Vuescan Epson 4490 Nightmare

[david.boreham]

I'd be very interested to hear from anyone that has been able
to scan negatives using vuescan with the 4490, while retaining
all their hair ! Mine is wearing very thin. Vuescan seems to want
to drive me crazy with random behavior. I should mention that
my troubles seem to all relate batch scanning. If I select the
crop manually one frame at a time, all seems well.

Some examples:

1. CCD Exposure changes randomly, and is often wrong (clipping).
Ok, so I can lock it and set the exposure manually, fixed that problem.

2. Final scan often is completely different from the preview.
Generally what I see in the preview display looks quite good.
However the final scan often has a horrible color cast that
was not there in the preview. At first I thought this was due
to scanner mechanical positioning errors between the
two scans, leading to border pixels being used (or not used)
in the color balance. However I realized the 'buffer' setting in the
crop tab could be used to 'fix' this problem if it were present.
The problem persists...
On one film, vuescan repeatedly scanned one frame as pure white
(the preview image for that frame was fine).

3. Ah Ha ! I found 'scan from preview' mode, which I figured would
obviously solve my 'preview is not the same as final scan' woes.
However scan from preview mode just doesn't work at all with
the epson 4490 , for me at least. It does the primary scan, which
looks ok on screen. Then it does the IR scan. Following that,
all I see on the screen in preview mode is the IR data !
Also, cropping X and Y are transposed between the regular
mode and the 'scan from preview' mode. This makes me wonder
if the crop locations are all screwed up in the regular scan,
which would certainly explain the crazy demonic exposure randomness
and color casts.

Of course, the epson scanner software is (even more) worthless
and Silverfast does not support ICE on this scanner (which is the
reason I bought this particular scanner model).

Arrrghghghghghghghgh....

Ed, please consider making vuescan open source. That way I could fix
the bugs myself.
Also make the documentation wiki-ized so we can improve that too.

 

[Don]

I'd be very interested to hear from anyone that has been able
to scan negatives using vuescan with the 4490, while retaining
all their hair ! Mine is wearing very thin. Vuescan seems to want
to drive me crazy with random behavior. ....
Arrrghghghghghghghgh....

That seems to be a common reaction to VueScan. ;o) It's very flaky.

I'm not a user, but the consensus seems to be to try and find a
version which works and then stick with it i.e. resist the urge to
upgrade. Also, when upgrading make sure you keep the old version, just
in case (e.g. rename the directory, do not delete). And let others
debug any new versions first!

The trouble is finding old versions (another FAQ here). So try asking
specifically for a version which does what you want with your scanner.
Maybe some kind soul can send it to you.

Of course, the epson scanner software is (even more) worthless
and Silverfast does not support ICE on this scanner (which is the
reason I bought this particular scanner model).

VueScan doesn't do ICE either, as far as know, because Kodak wants
royalties. VueScan's so-called IR cleaning is vastly inferior to ICE.

What specific problems are you having with native software? Usually,
native software can produce a decent raw scan (which can be edited
later). Native software most often fails when one tries to use all the
"bells and whistles".

Ed, please consider making vuescan open source. That way I could fix
the bugs myself.

Not at all likely. Then all the messy code would be there for all to
see! You'd be pulling what's left of your hair wading through it! ;o)

Seriously though, scanner software is really very simple (at its most
basic level it's just data acquisition). I wrote my own
quick-and-dirty program fairly quickly. However, it's very basic
because I scan raw, archive and then do the image editing later in an
external editor.

I don't know how easy it is to get the Epson SDK but with Nikon it was
like pulling teeth. However, once I got it, the SDK is excellent.

Don.

 

[david.boreham]

VueScan doesn't do ICE either, as far as know, because Kodak wants
royalties. VueScan's so-called IR cleaning is vastly inferior to ICE.

Actually (when it's working, which it isn't right now), vuescan's
IR defect cleaning works just fine for me.

What specific problems are you having with native software?

