Ingrid Halama wrote:
> Maybe you are not aware that the Chart offers an UNO
API (see
> ::om::sun::star::chart )? You can access this API via
OpenOffice.org
> Basic macros for example and thus automate tasks (see
> http://api.openoffice.org/docs/DevelopersGuide
/DevelopersGuide.pdf for
> more details).
>> -- only a limited number of programmers can do
this
> This is always the case. Having a scripting language
people need to
> learn that first also. And they need to be willing to
share their
> results and they need to be able to share their
results. So providing
> a scripting language does not heal all problems.
e.g. Can I build with the UNO API in a number of minutes a
Venn diagram?
... or some other custom plot?
I am afraid, this is NOT the case.
- as I mentioned, there are over 500 packages available for
R (mostly
developed by independent groups using the R-language)
- gnuplot:: there are 649,997 downloads only from
http://sourceforge.net/search/?type_of_searc
h=soft&words=gnuplot
- asymptote: 23,631 downloads, see
http://sourceforge.net/search/?type_of_sea
rch=soft&words=asymptote
-- this is quite a powerful descriptive vector graphics
language for
technical drawing, I highly recommend to take a look at it
All these programs have simple drawing commands, like plot()
or draw()
and I do not need to learn complex UNO APIs.
- I can sit down and write in 5 minutes a script that will
solve a
particular problem, e.g. draw a very unusual graphic/ chart
- and do not need to ask OOo to implement that particular
type of Chart
- the (mathematical, statistical) professionals will
continue to use
these programs, so why not *win their help* for OOo
- and let *them* write new /extend existing features
>> -- it takes eons to implement a new feature
> Be fair. This highly depends on the feature itself, on
the current
> state of the affected code and on the availability of
resources.
I see features pending for years. What chance has a very
fancy feature
to be implemented? It usually gets a WONTFIX. There are even
popular
requests not implemented yet.
Having the users writing their scripts also avoids the need
for
resources. BUT, for this to work, the scripting must be
plain easy and
familiar. I believe plot() is easier to understand than
::com::sun::star::chart.
Everything is about the choice. I currently have no choice
but to learn
the (quite limited) UNO APIs. If a scripting engine would be
implemented
like that in gnuplot, R, or asymptote (which are by the ways
very
similar), I do have choices. And chances are great that many
will know
already a lot of those languages. I have added on the chart
wiki page a
great number of chart-types (please follow the links there,
see
http:
//wiki.services.openoffice.org/wiki/Chart2): these
charts were all
done in R, and most did not require implementing a fancy new
feature in
R, but were rather done using existing R-scripting (as an R
package).
> Please provide more specific use cases. I asume you
don't want just a
> clone of R or gnuplot, otherwise you just could use
them. So it's hard
> to estimate what you are asking for.
I do use R-extensively. I have some limited knowledge of
gnuplot, too.
What I miss is the ability to code easily my own type of
chart. IF it is
not present in Chart, I do not have any possibility to make
it. In R
(and asymptote), I would have simply written some plot()
directives and
created my custom chart type. (for R actually I probably
would have
searched an existing package, chances are high that the work
has already
been done).
It is also much about resources. Why reinvent the wheel?
Sometimes we
may need a square wheel, but most of the time the time
proven round
wheel will be far superior. Some of the languages mentioned
are mature
and very robust.
The main issue is, that without such effective drawing
directives, the
OOo team is stuck *to implement any requested feature* on
its own, while
following my suggestion, this work would be largely done by
the users
themselves (and other professionals not necessarily
affiliated with OOo).
I hope that the developers take a careful look at this more
global
approach. Having the developers split on thousands of open
source
projects and copying a feature from one product to the next
won't help
the open source community.
Sincerely,
Leonard Mada
------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribe graphics.openoffice.org
For additional commands, e-mail: dev-helpgraphics.openoffice.org
|