Rendering transparent objects changed in M8
David Skulsky
that transparent objects (i.e., objects with Opacity < 1) were not
rendering properly under Mathematica 8; these same objects did render
properly under Mathematica 7.
I contacted Wolfram; they investigated the issue and promptly replied
with an explanation and what I expect will solve the problem.
Apparently the default method for rendering transparent objects was
changed in M8 "from a BSP tree to a depth-peeling method," supposedly
for speed purposes. Evidently this change can cause problems on some
systems (mine!).
Switching back to the BSP tree can be accomplished by adding a hidden
option to Style:
Style[Graphics3D[...], RenderingOptions->{"Graphics3DRenderingEngine"-
>"BSPTree"}]
I will not be able to test this on my system until next week, but
Wolfram was able to test it on one of their machines and confirmed
that it fixed the problem. They also advise that this workaround has
not been tested on other ATI Radeon graphics cards.
David Skulsky
truxton spangler
Shame there isn't a web page on their site with bug fixes and work
arounds. Apart from being useful for users it would reduce the load on
their tech support people I would guess, because we could check out
the fixes and work arounds page before contacting them.
TS
David Skulsky
wrote:
I like your idea and I hope you send it to Wolfram (and that they
implement it).
Meanwhile, I'll post any additional workarounds I learn about in this
group (perhaps we should include "workaround" in the subject to make
searching a bit easier).
David
Daniel Lichtblau
Might be a good idea. I gather there were some management misgivings regarding this in the past, but my recollection is sufficiently foggy that I cannot really address that, and in particular I do not know if this remains a current issue. I will make some comments on various social and technical aspects.
(1) It is difficult to navigate between bugs vs. intentional and justifiable changes (with new "features" falling in the middle). I generally can address such things when writing replies as an individual. I am not so sure how well this transfers to corporate "official responsa", so to speak.
(2) It is easy to get into awkward situations where some users will claim that changes in behavior are bugs, when we at WRI opine otherwise. While I agree that having a publicly available set of workarounds can be useful (and recently posted one such), I do not like the idea of appearing to endorse that some changes are bugs. Obviously, the ones we agree are bugs do not fall prey to this problem. But I can say from past experience that a lot of energy and patience can be spent on such issues, and I'd not want to see that sort of problem elevated to corporate status.
(3) It is not easy to get this information to users who might need it. We do not always recognize when a question involves an issue in this territory. And most people who have such questions are not likely to be MathGroup readers (specifically, your "we could check out the fixes" is far from "most users would..."). The upshot is we might often face ignorance on both our end (when we fail to recognize this type of an FAQ) and the user's. This of course does not make having a workaround list a bad idea: it simply indicates why it might be of limited use.
(4) It can be difficult to write an FAQ in such a way that it is clear to most readers. So we could end up doing work that has little applicability. I believe our experience is that this is an issue, but not a serious one (meaning, an FAQ can be generally useful as a first line of defense).
I myself tend to be neutral on this. I simply want to indicate reasons why it could cause more work than it is worth (both for WRI and users).
I should mention explicitly that I am not speaking for Wolfram Research on this matter.
Daniel Lichtblau
Wolfram Research
truxton spangler
> On Nov 26, 2:28 am, truxton spangler <truxtonspangle...@gmail.com>
> wrote:
>
>
>
> > On Nov 25, 9:55 pm, David Skulsky <edskul...@gmail.com> wrote:
>
d
> > > that transparent objects (i.e., objects with Opacity < 1) were not
> > > rendering properly under Mathematica 8; these same objects did render
> > > properly under Mathematica 7.
>
> > > I contacted Wolfram; they investigated the issue and promptly replied
> > > with an explanation and what I expect will solve the problem.
> > > Apparently the default method for rendering transparent objects was
> > > changed in M8 "from a BSP tree to a depth-peeling method," supposedly
me
> > > systems (mine!).
