The "Knows" and the "Know-Nots" plus Learning from Potatoes
The "Knows" and the "Know-Nots"
Here is quote from a memorable article in Scientific American by David PogueProbably there is no other field as affected by the disarmingly simple discovery that "there is no one type of tech consumer" as the high performance computing (including high throughput, Big Data and more).We recognize and name many of our demographic tribes: liberal and conservative, wealthy and poor, gay and straight. We classify ourselves that way, we watch out for oppression against one group or another, we pass laws to ensure equal treatment. But when will we recognize the existence of the two different types of technical consumers—the Knows and the Know-Nots?Over and over again, I run into products that have been tacitly designed for either group. The creators have a mental picture of a product's audience and the users' technical experience. You can tell from terminology, the amount of detail in the instructions, the number of steps required to accomplish anything.Unfortunately, there is no one type of tech consumer. Someone winds up unhappy. If the design and interface are too technical, novices feel incompetent, shut out and stupid; if the experience is too simple, tech geeks feel insulted and talked down to.
The "Knows"
Most of documentation, papers, users groups in performance computing are for the knows . They set the tone. I have experience with HTCondor-users groups, Grid Engine groups, Open Science Grid All Hands meetings. There are two types of knows: (1) those who really know and (2) the one who really are FalseKnows, i,e. Knows-Nots mimicking to appear as Knows.
The Know-Nots
In my experience, the Know-Nots are minimum 95% or more in a user population in HPC eager to use powerful resources. Now they feel humiliated, fed up, neglected, frustrated, discriminated to remain second class citizens among the research and engineering professionals.This is one of the reason we started the San Francisco High Performance Computing Meetup
Democratizing access to High Performance computing ?
Access? Which access? The Top500 supercomputers, not only are very hard to use, but almost impossible to reach. They are all owned by government, defense, security, spies, academia and a few are privately owned by large companies. They have no public access and are managed based on policies from from a hidden Castle from a Franz Kafka book, the one you can see and never get there.
The same applies to Open Science Grid. I talked about this subject before , see Make money with Entrepreneurial Performance Computing
So there is time to switch the focus from very expensive computers no one can access, or large grids using throughput computing administrated by policies. They are all modern Kafkian Castles
This is happening already
So there is time to switch the focus from very expensive computers no one can access, or large grids using throughput computing administrated by policies. They are all modern Kafkian Castles
Public Cloud and Hybrid Solutions
So before we start simplifying the access to HPC applications per se , lets focus on enabling public IaaS clouds, like AWS, \Azure, RackSpace to run more and more sophisticated HPC/HTC solutionThis is happening already
Dealing with Eastman Chemical
In 1998 as part of Genias Software, the company founded by Wolfgang Gentzsch which later became Gridware managed to get Eastman Chemical as a client for CODINE.
See a copy of the newsletter from December 1998
At the time, we did an effort to understand what Eastman does. We wanted to learn, we wanted to please, we wanted to make the extra mile effort. We were the underdogs, LSF was the market leader, and their people had a semi-permanent grin of superiority at each trade show we met. Dr. Bruce Wilson from Eastman , a very able researcher, liked not what we knew, but our desire to know.
What the customers want: Learning from potatoes
This is the secret of Eastman success. They have no set goals like discovering o sort of Higgs particle in twenty years. They want to deliver what the customers wants, now. There is art and magic here.
So the problem is, if you are a geeky Know, sitting at your desk, feeding in numbers, how do you satisfy the desires of a very sophisticated British CEO who needs to deliver special packaging for potatoes?
Comments