Over the history of the Internet, there have been many attempts made to use web-based technology to disseminate time-critical information, such as stock quotes and weather reports. One of the first of these attempts was the auto-refresh/client-pull feature built into the web browsers, which is accessed via the
A "client", in HTML-speak, usually means the web browser. "Client-pull", then, means that the web browser itself, without action initiated by either the user or the web server, can be programmed to automatically "pull" resources from the server at specified intervals.
META tag with its auto-refresh/client pull feature, a web programmer may set a delay of some number of seconds on an HTML page; once that page is loaded into a web browser, a delay of that number of seconds occurs, after which the web browser automatically pulls another given HTML page from the server, replacing the current page.
Have you ever been to a web site where, after a given number of seconds, you are automatically sent to another page? This behavior uses the
META tag auto-refresh/client-pull feature. It is frequently used for "splash" pages leading into a website, where you see a page with a single graphic (usually including some sort of short-cut link to the main page) which, after a delay, automatically takes you to the main home page for the website. It is also used with a "redirect" page, where you see a forwarding address (the new URL for the website) which, after a short delay, automatically takes you to that new site.
The original purpose of this auto-refresh/client-pull feature was to automatically call pages which contained constantly changing, time-critical information, such as stock quotes, from a web server. The idea was that, after 30 seconds or a minute of displaying this time-critical information, the browser would automatically pull that page from the server again, displaying the fresh, new information that the user needed.
The problem with this process was that the user actually spent more time staring at the blank browser window, waiting for the page to refresh, than reading the page itself in the brief intervals between refresh cycles. With some pages, the user could wait up to two or three minutes (or more) for the page to reload, after which the page would be readable for only 30 seconds or less before beginning the refresh cycle again. This was not a very practical means of getting volatile business or weather information to users.
Enter "server-push" technology, which a few years ago was hailed practically as the second coming of the Messiah to Internet developers, but which has since dwindled into obscurity. The idea behind server-push technology was that a user could subscribe to a service which would continuously feed time-critical information to a user's web browser or screen-saver; this information would run all day long across the user's computer, whether the user was there to appreciate the information or not. The good part of this technology was that the stream of information was continuous and unbroken, so stock tickers and weather information and news items could smoothly run across the screen, disseminating time-critical material in an efficient and seamless manner. Internet developers were thrilled because this server-push "broadcasting" was not only great at disseminating time-critical information, it was also great at disseminating ADVERTISEMENTS, which meant MONEY! The bad part of this technology was that it caused time-critical information to run smoothly across the user's screen all day long, whether the user was there to appreciate the information or not; this stream of information took up a permanent chunk of the Internet pipeline (the T1 or T3 line) leading into a business. With enough users in an office subscribing to these server-push services, the entire bandwidth of a business' Internet connection would be filled, making it impossible for anyone in the office to access the Internet, send email, surf the Web, etc. Subscriptions to server-push services were cancelled for this reason, leading to the demise of wide-spread server-push services.
This leads me to a topic which I must expound upon, in all good conscience. With all of the talk about broadband or DSL high-speed Internet connections into homes and home offices, we must not forget that the essential telecommunications pipelines, the T1 and T3 lines, connecting businesses and web servers to the larger Internet, are still relatively expensive, running about $1000 per month for a single T1 line. One T1 line can take only about 30-40 simultaneous users of (marginally) FM-radio-quality audio streams before becoming completely filled up, and maybe only half that many low-quality video streams; it could cost tens or hundreds of thousands of dollars per month to reach as many simultaneous streaming audio users on the Internet as you could reach with a low-powered college radio station. Until the cost of the pipelines goes down, and until more telecommunications infrastructure on the Internet as a whole is built, true internet-based broadcasting (or any extensive high-bandwidth usage from a single site) will be impractical, strictly from a cost standpoint. Just think how slow the Internet is during peak times with the current small file sizes and minimal downloads, and you can imagine how long it's going to take before there's enough telecommunications pipeline in the world to support video broadband Internet broadcasting on any scale at all and still allow people to access their email and surf the Web. And USER access to high-speed connections is still quite limited, even within large cities in the USA.
The cost of T1/T3 infrastructure (or similar technology) MUST come down soon, of course, with DSL and other services offering similar speeds for rock-bottom prices, but it's going to take more time. Everyone has been talking about "broadband" for years now, but true broadband services over the Internet are still many years away.
Now, back to the
META tag and its auto-refresh/client-pull feature.
Having said all of this about the
META tag and its original purpose, I can safely say that the auto-refresh/client-pull feature has been relegated to 1) making automatically advancing splash pages, and 2) making automatically advancing redirect pages or forwarding address pages.
In the next section, we'll look at the actual code involved in creating this type of
Copyright © 2001 Michael Masumoto. All Rights Reserved.