Everything_Computers_Logo.JPG (16666 bytes)

IWE Logo.gif (3354 bytes)

Nav Bar.GIF (5852 bytes)

Plugged In

December 2, 1996

Will Automatic Software Updates Encourage Laziness in Developers?

By Bob O'Donnell

As I discussed last week, automatic software updates via the Web are here and they're going to make a big impact on both end-users and IS departments. But frankly, no group is going to be more affected than software application vendors.

The Web is proving to be an ideal channel for distributing software, particularly when it comes to small bug fixes and the like. No longer do companies need to worry about informing all their users when a big (or even not so big) problem arises, sending out update disks, etc. They just include updating technology in their current app and let the Web's reach do the hard work for them. Users get the latest software and developers avoid the distribution problems associated with upgrades. It seems like a good deal all the way around.

But is it? Will developers use this as an excuse (as if they needed another one) to send out buggy software, thinking they'll just fix it in the first online rev? Imagine this scenario: Company A is working on an important product that's behind schedule, but is expected to provide critical income to meet Wall Street's predictions for the company's earnings. A decision is made that the product must ship three weeks before the end of the quarter to bring in the necessary income -- regardless of whether the program is finished or not. Management justifies the action because of the easy availability of automatic updates on the Web. The product ships at deadline, but end-users and IS managers waste countless hours trying to resolve problems resulting from the program's lack of polish and insufficient testing.

Sounds pretty plausible, doesn't it? The sad truth is, this probably already happens more than we know, even with companies that don't have the benefit of easy updates via the Web. I guess we should be thankful that automatic software updates can at least help some of us get past the problems that poorly implemented development practices have spawned. Call me ungrateful, but I don't think that's good enough.

Unfortunately, these policies seem to be a mere continuation of software development and distribution practices created by the Web. I remember, for example, when beta testing meant working with a targeted group of people who banged away at a product before it was shipped. Vendors on the Web have created a culture where the release of a new beta is bigger news than the shipping release, and using beta applications for critical work functions has practically become de rigeur. The fact that vendors release betas of the next version of an application a few weeks after they've released the previous shipping version just adds to the problem.

Easy Web-based software updates aren't all bad, though. Properly executed, they can bring real value to applications, such as by adding a popular feature in a small interim upgrade. Smart vendors can use such updates to greatly endear themselves to their customer base.

And for IS departments, a well thought-out strategy of working with automatic software updates could actually help solve Help Desk problems instead of just creating them. Castanet Tuner technology from Marimba, for example, could be used to ensure that everyone upgrades to the newest rev at the same time, avoiding the problem of multiple versions of an application running across an enterprise. On the negative side, however, Castanet can perform updates without any real warning, doesn't offer an easy way to cancel an upgrade, and can quickly suck up hard disk space, again, without much warning. What if an organization doesn't want its end-users to have a minor upgrade, or more importantly, what if an automatic upgrade in one app breaks another mission critical app? These and other similar questions still need to be resolved.

Like many changes in the computer industry, the move toward automatic software updates doesn't necessarily provide any easy answers or complete solutions to common problems. The concept is undeniably appealing, but it's not clear yet that the implementations will do everything we hope.


© Copyright 1996, by InfoWorld Publishing Corp., a subsidiary of IDG Communications, Inc. Reprinted from InfoWorld, 155 Bovet Road, San Mateo, CA 94402. Further reproduction is prohibited.

 

 


Home | Radio | Television | Books | Magazines | Consulting | What's New

Search | Feedback | Troubleshooting Guide | Audio | Site Map

Send mail to bob@everythingtechnology.com with questions or comments about this web site.
Copyright © 1997- 2005 O'Donnell Enterprises. All rights reserved.
Last modified: January 01, 2005
Web site hosting provided by Global Network Services