Notes for WWW Information Publishers at UO

Writing Documents for use on WWW

Most web publishing consists of making available files that contain the information you want to publish. There are many possible formats of information you might place in such files; whatever kind of information you publish, you should name your file with a "file type" (sometimes called the file "extension) appropriate to the kind of document it contains. Some commonly seen file formats on the web include:

ordinary ascii text    .txt
HTML hypertext         .html or .htm
GIF format images      .gif (good for icons, line art, etc.)
JPEG format images      .jpeg (good for photos)
sounds                 .au (8 KHz ulaw; other formats also possible)
movies (MPEG)          .mpeg
Adobe Acrobat (PDF)    .pdf

An important issue you need to be concerned with is copyright. Except for materials covered under "fair use" provisions of the copyright law, you must be careful to obtain permission for the use of any materials, especially those developed outside UO. Note, though, that creating a link to someone else's materials is just citation, not copying. The flip side of copyright protection is that you (or the UO, if it's a work for hire) probably hold copyright to anything you publish on WWW, but given current norms it is likely that your copyright will be routinely violated -- expect anything you publish to be copied.

Another important issue is accessibility. Particularly if you are writing UO official materials, it is important to make your work as accessible as possible, e.g. to blind "readers" who may use a web browser that speaks the text they are "viewing". The UO Adaptive Technology Lab provides assistance in making technology accessible to UO students with disabilities.

Once you have created your materials, it is a good idea to take a look at them with several different WWW browsers, since different browser software will present your materials differently. In particular, it is generally desirable to review your materials using both some version of Netscape and using Lynx, since the differences in appearance between graphical browsers and text-only browsers are often very significant. What may be less obvious is that things often look different on a 640x480 Mac from how they appear on a 1280x1024 X terminal.

Getting Started

• Introduction to HTML documentation from U Toronto [extensive]
• Creating HTML Documents from NCSA [introductory]
• HTML Quick Reference Guide from U Kansas [terse]
• Web 66: cookbook from U Minn [Mac-oriented]
• HTML Documents from Clarkson College [tutorial]
• HTML Tutorial from Case Western Reserve [tutorial]

• Composing Good HTML by J. TIlton [introductory]
• Yale C/AIM WWW Style Manual [advanced]
• Elements of HTML Style, by Jonathan Cohen [literate]

• NCSA list of style guides.

Many tools exist to ease authoring of HTML documents and conversion of existing documents to HTML format. For example, if you have a Microsoft Word or Wordperfect document, you can convert it to HTML. Specialized editors for HTML are becoming widely available, as are addons to standard word processors that generate HTML. The set of tools is changing very rapidly, so it is hard to make recommendations as to the best current tools. For an extensive list of web-related tools, see the Yahoo Web catalog from yahoo.com

Publishing

When publishing materials of general interest to the UO community, it is usually desirable to announce their availability with an email message to webmaster@www.uoregon.edu. Include in the message the URL of the main entry point to your materials, and your best guess as to where links to the materials belong in the UO web.

Some general references on Web publishing are:

• What every information provider should know by Meng Weng Wong
• Publishing on the World Wide Web by Gareth Rees
• Webiquette by Brigitte Jellinek [idiosyncratic]

Providing information for www.uoregon.edu

Most information on that server consists of links to other information on other servers at UO. If you have put together a body of related material that you believe is of general interest to the UO community, we can include a pointer to it. Send email to webmaster@www.uoregon.edu.

Providing information on darkwing and gladstone

1. Create a directory in your top level darkwing or gladstone directory called "public_html".
2. Create your web page in public_html.
3. Make sure your directory and the files in it are world-readable, and that your home directory, the public_html directory, and any subdirectories, also have "execute" i.e. Unix search, access.
4. To reference your work, tell people to look at the URL http://darkwing.uoregon.edu/~myusername/index.html
   where myusername is replaced by your darkwing user name, and index.html is replaced with whatever file name you created. If you were publishing on gladstone, the "darkwing.uoregon.edu" would be replaced by "gladstone.uoregon.edu."

