sage notebook work planning
William Stein
Mike Hansen and I just spent a while carefully going through:
http://wiki.sagemath.org/SageUsability
and updated it with more ideas...
William
--
William Stein
Associate Professor of Mathematics
University of Washington
http://wstein.org
Fernando Perez
>
> Hi,
>
> Mike Hansen and I just spent a while carefully going through:
>
> http://wiki.sagemath.org/SageUsability
>
> and updated it with more ideas...
I wonder if it would be totally crazy to consider using reST as the
native notebook format, or at least have it be optionally so.
I know that one can already execute 'live' sphinx docs in html, which
is super cool. But given that reST has support for input and output
code blocks (which it can even doctest), I'd really like to basically
edit in the notebook itself reST sources, with text cells being plain
reST, input cells being code blocks and output being 'text' code
blocks, as suggested here:
http://groups.google.com/group/sphinx-dev/browse_thread/thread/ecbcea0ab4d1ca78/86e3da68e1576f7a
I don't know enough about all the sphinx machinery yet to see if there
are any fatal flaws in this idea, and given the amount of existing
worksheets out there obviously such a change couldn't be done lightly.
But I'm so happy with using sphinx more and more, that being able to
simultaneously edit reST and execute code seems really appealing to
me.
Does this sound possible/reasonable/useful?
Cheers,
f
Jason Grout
Along these lines, I just started to write/adapt a short tutorial on
Calculus in Sage. In hopes that eventually would make it into some sort
of documentation or primer or something, I tried to write it somewhat in
ReST. You can see the result here: http://sagenb.org/home/pub/791/
My hope is that it will be easy to convert from a worksheet into a
document. I found it a wonderful way to write documentation, and kept
wishing I had a better way to structure the document (i.e., collapsible
headings, etc.)
Jason
--
Jason Grout
William Stein
>
> Fernando Perez wrote:
>> On Tue, Sep 8, 2009 at 6:45 PM, William Stein<wst...@gmail.com> wrote:
>>> Hi,
>>>
>>> Mike Hansen and I just spent a while carefully going through:
>>>
>>> http://wiki.sagemath.org/SageUsability
>>>
>>> and updated it with more ideas...
>>
>> I wonder if it would be totally crazy to consider using reST as the
>> native notebook format, or at least have it be optionally so.
>>
>> I know that one can already execute 'live' sphinx docs in html, which
>> is super cool. But given that reST has support for input and output
>> code blocks (which it can even doctest), I'd really like to basically
>> edit in the notebook itself reST sources, with text cells being plain
>> reST, input cells being code blocks and output being 'text' code
>> blocks, as suggested here:
>>
>> http://groups.google.com/group/sphinx-dev/browse_thread/thread/ecbcea0ab4d1ca78/86e3da68e1576f7a
>>
>> I don't know enough about all the sphinx machinery yet to see if there
>> are any fatal flaws in this idea, and given the amount of existing
>> worksheets out there obviously such a change couldn't be done lightly.
>> But I'm so happy with using sphinx more and more, that being able to
>> simultaneously edit reST and execute code seems really appealing to
>> me.
>>
>> Does this sound possible/reasonable/useful?
Jason below.
>>
>
> Along these lines, I just started to write/adapt a short tutorial on
> Calculus in Sage. In hopes that eventually would make it into some sort
> of documentation or primer or something, I tried to write it somewhat in
> ReST. You can see the result here: http://sagenb.org/home/pub/791/
>
> My hope is that it will be easy to convert from a worksheet into a
> document. I found it a wonderful way to write documentation, and kept
> wishing I had a better way to structure the document (i.e., collapsible
> headings, etc.)
>
> Jason
>
>
>
> --
> Jason Grout
>
>
> >
>
Jason Grout
At the same time, I think it would be big mistake to give up easy
WYSIWYG editing of text cells. Is it easy to convert between HTML and
ReST (both directions)? If so, I think it would be easy to make a text
cell toggleable between TinyMCE and a textbox that would let you put in
ReST.
Jason
--
Jason Grout
Brian Granger
At the same time, I think it would be big mistake to give up easy
WYSIWYG editing of text cells. Is it easy to convert between HTML and
ReST (both directions)? If so, I think it would be easy to make a text
cell toggleable between TinyMCE and a textbox that would let you put in
ReST.
Yes, I agree. For developer types, ReST is wonderful and quick. But for many potential Sage users
(like all of my students) ReST would be a huge step back from the current WYSIWYG editing.
Jason
--
Jason Grout
Fernando Perez
Certainly! What I'd love to have is reST as a native storage format
(so it can be easily edited outside the GUI with emacs) along with a
choice of visual/native editing in the notebook interface. It's just
that in switching from visual to native you'd go to reST as the 'raw'
format instead of pure html, and the html would be only rendered for
visual display.
This would also let us produce a pdf of any notebook with the full
sphinx machinery: every worksheet could have an optional
cell/tab/something to provide the equivalent of conf.py for sphinx
docs.
Just thinking out loud...
Cheers,
f
Rob Beezer
> Certainly! What I'd love to have is reST as a native storage format
> (so it can be easily edited outside the GUI with emacs) along with a
> choice of visual/native editing in the notebook interface. It's just
> that in switching from visual to native you'd go to reST as the 'raw'
> format instead of pure html, and the html would be only rendered for
> visual display.
other words, would jsMath still be used to interpret things like
output from show(), or the "typeset" box, or would math be written as
reST, such as `x^2+y^2` to be converted to images by a full-blown TeX
implementation?
The question could illustrate that I'm not sure I fully understand
what is being suggested. Generally, I am very interested in the
ability to put complex jsMath (output from the tex4ht translator) into
the space between cells and have it rendered nicely in a browser. The
HTML output of mathematics created by Sphinx is inline PNG's which I
think of as vastly inferior to what jsMath can do.
Rob
William Stein
>
> On Sep 8, 8:54 pm, Fernando Perez <fperez....@gmail.com> wrote:
>> Certainly! What I'd love to have is reST as a native storage format
>> (so it can be easily edited outside the GUI with emacs) along with a
>> choice of visual/native editing in the notebook interface. It's just
>> that in switching from visual to native you'd go to reST as the 'raw'
>> format instead of pure html, and the html would be only rendered for
>> visual display.
>
> What place would snippets of jsMath have in this arrangement? In
> other words, would jsMath still be used to interpret things like
> output from show(), or the "typeset" box, or would math be written as
> reST, such as `x^2+y^2` to be converted to images by a full-blown TeX
> implementation?
I am imagining a completely different supported mode that is very
Sphinx-ish. That would I think not necessarily have an impact on the
current look/feel of the notebook. Having tunderlying format be
HTML should always be the default since HTML is by far the most
powerful and flexible language in which to create... HTML, which is
what the notebook displays after all.
> The question could illustrate that I'm not sure I fully understand
> what is being suggested. Generally, I am very interested in the
> ability to put complex jsMath (output from the tex4ht translator) into
> the space between cells and have it rendered nicely in a browser. The
> HTML output of mathematics created by Sphinx is inline PNG's which I
> think of as vastly inferior to what jsMath can do.
Sphinx can generate jsmath. In fact, we now do just that for the
notebook's docstrings.
William
Rob Beezer
Thanks for the explanations.
> Sphinx can generate jsmath. In fact, we now do just that for the
> notebook's docstrings.
formatting. And it is good to know that Sphinx can emit jsMath!
Rob