To cite this article please refer to the printed edition of TrAC: Trends Anal. Chem., (in preparation).
ChemWeb, Inc., 34-42 Cleveland Street, London, England W1P 6LB
The use of the Internet as a first step in locating chemical information is increasing rapidly, although users are frequently frustrated by the lack of permanence and quality of many Internet sites. The enthusiasm for setting up a new site is often not matched by the investment in time and effort required to maintain the facility. Whereas users are accustomed to having free access to information on the web, major financial investment is needed to develop web resources that are internationally recognised, that provide the quality of information that users require, and that are regularly maintained and updated. In the field of chemistry many resources have been set up by learned organizations, such as the American Chemical Society ChemCenter site, originally to enable members to access information about membership, products and services, including conference information and details of publications. ChemCenter has expanded considerably to provide access to STN databases, a job bank and shopping centre and is recognised as a major chemical web site. Other sites have been developed by enthusiastic individuals who began by creating pointers to chemical resources relevant to their academic teaching programs. The ChemDex site, one such example developed by Mark Winter at Sheffield University, has become one of the major compilations of chemistry-based web sites, containing over 3000 links. Most other chemistry web sites provide a single type of content and service. The Chemistry Place is an excellent example of a site aimed specifically at the teaching of chemistry. Although it is subscription based, the rates are very reasonable and the site offers original tutorials, details of experiments and problem solving exercises.
ChemWeb, officially launched at the ACS meeting in San Francisco in April 1997, is one of the most recent resources under development and aims to be a comprehensive worldwide club for chemistry research and communication. It is being developed, with a $6 million investment, by ChemWeb, Inc., a joint venture of Current Science Group, London, and MDL Information Systems, Inc., and it combines electronic publishing, databases, journals and scientific forums. Not restricted to promoting the ideals or activities of a single institution, ChemWeb seeks to integrate useful applications with primary information, to provide the best possible service to members, and it welcomes collaboration with other information providers. ChemWeb is modelled on the highly successful sister venture, BioMedNet, which currently has over 100,000 members in the biological and medical communities, and is one of the world's most intensively used scientific web sites [1].
The aim of ChemWeb is to provide, in the form of an online community, all of the various facilities and services which a chemist would need in his or her day-to-day professional life. This includes, but is not limited to, information resources (books, journals and databases), a messaging system, electronic conferences and meetings, a conference diary, a job exchange, a shopping mall, an electronic magazine (news, reviews, etc.) and, eventually, computational services.
Membership of ChemWeb is open to anyone with an interest in chemistry and is currently free to all individuals. It is expected that the future development and maintenance of the facility will mainly be financed by the advertisers and sponsors whose advertisements appear on each of the web pages, and by the commission raised on products sold through ChemWeb.
It is unlikely that any single site would provide access to all of the information that a chemist needs, but the success of BioMedNet shows that many scientists throughout the world value the opportunity to have a wealth of information supplied by one source. ChemWeb does not intend to restrict the products and services it offers to particular vendors, and will collaborate with any interested companies in order to offer as wide a range as possible. All of the facilities that were developed for the BioMedNet site are also being offered on ChemWeb and these include:
All of the above options are searchable. In addition, a Meeting Room facility and an enhanced searchable version of ChemDex, ChemDex Plus, are also under development.
Providing journals in electronic form means that not only is the journal accessible to readers much quicker, but it is also possible to offer many capabilities that are only available electronically, such as graphics manipulation, search and retrieval, hyperlinks and online purchasing. The ChemWeb library currently contains around 30 scientific journals from a number of publishers, at present mainly from Current Science Group companies and from Gordon & Breach, but there are plans to extend the library to cover about 300 publications during the coming year. Members can search all the journal articles free of charge. The Tables of Contents, publisher's information and most abstracts can also be viewed without charge, but there is usually a charge before users can view the full text of articles. Some journals are available in PDF format, which is searchable. Users must have Adobe Acrobat Reader to read the .pdf files, and this can be downloaded through the ChemWeb Home Page.
All the major chemical online search services offer both text and structure searching of their databases, but up to now there has been very little of the latter available on the web. One of the first databases to be offered through ChemWeb is the NCI-3D database, and the Available Chemicals Directory will be available shortly. Both of these databases are structure searchable. ISIS/Draw from MDL Information Systems (or another compatible structure drawing progam) and the Chemscape Chime plug-in are used to draw and display the structures respectively, but links are provided from the ChemWeb Home Page to the ISIS/Draw and the Chemscape Chime web pages so that users can download the programs before structure searching.
