approach to information
533 integration. We assume the implementers have their own way of filing up their database with meta-data
534 describing simulations from their own efforts. The idea is that they write database views to provide
535 a virtual implementation of the SimDB/RDB schema. ADQL queries sent to their service can now still be
536 understood and replied to. The users should also be able to write custom code to produce the appropriate
537 XML documents based on a request for an identified resource, possibly by querying these same views.
539 <h4><a name="sec3_4"/>3.4 Meta data production pipe line</h4>
540 The SimDB data model is relatively comprehensive, which reflects itself in XML documents
541 of substantial size ad complexity for realistic cases.
542 For a registration scenario, i.e. one where a user is allowed to upload XML documents to a SimDB implementation,
543 one would prefer not to have to produce these documents by hand. By far the preferred manner in our opinion
544 would be for simulation and post-processing pipe-lines to produce compliant documents.
545 We have contacted authors of some of the most popular major simulation codes (Springel; Norman et al; more needed),
546 and they have agreed that this is feasible and are willing to participate in this effort.
548 <h4><a name="sec3_5"/>3.5 Client tools</h4>
549 One reason to produce a standard which uses ADQL on top of a standard data model is that client tools
550 can be written to query different such holdings. For example we could envision a tool such as VisIVO [..]
551 to offer some user-friendly interface for querying SimDB implementations retrieved from an IVOA Registry.
552 The user need to see any ADQL, that is all generated by VisIVO, but can be shown results and services.
553 In particular if a cut-out service is available, VisIVO could provide an interface for the user to decide
554 on the sub-volume, retrieve and visualise it. The advantage of having a standard data model
555 clearly is that the same ADQL can be sent to all SimDB services.
556 <em class="todo">@@ TODO contact VisIVO people to see whether this could be implemented @@</em>.
558 <!-- ++++++++++++++++++++++++ -->
561 <h2><a name="sec4"/>4 Analysis model</h2>
562 <em class="todo">@@TODO Gerard@@</em>
563 An <i>analysis model</i>, also called domain model, is an abstract, high-level representation of the
564 <i>universe of discourse</i> (UoD), the part of the world that our application deals with.
565 It is a UML model, with emphasis on the concepts and their exact relationships in the UoD, though details
566 such as attributes need not be completely filled in.
567 Importantly, it should not be influenced by application scenarios apart form knowledge of their UoD.
568 Here we describe the UoD and our analysis model. The model is strongly influenced by patterns
569 discovered in earlier work on a
570 <i><a href="http://www.ivoa.net/internal/IVOA/IvoaDataModel/DomainModelv0.9.1.doc">Domain model for Astronomy</a></i>,
571 co-written by one of the authors of the present note. We describe some of its main patterns below as well.
573 <h4><a name="sec4.1"/>4.1 Universe of Discourse</h4>
575 <h4><a name="sec4.2"/>4.2 Domain Model for Astronomy</h4>
577 <h4><a name="sec4.3"/>4.3 SimDB analysis model</h4>
578 <em class="todo">@@TODO create a version and add it to volute@@</em>.
580 <!-- ++++++++++++++++++++++++ -->
582 <h2><a name="sec5"/>5 Logical Model: SimDB</h2>
583 <p>
584 Here we introduce the core of our proposal, the UML representaiton of our logical data model
585 for our Simulation Database. The exact representation of this model is an
586 <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/input/SimDB_DM.xml">XMI file</a>,
587 which can be found in the <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm">snapdm section</a>
588 of the <a href="http://volute.googlecode.com/svn/">Volute subversion database</a> on Google code.
589 Other representations can be found in that same hierarchy, in particular check out the
590 <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/input/SimDB_DM.xml">HTML documentation</a> which we generated from the XMI
591 representation with the XSLT pipeline described in <a href="#appB">Appendix B</a>. This generated documentation file contains
592 the explicit description of all of the elements in the model and forms the reference documentaiton document for the model.
593 <h3><a name="sec5_1"/>5.1 Overview</h4>
594 <p>
595 The logical data model is a fully detailed model of the application domain. It is to form the basis of physical
596 models, representing the model in various computational environments.
