Was webAgent a good idea?

WI’ve often wondered if writing webAgent was really a good idea or not. This is especially true for webAgent 2.

It was a bad idea

The main idea here is that webAgent kept us isolated from the mainstream of web development and slowed our adoption of more standard web technologies. Without webAgent, we would have needed to adopt Java or Python or some such a lot sooner. In the meantime, we created a large collection of “legacy” applications that will have to be rewritten or replaced at some point.

It was a good idea

Without webAgent, it probably would have taken the University years to develop an administrative web presence. Whatever issues it might have caused for the future, having webAgent in our toolset meant we were able to provide many useful services to the University community that otherwise probably wouldn’t have happened.

Also, almost any solution we might have gone with in the 1990‘s would probably be obsolete and in need of replacement by now. Having webAgent gave us breathing room until solutions like Python and Django were available. In the meantime, we gained a lot of experience that helped us recognize the value of these tools when we were ready to evaluate them.

Verdict?

When the verdict was called for, the Jury declined,
As the word was so puzzling to spell;
But they ventured to hope that the Snark wouldn’t mind
Undertaking that duty as well.

Lewis Carroll, The Hunting of the Snark, Fit the Sixth

If you’d asked me a few years ago, when the PyPE project was in its early stages, I think I would have inclined more to the “bad idea” side. I was starting to worry that people were so content with webAgent that we’d never get motivated to move on. Now that PyPE has been successful for several years and less and less development is being done it webAgent, I’m feeling a lot more positive about it.

Keys to success

What factors led to webAgent’s success? It may seem counterintuitive, but I think the fact that it was never intended as a permanent solution contributed significantly. It helped keep the projects small and focused instead of bloated and complicated. This was particularly important for webAgent 2, which could easily have fallen prey to the second system effect.