Where to start ?? It takes forever to scan. It won't write its output
to a network mounted drive. It won't rotate the images. Its attempts
at balancing the old negative emulsions I'm scanning are horrible
(although they _are_ consistent, unlike vuescan).

Usually,
native software can produce a decent raw scan (which can be edited
later). Native software most often fails when one tries to use all the
"bells and whistles".

Yeah, problem is that I'm scanning hundreds of frames
and there's really no way that I want to be manually balancing the
color and exposure in all of them. If there were some offline automated
tool
that would do the color negative post processing steps
(e.g. imagemagick or similar), then yes I'd use vuescan to take
a full-page raw scan and post process. However so far I haven't
found such a tool.

Not at all likely. Then all the messy code would be there for all to
see! You'd be pulling what's left of your hair wading through it! ;o)

Bring it on. I work with multi-million lines of code projects every
day.
Nothing could possibly surprise me when it comes to messy code.
I've been through 'open sourcing' events on a few big projects. We
never
tidied up code (that'd break it, which is the opposite of what you
want).
Generally curse words in comments are removed though.

Seriously though, scanner software is really very simple (at its most
basic level it's just data acquisition). I wrote my own
quick-and-dirty program fairly quickly. However, it's very basic
because I scan raw, archive and then do the image editing later in an
external editor.

It's the image processing necessary to handle old faded color
negatives that embodies the value to me in this case. Such software
is absolutely not really simple !

 

[david.boreham]

BTW, I _think_ I have succeeded in achieving consistent and acceptable
output from vuescan (please cross your fingers everyone...).

Problem is I don't actually know how I did it. I deleted the .ini file,
then began with a non-batch single frame scan. That worked. Then
I changed the config, one little step at a time, taking a test scan
each time, until I had everything enabled that I want.

However, IR cleaning is now not working It did work before I fixed
the color casts and exposure problems. The IR scan is being done
correctly because I can see dust spots in the scan image shown on
the screen. For some reason vuescan is now ignoring the IR image
and produces a dirty scan.

 

[david.boreham]

However, IR cleaning is now not working It did work before I fixed
the color casts and exposure problems. The IR scan is being done
correctly because I can see dust spots in the scan image shown on
the screen. For some reason vuescan is now ignoring the IR image
and produces a dirty scan.

This appears to be a straightforward software bug : if I view a final
scan
in vuescan, then from the 'filter' tab, disable IR Clean, then
re-enable it,
the scan image refreshes to the cleaned state. Somehow the config
reflected in the UI is not being observed by the software when it
processes
a scanned image.

 

[Roger S.]

If you are scanning negatives, try the advanced workflow. This would
explain inconsistency from the preview and the final scan (not
Vuescan's fault- preview is done at a nominal exposure, final at the
real exposure unless you lock the exposure and film base color).

Re: the IR problem, probably need to contact Ed Hamrick. Did you
manually change the IR exposure?

 

[mark.thomas.7]

Did you log the issues with Ed?

If the 4490 is a relatively new scanner, there may well be bugs in his
implementation. Way back when I got my copy of Vuescan in about 2000,
it was for a then new Acer film scanner. There were a few minor bugs,
but within a couple of weeks, Ed had fixed them. And since then the
program has improved beyond recognition. As far as I am concerned,
Vuescan was the best $49 (I think it was) I ever spent - it's certainly
the only one-man-show piece of software I have anywhere my computer -
it works consistently and excellently, and I'm still getting free
updates.. And even while it was buggy for those few weeks out of 6
years, it was a damn sight better than the Acer software. You said
yourself it is better than the Epson equivalent..

So I suggest you log the issues, and be patient - Ed's only one guy
supporting 500 or more scanners! I've watched Vuescan's development
over the years, and I think it is a remarkable product.

Why this post? I just thought it would be good to balance Don's post.
You might want to check Don's posting history regarding Vuescan, and
regard his comments accordingly. Feel free to do the same with mine!
(O:

 

[mark.thomas.7]

And a specific reply to Don...