Conference listings are appearing in several web sites, and are a valuable resource for chemists, but many conference listings are capable of being browsed only by date, and not searched by keyword. One of the ChemWeb databases is a conference diary: a comprehensive database of chemistry-related conferences that can be searched to find details of relevant meetings. Members can submit details of forthcoming conferences to be added to the database. There are plans to set up facilities for online registration for conferences and also to host electronic conferences.
Currently, the following databases are also available on ChemWeb: the Investigational Drugs database IDdb; a preview of the ISI Reaction Citation Index and the OHS Material Safety Data Sheets. The charges for searching the databases are determined by the database producer in conjunction with ChemWeb. There is no charge for submitting data to, or searching, the conference database, and users can currently search the NCI-3D database free of charge. ChemWeb offers a limited free trial for searching the Material Safety Data Sheets (10 free searches), whereas users must pay a subscription to search the IDdb. It is not clear when you start to use ChemWeb exactly what is available free of charge and what is available only on subscription, but this could be rectified with an improved Help System, which is currently under development.
One of the advantages of a web resource over the traditional online services is that the users, or members, can not only retrieve information from the site, but can also contribute information. The Job Exchange is one area of ChemWeb which is created entirely by the information submitted by members and it contains listings of hundreds of jobs offered and jobs wanted throughout the world. Members can advertise an available position, or advertise themselves if they require a job. Job advertising and job applications are frequently confidential issues, and, to preserve confidentiality, it is possible to advertise a job anonymously and to reply to a job advert through a mailbox. The jobs are searchable by country, discipline, and type of position.
Organizations with software, books, equipment and services for sale can advertise in the Shopping Mall. Thus Shopping Mall entries are also contributed by members, and are classified into sections so that users can browse through a relevant section, but specific items can be retrieved using a full-text search.
In BioMedNet, the most frequently visited part of the site has been HMS Beagle, a weekly magazine for the biomedical community. ChemWeb has developed The Alchemist, a comparable weekly magazine for chemists. Launched in July 1997, it contains news, weekly research summaries, articles by top science writers, surveys of chemistry discussion lists, news from conferences and reports of valuable web sites. The news will eventually be updated daily, and this, together with the attractive light format of The Alchemist, is designed to encourage members to visit the ChemWeb site daily. Other chemistry webzines have been developed in more specific areas, such as Chemical Web Marketing & Technology, which covers chemical marketing issues, and CyberChem Webzine, which is an online news source for the chemical industry, but The Alchemist is currently the only broad-ranging chemical magazine on the web.
ChemWeb is designed to be a club for members and, as such, it is important that members have the opportunity to communicate with each other. The Meeting Room facility, currently under development, will allow members to hold on-line meetings, engage in real-time chat, and leave messages. Members will be able to search through the membership profiles for people with similar interests and invite those members to attend a meeting on a relevant subject. Any comments that members have about ChemWeb's content or services can be sent directly back to ChemWeb using the Feedback facility. Members are able to notify ChemWeb of any aspects that they would like to see improved or any facilities they would like to see that are not currently available.
ChemDex Plus will contain everything from the Sheffield University ChemDex site and, in addition, will be fully searchable. There are also plans to commission reviews for selected web sites and to publish these and associated comments within ChemDex Plus.
Membership is open to all individuals, who must join by completing a registration form before obtaining access. There is a set of membership rules to ensure that members respect copyright, pay their bills for using copyright material and do not abuse the club. Access to ChemWeb is by membership ID and password. Members have free access to much of the information on ChemWeb, but they must have a subscription to a journal or open an account before they can display the full text of the journal articles. There are currently over 9000 members on the database, and this is increasing at the rate of about 2000 members per month. This rate is expected to rise, however, such that a membership of around 50,000 is expected after the first year of operation, based on the statistics of the BioMedNet site.
ChemWeb is one of the first Web services to be powered by Silicon Graphics Inc.'s Origin 2000 server, which supports all ChemWeb.com databases and information management technologies. The Origin 2000 server was chosen for its exceptional speed and responsiveness and because it is the most appropriate platform to run the MDL ISIS software and Chemscape.
The live ChemWeb system operates on a Pentium Pro 200, linked to the Silicon Graphics Origin 2000 processor via an Ethernet link operating at 100 Mbits/sec. Users access ChemWeb via their service providers through a firewall, as shown in the following diagram:
The Pentium Pro 200 operates under Windows NT 4.0 and contains the Web Server, (live ChemWeb), the BiblioteK text searching software, and all software relating to the operation of the ChemWeb club. The 8-processor Silicon Graphics Origin 2000 contains the Oracle relational database, the Netscape server, ISIS Host, Chemscape, the OHS and ACD databases.
