| Re: high quality 'old' direct graphics [message #73721 is a reply to message #73717] |
Wed, 24 November 2010 12:36   |
![Paul Van Delst[1] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "Paul Van Delst[1] is currently offline")
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
David Fanning wrote:
> Paolo writes:
>
>> now I am totally confused.
>>
>> I haven't gotten to play around with IDL 8 (we just don't have it,
>> I guess astronomers just prefer to work with old software).
>>
>> From what I understood from the new object graphic system, the point
>> was that you could give various dot commands (if you create an
>> object myplot=plot(x,y) and then you would issue comamnds like
>>
>> ...
>>
>> only better looking (and that should apply for all graphic keywords,
>> with maybe a few new ones added).
>>
>> Is my understanding right?
>
> I think your understanding probably comes from listening
> to the marketing folks, but, yes, that's the general idea. :-)
>
>> If it is, why we need the wrapper?
>
> Uh, mostly to tide you over until 2018. :-)
>
>> If it isn't, what went wrong with the new plot objects?
>
> It's not that anything has gone wrong with the new graphics
> objects, it's just that not all that much has ever gone right.
Hmm. I think that was true for iTools/OG (i.e. too much code required to replace the DG "plot") but it's not as clear
with NG.
Most of the problems I've encountered have been due to me not wanting to shift my perception anchor (i.e. "Why doesn't
the NG stuff work the way DG did! Argh!" type of thing).
But, once one decides to make that change and invests the time to do it, it's not so bad. The NG learning curve is much
shallower than that for OG, but it's not flat.
My tipping point was the creation of barplots with legends. I have never been able to get good DG barplots (futzing
about with bar widths, the number of bars/category, colouring, etc). When I used the NG barplot function, it worked just
how I would expect it to work - that is: here is my data, plot it. Voila. It just worked.
> I still have hope, but not enough to stop publication of
> a book on traditional graphics, which I think people will
> still be using 10 years from now, when we are on the 6th
> (or will it be 7th) "new" edition of the New Graphics.
Yes. OG/NG is still waaaaaay too slow compared to DG to switch over completely. E.g.: I still find generating 1000's of
DG plots in a loop and watching the resulting "animation" instructive for visualisation of some types of data. No way,
no how that could be done that simply in NG.
And there is simply too much existing IDL code using DG for ITTVIS to discard that capability (without some sort of
backlash from users)
> Sooner or later, we are going to have to get something
> simple or we will never get enough bugs out of it to
> make it useful. IDL 8 graphics are clearly a step in
> the right direction. But I have my doubts about whether
> they are the final answer.
I haven't yet plumbed the depths of the NG capabilities so feel free to dismiss my opinion based on my blatant
misrepresentation where I use pulled-from-hat numbers as being quantitatively descriptive, but I reckon NG is about 90%
of the way to replacing DG (not taking into account any potential lingering NG bugs).
I think Paulo Penteado is a better choice for a more useful assessment in that regard since his posting history
indicates he clearly has a much more advanced, and quite nuanced, understanding of NG and how to use it. (Maybe you and
he should team up bookwise.....? :o)
> In the meantime, there will still be software available
> that gets the job done in a simple, elegant way. :-)
true that.
cheers,
paulv
p.s. Happy TG all!
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
