9305 -- News

May World-Wide Web News

The last W3 news was in November of last year. Since then so much has happened that as usual we haven't had time to write a newsletter. But here for those who aren't on the mailing list is a summary of what's been happening. In this issue:

NCSA's Mosaic for X

The National Center for Supercomputer Applications , (authors of the famous NCSA Telnet for the Macintosh) have a team of developers working on WWW software. Their first product, Mosaic for X , took the workstation world by storm as being the most polished and sophisticated W3 browser available for X workstations.

One of the delights of this product is that one only has to FTP the binary over, set it executable, and run it. There is no setting up to do. This has put thousands of people on the web. The binaries, and sources if you need them, are available by anonymous FTP from ftp.ncsa.UIUC.edu. Look in directory /Web/xmosaic.

Features of Mosaic, which after 15 or so beta releases is now at 1.0, include embedded graphics and launching of multimedia applications. Mosaic keeps track of all the documents you have ever read, and colours the link differently if it leads to something you have read before. This is useful not only for avoiding old things when looking for new, but also for following your tracks through unknown territory. The bookmark list, known as a "hotlist", is saved between sessions as a private list of interesting places. You can even add personal annotations onto any document, which will will appear each time you (but only you) read it. The look and feel is Motif (though you don't need Motif to run it, as the binaries come with the libraries linked in). Marc Andreesen, the author, has done a really good job here.

tkWWW Hypertext Editor

News is just in of a new release of tkWWW we have only just tried out today. TkWWW is a WWW client for X windows which uses no Motif, but the public domain tk/tcl tool kit.

The exciting thing about this release is that it is Joe Wang's first pass at a wysiwyg hypertext editor. This allows the direct editing of hypertext documents, including link creation. It is in the early stages, but developers are encouraged to try it and make suggestions.

TkWWW is available by anonymous FTP from info.cern.ch, in /pub/www/src as well as various other sites.

Macintosh browser released

CERN has now released a W3 browser for the Mac . This uses multiple windows, and linked text is coloured blue or underlined on black and white displays. The application is available on Appletalk zone LDI_2, machine "Robert Cailliau", either by public folder or by Appleshare (guest user). Outside CERN users use anonymous FTP from info.cern.ch, in directory /pub/www/bin/mac. You MUST use MacTCP version 1.1.1 or later with it, as there was a bug in the version many people have (available inside CERN from the same source).

NCSA also have a prototype of Mosaic running on the Mac, and this should be available later in the summer.

What about the PC?

The most frequently asked question about WWW is now, "What about my Windows machine?". I can tell you than I have on my notebook an advance copy of the "Cello" browser by Tom Bruce of Cornell's legal Information Institute, and very nice it looks too. The catch? Tom doesn't want to let it out until he has polished it. He plans a July release, with beta test versions available in June.

Meanwhile, several other people are working on alternatives, which are hot on Cello's heels, including a Windows version of Mosaic from NCSA.

CERN W3 Software Goes Public Domain

After much discussion, CERN has now released a certain subset of the W3 software as public domain code. This was done in order to further the spread of the web, and to ensure that the protocols were used in a consistent way. There had been a danger of developers having to all re-invent the protocol code, and remake incompatible mistakes. Now the "libwww" code library which forms the basis of many of the browsers is in the public domain, this can serve as a common basis. It deals with the HTTP protocol handling, and also Gopher, WAIS, FTP , local file, and News access, as well as parsing of the hypertext format (HTML), and negotiation of other available formats.

Also released into the public domain is the basic server, httpd, and the line mode client.

Easier to set up servers

The news release of the CERN server is now easier to set up. You just run it under the inet daemon (or stand-alone if you like). Giving it the name of a directory on the command line will simply export the directory tree.

The server automatically generates hypertext representations of you directory structure. Directories are listed with the subdirectories first, and then the files. The text of README files, if any, is included in the directory listing at the top, bottom or not at all depending command line options. The server is available in source and binary form by anonymous FTP from info.cern.ch in /pub/www/src/WWW_Daemon_2.05.tar.Z and /pub/www/bin/*/httpd_2.05. Take a later version if it exists. The rule file is still available for more complex server situations.

Servers are also available for VMS and VM/XA, the latter being written largely in REXX for flexibility.

Use that gopher data

If you already have a "Gopher" server on a unix machine, you will have created various special files to define your gopher tree. A new server from NCSA , also in the public domain, allows you to export all this information using the world wide web's "HTTP" protocol. Like the CERN server, this synthesises hypertext, but in this case uses the gopher-style .cap files, etc., as input. This will give your readers a better view of your information than when they go through the gopher protocol. You will be able to maintain one database for access though both servers in parallel. Of course, you might be tempted to write a bit of hypertext here and there too!

The NCSA server is available by anonymous FTP from ftp.ncsa.uiuc.edu.

A number of other servers are also available, allowing all kinds of neat features including access control by IP address, gateways to all kinds of things (including gnu "info" trees). Check the web for details .

Direct access to WAIS servers

The WAIS project uses a special protocol, a variant of the Z39.50 library access protocol, to talk to full text search engines. Until now, W3 browsers had to go through a gateway to reach WAIS servers. The gateways still exist, but it possible to compile a client to have the ability to talk WAIS protocol directly to these servers, which currently exist on about 113 hosts worldwide, some of which have many databases.

This feature allows the regular httpd server to act as a WAIS gateway just by setting the rule file appropriately.

Those wishing to compile clients in this way will also require the "freeWAIS" distribution from CNIDR.org, which although not public domain is freely available.

Multi-format documents

The full HTTP protocol used by most new W3 clients (not yet in Mosaic) allows the client to tell the server what data formats it can handle. This allows the server to chose a format if it has alternatives, or if it has conversion programs.

If you have the W3 server (httpd) version 2.05 or later, you can simply put copies of a file in different formats (such as xxx.tiff, xxx.gif, xxx.ps) in the directory, and make a link to xxx.multi. The server will return whichever one a client can handle.

This opens great new possibilities for sharing complex data, as the system is not limited to standard data formats: a group can declare its own private data formats so that members of the group can use their preferred applications to handle the source of data such as word processor documents, computer-aided design (CAD) drawings, etc.

The exploding web

Lots of new information is out there, with 64 servers registered at the last count. In the physics area, the database of -- was changed on request from durham in 2000/5/5 to durpdg.dur.ac.uk TBL --> particle properties and other good things from Durham and RAL, and many new sites on line, including more information from DESY in Hamburg, and from the Czech Academy of Sciences in Prague. See also the corresponding article "WWW in Experiments" by R. Cailliau.

Outside physics, examples of new information are the " Lysator " computer society in Sweden, MIT's oldest journal "The Tech" , NCSA's "Access" magazine , and High-Performance Computing information from Loughborough University, UK. The Legal Information Institute continues to roll out more US law in hypertext, while the University of Vermont's Scholarly Communication Project serves a number of journals.

The subject catalogue, also known as the W3 virtual library, as grown in various areas, but volunteers (librarians?) are still needed to keep it up to date in specific areas.

Searching FTP archive sites with Warchie

The famous "archie" index of (almost) all files available by anonymous FTP has a WWW gateway now called "Warchie". This software is currently running at Nexor Ltd in the UK where it was developed, on the end of a rather slow link, and so looking for another home. Any archie sites would be welcome to take this software and run it locally. Warchie his gives very easy access to the world of FTP archives. There are five linked indexes which search the database in different ways. Each returns a hypertext list which points directly to the archive sites found.

(previous issue: November 1992 )

Tim BL & RC