Re: [Texmacs-dev] Drawing mode feature request: Annotation of .eps graphics.

[Henri Lesourd]

> On Tue, 2005-11-08 at 00:24 +0100, Henri Lesourd wrote:
>  
>
> > > I'd like to create a diagram with some tool, and then use the drawing
> > > mode to put text annotations, etc. on it.
> > >
> > >      
> > An patch for exactly this feature exists, but
> > it had to be delayed due to problems for the
> > accessibility of the parameters of the <superpose>
> > TeXmacs primitive. I should finish this one
> > in the next few days (but you will need to
> > get the CVS version, if you want to use it
> > next week).
> >    
>
> You're amazing.
That is also my own opinion :) ...

>  Thank you for spending so much of your time developing
> these new features of TeXmacs.
>
>  
... but in fact it is also my job, I do research in the
domain of user interfaces to theorem provers. One of the
first steps was implementing a first, principled support
for graphics into TeXmacs.

In practice, what I do is that I have a list, and then I
implement the features one after the other (or sometimes
I don't add such or such a feature, because i.e. I discover
that I would need some other thing implemented first).

Thus sometimes I already know the problem (like this image
annotation feature : indeed, this one has in fact explicitely
been added to the list due to a suggestion of Joris), but
sometimes not. So I will never be unhappy if people keep
asking for new features, problems, etc.

> As a study, I created two vector diagrams for my last Introductory
> Linear Algebra assignment hand-in.  One diagram was created using
> 'eukleides', and the other with 'drgeo'.  The first took more than 3
> hours to complete, even after having had some experience with eukleides
> from past use.  The second took around an hour, including time
> experimenting with DrGeo, since I've never used it before.  Certainly
> the interactive drawing editor is easier to use and faster than purely
> scripted one.
>
>  
It would be nice to know the time spent for doing the very same
diagrams using TeXmacs, then.

> Some things I noticed:
>
> DrGeo needs love and attention.  It's not working right, and needs
> development time spent on it.  For the most part, it's very good, and I
> liked using it.  It's fun to learn to create geometric constructions
> like that.  What I don't like is that I have to click and modify
> attributes of each line at a time.
In the TeXmacs graphical editor, you can definitely group changes.

>  I like the XFig feature where you
> put it into an update attributes mode and then click around on things to
> set them with one click.
Cf. previous answer.

>  I like how all the settings are displayed on a
> panel at the bottom.
This one is missing in TeXmacs, and it will
not be extremely easy to add (need to implement
other widgets :-(... ).

>  Certainly there are features of XFig worth mining
> a few nuggets from.  I had trouble getting a usable .eps from DrGeo; I
> found that the best route was to export to LaTeX, run that to get
> a .dvi, then run dvips and finally pstoepsi then rename the .epsi
> to .eps so that TeXmacs could see it.  Round-about, but it works.
>  
Again a problem you wouldn't have with TeXmacs.

> Eukleides is just as interesting, but having to script it makes creating
> a diagram into a very slow iterative process.  Something that combines
> the two approaches in some way might be cool... though I suppose DrGeo
> already does this, since it has Guile built into it.  I like the idea of
> having the potential to automate some kinds of diagramming or graphing
> using Scheme or some mathematical language built from a Scheme
> framework.
>
>  
I definitely agree on this.

> I would really love to have the ability to easily and interactively
> create 3-dimensional geometric constructions that can be viewed from any
> of a number of vantage points.  I'm not sure how the interface to such a
> thing would work...  My intuition is telling me I should get 'blender'
> and see how they do it.  I want to make diagrams like the ones in my
> calculus textbook.
>  
That one is probably for later, although as soon as you have
the ability to draw lines and filled polygons (currently,
you have everything you need for this purpose in TeXmacs),
it should not be very difficult to implement a routine
to calculate the projection of a set of 3D objects into
a TeXmacs graphics. Of course, interactive edition is
a completely different matter, then you would go into
the problem of defining a full-blown 3D editor, with
all the user interface problems that one could imagine.