“Simplicity” at Spacetelescope.org
– an innovative web scheme for science communication

By Lars Lindberg Christensen, Lars Holm Nielsen (freelance) & Erik Nordström Andersen (freelance)

Abstract
Designing any web site is far from trivial. Designing web sites that are both user friendly and easy to maintain is a real challenge.

Here we describe the ideas behind Spacetelescope.org, the new public and press web site for the NASA/ESA Hubble Space Telescope in Europe, and the general structure and workings of Simplicity, an innovative scheme for producing web sites which we developed to build Spacetelescope.org.

Simplicity provides an efficient alternative to existing commercial content management systems and is free for everyone to use.

The web was not a fad...
A recent billboard advertisement read “The web wasn’t just a passing fad”. No one would argue with that statement. In science communication the web is one of the most frequently used ways of distributing popular information about science to the media, the public and decision makers today. Although the web is still more of a ‘professional medium’, it is also increasingly a layman’s tool. For several years the web has been the preferred tool for journalists to conduct story research and therefore a proper web site must be a very high priority for any public information office.
The European outreach and education office for both the Hubble Space Telescope and the James Webb Space Telescope, the Hubble European Space Agency Information Centre (HEIC), has existed since 1999 to disseminate public information about these two missions. More details about our work can be found on the web at: http://www.spacetelescope.org/about_us

At HEIC we see the web as an excellent tool for the distribution of outreach products and for product archiving (as a repository), while also providing a service that is available at any time as well as a search engine for the rapid retrieval of relevant material. The most critical issue for us is time management — an efficient outreach office needs to spend most of its time producing material, and very little time on actually distributing it.

Early in 2004 we began designing our new web site and we naturally wanted to exploit the positive features of the web and to produce a web site that fulfilled our particular needs for efficiency. We quickly realized that the need to reduce manpower consumption for web maintenance was a general one and we extended our methodology into a general scheme for building efficient web sites. A scheme we now offer to the community.

The result of our efforts is the new public and press web site for the NASA/ESA Hubble Space Telescope in Europe and the web scheme Simplicity that combines ease of use for visitors with a simple and effective strategy for maintenance.

Put yourself in the place of others
Targeting a website to its primary customers is essential to make it successful, and the front page of a web site is undoubtedly the most important page of all. In our preparations for an effective front page we devised a so-called “Front page priority matrix” where we collected an overview of different target groups. The importance of different elements on the page was assessed — with us trying to imagine the preferences of the target group in question. Following this we gave each of the groups effective weights calculated from how big a target group they represent and from their ‘importance’ (as judged by our own particular criteria).

Front page priority matrix

The study of this matrix led to the following conclusions:

A simple page overview is the most important.
News must have top priority.
Hubble images have to be prominent.
We can’t allocate excessive space for flash animations.
We can’t allocate excessive space for design components.

We ended up with the design for the front page seen in Figure 1. For us this is the compromise between efficiency, searchability, science, visual appeal, ‘action’ and overview that we think caters for the needs of most target groups and the most important users.

For Spacetelescope.org there were three main areas of focus in the planning phase. Firstly to plan the functionality of the web site, then to plan the sitemap — ie, the structure (files and directories) and finally to plan file formats and sizes for images and videos — ie, the structure of the metadata.

Anyone implementing a web site using the Simplicity scheme that is described below will need to go through the same three planning steps and to commit one or two days to set up and adapt the Perl scripts to specific needs.

Some of the main features of the finished Spacetelescope.org site are:

The complete collection of ESA and NASA Hubble images
Huge volumes of animations and films (~35 GB)
Optimal image quality (best available tiff/jpeg files, sizes up to 250 MB)
Interactive flash applications and
Last, but not least, a well-functioning hardware, software and network infrastructure.

Requirements for Simplicity – the web scheme
Technically the general web site scheme behind Spacetelescope.org was built to satisfy several requirements.
Firstly it will produce user-friendly web sites that are easy to navigate, have consistent, attractive designs and are extremely fast for the customers. In today’s information overloaded society it is crucial to provide search capabilities that enable the user to sift through the vast amounts of information swiftly and receive an instant response to each query.

Secondly the technology behind the site juggles huge data files — images and videos — (from 10s to 1000s of MB) in archives containing thousands of items each represented in up to 15-20 different formats (eg, thumbnails, wallpaper, originals etc. for the images), without impeding function or maintenance.

