- From: Craig R. McClanahan <crm@perpetua.dat.com>
- Date: Thu, 18 May 1995 16:56:34 +0800
- To: cdackus@knoware.nl
- Cc: www-html@www10.w3.org
>>>>> "Christian" == Christian Dackus <cdackus@knoware.nl> writes: Christian> I was wondering, for a week I went to visit a big dutch Christian> publishing business. They are on the web too Christian> (http://vnu.ib.com). Our firm has been bought by this Christian> so called VNU. Christian> What I saw there was a demonstration via Netscape, of Christian> some magazines they are planning to publish online very Christian> soon. The lay-out was like this: Christian> Christian> Index: Pull-down menu with all the chapters/articles Christian> Very nice, but the interesting thing was, that by Christian> choosing one option of the pull-down menu, the browser Christian> skipped to the requested chapter. I tried to access this page, but the host never answered -- so I don't know for sure what they did. However, I can conceive of how you might create this kind of visual appearance using image maps. This is a technique that displays an image with the ISMAP attribute. The client browser then lets the user click on the image someplace (probably on what looks like the menu bar), and sends back to the host the mouse coordinates (relative to the corner of the image) where the click occurred. Thus, if your image had a line that looked like the Netscape menu bar: ________________________________________________________________________ | | | File Edit View Go (and so on ...) | | | ------------------------------------------------------------------------ Then, if the user clicked somewhere near the word "Edit", it would send back the coordinates to a script. The script would figure out which "option" was "selected", and send back a new image that has the same menu bar at the top, plus the opened "Edit" menu, listing all of the options. You could extend this technique to do a pretty good imitation of a menu bar's user interface, simply by having a separate image file for each possible state, and a script smart enough to figure out which image to display next. For more info on image maps, you can consult one of the many online HTML references (I always start at the NCSA home page when looking for stuff like this). You'll probably also need to review how to set up executable scripts, which should be discussed in the docs for whatever HTTPD server you're running. A common standard for interfacing between an HTTPD server and your script is called the Common Gateway Interface (CGI) standard. On a UNIX system, the input from an image map (or a form, for that matter) comes in as a bunch of environment variables. The "script" itself can be a shell script, a PERL program, an executable written in C, or whatever you need that can manipulate environment variables and write text to standard output. Christian> Now I was not able to see the code, so I tried myself Christian> to create it, I did it like this: Christian> <form method=post action=""> <select> <option><a Christian> href="..">Chapter 1</a> <option><a href="..">Chapter Christian> 2</a> </select> </form> Christian> Well, don't try it, it won't work. So now my question Christian> is, what kind of code do I have to use. Is it some new Christian> feature of HTML 2/3, is it only possible with the Christian> commercial version of Netscape, or I am to stupid to Christian> find out myself how easy it is (which I really believe) Christian> to make a pull-down with links? Christian> Who can help me out? Christian> I am sorry, but that is NOT all. I have another Christian> problem, namely something of which I think that it is Christian> difficult indeed. I need to make a script, I think, Christian> that can put some names (taken from the name of some Christian> folder/directory) into a standard HTML document. Christian> p.e. <title>hi</title> <h1>hi</h1> <p> <img Christian> src=".......1.gif"> <p> img src="......2.gif"> Christian> etc. I need to fill in the names where the dots Christian> are. TIPS? One approach would be to explore whether your server knows how to display directory contents that it builds dynamically for itself. For example, the CERN server does this -- it also lets you define a little icon to go next to each filename, based on it's extension. When the user clicks on one of the filenames in the list, it is automatically retreived (i.e. the filenames are really hypertext links). This is the way most servers deal with FTP type requests. Another way to do what your talking about relies on using a script, as was discussed earlier in relationship to image maps. The key concept is that you can output anything you want from a script, including dynamically created HTML text! A very trivial skeleton of this might be a C-Shell script that works like this: echo Content-type: text/html echo echo '<head>' echo '<title>Image List</title>' echo '</head>' echo '<body>' echo 'Here are the images that matched your request:' echo '<p>' cd /your/directory/goes/here set Filenames = `ls *.gif` foreach Filename ($Filenames) echo '<img src=\"' $Filename '\">' echo '<p>' end echo '</body>' As you can see, the actual text of the page is constructed dynamically, based on the names of the GIF files in the directory at the time the script is executed. You could easily extend this to make the filenames hypertext links instead of the images themselves, by dynamically constructing an anchor element (<a href=xxxxx>xxxxx</a>) format, instead of <img src>. Christian> Regards, Good luck! Christian> __________________________________________________________ Christian> Christian Dackus # # # # # # Manager On-line Services Christian> Limburg on-line # # # # # # cdackus@knoware.nl Christian> __________________________________________________________ ---------- Craig R. McClanahan EMAIL: crm@dat.com DAT Services Phone: 503-643-4331 Beaverton, OR, USA Fax: 503-526-6442
Received on Thursday, 18 May 1995 19:53:37 UTC