Towards a user environment integrating hypermedia browsers, scientific visualization programs and numerical simulation programs

Andrea Leone and Antonio Ticca

CRS4 - Cagliari

Abstract

The success of projects like CERN World Wide Web has spurred the development of many hypermedia browsers on Internet. One of them, NCSA Mosaic, has a rather sophisticated ability of supporting user defined layouts of widgets, comparable to what is available in simple interface builders, and can be easily interfaced to other programs. In this article we will describe an experimental set-up that tries to explore the possibility of integrating, under a single user interface: a) hypermedia document browsing b) data feeding to scientific visualization programs c) parameter feeding and running numerical simulation programs

Introduction

Internet, the world-wide educational computer network, is quickly percolating to all educational institutions, research centers and, more recently, to commercial companies.

Many of the institutions connected to Internet provide informations to the network community. However, each of these information providers had to be accessed using a system-specific protocol. World-Wide-Web (WWW) [2] was specifically designed to overcome this kind of problems by using a so called Uniform Resource Locator (URL), a unique document addressing scheme.

Between the many browser available for WWW, NCSA Mosaic [1] is definitely the most powerful, allowing an easy user access to hypertextual and multimedia documents.

One of the most interesting features of Mosaic is its ability to be easily support new data types by using the same mechanism for Metamail (MIME) configuration. Another useful feature present in Mosaic is a reasonably powerful support for the user to define forms based on Motif widgets. Together these two features make possible to utilize Mosaic, not only for data browsing, but also as a general front-end for end-user applications.

In this short article we will describe how one can configure Mosaic as a scientific data browser and an interface to simulation/analysis programs.

CERN World Wide Web

The WWW projects developed at CERN, is a practical project to bring a global information universe into existence using available technology.

The basic idea is to represent non homogeneous information stored in large databases into an homogeneous form with the help of a markup language called HTML (Hyper Text Markup Language).

The WWW architecture require a client/server model communicating via a "logical information bus". The client has the responsibility of resolving a document address into a document using its repertoire of network protocols. The server provides data in a simple hypertext or plaintext form, or, by negotiation with the client, in any other data format. The first requirement of an Information bus is the possibility of the universal identification of a document through a set of protocols and a set of data format.

The success of WWW spurred achademic institutes as well as scientific organization to produce new integrated tools for browsing and accessing the whole information universe. One of them, NCSA Mosaic is the most diffuse of all the search engines across a wide-area network.

NCSA Mosaic

NCSA Mosaic is a networked information discovery, retrieval, and collaboration tool and World Wide Web browser developed at the National Center for Supercomputing Applications. Such system integrates into existing network protocols, formats, data sources, environments, and provides powerful new capabilities for using and sharing information.

The NCSA Mosaic clients have the following functionality:

Graphical display of plain text, rich (formatted) text and hypertext, as well as inlined access to graph, images, audio clips, video sequences, and scientific data in multimedia and hypermedia documents.
A graphical user interface by which one is able to navigate into the information space and to keep track of the global hystory.
In-document and network-based search capabilities, including those capabilities required to access WAIS datacases.
Full TCP/IP based communications support, including HTTP, Gopher, FTP and NNTP protocol support, and gatewayed access to WAIS, Techinfo, TeXinfo, Archie, finger, Hyper-G, X-500, Whois and other data sources.
NCSA Data Transfer Mechanism communications support, for integration with NCSA Collage, SGI Explorer and other network-based DTM [3] clients and information servers.
Fill-out forms. As per the current HTML+ spec, documents can specify interactive fill-out forms, with input elements including text entry areas, toggle buttons, selection lists, popup menus, etc. Mosaic visualizes such fill-out forms as sets of Motif widgets embedded inside the document. In this way it is possible to build easily sophisticated front-end interfaces to databases and search engines.