597 The logical model is represented as a set of UML diagrams, which we created using MagicDraw Community Edition 12.1 and stored as an
598 XMI file in the GoogleCode
599 SVN repository: <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/input/SNAP_Simulation_DM.xml">
600 SNAP_Simulation_DM.xml</a> <em class="todo">@@TODO should change all occurrences of names with SNAP to using SimDB@@</em>
601 JPG representations of the model can be found in <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/input/images/">this</a>
602 directory. <em class="todo">@@TODO find proper representation image of the complete model. Possibly color packages differently.@@</em>
603 </p>
604 <h3><a name="sec5_2"/>5.2 Normalisation</h4>
605 <h3><a name="sec5_3"/>5.3 Model contents</h3>
606 <h4><a name="sec5_3_1"/>5.3.1 Resource hierarchy</h4>
607 <p>
608 At the root of the SimDb data model is an abstract class called Resource, in the rest
609 of this document we will refere to this as SimDB/Resource.
610 It represents the different types of highest level meta-data objects to be stored in a SimDB.
611 Examples of this are represented as subclasses. First Experiment (SimDB/Experiment), which represents
612 different types of experiments that have been performed (run/executed/...) and have produced the results
613 that SimDB users may be interested in. Examples of SimDB/Experiment-s are first simulations,
614 but also the various post-processing operations transforming simulation results into other products
615 such as halo catalogues, density fields etc.
616 </p>
617 <p>
618 The second major type of SimDB/Resource is the SimDB/Protocol.
619 This concept represents a <i>formally prescribed way of doing an experiment</i>.
620 It is derived from the concept with the same name in the domain model, which itself was inspired
621 by the concept with the same name in Chapter 8.5 in <a href="#r_AnalaysisPatterns>[3]</a>.
622 In the SimDB/DM this concept has concrete representations in the computer programs that are being
623 used to run simulations and post-processing etc. As such it defines the possible input parameters,
624 possble algorithms, the kind of results that can be produced by the code. Every SimDB/Experiment must
625 indicate which SimDB/Protocol was used and for example provide values for the input parameters, indicate
626 which physics was used
627 </p>
628 <p>
629 The SimDB/Resource concept is clearly similar, but in general <i>not equivalent</i> to the Resource Registry's Resource concept.
630 In data modeling terms, it is not true that a SimDB/Resource <i>is a</i> Registry/Resource.
631 Often the reason is similar to the reasons that a single image is not a Registry/Resource, whereas a SIAP-compatible service is.
632 The granularity of a SimDB will be higher than a Registry and many simulations on their own will be too small.
633 The SimDB itself will have to be registered (see <a href="#">section ???</a> for a further discussion
634 <em class="todo">@@ TODO add propoer section and href@@</em>),
635 i.e. a SimDb service <i>is a</i> Registry/Resource. In discussion with Ray Plante (IVOA Interop May 2007, Beijing)
636 on this issue it was proposed that some part of the contents could also be registered in a Registry directly.
637 I.e. we should be able to identify Registry/Resource-s in SimDB. Considerations to decide on how to make this identification would be for example
638 that all data products resulting form a well defined (and published) scientific project could qualify.
639 To represent such a possibility for now we have introduced another subclass of SimDB/Resource: SimDB/Project.
640 This is not much more than an aggregation of experiments, with some additional atrributes describing the motivation etc.
641 The metadata of a SimDB/Project is not the same as that of a Registry/Resource, however we propose that we should be able
642 to define a transformation (possibly implemented again in XSLT) to transform a SimDB/Project and produce a Registry/XML representation.
643 Some more thoughts on this subject will be given in <a href="#">section ???</a> <em class="todo">@@ TODO add proper section and href@@</em> mentioned above.
644 </p>
646 <h4><a name="sec5_3_2"/>5.3.2 Target</h4>
648 <h4><a name="sec5_3_3"/>5.3.3 Characterisation</h4>
650 <h4><a name="sec5_3_4"/>5.3.4 Semantics</h4>
652 <h4><a name="sec5_3_5"/>5.3.5 Units</h4>
655 <h4><a name="sec5_3_6"/>5.3.6 Services</h4>
656 The goal of the SimDB specification is to define a protocol for querying interesting simulations and related SimDB/Resource-s.
657 Once these have been identified the user should be able to access these simulations.
660 <h2><a name="sec6"/>6 Physical models</h2>
661 Here we describe how we create physical models out of the logical model.
662 A <i>physical model</i> is (see <em class="todo">@@TODO reference to some standard reference on data modelling@@</em>)
663 a representation of the logical model that is adapted to a particular software environment.
664 The DM WG has mandated (IVOA interoperability meeting, Cambridge, UK, May 2003) that one
665 such representation should be an XML schema. This is to be used to define the structure of XML documents
666 used in message to communicate instances of the SimDB Resource type.