Thirdly the maintenance of the web system is extremely easy. Design changes are made in just one place so that the web master is not forced to manually update hundreds of pages. Structural changes such as the addition of new archives are also possible with relatively small changes to the scripts.

Finally, the website is relatively ‘CPU light’ and can, on standard server hardware, handle many concurrent visitors.
Components of the web scheme
Apart from the data itself (images, videos etc.) Simplicity consists of three main components:

Dreamweaver templates for the ‘wrapping’ of the design
Perl scripts to execute various search and display queries
Excel, or comma separated, files for the metadata, or information, attached to images, news stories, videos, posters etc.

Dreamweaver MX is used to edit the html-scripts. Dreamweaver is a simple and visual commercial editor that enables the easy editing of web pages, and also provides a template scheme. The templates define editable areas of a web page, making it possible to keep a consistent design on all web pages. Any changes made to a template will cascade to all web pages that are based on it, and so design changes need only to be made in one place. To further ease the maintenance load, so-called nested templates are used, which are templates based in turn on another template. This makes it possible to define the global design of the website in one template and create templates for the different sections based on this global template to hold the individual section design items and menus.
Archives
One of the pillars of the Simplicity scheme is the concept of “archives”. An archive can be any collection of data and metadata (images and information). The content of an archive can be formatted in any way that is needed, and in different ways in the different parts of the website.

Screenshot from a metadata file being edited in Excel.

Instead of using off-the-shelf database solutions that have problems dealing with huge files, a large maintenance overhead and a potentially slow response time, the choice for Simplicity fell on lightweight Perl scripts as the ‘engine’ to create the dynamic web content.

In Simplicity the Perl scripts search and show excerpts of the archive metadata that are stored in Excel files (in reality comma-delimited text files). These files hold all the data for each news item, image and video and are edited with Microsoft Excel (or any column-based editor). The metadata include the object ID (for instance the image filename), title and caption. Excel, although not traditional for this type of work, has maintenance-friendly features such as spell checking and is familiar and easy to use.

The Perl scripts can format the data/metadata content of the archives in many different ways and also make refined searches possible. Other features are also built into the Perl scripts:

Publishing at a pre-determined time
Generating static html pages for the data and metadata contents that are changed rarely (news, images, videos etc.)

Search functions are provided through a freely available search engine called “ht://Dig” that indexes all web pages on a regular basis, and can be customized to fit into the overall design. Furthermore a freely available link-checker can be implemented into the scheme to make sure that all links on the web site are valid.
Pros and cons
Why choose a ‘home-made’ low-tech solution over one of the many content management systems (CMS) (A CMS is a large database driven tool that helps to structure information in the form of text, images and animations and place it on the web in a predefined way) on the market? A CMS can certainly be adapted to most common user demands, such as ease of maintenance and a consistent design, but when it comes to performance and handling of huge image files, we believe most CMS fail.

In science communication there is an extreme need for flexibility, and this implies the fully autonomous control of a web scheme and its technical maintenance and flexibility to quickly adapt to new ideas. Most CMS do not provide this. In addition most CMS do not provide the lightning fast response needed.

On the down side Simplicity is not a foolproof scheme. No web system is ever 100% foolproof, but our scheme is probably more open to error, especially when used by non-technical staff. Simplicity is also not a multi-user system, in the sense that only one person at a time can edit the page design, or update the individual metadata files. In a normal outreach office none of these issues should present major worries as it is usually staffed with technical personnel and there is no need for workflow control, approval control and version tracking.

As a happy side effect, the construction of Simplicity (including the implementation of Spacetelescope.org with all its data and metadata) only required three man-months of work, compared to an estimated two or three times longer for off-the-shelf CMS (with less functionality). Some of this time was naturally invested in an integral knowledge of the scheme which in the long run will contribute towards a reduced total cost of ownership. The total implementation costs were about 13 kEuro.

Simplicity’s low-tech solution has already proven its performance capability. Spacetelescope.org is running on a single Apache web server and was able to cope with 2.3 million hits per day (50-60 requests/sec peak load) and the delivery of up to 180 GB of data per day during the first weeks of operation.

References

http://www.spacetelescope.org
http://www.spacetelescope.org/projects/web
Web editor: Dreamweaver – http://www.macromedia.com
Free search engine: ht://Dig – http://www.htdig.org

“Simplicity” at Spacetelescope.org – an innovative web scheme for science communication

“Simplicity” at Spacetelescope.org
– an innovative web scheme for science communication