I once worked for a University group doing medical research. We had literally 10s of millions of records of patient treatments and 100s of millions of records of perscription data to analyze. A researcher would come to us with a question, such as "What drugs are being used to treat AIDS patients, what is the length of treatment, and what is the average cost of treatment for each drug used?"
We would extract out a subset of the data - say 1000 records or so - and design a query that should give the data needed. We would then run it against the subset records and verify that it worked. Once we were satisfied, we would run it against the main database. Then, for the next two days, we would read, go get coffee/sodas, talk, or take classes.
Eventually, the answers would come back. Then, we had to bring the numbers into something that could produce the pretty graphs and charts that the researchers wanted to insert into their research papers.
Often, the end result would be the researcher looking at the charts and saying, "Oh... I meant to have you restrict the query to drugs in this class rather than all drugs!"
When I first started my career as a professional programmer,
I worked for a large company that had that exact same
disclosure statement in its employment contract. When I
asked them about modifications, they had a simple answer: "
You don't like the clause? Don't sign and we'll move on
to our next candidate."
One thing that I noticed in the article that was glossed over was the statement that these "independent" evaluations were done using Microsoft's guidelines for comparison.
So... if I commission a study and tell you that in doing the cost/benefit comparison a nice box is more important than low cost...
I have written several device drivers for both Windoze
and Linux. In my opinion, the Linux drivers were
far more straightforward to write and maintain. As
long as I could get _complete_ information for
the hardware, I would choose to work on a Linux
driver over a Windows driver if I were given the choice.
Slashdot Mirror
We would extract out a subset of the data - say 1000 records or so - and design a query that should give the data needed. We would then run it against the subset records and verify that it worked. Once we were satisfied, we would run it against the main database. Then, for the next two days, we would read, go get coffee/sodas, talk, or take classes.
Eventually, the answers would come back. Then, we had to bring the numbers into something that could produce the pretty graphs and charts that the researchers wanted to insert into their research papers.
Often, the end result would be the researcher looking at the charts and saying, "Oh... I meant to have you restrict the query to drugs in this class rather than all drugs!"
Sigh... Back to the drawing board.
When I first started my career as a professional programmer, I worked for a large company that had that exact same disclosure statement in its employment contract. When I asked them about modifications, they had a simple answer: " You don't like the clause? Don't sign and we'll move on to our next candidate."
One thing that I noticed in the article that was glossed over was the statement that these "independent" evaluations were done using Microsoft's guidelines for comparison. So... if I commission a study and tell you that in doing the cost/benefit comparison a nice box is more important than low cost...
I have written several device drivers for both Windoze and Linux. In my opinion, the Linux drivers were far more straightforward to write and maintain. As long as I could get _complete_ information for the hardware, I would choose to work on a Linux driver over a Windows driver if I were given the choice.