667 Together with this we also create a relational database schema.
668 We propose this model as we want to use the ADQL standard under development in the VOQL WG
669 in the protocol for querying SimDB-s.
672 <h3><a name="sec6_1"/>6.1 Identifiers and References</h3>
674 We want to be able to identify each instance of each concrete type explicitly in a globaly unique way.
676 To this end we need to be able to assign identifier on each
679 <h3><a name="sec6_2"/>6.2 RDBM Schema</h3>
680 The public schema, i.e. the view the outside world has of a SimDB, is a relational schema.
681 This will be formally defined using VOTables containing the appropriate TABLE definitions
682 <ul>
683 <li>object types are mapped to tables, one table per object type</li>
684 <li>Inheritance hierarchies: JOINED strategy as defined in JPA, i.e. each table only has columns for the attributes and references defined on the corresponding type.
685 Also an ID column that is a PK and also a FK to the ID of the base class' table. Possibly a container column (see below)</li>
686 <li>Primary key column: <tt>ID NUMERIC(18)</tt></li>
687 <li>Foreign key to container: <tt>containerId</tt><br/>plus foreign key and index declaration</li>
688 <li>References: &lt;referenceName&gt;Id<br/>plus foreign key and index declaration.</li>
689 <li>Using topological sort of object types based on (extends|container|reference) relations we generated
690 create table statements and ther indexes and foreign keys in blocks. drop table statements in opposite order.</li>
691 <li>For each class we create a view named "v_&lt;class name&gt;<br/>returns all columns for that class; uses join to base class's view.</li>
692 <li>generate a discriminator column on table for root in inheritance hierarchy, stores name of class (must be unique in inheritance hierarchy!)</li>
693 <li>attributes mapped to single column if their type is simple (i.e. primitive, or enumeration)</li>
694 <li>if attribute's type is dataType mapped to as many columns as the dataType has attributes,
695 with column names the name of the dataType's attributes, prefixed by &lt;attribute-name&gt;_</li>
696 <li>For PK columns we use the
697 </ul>
699 <h3><a name="sec6_3"/>6.3 XML Schema</h3>
701 <h3><a name="sec6_4"/>6.4 UTYPE-s</h3>
702 <p>
703 It is generally the case that contents of databases may be represented in ways that do not
704 conform to one of the standard serialisations. Nothing prevents services to be developed on
705 top of SimDB that represent SimDB/Resource-s or even fragments of these in another form.
706 The standard example would be to have VOTables storing the results of a generic ADQL query of the SimDB/RDB representation.
707 VOTable first introduced the option to have a UTYPE attribute in FIELD definition tags store
708 a pointer to an element in a data model that the column represents.
709 </p>
710 <p>
711 The <a href="#r_SpectrumDatamodel">Spectrum data model</a> was the first to add explicit
712 UTYPE-s for each of the attributes in its model and the <a href="#r_CharacterisationDM">Characterisaiton data model</a>
713 has followed that example. As long as the precise usage and relation of the syntax of the underlying data model is
714 is not defined, we will follow these exmaples by assigning UTYPE-s explicitly to all elements in the model.
715 However, we will follow a fixed set of rules to makes this assignment and implement these in XSLT.
716 If a similar approach is at some time accepted within the IVOA, possibly in an alternative form, it will be straightforward
717 to adjust our definitions.
718 </p>
719 <p>
720 Our assumption is that the UTYPE should be able to uniquely represent any element in the data model, and in a manner
721 that is also easily interpreted. For now the elements that we assume need to be able to address are those that can be
722 represented by a single value in a column. This leaves us to requiring to be able to derive UTYPE-s for the following
723 model elements:
