Am 19.05.2010 23:22, schrieb nemesis:
> clipka escreveu:
>> BTW, I just had a look at the Windows front-end code, to find that it
>> would be quite easy to change it to use a prettier scaling method -
>> all it would take would be to change a scaling-mode constant, and
>> slightly adjust some code called whenever an SMP square has been
>> rendered. However, that change would apparently break compatibility
>> with Windows 95/98/Me; I don't think such a minor improvement would be
>> worth that price.
>
> who the freak still uses those ancient interfaces?!

How about people who have some spare computer they want to use to run 
POV-Ray renders, and choose to install neither XP or newer (because the 
computer might be too slow, or they might want to save money on a 
license and re-use some old Win98 license for that purpose), nor Linux 
(because they don't grok it well enough)?

I'm not saying breaking compatibility would be a horribly dear price to 
pay /per se/, but I think /this/ particular issue is not worth paying it.

Post a reply to this message

On 05/19/2010 07:01 PM, clipka wrote:

> <snip> nor Linux
> (because they don't grok it well enough)?

ha-ha ... I'm thinking of those tv commercials that say "so easy even a
caveman can do it"

> I'm not saying breaking compatibility would be a horribly dear price to
> pay /per se/, but I think /this/ particular issue is not worth paying it.

concur ... also I would think this issue would probably be pretty low on
the priority list to boot!

Post a reply to this message

"clipka" <ano### [at] anonymousorg> schreef in bericht 
news:4bf45fb9$1@news.povray.org...
> Am 19.05.2010 23:22, schrieb nemesis:
>> who the freak still uses those ancient interfaces?!
>
> How about people who have some spare computer they want to use to run 
> POV-Ray renders, and choose to install neither XP or newer (because the 
> computer might be too slow, or they might want to save money on a license 
> and re-use some old Win98 license for that purpose), nor Linux (because 
> they don't grok it well enough)?

Exactly. I still have an old PC with Me and an even older portable with 98. 
Once in a while I use them.  :-)
And I don't grok Linux. ;-)

>
> I'm not saying breaking compatibility would be a horribly dear price to 
> pay /per se/, but I think /this/ particular issue is not worth paying it.

I agree.

Thomas

Post a reply to this message

> I don't see anything in there that would deserve the term "ugly", except 
> maybe the scaling of the grey/white checkered pattern indicating the areas 
> that haven't been rendered yet. That, and maybe the fact that it's not 
> using any form of interpolation. But I'd consider neither of those 
> seriously unfit for a preview window.
>
> Though of course if you'd volunteer to improve it, I guess you'd be 
> happily invited ;-)

I'll download the source again and take a look.

> (Then again, maybe you're seeing something I failed to see in my simple 
> experiments I just did.)

Take a loot at the attached (left maximised, right 100%).  I think this is 
some kind of "worst case" because it uses diagonal lines and the scale 
factor is probably around 99% to fit it full screen.

Post a reply to this message

Attachments:
Download 'image3.png' (44 KB)

Preview of image 'image3.png'

> BTW, I just had a look at the Windows front-end code, to find that it 
> would be quite easy to change it to use a prettier scaling method - all it 
> would take would be to change a scaling-mode constant, and slightly adjust 
> some code called whenever an SMP square has been rendered.

That's what I thought.

> However, that change would apparently break compatibility with Windows 
> 95/98/Me; I don't think such a minor improvement would be worth that 
> price.

Couldn't it be a variable rather than a constant, set by a simple check for 
OS version?

Post a reply to this message

Am 20.05.2010 09:45, schrieb scott:

>> However, that change would apparently break compatibility with Windows
>> 95/98/Me; I don't think such a minor improvement would be worth that
>> price.
>
> Couldn't it be a variable rather than a constant, set by a simple check
> for OS version?

Indeed. I'm not familiar with testing for Windows version though, so I'd 
file that under "non-trivial change" - I guess there might be quite some 
pitfalls to that, given the diversity of past, present and (probably) 
future Windows versions (for instance, XP Pro x64 Edition is likely to 
identify as Windows 2003 instead).


Another option would be to activate the nicer smoothing mode in 64-bit 
versions only, as that's a safe bet for XP or newer.

Post a reply to this message

Thomas de Groot escreveu:
> "clipka" <ano### [at] anonymousorg> schreef in bericht 
> news:4bf45fb9$1@news.povray.org...
>> Am 19.05.2010 23:22, schrieb nemesis:
>>> who the freak still uses those ancient interfaces?!
>> How about people who have some spare computer they want to use to run 
>> POV-Ray renders, and choose to install neither XP or newer (because the 
>> computer might be too slow, or they might want to save money on a license 
>> and re-use some old Win98 license for that purpose), nor Linux (because 
>> they don't grok it well enough)?
> 
> Exactly. I still have an old PC with Me and an even older portable with 98. 
> Once in a while I use them.  :-)

well, guess that already makes up for 10% of all povray users. :P

True, keep the old code in.

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

scott escreveu:
>> I don't see anything in there that would deserve the term "ugly", 
>> except maybe the scaling of the grey/white checkered pattern 
>> indicating the areas that haven't been rendered yet. That, and maybe 
>> the fact that it's not using any form of interpolation. But I'd 
>> consider neither of those seriously unfit for a preview window.
>>
>> Though of course if you'd volunteer to improve it, I guess you'd be 
>> happily invited ;-)
> 
> I'll download the source again and take a look.
> 
>> (Then again, maybe you're seeing something I failed to see in my 
>> simple experiments I just did.)
> 
> Take a loot at the attached (left maximised, right 100%).  I think this 
> is some kind of "worst case" because it uses diagonal lines and the 
> scale factor is probably around 99% to fit it full screen.

BTW, Blender has this same issue, even though being OpenGL rather than 
Windows-specific.  Don't know about 2.5.

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

clipka wrote:

> Indeed. I'm not familiar with testing for Windows version though, so I'd 
> file that under "non-trivial change" - I guess there might be quite some 
> pitfalls to that, given the diversity of past, present and (probably) 
> future Windows versions 

It's actually rather straight-forward as the internal version
numbering is different from the marketing names. The GetVersionEx
API call will fill data into a OSVERSIONINFO struct you pass to
it. If dwMajorVersion >= 5 you're on Windows 2000 or higher.
If the improved scaling method is not yet available in 2000
then the test should be

   dwMajorVersion >= 6 || (dwMajorVersion == 5 && dwMinorVersion >= 1)

for XP or higher. See

   http://msdn.microsoft.com/en-us/library/ms724451(v=VS.85).aspx

for more information.

Post a reply to this message

Am 20.05.2010 09:45, schrieb scott:
>> BTW, I just had a look at the Windows front-end code, to find that it
>> would be quite easy to change it to use a prettier scaling method -
>> all it would take would be to change a scaling-mode constant, and
>> slightly adjust some code called whenever an SMP square has been
>> rendered.
>
> That's what I thought.

Hey-oh! You gotta love Microsoft...

Now there is this STRETCH_HALFTONE mode they have for displaying scaled 
bitmaps.

Seems like it does what you'd expect it to, when scaling bitmaps /down/ 
to fit large images into small windows.

Seems like it is /also/ capable of smoothing bitmaps when scaling them 
/up/ to fit small images into large windows.

Well, yes, /capable/ it apparently is - but does it /do/ it?

Looks like Microsoft hid some mean little algorithm in there that gets 
to decide which image will be smoothed and which will just be scaled 
(seemingly depending on how many different colors you have in there)...

Oh bugger!

Post a reply to this message