That seems to be a common reaction to VueScan. ;o)

Don's opinion, oft repeated, by Don. (Hmm, why the smiley? Does he
also recognise how silly he sounds?)

It's very flaky.

More Don opinion. Do your own survey - just search on 'Vuescan' and
see what you find. BUT, gather all the posts by Don together and make
sure you count him as one person only.. I use Vuescan. All the time.
Not on many scanners, admittedly, only 6 so far over 6 years, and only
two currently (an Acer film scanner and an Epson flatbed). It works
flawlessly, and gives far more control and far better results than the
manufacturers software. Plus I don't have to learn a new interface,
and I very much like the UI Ed has created. Unlike Don, I don't
pretend to speak for the multitudes, this is all just my opinion.

I'm not a user

I *am*. Who would have more cred, I wonder? If I was not a user, I
wouldn't presume to post herewith and offer uninformed opinions. Don
does, and that is his choice, of course.

but the consensus seems to be to try and find a
version which works and then stick with it

Whose consensus? Like I said, unlike Don, I won't pretend to speak for
the masses (no JC syndrome here), and invite you to check on whether it
is indeed a 'consensus'.. I haven't heard this recommendation
frequently applied specifically to VS, except from Don. But then I
don't stalk every Vuescan post, either. I just upgrade when I think of
it. And I've never been bitten with any new bugs that way. I do
always keep an archive of software for my PC anyway, but in this case
it doesn't bother me as I haven't found the product to be in any way
flaky. Don's mileage may vary (well.. - if he was a *user* it
might...)

Just as an aside, (if I *was* a consensus-seeker (O my reading of
these forums would indicate that the flakiest software is, by a country
mile, Nikonscan... Even Don lambasts it and the Nikon software
designers.... But I've not done accurate stats, so just ignore that
opinion. (O;

VueScan doesn't do ICE either, as far as know, because Kodak wants
royalties.

ICE is a trade name. The process is called IR cleaning. Of course it
doesn't do 'ICE'.

VueScan's so-called IR cleaning is vastly inferior to ICE.

More 'Don opinion' from that non-user. Please post examples, Don, or
STFU.

If you actually try to research this, you will find an endless tirade
of usenet posts about how bad Vuescan's IR is... but they are *all*
from Don!! (and perhaps one other lunatic who posts using forged
identities). Trying to find at least one that is relevant, his one is
interesting:
http://groups.google.com.au/group/comp.periphs.scanners/browse_frm/thread/2eed4f107f5df18c/
- someone posts about a bugfix, and another gives Vuescan IR the
thumbs up.. but Don chimes in with his vitriol AGAIN.

Not at all likely. Then all the messy code would be there for all to
see! You'd be pulling what's left of your hair wading through it! ;o)

<sarcasm>
Yes, Don, *of course* that is the *only* reason that Ed wouldn't post
his code for everyone to see...

Seriously though, scanner software is really very simple (at its most
basic level it's just data acquisition). I wrote my own
quick-and-dirty program fairly quickly. However, it's very basic
because I scan raw, archive and then do the image editing later in an
external editor.

If it's crap, why bring it up? If it's good, why not post *your* code?
Can you spell 'hypocrite'?
Oh, yes, it's absolutely *easy* and *trivial* to write a program that
does all that Vuescan does, for 500 different scanners with different
hardware, on-board programming, etc.. That's why there are soooo many
alternatives you can use, like... er.... Silverfast, and, and .....
What's your recommendation, there, Don?

And yes, it's easy to mouth off and *imply* that Don could do sooo much
better. Given the time he spends sniping, why hasn't he just *done*
it? Could it be that he isn't quite the programmer he thinks he is, or
that he hasn't considered that the software supports more than just one
scanner..? Could it be that he is really pissed that Ed makes a
decent living from it? (O;

When Don actually shows his expertise, people might listen a bit more
carefully, but I haven't seen any evidence, just talk. Often that talk
has been a giveaway - a classic example being a 'discussion' I just
found between him and 'UrbanVoyeur', on color management. Don's
'approach' and his comprehension of that topic is.. er.. interesting,
and *very* telling.

