On Thu, Apr 10, 2008 at 3:29 PM, Joris De Ridder <[EMAIL PROTECTED]> wrote: > > On 10 Apr 2008, at 23:23, Travis E. Oliphant wrote: > > > Cool. I started scipy.misc.info a long time ago to try and do > > this. I > > didn't advertise it well enough ;-) > > Yep, I also started to write my own docsearch tool but neglected to > advertise it. > > > > > On 10 Apr 2008, at 23:58, Gael Varoquaux wrote: > > > Numpy is not the place where this should go. This should go in eg > > ipython, which provides an interactiv frontend. One of the reasons is > > that this is certainly not numpy-dependent. > > I kind of disagree here. Such a tool can be made much more intelligent > and thus useful if it really hooks into NumPy/SciPy. Provide "See > also" functionality, give info on categories of functions, use magic > words, etc. None of these things is too difficult to implement. > However, such capabilities are beyond the more general docstring > search tool that would be suitable for IPython. Of course, no problem > if IPython is willing to special-case NumPy/SciPy.
Given how I happen to make a living with numpy/scipy, the answer would be yes ;) A slightly more proper answer: we can have this special-casing easily available without 'polluting' the ipython core code, since ipython can load fairly self-contained features, even automatically. So this extension (which Paul just provided as an ipython ticket, thanks!!) can be loaded only if numpy is around for use. Cheers, f _______________________________________________ Numpy-discussion mailing list [email protected] http://projects.scipy.org/mailman/listinfo/numpy-discussion
