May 20, 2018
I have a LOT of partially-written blog posts, but am struggling to get any of them finished (obviously). Much of the problem is that they have so many dependencies on each other. Clearly, then, I should consider refactoring my writing plans. 🙂
So let’s start with this. Here, in no particular order, is a list of some things that I’ve said in the past, and which I still think are or should be of interest today. It’s meant to be background for numerous posts I write in the near future, and indeed a few hooks for such posts are included below.
1. Data(base) management technology is progressing pretty much as I expected.
- Vendors generally recognize that maturing a data store is an important, many-years-long process.
- Multiple kinds of data model are viable …
- … but it’s usually helpful to be able to do some kind of JOIN.
- To deal with the variety of hardware/network/storage arrangements out there, layering/tiering is on the rise. (An amazing number of vendors each seem to think they invented the idea.)
2. Rightly or wrongly, enterprises are often quite sloppy about analytic accuracy.
- My two central examples have long been inaccurate metrics and false-positive alerts.
- In predictive analytics, it’s straightforward to quantify how much additional value you’re leaving on the table with your imperfect accuracy.
- Enterprise search and other text technologies are still often terrible.
- After years of “real-time” overhype, organizations have seemingly swung to under-valuing real-time analytics.
3. Outside traditional enterprises, the accuracy problem can be even worse, and the consequences of analytic inaccuracy can be severe. In some cases this is well understood; autonomous vehicle researchers, for example, seem properly attentive to the challenge of not-killing-pedestrians. But in others it’s a mess. For example, I don’t think the “fake news on social media” challenge will be resolved without new technical approaches that, to my knowledge, aren’t yet even being tried.
4. More generally, I’ve long argued that the technology industry would someday have to deal with a variety of public policy and social concerns. That day has come. In anticipation, I wrote at length about privacy/surveillance, and a little about some other areas, including net neutrality, patents, economic development, and public technology spending. Missing subjects include censorship (private and public alike), and perhaps also at the efforts to tie data ownership into anti-trust policy.
5. Given all the tech-specific public policy work that’s needed, I’m pulling back from some my broader political efforts. However, I stand by my overview opinions of last February, and I delivered on some of its IOUs in a two-part series on persuasion.
6. The ongoing rise of “edge computing” and the “Internet of Things” fit into the general trend that in 2013 I summarized as appliances, clusters and clouds.
7. I continue to think that a huge fraction of analytics is properly characterized as monitoring. That ties into a number of areas of interest. For example:
- Platform technologies — including distributed data management — are often compete on the maturity of their built-in monitoring.
- My complaints about BI inaccuracy commonly relate to use cases in monitoring.
- Privacy/surveillance issues are commonly about monitoring. It’s common to worry that such monitoring is actually too accurate.
- But I also worry that privacy/surveillance monitoring isn’t accurate enough … and hence that it leads to people being discriminated against who absolutely don’t need to be.
- Edge computing involves a lot of devices that need to be monitored.
- Censorship obviously has a lot to do with monitoring.
8. And finally for now, my core precepts for strategic messaging haven’t changed.
- As you may have already guessed, the title of this post is based on a classic song.