My experience with Vuescan and Ed has been excellent. Yes, Ed has a
short fuse and I have seen him 'go off', but usually it has been with
good reason. Because of people like Don. And I imagine that is why
you don't see him around here much - he can much better spend his time
developing the software and producing a better product for the
*genuine* users of it, than waste time arguing with persons unburdened
with facts (let alone a registered copy of the program they know *so
much* about).

I don't know how easy it is to get the Epson SDK

Exactly, every word is a giveaway - not knowing never stopped him
before...

(O;

To the OP (as you seem to have done) - go back to all defaults,
identify when the problem/s kick in, and send a brief email and your
log file to Ed's support email address. He *will* help.


And do post back here and tell us what transpires. Genuine user
comments and experiences are very welcome. They balance out the crap.
(O:

 

[Don]

Actually (when it's working, which it isn't right now), vuescan's
IR defect cleaning works just fine for me.

ICE uses a much more sophisticated algorithm but if you're happy with
Vuescan's IR cleaning then that's all that counts. It's just that I do
remember people complaining here repeatedly about IR cleaning bugs.

Where to start ?? It takes forever to scan. It won't write its output
to a network mounted drive. It won't rotate the images. Its attempts
at balancing the old negative emulsions I'm scanning are horrible
(although they _are_ consistent, unlike vuescan).

The inability to access a network drive is inexcusable. However, as to
it being slow, all native software is "slow". But that's because they
do things "properly" (read, being very conservative and going by the
book). This is so the software is reliable and consistent, and they
have less to worry about after-sales service.

Vuescan, by contrast, cuts a lot of corners and pays the price. That's
why it's so flaky and versions are tripping over themselves to keep up
with the bugs. So it's really essential to find a Vuescan version
which works and then stick with it.

Yeah, problem is that I'm scanning hundreds of frames
and there's really no way that I want to be manually balancing the
color and exposure in all of them. If there were some offline automated
tool that would do the color negative post processing steps
(e.g. imagemagick or similar), then yes I'd use vuescan to take
a full-page raw scan and post process. However so far I haven't
found such a tool.

I think any software that can do batch processing will probably work.
I use Photoshop and there you can set up so-called "Actions" (scripts,
really) and then apply them to all files in a directory, for example.

Since you're happy with Vuescan output another option may be to use
native software to actually scan and then use Vuescan afterwards as an
image editor. If memory serves it's the so-called "scan from raw".

Bring it on. I work with multi-million lines of code projects every day.
Nothing could possibly surprise me when it comes to messy code.
I've been through 'open sourcing' events on a few big projects. We never
tidied up code (that'd break it, which is the opposite of what you want).
Generally curse words in comments are removed though.

I've been at this since the IBM mainframe days and sometimes it's just
better to toss the lot and start from scratch. I know... The bean
counters shriek in horror when that's suggested, not realizing it's
often cheaper to do that instead of throwing good money after bad.

I disassembled parts of Vuescan a while back and it was pretty awful.
But even without that, just looking at its track record says it all. A
bug shows up, disappears, shows up again, disappears, etc.

It's the image processing necessary to handle old faded color
negatives that embodies the value to me in this case. Such software
is absolutely not really simple !

I see! That's a different story altogether. Maybe, as I suggest above,
(if all else fails) using native software to actually scan and then
use Vuescan only as an image editor afterwards could do the trick?

Don.

 

[Noons]

Problem is I don't actually know how I did it. I deleted the .ini file,
then began with a non-batch single frame scan. That worked. Then
I changed the config, one little step at a time, taking a test scan
each time, until I had everything enabled that I want.

However, IR cleaning is now not working It did work before I fixed
the color casts and exposure problems. The IR scan is being done
correctly because I can see dust spots in the scan image shown on
the screen. For some reason vuescan is now ignoring the IR image
and produces a dirty scan.

