On Tue, Jun 5, 2012 at 7:44 PM, Jason Grout <jason-s...@creativetrax.com> wrote: > To be fair, though the notebook allows for code and documentation to be > intermixed, it's not the traditional literate programming tool that allows > you to, on the one hand, produce nice documentation, but on the other hand, > produce a single runnable file. And it's a far cry from something like > noweb, that can reorder your code when it "tangles". > > The ipython notebook is closer to this, since you guys explicitly export to > a python file, and also export (or will export) other formats that nicely > present everything.
And to me that's very deliberate. I've never really liked 'real' literate programming in that the very concept of it is completely batch oriented: you write code and text in one place, then you run it via something that produces a program to be then executed. Even though the results look often excellent, my preferences are firmly set on the mathematica/ipython/sage/etc approach that keeps you working *inside* the environment where you *execute* the code. I want to have my objects in memory while I play with them, refining each cell as I go and gradually molding the code into what I need as I both refine the code and understand the problem better based on the results. It's possible that some of the newer literate programming environments from the R people have changed in this regard, I have to admit I haven't followed that world very closely. But the classic sweave approach is most definitely *not* my cup of tea, and not something we're trying to emulate in any way in IPython. We do make it possible to produce importable code automatically from a notebook (and have ideas on how to improve that further), and our latex, rst, html, etc conversion machinery slowly but surely is improving, so I'm not saying that we have something that we consider finished and ready. But I do have a pretty clear vision of the kind of computational environment we're trying to create (and one where the parallel extensions fit very naturally in a way that I don't find in these other models and which I think is increasingly important). cheers, f -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