>
> > > Switching back to the BSP tree can be accomplished by adding a hidden
> > > option to Style:
>
-
>
> > > >"BSPTree"}]
>
> > > I will not be able to test this on my system until next week, but
> > > Wolfram was able to test it on one of their machines and confirmed
as
> > > not been tested on other ATI Radeon graphics cards.
>
> > > David Skulsky
>
> > Shame there isn't a web page on their site with bug fixes and work
> > arounds. Apart from being useful for users it would reduce the load on
> > their tech support people I would guess, because we could check out
> > the fixes and work arounds page before contacting them.
>
> > TS
>
> I like your idea and I hope you send it to Wolfram (and that they
> implement it).
>
> Meanwhile, I'll post any additional workarounds I learn about in this
> group (perhaps we should include "workaround" in the subject to make
> searching a bit easier).
>
> David
People have asked for a bugs list for ages and been ignored so I do
not see any point in contacting them. Likewise a resource for
undocumented or insufficiently documented functionality has been
called for before. Obviously the actual documentation is the place for
this but a about this -- I've never understood why all function
options are not documented with usage examples. Even though there is a
business case for it (reduce tech support time, get relevant user
required documentation to include in the next version) the company
prefers to keeps its cards close to its chest.
Helen Read
>
>> "David Skulsky"<edsk...@gmail.com> wrote:
>> truxton spangler<truxtonspangle...@gmail.com> wrote:
>> [...]
>>> Shame there isn't a web page on their site with bug fixes and work
>>> arounds. Apart from being useful for users it would reduce the load
>>> on their tech support people I would guess, because we could check out
>>> the fixes and work arounds page before contacting them.
>>>
>>> TS
>>
>> I like your idea and I hope you send it to Wolfram (and that they
>> implement it).
>>
>> Meanwhile, I'll post any additional workarounds I learn about in this
>> group (perhaps we should include "workaround" in the subject to make
>> searching a bit easier).
>>
>> David
>
> Might be a good idea. I gather there were some management misgivings
> regarding this in the past, but my recollection is sufficiently foggy
> that I cannot really address that, and in particular I do not know if
> this remains a current issue. I will make some comments on various
> social and technical aspects.
>
[snip]
> (3) It is not easy to get this information to users who might need
> it. We do not always recognize when a question involves an issue in this
> territory. And most people who have such questions are not likely to be
> MathGroup readers (specifically, your "we could check out the fixes" is
> far from "most users would..."). The upshot is we might often face
> ignorance on both our end (when we fail to recognize this type of an
> FAQ) and the user's. This of course does not make having a workaround
> list a bad idea: it simply indicates why it might be of limited use.
>
> (4) It can be difficult to write an FAQ in such a way that it is
> clear to most readers. So we could end up doing work that has little
> applicability. I believe our experience is that this is an issue, but
> not a serious one (meaning, an FAQ can be generally useful as a first
> line of defense).
It seems to me that the Wolfram User Portal would be the perfect place
for some kind of FAQ / Wiki / Forum. I set up a Wolfram ID a few months
ago, when I signed up to be a Beta tester for v.8. At the time the
Portal was pretty bare bones, and I assumed that more features would be
added before the release of Mathematica 8. Here we are months later, and
still the *only* thing the Portal seems useful for is for obtaining
Activation Keys and/or downloading the software. It would be nice if
there were some actual content there.
I know plenty of Mathematica users/enthusiasts (students and colleagues
of mine) who don't have any idea how to get to Math Group, but would
seek out a FAQ or Wiki or Forum if such things were readily available
via a tab in the User Portal. It would be great if I could tell someone
who runs into something weird in Mathematica, "Log into the User Portal
and look around the FAQ..." or "Log into the User Portal and post a
question on the Forum."
--
Helen Read
University of Vermont
David Skulsky
This is just a confirmation that the workaround provided by WRI did,
in fact, solve the rendering problem on my machine.
David Skulsky