So that there is no interruption to ChemWeb during maintenance and updating of the databases and programs, there is an additional Pentium Pro 200, "SWAP", operating in parallel to the live server. All updates and amendments to the Web server and to the software are uploaded on to this parallel system. When the updates are fully checked, the parallel system is switched over to become the live system and the previous live system becomes the new "SWAP" system. In the event of a failure in the live system, the SWAP system can immediately be used as a back-up. ChemWeb development is carried out on a separate Pentium Pro 200 linked to a Silicon Graphics O2.
Text searching is performed by the purpose-written search software, BiblioteK. BiblioteK was originally written for BioMedNet, but has been further developed for ChemWeb. It uses multi-level, tree-structured indexing to ensure fast searching, and it enables users to display increasingly detailed levels of information. The search system has been developed such that increasing the size of the databases will not increase search times. Server and Collector technology is used, whereby several PCs are used for searching in parallel. Prior to searching, a database is segmented into parts. Each part is distributed to a separate PC (Server) for searching and the results are recombined by a Collector. This system is used both for searching ChemWeb databases and also remote databases. This enables any increase in the size of databases to be accommodated by using more Servers for searching without modification of the search programs. Where chemical structures are available on a database, structure searching is performed by MDL software, ISIS/Host and Chemscape.
A new programming language, with the code name MASH [2], has been developed by Electronic Press Ltd (part of the Current Science Group) for the display of the web pages. It is a general data manipulation language, almost indistinguishable from HTML, but it allows templates to be imported and nested to construct and display the web pages. The main advantage of MASH is that the information in the web page is separated into a series of building blocks, so that changes to one template will immediately be reflected in all the related web pages throughout the site. This facility will be invaluable as the ChemWeb club develops.
The need to deliver journals electronically was recognised in the early Eighties, and ADONIS BV, who has been distributing the full text of journals on CD-ROM since 1991, has demonstrated the advantage of making articles readily available in electronic form. According to a quotation in the June 1997 Issue of Information World Review [3], David McNeillie, IP Services Manager at Zeneca Pharmaceuticals, thinks that access to articles in electronic form is most easily accomplished from a single source. He says: "Zeneca feels that ADONIS still has a better coverage of the relevant literature than any other service". ChemWeb has some way to go before it is able to match the 800 journals offered by the ADONIS Document Delivery Service, but ChemWeb aims to provide the opportunity for publishers to distribute their journals on the Web through a single source while allowing them to maintain control over the access to the publications. Each publisher has a home page, or "gateway", apparently within ChemWeb, but actually situated outside the firewall and therefore not in the secure zone. The publishers set the price of their products, and users are assigned a password for use with each publication. Access can be granted if the user has a subscription to the journal, or users may pay to view individual articles. If there is a corporate subscription to a journal, the publisher can specify the number of users that have access to the publication through a given password. Access is then denied if more than the permitted number of users try to use the password.
Currently, all journal articles are stored by ChemWeb, but it is possible for publishers to store all their own publications and for ChemWeb users to be able to search the journals through ChemWeb, regardless of their physical location. Remote storage of publications will be encouraged to reduce the information stored on the ChemWeb servers.
Although a large investment has enabled ChemWeb to be set up, it is intended that ChemWeb will become self supporting within two years, through the revenue received from the advertisers and sponsors, and from commission received from publishers and database producers.
Advertisers have the option of subscribing to an advertisement "button", or of advertising products or services directly in the Shopping Mall. There is a choice of advertising for 6 or 12 months in the Shopping Mall, and there are different rates for an advertisement button according to whether the button is at the top of the ChemWeb home page, or on a less frequently visited page. Information from the advertisers can be put directly on to the ChemWeb server, or ChemWeb can just provide a link to the advertiser's own server.
Sponsors can choose keyword sponsorship, or speciality sponsorship or, alternatively, they can sponsor specific words in library articles. In the case of keyword sponsorship, the sponsor may choose up to six keywords and when any of these are selected as search terms by a member, the sponsor's advert is displayed as a banner at the top of the search results screen. Sponsors can also have a hypertext link to their own home page from a chosen word in any of the journal articles. In addition, sponsors can choose to place an advert which appears only to members with a particular specialisation.
In the words of Sir John Maddox, former editor of Nature, "ChemWeb promises to provide the interlocking data, journals and search services that are only possible in a well-designed electronic medium" [1]. It is, however, first and foremost a club for chemists who not only have access to a vast collection of information, but also have easy contact with thousands of fellow chemists, many of whom will be working in a similar area of specialization.
References
1. Information Today, October 1996, p.48
2. Developed by Andrew
Witbrock at Electronic Press Ltd.
3. Warr, W.A. "Focus", Web offerings: hype or hope?
Information World Review, June 1997.
Last update: 9 September 1997
© Copyright 1997, Elsevier Science, All rights reserved.
