My IT experience was with the network of a large utility. There were applications in use which contained millions and millions of records. They ran 24/7 with the exception of downtime for nightly back-ups. None of them took more than several minutes to restore service to the users after those back-ups. Below is a quote of mail received from one of my members who has made a very good living the the field. I believe that he knows of what he speaks.

73, Bob Vallio -- W6RGG

------------------------------------

ARRL promotes LoTW as a "service product"; a "feature", something they are proud of, something they offer to the amateur community, and something that is supposedly a benefit of ARRL membership (e.g., the ability to get rewards without handling paper QSLs). It is NOT like many other amateur radio computer products in the realm of "I built this for my own use. You are welcome to use it, but no guarantees."

I have designed and built many PRODUCTS, including hardware and software. Some of them have shipped in quantities of tens-of-millions. I don't ship any "works in progress." That doesn't mean that my products have never had problems that needed fixing, only that they were thoroughly tested BEFORE release, and subjected to extensive design review BEFORE even being built. They are the result of my best tradeoffs among features, cost, reliability, and time-to-market. I put them out there, and am willing to be measured by their success or failure.

It is fairly clear, both from discussions I have had with ARRL and others, and from experience, that LoTW began as a "weekend project" by some ambitious programmer, and that it caught on and grew. In fact, it grew way beyond the capabilities of the original "design." (I believe a lot of the problems with LoTW have to do with the choice of database, and to a certain extent with the amount of resources that ARRL is willing to devote to the project.) Good designers build prototypes, and test them extensively. Then they start from scratch to build a PRODUCT from what was learned in the prototype stage. LoTW appears to be a "prototype" that was "shipped".

Of course, there will be problems. However, ARRL has indicated many times that they are either oblivious to them (e.g., log processing delays: "What problem?") or they consider turnaround times of days or weeks to be an acceptable design decision. (In contrast, log processing response times on both eQSL and Clublog are nearly instantaneous, even after a big contest weekend.) They also don't consider it important to provide any feedback to their user community on the progress of repairs, or the progress of log processing. (A simple indicator of the number of logs in the queue, or the expected processing time, would alleviate a lot of angst.) A message that "The system is down. We're working on it." is not very reassuring. Neither is the occasional cryptic error message from their database engine that occasionally comes up when you try to log into the system. (Sort of like hitting the UP elevator button from the lobby of a beautiful hotel, and instead getting dumped in the base
ment with the steam pipes.)

Yes, there may be times when the system must be brought down for major reconfiguration. But any reasonable service provider would inform their user community of such an event BEFORE it happened, and not respond only after users complain about the disappearance of the service.

Now, if the problems were the result of something totally out of ARRL's control, then all bets are off. For example, the IEEE last week had major problems with their e-mail forwarding system, due to outages caused by Sandy. Other than not incorporating a hot standby backup system, that is not a problem with the IEEE's system design or operation. But many of us believe that LoTW is in need of a re-DESIGN, and that is from a rather knowledgeable user community. So far, ARRL has been non-responsive.

That said, I agree that *anything* is better than nothing. But as long as ARRL touts their system as a "real" service, then they deserve the criticism that goes with that. If they want to call it an "extended beta test," that's fine too, and the expectations will be much lower. They have not chosen to do that, though."