From 2ba3d3b802bfa9ca330f1dc2e54653b3bcceba65 Mon Sep 17 00:00:00 2001
From: Claudio Sacerdoti Coen The work is articulated in three, conceptually sequential
tasks: Architectural Design of the Distribution Model. The big
issue it to find the right compromise between two opposite
requirements: distribution (in the sense of the
Web: few rules, no central authority) and coherence
@@ -31,26 +33,27 @@
mapping to Uniform Resource Locators, and the integration
of databases in the distribution model. The final aim is
to have a phisically distributed library with a
- single logical view.
-
Prototype implementation. First prototyping implementation + of the distribution layer.
+Integration with the Consultation Engine. First implementation of the library as a distributed repository. Distribution should be completely transparent to users of - the Consultation Engine. - + the Consultation Engine.
+The main milestone is the release of the advanced MOWGLI prototype, for the fourth meeting of the project, at month 24. The advanced prototype will integrate the previous one with the new distribution facilities offered by the distribution layer.
-Criteria for evaluating MOWGLI's dissemination and its technological impact will be detailed in the Dissemination and use Plan. The Project Dissemination & Exploitation Committee will be in charge to monitor the state of advancement of the Project according to these criteria.
-The work is organised in the following tasks:
-MathML rendering/browsing engines. Our privileged rendering language will be MathML. In order to be able to test the presentational stylesheets,(Task 2.5) we need a MathML compliant browser, that will be developed as a part of the project (starting from a previous - prototype).
Consultation Engine (archiving, searching and retrieving). The task is devoted to the architectural design and implementation of the main functionalities for the - consultation and the management of the library.
Assisted Annotation. This cover an additional functionality of the interface, aimed to support the user in the process of annotating a proof in natural - language.
LaTeX-based authoring tool. A tool supporting automatic generation of Content-MathML from a suitably - (macro-)enriched version of LaTeX.
The main milestone is the release of the first MOWGLI prototype, at month 18.
-Use, meaning and classification. This task, based on the previous documents D1 and D2, is aimed to provide a precise definition and classification of all metadata required for an effective and efficient management of the - library.
Modelling. This is devoted to the definition of a precise markup model. To this aim, we plan to use the Resource - Description Framework of W3C.
The definition of the Metadata model is amain milestone, since the architectural design and implementation of the consultation engine (Task 4.2) will be essentially based on it. Similarly, it is required for Task 2.4 (automatic extraction of metadata).
-Main milestones are the periodic meetings, at month 6, 12, 20, 24, 30.
-The work plan is naturally organised in subtasks according to the different basic kind of interactions and manipulation to be considered, namely:
-Mathematics and the Web. State of the art, standards and + tools.
+Structured and Formal Mathematics. Delineation and layering of Semantic Components. Requirements for the interaction with tools for the automation of formal - reasoning.
Metadata. Classification and data mining for content-based mathematical documents, and key architectural guidelines - for the metadata model.
Searching and Retrieving. State of the art, use cases and + application scenarios.
+Distribution. Distributed document repositories and + peer-to-peer interoperability.
+Document Authoring. State of the art, use cases and + application scenarios.
+Part of the activity, covering Tasks T1.1-2-6 should be
concluded within the first three months, with the
preliminary report D1.a. The main aim of this phase is
@@ -71,7 +74,7 @@
The preliminary report D1.a is not a realmilestone:
it should be considered as a first internal draft summarising
the main functionalities required by the system and the basic
@@ -80,5 +83,5 @@
eventually detailed in D1.b and D1.c, which are real
Milestones, since the rest of the work will be largely
driven by them.
Large scale testing and validation will start after the release of the first MOWGLI prototype, at month 18. We shall consider three main validation tests:
-Education. Full development of a fragment of the library covering a typical undergraduate course in algebra of - analysis.
Certified code. The aim is to be able to present the formalization and the demonstration of some security properties related to the code embedded into a smart card. The presentation must be in a format understandable by the company in charge of the evaluation of the code and - in accordance with the Common Criteria standard.
Electronic Publishing. The aim is to test the LaTeX-based authoring tool and to demonstrate how an electronic physics journal benefits from the exploitation of - content markup in journal articles.
We expect a lot of feed-back during this phase, that will be taken into account for the definition of the advanced and final versions of the MOWGLI prototype.
-The work package is articulated in the following tasks:
-XML exportation. The task is devoted to the translation of the standard library of the COQ Proof assistant into a suitable XML dialect, and to the definition of a low-level DTD for the terms of the Calculus of Inductive - Construction (the logical system used by COQ).
Stylesheets to intermediate representation. Implementation of a bunch of stylesheets transforming the low-level logical description of COQ-expressions into a ``standard'' intermediate, content-level representation - such as MathML content.
Proof transformation. Similar to the previous task, but for proofs. The delicate point, here, is the fact that no ``standard'' intermediate representation currently exists, - and thus it has to be defined.
Automatic extraction of metadata. Relevant metadata such as list of identifiers in critical positions inside statements can be automatically extracted from the fully structured representation of mathematical objects. This information can then be exploited for searching and retrieving. The precise list of metadata will be defined - in Work Package 3.
Presentational Stylesheets. Implementation of a bunch of stylesheets transforming the intermediate content representation into a suitable rendering format (MathML - presentation, HTML, etc.)
Automatic Proof Generation in Natural Language. Similar to the previous task but for proofs. In this case, a fully automated approach is unlikely to produce really satisfactory results, and the process should be possibly integrated with some mechanism for interactive annotation - (see Task 4.3).
The exportation module D2.a. is our first milestone: without a large amount of available documents it would be impossible to test the transformations. Similarly, @@ -92,5 +95,5 @@ particular, for MathML).
For the end of month 18, we expect to have a first working prototype of the whole application.
-