724 <ul>
725 <li>Attribute</li>
726 <li>Reference</li>
727 <li>Collection</li>
728 </ul>
730 </p>
733 <h3><a name="sec6_5"/>6.5 Java/JPA+JAXB (non normative)</h3>
735 <h2><a name="sec7"/>7 Query Protocols</h2>
736 <h3><a name="sec7_1"/>7.1 ADQL</h3>
738 <h3><a name="sec7_2"/>7.2 REST</h3>
739 <p>
740 Under this heading we mean a protocol whereby data products can be retrieved through
741 HTTP GET requests. Possibly also they can be POST-ed, or PUT.
742 This needs to be discussed further, but maybe can be punted until a future release.
743 The GET will always only be able to get a complete SimDB resource, serialised to SimDB/XML, similar to the Registry.
744 </p>
745 <h3><a name="sec7_3"/>7.3 TAP?</h3>
746 Issues:
747 <ul>
748 <li>How does TAP deal with units?</li>
749 <li>In TAP, does a table column containing values always has a single UCD and a single Unit?</li>
750 <li>Is TAP suited for this kind of metadata databases?</li>
751 </ul>
753 <h2><a name="sec8"/>8 Next Steps</h2>
754 <h3><a name="sec8_1"/>8.1 Reference implementations</h3>
755 <h4><a name="sec8_1_1"/>8.1.1 France</h4>
756 <em class="todo">@@ TODO Laurent @@</em>
757 <h4><a name="sec8_1_2"/>8.1.2 Germany</h4>
758 <em class="todo">@@ TODO Gerard @@</em>
759 <h4><a name="sec8_1_3"/>8.1.3 Italy</h4>
760 <em class="todo">@@ TODO Patrizia @@</em>
761 <h4><a name="sec8_1_4"/>8.1.4 USA</h4>
762 <em class="todo">@@ TODO Rick @@</em>
764 <h3><a name="sec8_2"/>8.2 Generating XML from simulation pipe lines</h3>
766 <h3><a name="sec8_3"/>8.3 SimDAP services</h3>
768 <h2><a name="appA"/>Appendix A: Data modelling specifics</h2>
769 Here we describe various aspects of UML modelling as we applied it to the current
770 problem area.
771 <p>
772 UML allows communities to create a domain specific modelling language through its Profiling capabilitites
773 <em class="todo">@@ TODO is this the proper term ?@@</em>.
775 We have an initial implementation of a UML profile as created by MagicDraw available under <a href="">this link</a>.
776 Here we list the main elements and give a a short motivation for their inclusion in the model/.
777 It is our opinion that the DM working group should be ultimately responsible for a profile such as this,
778 defining a domain specific language for all IVOA data modelling efforts.
779 </p>
780 <p>
781 As first step in our simulation pipeline we generate an XML document that represents the data model in a form
782 that is more easily interpreted, both by human readers and by XSLT scripts, than the XMI representation.
783 This document itself is structured according to an XML schema that
784 represents the UML profile rather directly and that we here shortly describe.
785 </p>
786 This schema is located in
787 <a href="http://volute.googlecode.com//svn/trunk/projects/theory/snapdm/input/intermediateModel.xsd">
788 http://volute.googlecode.com//svn/trunk/projects/theory/snapdm/input/intermediateModel.xsd</a>.
791 We introduce our own XML format, defined by the XML schema in
792 <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/res/intermediateModel.xsd">intermediateModel.xsd</a>,
793 for representing the logical model. For the time being we call this the <i>intermediate representation</i>.
794 The first step in the generation pipeline is a translation of the XMI to an XML document following this format.
795 This transformation is implemented in the
796 <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/res/xmi2intermediate.xsl">xmi2intermediate.xsl</a>
797 XSLT script. The latest version of the intermediate representation for the SimDB data model can be found in
798 <a href="http://volute.googlecode.com/svn/trunk/projects/theory/snapdm/output/SNAP_Simulation_DM_INTERMEDIATE.xml">this location</a>.
799 All other generation scripts work on this intermediate representation, not on the XMI document.
800 Variations in tool-generated XMI, or different versions of XMI can now be supported by an appropriately adjusted
801 XSLT script.
802 One reasons why this may be useful is that are different tools may produce different versions or different
803 dialects of XMI. Another reason for this representation is that XMI is a rather complex representation of a UML
804 model. Since we are using a rather restricted <a href="#profile">profile</a> we do not need this generality, and
805 this allows us to represent the model using XML documents that are much easier to handle with XSLT.
808 <p>
809 We illustrate out UML profile using an example data model
810 derived form the SimDB/DM, shown in the following diagram:<br/>
811 <img src="img/example.jpg"/>
812 <br/>
813 We now describe the individual elements.
814 some of these are standard, some of these are domain specific extensions following
815 standard UML profile <i>stereotype</i> extension elements and associated <i>tag definition</i>.
