Distributed Engineering Data Management (EDM) using HyperText
Authors
Introduction
The large number of institutes and design teams involved in the projects
for the LHC experiments need efficient exchange of information and
technical data. Due to the fact that different Institutes and departments have
their own way of organizing engineering data, an information sharing system
which allows
management of distributed data in a heterogeneous environment is required.
The data management system should be robust enough to cover the whole
life-time of an experiment.
This paper describes an Engineering Data Management (EDM) system
for the LHC Detector Design projects. The aim of distributed EDM is to
integrate collaborative institutes and
companies connected to Internet and to provide them with tools for exchanging
information between the remotely located design teams. The system is based
on publicly available (i.e. no license costs involved) software products and
networking facilities. This makes the EDM system easy to accept and join,
which is a key factor for succesful implementation.
The feasiblity of such a system is proved by pilot
projects implemented within the ATLAS and
CMS
collaborations.
Principle
An information retrieval system based on hypertext
through the World Wide Web
(WWW)
is used to assist distributed design activities. The World Wide Web is a
global hypertext system originating at CERN which can handle text and graphics
information.
The overall data management task is divided into managing separate design
projects, which each take care of presenting the status of
their own work. This is done by using explanatory WWW HyperText documents with
links to design files and to information
server directories (see FTP -archive below) containing more dynamic design
data.
Figure 1. The distibuted engineering information mapped into one
logical entity under WWW.
The EDM system itself is based on a "metadata" concept. It basically means that the
management of the actual engineering data is based on data about the actual data, the
"metadata". Metadata might be e.g. short descriptions of drawings, release dates,
names, owners or computer-oriented information like file sizes, types, creation dates
etc.
The EDM system described is built using
metadata in
HyperText documents. The HyperText documents are linked together to form a
framework for the management of the data. The HyperText documents can be
edited and linked freely to conform to each data management need. In addition
to fixed HyperText documents we use the so called "directory browsing"; a
feature of the browsers to present a dynamic view to an FTP archive containing
engineering data. This gives an access to non-released, frequently changed
design data.
Since the actual engineering data is kept in standard formats and the
EDM system is actually a layer of HyperText documents on top of the data,
future changes are simplified. The data can be
quite easily remapped, e.g onto an industrial PIM (Product Information
Management) system.
HyperText documents for EDM
The HyperText pages guide users to the information they are looking for.
Each project will have a default "Home Page" which contains short textual
descriptions and links to sub-projects. The projects and sub-projects
use :
- Links to a HyperText or Graphics gallery of released (and
therefore stable) drawings. This can be organized as a common database
containing (links to) the project's official design data and updated by
authorized EDM adminstrators only.
These documents present released drawings in
either text only or with embedded graphics. Text only presentation can be
followed with any browser.
The graphics gallery is meant to be browsed with NCSA XMosaic, or another
browser which handles embedded .gif images.
- A link to text documentation (.html files in a "docs"
directory).
- A directory browsing link to design data on an FTP server. This can be more dynamic
design data, typically CAD exchange files during the product development phase.
Or a dynamic archive for a distributed design team on a common design task.
- Links to contact persons and their coordinates.
Example
An example of how to structure design data for easy access through WWW is
available for ALICE Detector design.
Another example from the CMS collaboration is the Graphical overview of General Parameter drawings.
This is especially directed for new viewers learning about the designs, but can as well be used to archive of released (and thus stable) design versions.
Users
The users of the EDM system can be grouped into three categories with
different objectives:
- Anyone interested to browse the EDM pages or anyone interested to use
the FTP archive. In this case the user has only read access and can take a
copy of
all of the visible documents.
- EDM users within a project. The system manager can give persons in a
project user-id's with
read/write permissions in their FTP area so that they can maintain their own
FTP directories.
- The administrators, who have full permission to the
whole EDM system, have the principal responsibility to update the HyperText EDM
documents. These persons can also be a project member assigned to look after
EDM information.
Recommended data types
Previewing of data is limited by the capabilities of the browsers, but any
file format can be
transferred using FTP. Currently access to native CAD-data bases will
require human intervention, but
the hypertext can contain information about how to obtain the original data.
The data in the HyperText system can have following formats:
- HTML
- The HyperText EDM documents are text in HTML format. This format can
be used to present any textual information, given that there is a
conversion from a native file format to HTML. Small documents can be converted
manually (or drafted in HTML in the first place).
- ASCII
- Ascii text can be imbedded in it's original format in an HTML
document.
- Postscript
- Design data should as a default be put on the FTP server in Postscript
(.ps) format for remote previewing.
- IGES
- CAD models should be available in a neutral format like IGES or SET, which can be
retrieved and read into other CAD-systems. (The ISO STandard for Exchange of Product data, STEP should be
used in the future.)
- Other Graphics formats
- XMosaic can browse several bitmap formats. GIF
files are most common since that is also the format of the imbedded images in
an HTML document.
The recommended tools: MOSAIC, WWW and FTP
- Hypertext browser
- The hypertext browser is the user interface to the system. Through the
browser one can
browse and read documents on the network. Advanced browsers handle also
graphics by activating previewers to consult graphics files. Several
browsers exists in the Public Domain. We recommend the use of NCSA
MOSAIC in an X-window environment, as this is a powerful WWW browser with
many other useful functions. With XMosaic, previewing of remote PostScript
files (and several other formats) is possible. Extensions with previewing of
CAD-files by starting up
the CAD-application directly (for example AutoCAD by the file extension .dwg)
from XMosaic can be envisaged.
- Line-mode browser
- All systems should also have the WWW Linemode browser installed to give
access also for terminal users. This is the most basic way to access the EDM
system and is available virtually for any piece of networked computer
equipmment.A WWW
Telnet
server is another program that gives access for anybody over telnet, not
having a user-id for the server.
- FTP archive
- FTP (File Transfer Protocol) provides a simple way to exchange files
between
platforms. The FTP archives should be used as
exchange areas for CAD design files
in neutral formats like IGES.
(CAD files native to a given system can also be exchanged this way)
An anonymous FTP server can be setup
with public access for general design data, or with group access to some
directories for specific projects. We recommend the Wuarchive FTPD package
- WWW server
- The Engineering information hypertext pages need to reside on a WWW
server. Such a server can be
set up on any workstation with free disk space, since it does not require much
CPU time. Still, if possible, a separate server is recommended for easier
maintenance, better availability and system security.
Platform requirements
To access information on the EDM system you need the following:
- Network access to Internet.
- On Unix or VMS Workstations:
- PC with Microsoft Windows to run NCSA Mosaic for Windows
- Apple Machintosh to run NCSA Mosaic for MAC
- An X-terminal, PC (XVision) or a Macintosh (MacX) can be also be used to connect to a server that has
XMosaic installed.
- The WWW Linemode browser runs on most platforms and be accessed from any terminal.
User commitment necessary
How well the system will work in the long run depends on the users'
commitment to use
it properly. Each project needs some responsible person for updating information
pages on the Web.
This maintenance should not require more than say, 1 man-hour pr. week, but is
essential for successful use of the system.
With dedication from the users it is clear that this system provides an
efficient tool for exchanging information in the
LHC collaborations and will certainly speed up the engineering work in the
project.
For more information, please contact the authors.
If you want a paper version of this document, you can pick up
this Postscript file or get it
from the library .
November 1993. Updated 05.95 NjH