The Student Union at Manchester Uni used to do a great fusion meal comprising a yorkshire pudding filled with chicken jalfrezi. Top nosh!
9 posts • joined 26 Jun 2009
I may be in a small minority here owning two Lumias, a Surface Pro 3 and a Band (1) but I'm really happy with them all. That's enough about me. Over the last 18 months I've completely trashed my Band as I clock up a lot of mountain and street miles through running and cycling. I've been waiting expectantly for Band 3 and have been getting ready for the let-down for a while - you don't get to be as big a Microsoft (and Nokia) fan as I am without being ready for the let-downs!
So, what will I do? Well, I want an activity logger that does everything that the Band does, as well as the swimming thing. I want the steps, the hearbeats, the notifications, etc, etc, etc. Band 3 already exists. It's called the Garmin VivoSmart HR+ and has been around for a few months. It even works on Windoes (most of the time). I'm buying one tomorrow. Microsoft were using Band to learn about wearables and IoT. As the article says, they (bizarrely) nailed it. Apart from the economics of being a Windows-based platform.
For about the last 20 years, Walmart has been more about shifting inventory and now has a shared private b2b cloud with most of its major suppliers to optimise this. Even old Sam Walton himself took this view - here's a quote of his:
"“People think we got big by putting big stores in small towns. Really we got big by replacing Inventory with Information.”
He was doing the Amazon thing for groceries long before Amazon with the first terabyte-scale retail/supply/logistics database in the mid 90s.
A good train of thought and I like the “atomic” theme but I think it’s over-polarising the issue a little.
It is possible to achieve the image analysis examples as User Defined Functions in Teradata and Oracle have a massive range of industry-specific “cartridges” that achieve this type of functionality. It depends a lot more on the business reasons for applying this type of functionality. In discovery mode, a more forgiving schema-lite environment is much more flexible for getting to grips with your problem (e.g. am I developing the best plane-spotting algorithm). Once the approach has been hardened for use in an operational environment (e.g. face scanning at an airport) then it goes on to the RDBMS platform to be integrated into broader analytics.
It's a neat piece of work (speaking as a geophysicist) but is it really Big Data (speaking as someone who works on Big Data in the geophysical world)? "Big data" usually has connotations around the need to crunch the data to produce some insight in a challenging timeframe and whose form (and maybe content) changes more rapidly than a system and data architect might be comfortable with. This is a "big" volume of data but surely it's just processed as a batch job every hour in a predictable and consistent way? Thanks for posting the article anyway!
Biting the hand that feeds IT © 1998–2019