Hmmmm, you must be using the latest version, 8.3.59 or so?
My 4990 developed a similar problem with these latest versions,
was fine until 8.3.56...


This is what I did to get rid of the problems:

1- delete the ini files I was using.
2- set colour correction to neutral.
3- select "Default options" from the File menu.
4- Re-enable all the advanced bits I was using before.

the above came out of a mail exchange with Ed to
find out how to default Vuescan to a vanilla init
after I'd been farting around with advanced settings.

5- In the Input tab, click on 'Transparency". As in physically
select that, then CLICK on the word. You'll notice the
screen flickers.
6- Now check all the crop tabs, the filter tabs and the colour and
output tabs and make sure all the options are still as you set.
7- Save a new ini file.
8- Make a few test scans with Preview, then scan an image near
the bottom of the scanner - the ones that force the carriage to
travel further. Check the crop is fine.
9- For the first batch scans, make sure you tick the
"display raw scan" option in Prefs: that lets you see what the
heck is being scanned and check straight away if there is a crop
error.

I think something changed in ini file handling in the latest versions
and that's what's causing the problems with the Epsons.
Wrong crop as the frames approach the bottom is the main
problem, there are also a few with exposure calculations off
the preview: when in doubt, lock exposure and select 1.5
duration for a normal density colour negative, if it is too dark
(overexposed negative) then try 2 or at the most 2.4. Values
in between are good as well, depends on what produces best
combination for your scanner. Note that mine is the 4990!

I tend to scan into a raw 48bit scanner file with just IR being
applied from original. For this I need to set in input tab
the Auto Save "Scan" option, then in output tab I set Raw
Output With "Save" option and 48bit output file size.

Then I rescan the raw files - different ini file! - and do all colour
adjusts into output 24bit tiff or jpg files, which I then clean up,
final tune and sharpen with GIMP. The raw files are saved on
CD-R so I can go back and redo colour balancing if later on I
feel like it.

Usual disclaimers apply: YMMV, IME, don't blame *me* if
this doesn't work.

HTH

 

[Mal]

I disassembled parts of Vuescan a while back and it was pretty awful.

That statement alone shows you are either an idiot or a liar or both.

Disassembling only translates machine code (from a library or
executable) into assembly language. It does not and cannot generate the
original high-level language that virtually all application programs are
written in.

Furthermore the structure of the assembly language code is a function
both of the compiler used and the structure of the high-level language code.

Vuescan was almost certainly written in C and/or C++. It is impossible
to judge how elegant (or otherwise) Ed Hamrick's original code is by
disassembling code in the final executable.

That doesn't stop you from making fatuous statements about it though.

Mal

 

[Raphael Bustin]

That statement alone shows you are either an idiot or a liar or both.

Disassembling only translates machine code (from a library or
executable) into assembly language. It does not and cannot generate the
original high-level language that virtually all application programs are
written in.

Furthermore the structure of the assembly language code is a function
both of the compiler used and the structure of the high-level language code.

Vuescan was almost certainly written in C and/or C++. It is impossible
to judge how elegant (or otherwise) Ed Hamrick's original code is by
disassembling code in the final executable.

That doesn't stop you from making fatuous statements about it though.

Thank you, Mal, for stating what I've wanted to say for
a while about Don's "disassembly" exercise.

Last time I disassembled anything was in the early 1980s
or so, and that was stuff that had been *written* in
assembler, and it was hard enough to do back then.

I just can't imagine disassembling any real, current application
(eg. NikonScan) from the machine/assembly code that a
typical disassembler provides.

Anyway, even if one had Nikon's source, with English
comments (!!!) it would be a daunting task.


rafe b
www.terrapinphoto.com

 

[Noons]

I disassembled parts of Vuescan a while back and it was pretty awful.

er.......
that was not a good moment, Don. The "prettyness" of disassembled
code proves nothing about how the original was written. By any stretch
of the imagination.