You can create HTML web documents (names ending in .html), GIF images (.gif), plain text (.txt), MPEG movies (.mpeg), scripts that generate documents (.cgi), etc. You can also create subdirectories of your main public_html directory.

Once you've created a web page on darkwing or gladstone, you can request that the Computing Center include your index page (index.html) in the list of web authors by sending email to webmaster@www.uoregon.edu.

Providing information on the VMScluster

1. Create a subdirectory in your top level directory called "WWW". For example, you could use the command
   CREATE/DIRECTORY [.WWW]/PROTECTION=W:RE
2. Create your web page in your [.WWW] directory, e.g. index.html.
3. Make sure your directory and the files in it are world-readable. For example,
   SET FILE [.WWW]*.*/PROT=W:R
   Note that every time you create a new file in your WWW subdirectory you must change its protection, since the default protection on the VMScluster is no public access.
4. To reference this, tell people to look at the URL http://www-vms.uoregon.edu/~myusername/index.html where "myusername" is replaced by your VMScluster user name, and index.html is replaced with whatever file name you created.

You can refer to files in subdirectories of your WWW directory by specifying a path in Unix-like syntax. For example, a file [MYUSERNAME.WWW.IMAGES]PHOTO.GIF could be found at http://www-vms.uoregon.edu/~myusername/images/photo.gif. Although file and directory names on the VMScluster are case insensitive, it is recommended that you use only lower case in VMScluster URLs.

Once you've created a web page on oregon, you can request that the Computing Center include your index page (index.html) in the list of web authors by sending email to webmaster@www.uoregon.edu.

Providing information on a departmental web server

Managing a WWW Server

• "Web Etiquette" for information providers (CERN)
• World Wide Web: HTTP Servers (Yahoo)
• Primer on WWW Servers

Some Frequently Asked Questions

Question: What's wrong with my HTML file?

Question: My images don't load. What's wrong?

• image not in GIF or JPEG format, perhaps because you copied the file from your PC in something other than binary mode (in Fetch, that's "Raw Data")
• errors in your IMG tag's SRC value, e.g. no quotation marks around URL, or reference to a file name that doesn't exist
• file protection problems (is the image file publicly readable?)

Question: why can't I access the files I've published on OREGON?

Question: How do I include a vistor count (odometer) on my page?

Question: How do I use imagemaps?

Question: can I avoid learning HTML by buying a web editor like PageMill?

Question: I'm trying to write a CGI script, but when I test it I get
500 Server error

500 Server Error

The server encountered an internal error or misconfiguration and was
unable to complete your request.

A common cause of this problem is that the script, when run, generates a compiler error rather than the output you (and the server) expected.

For example, suppose I have a perl script that used to work fine (using perl 4.1). The current version of perl is more picky, and finds errors in my script. Perhaps the script has a line in it like:

$user = "jqj@darkwing.uoregon.edu";

darkwing% ./test.cgi
Literal @darkwing now requires backslash at ./test.cgi line 8,
within string
Execution of ./test.cgi aborted due to compilation errors.

$user = "jqj\@darkwing.uoregon.edu";

Question: How do I create and maintain a log file using cgi scripts without making that file wide open to the world?

1. instead of writing to log files, have your CGI script generate email. If you then want to be sophisticated in processing the email, consider using procmail at the receiving end (which runs as you, not as an anonymous user).
2. instead of using darkwing, use a departmental or personal web server where you have more control over the execution environment and where security is less likely to be a major issue since your customers won't have accounts.
3. make your CGI script setuid to yourself using chmod. For example:

   #!/bin/sh
   # this is ~jqj/public_html/foo.cgi, and is  chmod 6755 foo.cgi
   echo "Content-type: text/plain"; echo; id 

   When run from the net, this produces:

   uid=60001(nobody) gid=60001(nobody) euid=9301(jqj) egid=160(lib) 

   Setuid programs are potentially dangerous.to your personal file security if you make a coding error. Perl is a good choice of language for setuid programs, since it includes lots of code designed to detect common coding/security errors. Despite the example above, you should never use setuid shell scripts, since they are very insecure.