817 <ul>
818 <li>Model<br/>
819 (no visual counterpart)</li>
820 <ul>
821 <li> &lt;&lt;model&gt;&gt; </li>
822 <ul>
823 <li>TagDefinition: author</li>
824 <li>TagDefinition: title</li>
825 </ul>
826 </ul>
827 <li> Package <br/><img src="img/package.jpg" />
828 <ul>
829 <li>package containment</li>
830 <li>package dependency</li>
831 </ul>
832 </li>
833 <li> Class <br/><img src="img/class.jpg" />
834 <ul>
835 <li>isAbstract<br/>
836 </li>
837 </ul>
838 </li>
839 <li> DataType <br/><img src="img/datatype.jpg" /></li>
840 <li> Enumeration <br/><img src="img/enumeration.jpg" /></li>
841 <li> Property: attribute<br/><img src="img/attribute.jpg" /></li>
842 <ul><li>&lt;&lt;attribute&gt;&gt; </li>
843 <ul>
844 <li>TagDefinition: minLength<br/>
845 </li>
846 <li>TagDefinition: maxLength<br/>
847 </li>
848 </ul>
849 <li> &lt;&lt;ontologyterm&gt;&gt; </li>
850 <ul>
851 <li>TagDefinition: ontology<br/>
852 A URL locating a standard (RDF|SKOS|OWL|???) document containing
853 a list of terms from which the value for this attribute may be obtained.
854 It is our opinion that the Semantics working group should be responsible for the
855 definition of relevant ontologies (or semantic vocabularies, or thesauri, or ...)
856 required for a given application domain, though the contents should be decided in
857 cooperation with domain experts.
858 </li>
859 </ul>
860 </ul>
861 <li>Inheritance
862 <br/><img src="img/inheritance.jpg" /></li>
863 <li>Binary association end: collection
864 <br/><img src="img/collection.jpg" /></li>
865 <li>Binary association end: reference
866 <br/><img src="img/reference.jpg" /></li>
867 <li>Binary association end: subsets
868 <br/><img src="img/subsets.jpg" /></li>
870 </ul>
872 </p>
875 <h2><a name="appB"/>Appendix B: XSLT pipe line</h2>
876 <em class="todo">@@ TODO Laurent @@</em>
878 <h2><name="glossary"/>Glossary and Acronyms</h2>
879 <dl>
880 <dt><a name="g_SimDB">SimDB</a></dt>
881 <dd></dd>
882 <dt><a name="g_SimDAP"/>SimDAP</dt>
883 <dd></dd>
884 <dt><a name="g_SimDB/DM"/>SimDB/DM</dt>
885 <dd>The logical data model defining the structure of <a href="#g_SimDB">SimDB</a>.</dd>
886 <dt><a name="g_SimDB/RDB"/>SimDB/RDB</dt>
887 <dd>The representation of the SimDB/DM as a relational data base schema.</dd>
888 <dt><a name="g_SimDB/XML"/>SimDB/XML</dt>
889 <dd>The XML representation of the SimDB/DM</dd>
890 <dt><a name="g_SimDB_resource"/>SimDB resource</dt>
891 <dd>A top-level data product stored in a SimDB.
892 A SimDB resource can be described in a SimDB/XML document, but none of its constitutents can.</dd>
893 </dl>
895 <h2><a name="references">References</a></h2>
897 <p><a name="r_UML">[1] ???, <i>UML standard</i>
898 <br/><a href="http://">http://</a>
899 </p>
900 <p><a name="r_XMI">[2] ???, <i>XMI standard</i>
901 <br/><a href="http://">http://</a>
902 </p>
903 <p><a name="r_AnalaysisPatterns">[3] Martin Fowler, <i>Analysis Patterns</i>, 1997, Addison Wesley.
904 <br/><a href="http://">http://</a>
905 </p>
906 <p><a name="r_TheoryinVO">[4] Lemson & Colberg, <i>Theory in the virtual observatory</i>
907 <br/><a href="http://">http://</a>
908 </p>
910 <p><a name="r_Characterisation">[5] ???, <i>Characterisation DM</i>
911 <br/><a href="http://">http://</a>
912 </p>
914 <p><a name="r_informatonIntegration">[6] <em class="todo>@@ TODO @@</em>references on global-as-view and information integration
915 <br/><a href="http://">http://</a>
916 </p>
918 <p><a name="r_visivo">[7] <em class="todo>@@ TODO @@</em>reference to VisIVO
919 <br/><a href="http://">http://</a>
920 </p>
922 <p><a name="r_SpectrumDatamodel">[8] <em class="todo>@@ TODO @@</em>reference to Spectrum data model
923 <br/><a href="http://">http://</a>
924 </p>
926 </body></html>

ViewVC Help
Powered by ViewVC 1.1.26