Sure: vuescan has bugs. Sure they appear, go away, then come back.
Hey: it's a single guy, charging a small price for what's there
already.
A LOT better IMHO than the crap that comes out of Microsoft and
even IBM, for a much higher price...

Again: no financial interest or otherwise in vuescan. Heck, I paid
the same for the pro version as anyone else. It's not bad at all,
all things considered. Steep learning curve, granted. Weird-ish
UI, granted.

But for the price? I'll take it anytime over Silverfast, for example.

 

[mark.thomas.7]

Off topic.

I've been at this since the IBM mainframe days

Yes, *of course* you have, Don, and then you..

..disassembled parts of Vuescan a while back and it was pretty awful.

As noted by others, what a gorgeous piece of foot-in-mouth that was!!!
You are now proven as a liar. It's a pity, because some of what you
say is worthwhile, but when it is mixed up with paranoid schizophrenia
and then you toss in a complete fabrication about reverse engineering
by 'disassembly'... It would be funny if it wasn't sad.

What a moron. You're no IBM programmer (unless perhaps a cleaner, or
maybe you think using Basic on an IBM-compatible PC makes you one..?),
and your color management knowledge was blown out of the water in other
threads. But keep posting! - it's fun to watch and there are plenty of
folk now who will keep you honest and remind listeners of your
background.

Hey, why not prove us wrong and show us that disassembled code, along
with your pertinent comments proving it's 'awfulness'? (O;

 

[Don]

That statement alone shows you are either an idiot or a liar or both.

And that statement alone shows you seem incapable of a civilized
exchange.

Therefore, giving you the benefit of the doubt just this once (i.e.
assuming you're new) this is only to inform you that I do not reply to
insults and flame baits, as a matter of course.

So don't be surprised if your future "contributions" of a similar kind
are ignored and don't draw any false conclusions from being ignored.

Don.

 

[Mal]

That statement alone shows you are either an idiot or a liar or both.

[Self-important drivel snipped]

Don.

I don't give a damn whether you respond to or even read my posts - in
fact I suggest you put me in your kill file right now to avoid having
your delicate sensitivities ruffled further.

The rest of this post is thus addressed to the group generally.

The purpose of my previous post was to point out to the non-programmers
who read this group that Don's claim to have "disassembled" Vuescan and
found the code "messy" is totally and utterly false. [The same point was
made in the post by (e-mail address removed)]. Furthermore, no competent
software developer would ever make such a claim since they would know
that disassembling does not produce the original high-level language.
Ergo: Don is not a competent software developer.

A brief scan of the google groups archives shows that Don's programming
abilities seem to be limited to Visual Basic - an abomination of a
language aimed at those who can't actually program properly. Even at
this level, judging from his previous posts to
microsoft.public.vb.general.discussion he clearly has problems.

So, in short, if you have no knowledge of programming/software
development, don't believe anything Don says on the subject. If you do,
you'll know what a **** he is anyway.

Mal

 

[Don]

The rest of this post is thus addressed to the group generally.
Ditto.

Furthermore, no competent
software developer would ever make such a claim since they would know
that disassembling does not produce the original high-level language.

Riiight... "Mal" has clearly missed the point. Several, actually...

If "Mal" had 25-odd years of programming experience starting with IBM
mainframes (360-148) maybe "Mal" would understand. Or maybe not...

Optimization notwithstanding, object code is what actually runs. And
it doesn't matter what (high level) language was used to generate it.
In other words, calls to the OS are the same, regardless. Therefore:

When - in spite of this optimization (!) - these OS calls are all over
the place, it doesn't take a rocket scientist to deduce that the
underlying source code is a mess. And one doesn't need to recreate the
source code 100% to see this. Simple, unemotional, objective fact.

Not to mention the unending avalanche of recurring (!) Vuescan bugs!

