Chronological Thread 

< Chronological >      < Thread >

---

>> Three queries:
>> 1. Regarding the saved session facility - if it is now possible to
>> save a session that includes the model as well as the graph layout,
>> could we not just load a session containing the model when you click
>> on the website link to a model?
> Hi Peter,
>
> I have already asked Tommy about this, but he was concerned that this
> would be difficult to do with the current repository.

one question - if we are planning on providing PCEnv specific data
within the CellML model repository, are we also happy storing and
providing access to other application specific data? And if so, what are
the requirements such applications would need to meet?

> One remaining issue relating to graphs which could possibly be improved
> is the axis labeling. The current system ensures that the distance
> between ticks is either a power of ten, or half of a power of ten, such
> that there is between 2 and 12 ticks on the graph. It then puts the
> first tick on a multiple of the distance between ticks. This works very
> well if you zoom in around the origin. The problem is, if you zoom in,
> say, around 1, the positions of the ticks might be.
>
> 1.0000001, 1.0000002, 1.0000003, 1.0000004, 1.0000005.
>
> We can't write out the 9 character representation of these numbers
> besides the ticks, because they would overlap each other, so I rounded
> them to two significant figures. This means that you get 1.0 beside
> every tick, which is again confusing. There is not really any standard
> way of addressing it. We could write something like...
>
> 1.000000 + x beside the axis (where we currently write the multiplier),
> and then write 1, 2, 3, 4, 5 beside each tick, although this could be
> counter-intuitive.

another option might be to remove intermediate labels when an overlap is
detected? so you might end up with just either end labelled - which
could still get messy if the user wants to keep scrolling in to
1.0000000000000000001..1.00000000000000000005. I think also if a user is
zooming in to that fine a resolution you probably don't need to worry so
much about labelling the axes, as long as there is some way to click on
the graph and see the actual coordinates at that point.

It still seems strange to me that all this work is being done by PCEnv
when there are so many graphing libraries around that have already
overcome these types of issues to some extent. Is there really nothing
suitable that PCEnv could be using? And if not, has any thought been
given to extracting the graphing part out of PCEnv to make it available
as a general graphing utility for use with XUL (mozilla?)? Seems like
something that might be useful to other application developers.


David.

---

• [cellml-discussion] PCEnv questions, Andrew Miller, 03/23/2007
  • [cellml-discussion] PCEnv questions, David Nickerson, 03/23/2007
    • [cellml-discussion] PCEnv questions, Tommy Yu, 03/23/2007

---

Archive powered by MHonArc 2.6.18.