Pretty much every cloud service failure these days results in a predictable volley of coverage; the unbelievers crow over another purported example of the perfidy of over-hyped utility computing solutions, while advocates point out that software failures in any venue are hardly unique and that economics, if nothing else, dictate that these solutions will improve and dominate the future of software provisioning. Google AppEngine's relatively short outage yesterday was no different, but it is different from the last AppEngine outage: this time, Google didn't get a pass on the negative press.

I think that is a good thing, despite the spirited defense that Clint Boulton mounts in his Googlewatch column over at eWeek. My impression the last time around was that Google was reaping the benefits of its well-rehearsed expectations management game... labeling the service a beta, keeping it free, orienting it less at the enterprise and more toward individual developers, who are more forgiving and understanding in these matters. That may make for good press, but it doesn't help Google put together a service that the enterprise is going to feel comfortable with. Boulton takes on the argument that Google is unprepared to develop and support enterprise-grade services by citing the citing the enterprise backgrounds of the technical leadership at the company, but that is a straw man. No one (certainly not Jason Hiner, who most of his criticism is leveled at) is suggesting there is a lack of experience or expertise at the company. The problem is that of intent. Google isn't really trying to deliver enterprise services, and the culture of the company (regardless of where individual employees came from) reflects that. They are famously reluctant to put out anything and commit to standing behind it in ways that the enterprise can count on. Getting an effective SLA out of them for Apps Premier edition was like pulling teeth, and didn't fully materialize until last year. They may well know the enterprise; they certainly are not catering to it.

That's why I think this higher level of criticism is good. I believe the company has been hoping to slide in under the radar, coming into the enterprise more from the bottom up than the top down (and I think that can be a valid strategy in some cases). But taking that route avoids taking much responsibility for your product, and that ultimately leads to weaker products. While I think Google generally develops pretty valuable software, it lacks polish and often stability. Those aspects are the last mile problem of software development, and without a strong push from somewhere, most developers never conquer them. A little public critique might be exactly what Google needs to take Apps, and AppEngine, from good to great.