Mosaic is based on an Uniform Resource Locator (URL) mechanism. With this is possible to encapsulate virtually any information system that accepts remote network queries. URL can point to documents residing on FTP or HTTP servers, queries to WAIS databases, documents or search engines residing on Gopher servers, a specific sub-part of a document residing on any server, a set of meta-data or almost anything else thast is available somewhere on the network.

User forms are composed by placing input fields within paragraph, preformatted /literal text, lists and tables. This gives considerable scope in designing the layout of forms. Each field is defined bi an input element and must have an name attribute which uniquely names the field in the document. Additional optional attributes can be used to specify the type of the field (default to free text), its size/precision, its initial value and wether the field is currently disabled or in error.

A scientific data dynamic front-end

Historically, hypermedia systems, and networked information systems in general, have not supported scientific data as a basic information type. Since NCSA Mosaic is specifically oriented towards scientific and academic research, it supports scientific data via the following functionality:

Scientific data can be included within a hypermedia document in the same manner as images, audio, and video sequences. The scientific data format of choice for NCSA Mosaic is NCSA's Hierarchical Data Format (HDF) [4], a machine independent file format that allows arbitrary grouping and annotation of heterogeneous data elements.
Data elements in format other than HDF can also be included in hypertext documents, with display and processing of received data dependent on local system capabilities.

Those scientific data can be shared among different application via a Data Transfer Mechanism (DTM) implemented in NCSA Mosaic that is supported by other synchronous collaboration tools for scientific data analysis and manipulation as well as with Polyview, SGI Explorer, NCSA Collage, and AVS.

We experimented these capabilities writing a specific Mosaic browser for scientific data stored in TH format (a CRS4 extension of HDF), able to send data across the network (via DTM facility).

TH is a data format that provide support for time dependant, multi-blocks data defined both on regular and irregular grids. Furthermore the user can annotate all the objects contained in a TH file.

Given a document address pointing to a TH file, and exploiting the facility to define document types and associate them a corresponding "viewer" (according to the MIME standard), the TH "viewer" will be able to examine the data structure and generate an hypermedia document containing a high-level description of the data (see Fig. 1).




												Fig. 1

Hot-links into the generated document will allow the user to directly explore the file in depth, selecting a particular time for the data and a particular variable (see Fig. 2).




												Fig. 2

At that time numerical data are available to be sent over the network via a DTM call, and can be analized by any software tool supporting DTM protocol.

In this example (see Fig. 3), NCSA Collage is used to get data sent from Mosaic. This tool allows scientific data analysis and visualization and collaborative work. In this session the authors are both connected to the same Collage server, sharing data received from the DTM port.

An image can be built from data and visualized to all participants. They can overlay text and sign to the image, to emphasizes region of particular interest or to comment data.

A data analysis program front-end

We are currently experimenting the form capability of the HTML+ language to control numerical simulations and other analysis programs, thus taking advantage of being under the same environment. An additional advantage of such a language is the possibility of using an ASCII terminal oriented browser, like LYNX, so that the user can control the program not only from a graphics workstation, but also from an ASCII terminal on a serial line connecting (e.g. from home).

Fig. 4 illustrates a front-end for an image processing module of a data analysis and reduction system for a molecular biology applicationi, that's being developed and used in a joint collaboration by the Computational Genetics Project of CRS4 and an experimental group at the Imperial Cancer Research Fund (London).




												Fig. 4

Conclusions

The point of this article is that it's possibile, with a limited effort, to set up a system having features comparable to simple interface builders, but fully integrated with hypertextual/multimedia document browsers and scientific visualization programs.

References

1: Marc Andressen. 1993. NCSA Mosaic Technical Summary National Center for Supercomputing Applications, Champaign, Illinois.
2: T.B. Lee, R. Cailliau & others. 1992. World-Wide-Web: An Information Infrastructure for High-Energy Physics. CERN Geneva, Switzerland.
3: The software Development Group, The DTM Data Transfer Mechanism. NCSA Technical Report, 1993.
4: The software Development Group, NCSA HDF Specification. 1990. NCSA Technical Report, Champaign, Illinois.