from Stephen Hedges, OPLIN Director
After yesterday's stunning announcement about the demise of the OPLIN email service, people are understandably beginning to wonder, "How was this decision to terminate the OPLIN email service made?"
First, let me emphasize that the OPLIN Board did not have an opportunity to discuss this decision. In fact, because of an email problem, they did not even receive notification of the decision until after it had been published on the Internet. This was my decision, and here is how it happened.
After days of unsuccessfully trying to restore the service (see the "gritty mail server failure details" posting elsewhere on this site), it was clear we needed to explore other options. My assumption was that action needed to be taken quickly, because we all have become very dependent on email to conduct our daily business.
Option 1: Restore the service to its previous state from backups.
That didn't work. Defying all logic, the previous version of Zimbra running on the previous operating system refused to restart when supplied with backups of the previous files. When this option failed, we knew we had a very, very serious problem.
Option 2: Hire someone to help us fix it.
We contacted Zimbra Inc. support on Sunday and they eventually called back. The conversation was not encouraging. We described the situation and their response was basically, "We'll think about it and get back to you." As of Tuesday morning, we have not heard from them again. We seem to have no hope of quickly getting things back to their original state with this option. (We still have some hope that we might be able to hire them to try and recover our old data -- eventually.)
Option 3: Start from scratch with a new service hosted elsewhere.
This option would probably take at least a week to implement and cost OPLIN between $150,000 and $390,000 per year, depending on whether or not we would be required (as a state agency) to use the state's email system. The Board actually discussed this option last April and rejected it. (Yes, we've been concerned about supporting the email system for at least a year now.)
Option 4: Start from scratch and build a simpler mail server, like the one before Zimbra, that's not as complex.
Though people were not big fans of the old, simple OPLIN email, we decided to try this anyway. The result was a little surprising.
First, you need to know that an email system consists of two basic components: a mail transport system, and a user interface.
Mail transport systems have not changed a lot since the early days of the Internet, and are fairly standardized. They only handle the delivery of email from one machine to another. That's why we were able to quickly build the temporary transfer server.
User interfaces, on the other hand, continue to get more complex. The user interface is the part of the system that allows you to retrieve email from the server, read it, save it, organize it, search for emails, and these days save emails as calendar appointments and shared documents and any other bells and whistles the software providers think might give them an advantage in competing for user business. This complexity makes it hard for anyone to maintain a particular system who is not a dedicated expert.
The current version of our old user interface is no longer simple and has become so complex that it would take us a while to figure out how to make it work and longer to figure out how to manage it. So in the interest of getting something up quickly, and not ending up with another complex system to maintain, I rejected this option.
Option 5: Do Not Resuscitate
Admit failure, pull the plug, and communicate quickly to libraries the need to arrange for different, more dependable email service. Meanwhile, work on some simple tools to transfer incoming mail to new accounts (done) and attempt to retrieve the old email data (in process).
The last option is the one I chose, as the other options crumbled away. Did I exceed my authority? Unquestionably I did, and will have to answer to the Board for implementing my decision. In my defense, though, it felt less like a decision and more like a surrender.