A brief scan of the google groups archives shows that Don's programming
abilities seem to be limited to Visual Basic

Another brief scan also shows that the "rabid Vuescan mob" i.e. "the
Bart and Ed show" (not to be confused with reasonable Vuescan users)
have been busy forging Don's messages all over the place.

Here's a newsflash: You can't believe everything you read on the Net.

One thing we can believe, though, is that Mal's raging outburst was
his very first post in this newsgroup! Not an auspicious beginning.

It's also *very* (!) notable that someone who has just joined would:
- be so rabid from the word go
- be so obsessive about Don's post in particular

Hmmm? I wonder? Chip on shoulder, perhaps? In any case, very odd and
certainly food for though!

Mal

Mal, indeed! As in mal-intent, malign, malice, etc. How very apt!

Don.

 

[Raphael Bustin]

Optimization notwithstanding, object code is what actually runs. And
it doesn't matter what (high level) language was used to generate it.
In other words, calls to the OS are the same, regardless. Therefore:

When - in spite of this optimization (!) - these OS calls are all over
the place, it doesn't take a rocket scientist to deduce that the
underlying source code is a mess. And one doesn't need to recreate the
source code 100% to see this. Simple, unemotional, objective fact.

Umm, Don? With all (cough, cough) due respect:
you don't know what you're talking about.

"OS calls all over the place." The horror!

Umm, Don. That's what happens. That's the way it's done.
Now I grant you some folks over-use OS facilities, but it would
be hard to prove that from the object code. It's hard enough
to do even if you *have the source.*

I challenge you to demonstrate how any one of the calls you
supposedly spotted is spurious or unwarranted. I'd be amazed
if you could do that from an object-level disassembly.

You can't claim to know anything of the structure of a modern
application (like NikonScan or Vuescan) from the object code.
There are layers of abstraction and hierarchy that would take
a lifetime to divine.


rafe b
www.terrapinphoto.com

 

[Mal]

Umm, Don? With all (cough, cough) due respect:
you don't know what you're talking about.

"OS calls all over the place." The horror!

Umm, Don. That's what happens. That's the way it's done.
Now I grant you some folks over-use OS facilities, but it would
be hard to prove that from the object code. It's hard enough
to do even if you *have the source.*

It is nigh on impossible to determine from a disassembly where OS
calls are made. Even if one could, the vast majority will be nothing
to do with the application (Vuescan in this case) but will be an
indirect result of the various libraries Vuescan is linked with (as
you state in different terms later on).

It is virtually certain that Vuescan will not make more than a
few (if any) direct system calls - applications like that have no
need to or use for them.

Finally, of course, whatever is the pattern of systems calls that
exists in the final object code is determined by the compiler and
the linker. They structure this code in a way that only the
compiler/linker writers would have proper knowledge of. And this
structure will change depending on what compiler flags were used.
So any pattern of system calls (or any other type of call - eg graphic
library calls) has no correlation with any pattern there may
be in the original source code.

I challenge you to demonstrate how any one of the calls you
supposedly spotted is spurious or unwarranted. I'd be amazed
if you could do that from an object-level disassembly.

So would everyone, because it's impossible

You can't claim to know anything of the structure of a modern
application (like NikonScan or Vuescan) from the object code.
There are layers of abstraction and hierarchy that would take
a lifetime to divine.

Precisely - which is where we came in and why Don's original claim

"I disassembled parts of Vuescan a while back and it was pretty awful."

shows either ignorance or is a falsehood.

Mal

 

[Mal]

If "Mal" had 25-odd years of programming experience starting with IBM
mainframes (360-148) maybe "Mal" would understand. Or maybe not...

I'm afraid I don't have the 25 years experience you claim, I only have
a real 40 (that's four zero, forty) years. Bummer, eh?

As to the rest of your post, rafe b has adequately dealt with your
further Vuescan disassembly claims and I've added a few comments of
my own to his post.

The attempted Red Herring on the recent forged posts and the
ad hominem attack are irrelevant and ignored.

Mal