Fachbereich Chemie, Freie Universität Berlin
First a short overview of the system shall be given, which consists of servers, documents, and browsers. Then several examples with particular emphasis on the requirements found in the area of chemistry will be shown, and finally an outlook on future developments will be given.
The application of the system shall be demonstrated by means of a typical example. A hypertext contains active words or phrases which serve as anchors or links to other documents. These active words are clearly marked, in the example by blue color and a line underneath. Now if you put the mouse pointer on such an active word and click, the current document is replaced by the document the link refers to. (Alternatively, a new window might pop up.) This behavior is well-known to everyone who ever used, e.g., the help system of MS Windows on a PC or a similar hypertext system on a Macintosh. Hypertext can be defined as text which is not constrained to be linear (Ted Nelson, 1965).
However, the WWW system is not restricted to text but offers full multimedia support such as graphics, pictures, sound and video. In the example, a larger picture would pop up if you click at the seal. On a local or local network basis, these features are also found on modern multimedia personal computers. As an arbitrary example, let me quote "Microsoft Encarta" which is a multimedia encyclopedia on CD ROM for MS Windows.
Now the particular feature of WWW is the fact that the documents are accessible worldwide and are easily retrievable by anyone who has Internet access. Browsers are available in the public domain for all important types of workstations or personal computers and operating systems (Unix/X window, MS Windows, OS/2, Macintosh etc.). Line-mode browsers suitable for text terminals (e.g., access via modem) are also available. In principle, every participant in the Internet might also provide information since servers are available for most platforms. Thus, the construction of a worldwide encyclopedia, a so-called interpedia, is under discussion.
The following examples illustrate how headings and highlighting, structured lists, umlauts or special characters and embedded pictures, and anchors are defined. Formatting (markup) is introduced by tags which are enclosed by angular brackets. Note that HTML defines neither the fonts to be used nor the page size. All anchors are defined by the same scheme, regardless of whether the hyperlink refers to plain text, hypertext, PostScript, pictures, sound or video. Information about the document type is submitted to the browser by means of the file extension (e.g., txt, html, ps, gif, au or mpeg).
Finally, various utility programs for converting documents written with some kind of word processor or as TeX/LaTeX documents are available (e.g., latex2html, fm2html (from FrameMaker MIF), rtftohtml; note that the Rich Text Format, RTF, is an optional output of, e.g., "Word for Windows".)
WWW is an excellent medium for providing online documentation of software; three nice examples are the manuals for MolScript, Mathematica and Rasmol. This type of application is easily extended to teaching material in general.
For example, there are a paper dealing with the "Simulation of EPR Spectra", a collection of refereed publications by Henry Rzepa, and a preprint collection of the School of Chemistry at Leeds University.
The journal Chemical Physics offers a full-featured automated electronic archive and distribution server of its preprint database. Hypertext abstracts can be viewed online, and the source of the full paper can be downloaded easily. The journals J. Am. Chem. Soc. and Chem. Rev. offer supplementary material on the ACS gopher server. Tables of contents and abstracts are available for the Applied Spectroscopy journal. Springer Journals offers a preview service requiring payment, tables of contents are distributed free of charge. The first issue of the electronic Journal of Molecular Modeling has been announced for January 1995.
Several online journals and periodicals are available in biology and medicine. True electronic journals are found in the fields of mathematics and physics, e.g., the Electronic Journal of Combinatorics, the Electronic Journal of Differential Equations, the New York Journal of Mathematics and the Journal of Artificial Intelligence Research (also available in print).
It should be emphasized that although the publication of preprints on an institutional WWW server offers a quick means of informing colleagues about new research results, the services of (electronic) publishing companies are still indispensible. Only a professional organization can guarantee quality standards by the established peer review system, authenticity of the documents, proper archiving and continuity of the service. Moreover, there is certainly a need for supplementing the online publication by permanent media such as CD-ROM or printed copies.
Besides papers, conferences are an important means of spreading scientific information. Saving the costs and troubles, but also the more pleasant features of conferences, they may be held electronically. An example is the First Electronic Computational Chemistry Conference held in November 1994.
