the web as a database

[Howard Gibson]

On Wed, 20 Apr 2005 15:49:36 -0400
Zbigniew Koziol <zkoziol-Zd07PnzKK1IAvxtiuMwx3w at public.gmane.org> wrote:

> 
> Just an idea. May be someone will want to comment?
>
> ...
> 
> Some may suggest the use of XML. Probably a good idea. I do not have 
> however a general understanding of what is behind XML. In principle, I 
> imagine, web sites could have a special file hosted on their server that 
> would contain a detailed information about the content of these web 
> sites, or at least about the company. Like in the example above. A sort 
> of like now /robots.txt is used, or newsfeed.xml .
> 
> Is there no other way?

Zbigniew,

   A couple of years ago, I wrote a Perl program for managing a hiking club schedule.  I generated the schedule as a set of SGML files.  Based on the date, and what the user requested, the Perl program either displaying the schedule, or the current hike.  I thought I was pretty clever.  

   XML was developed from SGML, and my process was probably closer to XML in that my tags are case sensitive.  The thing never went into production.

   Someday, when I have to pass the hiking club web page on to a new user, I will have to teach them HTML.  HTML is a simple enough language that the web page can be maintained by a hiker.  Hiking club no hiking club, HTML is a useful thing to know.  Learning it is not a waste of time.

   My SGML DTD is useful for maintaining hiking club schedule websites.  It will work for other schedules as long as they like the way I formatted the HTML.  Only I know the DTD.  There are no books on it, particularly not the kind for dummies.  There is no software for displaying it other than my Perl program.  The only authoring software is a text editor.  At least (x)emacs can read my DTD, and tell you what the next legal tag is. 

   The main outcome of my project is that I learned some Perl.

   How about a methodically structured HTML file?

   For my purposes, an HTML file would contain a table with a specific number of columns, with the first one being the date, and second one being the destination, etc.  A program reads through the HTML file searching for the first table, and then the first <td> tag.  The contents becomes the first date.

   This HTML file can be prepared using vi.  It can be prepared using any HTML editor.  It can be prepared using Microsoft Word!  The HTML file is functional on its own, without my program.

   In your case, you need to arbitrarily define a <meta> tag, which tells a searchbot that is is the web page of a physics lab.  You can supply a couple of keywords in another <meta> tag.  You can create a summary paragraph, flagged by a CSS class. 

   Again, all of this can be maintained using existing software.  The <meta> tags and CSS classes are all supported by HTML authoring tools.  The physics labs that already have web pages can do a minor edit on them, and they work with your tool.

-- 
Howard Gibson 
hgibson-MwcKTmeKVNQ at public.gmane.org
howard-42qnO8ePF9cV+D8aMU/kSg at public.gmane.org 
http://home.eol.ca/~hgibson
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml