Internet Messaging

Feature
Internet Messaging I

Adapted from the Prentice-Hall text Internet Messaging, From the
Desktop to the Enterprise, by David Strom and Marshall T. Rose.

Push Technology

If we had asked the question of being your own newsletter publisher in early 1997, the popular answer would have been to make use of a “push technology” product. Defined as automated delivery and/or notification of information via the Internet, many companies came out that winter with push products. What a difference a year makes: Many of these companies are either out of business or rapidly trying to redefine their purpose. Push technology has gone from hypeful to hopeless in less time than high fashion changes.

The idea behind push is as simple as our posed problem: sending content, whether it is a page of HTML or a series of advertising images in the form of animated screen savers, automatically to a given audience of users’ desktops. The particular stream of content is called a channel, and many of the push products have both a means of creating your own channel (such as the Example Corporation’s Human Resource Newsletter) and becoming a subscriber to other people’s channels. Channels could be private, within the corporate intranet, or public, such as the ESPN sports scores and Dow Industrials average. These channels vary tremendously. They could contain pointers to particular Web pages out on the Internet, or contain the actual information, such as graphics and text, that would be saved to your own hard disk.³

Yet the implementation is all but easy. Push products really didn’t have the publishing tools at all. You often didn’t know who your audience was, couldn’t tell what software they used to view your content and preparing content often took loads of time and was a “hit or miss” proposition. Most of the products couldn’t even tell you whether your readers actually received your content, let alone if they spent any time reading it.

Push suffered from several issues. First and foremost, you want to get your information delivered to your e-mail inbox, whatever and wherever that may be. While many push technologies install special software on your desktop or augment your browser with plug-ins or other software, most of us use our inboxes as ways to order our day’s priorities. Given our already bloated hard disks full of other software, the incremental piece of push software was too much for many of us to handle. It seems like a small point, but it isn’t. And we saw lots of users tired of getting so many cutesy screen-saver animations and other digital effluvia. They quickly turned the push channel off, uninstalled the software, and went back to using their e-mail for some real work.

Most of the push products didn’t really take advantage of e-mail at all—they 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 was a problem, because everyone uses different browser versions, different platforms and different configurations. Installing browser plug-ins is not always simple, and not every push vendor supported a wide enough range of operating system platforms either.

There are plenty of other problems with push implementations: How many push screens does it take to send a page of HTML to a particular desktop? How do push clients maintain and consolidate standard Web server log information in a way that can be useful to the Web server analyzers? Can a push publisher make any valid claims equating the value of desktops receiving their messages to the number of ad impressions delivered via Web banners?

Contrast this browser situation with e-mail now. E-mail is pretty much a bread and butter application—this is why you are reading this book. Everyone’s e-mail does work differently, but getting messages sent doesn’t require you to install extra software on your machine. You just send the message. Most e-mail programs don’t have plug-ins or extensions⁴ that require you to become a part-time software installer and troubleshooter.

Second, push ate bandwidth like nothing else, and became a pox upon the network. Products like PointCast behaved so badly that any mention of them could cause a long string of curse words from many IS managers’ lips. It wasn’t unusual to see 15% or more of a corporation’s overall T-1 circuit consumed with the network traffic resulting from PointCast software.

Push was hungry for bandwidth largely because of some sloppy programming on the part of its creators, who worked with fat Internet pipes and in small companies. Once this network faux pas was realized, the push creators moved quickly to cut their bandwidth consumption. However, by then the damage had been done, and corporate IS managers wanted little to do with some of these products.

Push also had no real standards to build upon—every vendor had its own scheme for notification and delivery of pushed content. This was especially true for Microsoft and Netscape, which developed their own incompatible software, protocols and systems. Some of the push prowess depended on the Web and HTML. Some worked at lower-level TCP/IP protocols. The wide variety of push differences continue to bedevil the push players, and even the applications developed for one company’s early software versions aren’t compatible with later ones.⁵

To make matters worse, few could agree on what push really means. What about products that didn’t really send any content at all but polled a particular server at specific intervals: Shouldn’t that be called scheduled pull? What about products that just organize Web sites that you have already visited: Shouldn’t that be called something else?

We had an opportunity to use many of these products and personally meet with the CEOs from many push companies. We asked them the same question: How do you want to receive information from us, via your own software or via e-mail? The universal answer was e-mail. Not eating one’s own dog food is the best reason we can find for avoiding the whole push arena entirely.

So what alternatives remain in a world without push? E-mail. And any notion of sending groups of e-mail messages brings us to the problems with using mailing lists. Mailing lists are relatively simple to set up, as we’ll show later in our section on solutions. But maintaining them and keeping their addresses current is another story entirely.

The real advantage that e-mail brings to the push party is universal notification. We’ve seen an explosion of such services and uses of e-mail by a wide variety of commercial and noncommercial vendors. For example, our favorite online bookstore Amazon.com can send you e-mail when a new book matching certain criteria or from certain authors is published. GreetSt.com, an online greeting card company, can send you e-mail to remind you not to miss sending that certain someone a birthday card.

We get daily e-mail telling us the closing prices of our investment portfolio, and other e-mail with news digests related to particular technologies. Most of the airlines have mailings set up to remind you of travel bargains, including American Airlines, which has different mailings for domestic and international fares. There is a site called RemindMe that reminds people when to move their cars to avoid getting parking tickets. The site can also send reminders for other events at varying intervals before any particular date.

All of this has implications for managing mailing lists, which is our next topic.

3 A different application that has some similarity to push is the use of Internet newsgroups to send messages to a group of users. This is outside the scope of this book.

4 A notable exception is Eudora. Qualcomm started using a spell-checker plug-in for Eudora a few versions ago, and now there are several plug-ins that are available for the product. Luckily, getting them set up is a lot easier than trying to install browser plug-ins, and many of them come bundled with the main Eudora software itself, so you don’t have to try to install them separately.

5 A good example of this is Intermind’s Communicator, one of the earliest push players. For those who publish their own channel, Communicator stores a series of pointers to various Web pages on your hard disk. The database format for the early version couldn’t be used for the next version, and even after the company got behind a single format, you still couldn’t move this database from one machine to another without a great deal of effort.