A searchable index allows the user to enter a search expression. Three typical examples are: an acronym database (look for abbreviations, e.g., "COSY"), abbreviations of chemical compound names (e.g., "DMSO") and our searchable chemistry index.
A lot more flexibility is offered by forms. One typical application of forms (using the "post" method) is to supply information, e.g., a message that is mailed to the administrators of a WWW server. An example is our form for the registration of documents related to chemistry which may be used to enter pointers into our chemistry index. Another typical application (using the "get" method) is to query some kind of database, for example, safety data of chemicals. In this case, we have developed a CGI script and a special-purpose database program which yields HTML output providing further links. Thus, textual explanations for the numbers of R and S safety sets are given on clicking at the respective anchors.
There is an excellent WWW interface available for querying the Brookhaven Protein Data Bank (PDB), making use of a form (PDB Browser). It stands to reason that capabilities for substructure searches via WWW are lacking so far, but the development of an appropriate interface should be feasible and might have a big impact on professional online database searches.
Another application, which is not restricted to chemistry, is offered by online library catalogs; in this field, WWW interfaces are gradually replacing the unpleasant "telnet" facilities. Actually, any resources of the WWW server may be used by means of the CGI concept, e.g., there is a form for the conversion of units calling the Unix utility program "units".
A further way of user interaction with a WWW server is provided by clickable maps, allowing the user to point at a particular spot or region of an image to obtain specific information. An obvious application is found in geographical maps, e.g., the map of German WWW servers. Our welcome page allows a quick selection of one of the main topics by clicking at the corresponding key word or icon. The citric acid cycle represents a nice example of an application in chemistry.
Another extension refers to spectra or chromatograms. It is often desirable to expand a certain portion of, e.g., an NMR spectrum. Clearly, a simple magnification of an image stored as pixel graphics would not solve the problem. Instead, the xy coordinates should be supplied to a plot program such as xmgr or xgraph. Again, an example may be found on our page about pinene.
MIME extensions for chemistry must be installed on the server side as well as on the browser part, e.g.,
chemical/x-pdb pdb chemical/x-xyz xyz chemical/x-mol mol application/x-xy xy
chemical/x-pdb; rasmol %s chemical/x-xyz; xmol -readFormat xyz %s chemical/x-mol; xmol -readFormat alchemy %s application/x-xy; xmgr %s
Another example for "chemical MIME" may be found on our pages dealing with amino acids, and there is a nice demonstration of hyperactive molecules by Henry Rzepa and Benjamin Whitaker. Moreover, a viewer such as xmol even allows the display of molecular animations, e.g., reactions or vibrations.
For the sake of commercial providers, several schemes for access restrictions are available (e.g., by host, by site or by password) or under development.
Currently the WWW system, which is still in its early stages, is in rapid development and growing exponentially with regard to the number of servers, documents and users. However, it must be admitted that presently several shortcomings and disadvantages are encountered. Thus, HTML 2 is rather restrictive; it does not provide support for sub- and superscripts, mathematical formulas or tables. However, these features will be present in HTML 3 (or HTML+); the first browsers suitable for that format, e.g., arena, are in their test stage.
Another problem is of technical nature. The promised "information superhighway" is not yet reality. In practice, access to remote servers is often slow, particularly at peak hours, or temporarily impossible. Unfortunately, quite a few WWW servers have a bad habit of shifting around their documents, or names of servers are changed, so that document URLs change. Therefore it would be advantageous if the present way of explicitly addressing URLs could be replaced by some indirect or symbolic link to an information server.
The third big problem concerns the question of "indexing the Web", i.e., providing means of quickly finding a particular piece of information. In principle, such indexes may be compiled manually, by schemes such as aliweb (list-based), or by robots that traverse the Web automatically (spider-based). Clearly, none of these methods is really satisfactory. For particular purposes such as the development of distributed databases, an alternative to WWW such as Hyper G might be advantageous; however, Hyper G is not scalable, more difficult to implement and less flexible than WWW.