Web Informant #174, 1 November 1999:
Son of (return of) push

Remember push technology, that darling of the moment of 1997? Back then, dozens of vendors had created products that could notify you periodically of new content, and some actually delivered this content to your humble desktop. At one point in time, I experimented with delivering these essays via push, only to return to tried and trusted email.

Push went from hype to hopeless within a few months. Yet despite this experience there are several new attempts to parade push to some unsuspecting new marks. I've been using two products lately: HP's Instant Delivery and another product from Entrypoint. Both are bad ideas and neither company has learned much from the original lessons of why push failed.

Let's look at this historical context for a moment. I'll summarize the notable problems of push and number them for subsequent reference:

1. The products really didn't have the publishing tools to manage the look and feel of the pushed content.
2. As a push publisher, you didn't know who your audience is and couldn't tell what software they used to view your content.
3. Early push products ate bandwidth like nothing else, and became a pox upon the network.
4. Most of the push products used the browser either as the control panel to tune in to a particular channel or as a container to deposit the information itself. This sounded good in theory, but was a nightmare to manage with the various different browser versions.
5. Even though they used a browser for controls or containers, push required additional software to be installed on your desktop. For a variety of reasons, no one wanted to deal with this extra download.
6. Finally, push also had no real standards to build upon for its channel definition and was caught in the cross fire between Microsoft and Netscape.

So let's look at push c.1999. HP Instant Delivery is an interesting idea: you have a PC with a printer, both of which you leave on all the time and connected to the Internet. You download some software (5) that manages what channels or publications you subscribe to and what frequency they are delivered. These take the form of printed documents that appear magically in your printer once you set everything up (1). Anyone can become a publisher: you just set up a rather lengthy form pointing it to particular web content (2). HP admirably provides guidelines on what potential publishers should do to improve the look of their content, suggesting steering clear of frames and animations and keeping your HTML simple (6).

My first attempt with HP-ID was troubled: I tried to obtain the latest Dilbert comic strip, only to find an empty frame at the top of the printed page. I thought at first that it was a conspiracy, since I was printing it out on my Lexmark laser (rather than an HP), but when I viewed the content online I also saw an empty frame. Subsequent Dilberts (and other things I subscribed to) came through fine. It was a little spooky watching my printer start up and generate the pages I requested during the day, though.

Entrypoint is the result of the marriage of Pointcast (one of the original push players) and eWallet. The idea here is to have a single place to go with news (3), stock quotes, and all your wallet-type information you need when you shop at a web storefront. It takes the form of a small control panel that sits on top (or bottom) of your screen, and displays information in your browser (4). You can store passwords and user names for various sites too, and even play your audio CDs and MP3s. It seems like the Internet version of the Ginzu knife. (It slices, it dices, it pushes!)

The real issue here is the setup time. There is a lot of stuff to type in and to get just right, and sometimes the eWallet feature just doesn't work properly in filling out the forms at certain storefronts when you have several browser windows open.

Of course, with both HP-ID and Entrypoint, you'll have to download some software (5). The Entrypoint client requires a browser to get it configured and working properly, HP-ID uses its own client in conjunction with your browser to select your channels. Both are lacking when it comes to publishing tools to report and track on what users are doing (1), (2). Both ignore the existing channel standards set up by Netscape and Microsoft (6): even though Entrypoint makes use of Pointcast technology, I could find no documented way to add my Web Informant channel.

Add up all these problems and we have a pig in a poke once again. Neither push product, in my opinion, has a bright future ahead. Neither product will make push more popular: it is far easier for people to use email for notification and delivery of content. It is far easier to set up bookmarks to return to sites you want to visit regularly. It is far easier to set up one of the many custom "My" pages (MyYahoo, MyNetscape, MyExcite, etc.) to display the information you need without having to trouble with downloading any additional software outside of your browser.

If you are interested in reviewing more of the historical record of the various push players, you can go look at lots of links to the vendors and articles and analyses written back in 1996 and 1997.

To subscribe, send a blank email to
webinformant-subscribe@egroups.com

To be removed from this list, send a blank email to
webinformant-unsubscribe@egroups.com

David Strom
david@strom.com
+1 (516) 944-3407
back issues
entire contents copyright 1999 by David Strom, Inc.
Web Informant is ® registered trademark with the U.S. Patent and Trademark Office.
ISSN #1524-6353 registered with U.S. Library of Congress.