surprise from IPython’s evolution
I watched the video and tried to write down what he said. Maybe someone can help me improve it.
Q: You created IPython back in 2001, what problems were you solving
at that point?
A: Basically I was trying to switch away from using property tools to using
open tools and I want to simplify a lot of my workflows from many many
languages to use less, when I discovered Python, I realize it can replace
a lot of different tool which just one language can make it easier keep
kind of ?move workflow
Q: So you make your own things, right?
A: Yes, but what was missing in Python was a good interactive computing environment.
Python has a basic interactive shell, but doesn’t fit the kind of workflows
that needed in scientific computing.Or You want to look at script, you wanna look at
data, you wanna look at plots variables, you want change the code little bit, a
very expiatory workflow and so that being a graduate student looking for excuses
not to work on his dissertations and instead to do something more fun were the perfect
combination to start writing IPython.
Q: What surprise that you feel about IPython evolution?
A: I guess the facts precisely that something began as such a simple personal fix for
a problem in my own workflow has grown into such a large project. IPython was first
Python Program that I ever wrote and for the looks it’s going to be the last I have ever
write. Because it began as an afternoon hack and here we are at 13 years later,
and fully literature, so the fact that would be able to learn from that one use case
of a personal scientific workflow
and begins to abstract and abstract way and building extremely broad aspect tool
now serves other needs including education, publication, scientific computing at that
scale has been a really surprising and interesting path that something that was born as
?? Both and community sense. That community has been so eager to work with ?? and adopted
the projects really not my work but actually the product
of our entire community of people more talented that I am and who do all hard work.
How do IPython Notebook come about?
That was something that the current iteration that everyone uses is about our 6 times
attempt… That came from tools that existed that very widely used in scientific computing community. I was a heavy user of sistance?? Maple and Mathematica, both of which have notebooks like environments, the Mathematica once called notebooks, the Maple called work sheets, that combine the idea of having texts, mathematics, code and the results of the code all in one documents, and that’s a very natural for scientists. So from day one, when I
started the very first mini script that was IPython 001 which is about 250 lines of codes
are ready to have mentions of mathematics in it. And I was a heavy Maple and Mathematica user so we were after those ideas from early on. But neither of our technology, nor the abstractions,
nor our own understanding of our problems were ready in those early 5 prototypes were missed steps but along the way, we learned many many things, and eventually both our ??capitalizing of kind essential ideas with the framework of the moral web with taking the bad notion and moving it over to web browser, with a lot of inspirations from project code sage, that build something in 2006 for mathematics, kind of took us to the moral?? notebook,… so it was a long road but the intention was there from the beginning, and in 2011, Brian Granger who is my Co PI in a lot of projects … was able to prototype what became the current version over across the summer, joining the technology that was just came out online, like web sokate, like JQuery, like tornado and Z&Q, which had the time all very new, but they kind of fit perfectly in having tried So what appears like very new solution to many people is something we actually been work with and thought over 15 years.
The intention was there since the beginning as I said,But obviously it was too a big problem for me as a graduate thinking, I want something like that there was no I am gonna do, so the first cut terminal basic environment, but the first prototype
5 prototypes all of which dead ended
those dead ends were very productive, they showed us all the wrong decisions and all the things
critical the facts that tools like Gmail showed everyone that you could build highly responsive,
highly interactive environments in the web that
difference because it was machine … was really small team to build something that exactly
production?? for many scientists.
How much programming do you feel that scientists need to know?
It varies quiet a bit depending on exactly what were doing, I think it’s important to know that
for scientists, programming has very
different flavors that adores software developers , programming is inertially exploratory, it’s
really about solving a question, it’s okay to write one scripts, and at the same time,
but scientists also need to learn how to
gradually understanding when they do need to abstract a little bit, when that data pattern
that data set going to be similar in other problems they should not be copied pasted
50 time but they should beginning moving a little bit functions and having a little
base finding a way.
Q: What you take on the notion of everyone should learn the code?
A: Why do we code? We code because we are trying to communicate with machine which does not have human
style natural intelligence how to accomplish task, and in addition I mean a dump machine, in coding
we have to
clarify to the point that dump machines can understand our thoughts process what it is we are going to do,
It means that we have to organize the process in a very precise and unambiguous way, we have to perhaps
abstract basics can be done
by using the reusable components etc, and that process of logical system organized abstract computational
thought that I think is a useful valuable skill in many context in life…